ABSTRACT
Title of Dissertation: REFRAMING RESPONSIBILITY FOR ACADEMIC
SUCCESS: A CAUSAL MODEL MEASURING THE
IMPACT OF STUDENT ATTRIBUTES IN THE FIRST
YEAR OF COLLEGE
Michele C. Murray, Doctor of Philosophy, 2006
Dissertation Directed By: Jeffrey F. Milem, Ph.D.
Department of Education Policy and Leadership
The purpose of this single-institution study was to investigate the predictive
power of student attributes in a path analytic model for academic success in the first year
of college. Student attributes were defined as academic self-concept, social self-concept
and self-determination; academic success was measured by cumulative college grade
point average. The conceptual model tested in this study blends psychological theories of
student attributes with Astin?s (1991) input-environment-outcome (I-E-O) model, a
sociological model of college impact. Using descriptive and path analytic techniques,
this study contributes to assessment philosophy by demonstrating that student attributes
predict academic success beyond what can be explained by prior achievement and
involvement.
By examining the contributions of student attributes to academic and social
involvement and to subsequent achievement, this study describes higher education as a
partnership between student and institution for which both have responsibility. The
findings of the study suggested at least through conclusions. First, accounting for student
attributes contributes to an understanding of academic success. Rather than focus on the
institution?s responsibility to engage students, this study demonstrates that academic and
social involvement and achievement are products, at least in part, of students? academic
self-concept and self-determination. Second, results from this study indicate that
measurable change in student attributes occurs during one year, a portion of which is
attributable to students? academic and social involvement. These findings substantiate
previous research on the impact of involvement on students? personal development
(Astin, 1994; Berger & Milem, 1999) and affirm the benefits of college attendance.
Third, this study demonstrates that the effects of the environment within the classic I-E-O 
model (Astin, 1991) are mediated through academic self-concept.
These findings reframe responsibility for student success by highlighting
students? dispositions toward the academic enterprise as the strongest predictor of
involvement and success. Consequently this study offers a different perspective of
students? academic and social involvement. Rather than referring to involvement as an
indication of the environment (Astin, 1994; Kuh, 1991), this study suggests that
involvement behaviors are a measure of students? responsibility toward their collegiate
experiences. The findings of this study have implications for future research, practice,
and policy.
Reframing Responsibility for Academic Success:
A Causal Model Measuring the Impact of Student Attributes
in the First Year of College
by
Michele C. Murray
Dissertation submitted to the Faculty of the Graduate School of the
University of Maryland, College Park in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
2006
Advisory Committee:
Associate Professor Jeffrey F. Milem, Chair
Associate Professor Robert G. Croninger
Associate Professor Sharon L. Fries-Britt
Assistant Professor Karen Kurotsuchi Inkelas
Associate Professor Marylu K. McEwen
?Copyright by
Michele C. Murray
2006
ii
DEDICATION
For my parents, Dwight and Elodie Murray, who have always encouraged me.
and
For my husband, Chris Lewers, who helps me believe that anything is possible.
iii
ACKNOWLEDGEMENTS
Many people say that writing a dissertation is a lonely, solitary project?well,
they are only partially correct. I am grateful for all the supporters who have surrounded
me these many months?each one has made this arduous task endurable and even
enjoyable.
First, many thanks go to my good and gracious God, who has given to me so
generously, not the least of which are the opportunity for education and people with
whom to share the journey.
Now I would like to thank Jeff Milem, my chair, advisor, and friend. He has been
an advocate and mentor from my first days in the program, and I am honored to have him
by my side as I finish. It has been a privilege to learn from him, and I am appreciative for
the ways he has guided me to answer my own questions. Many thanks also go to the
members of my committee, Bob Croninger, Sharon Fries-Britt, Karen Inkelas, and
Marylu McEwen. Their insightful comments and continuous encouragement helped
strengthen this study and create a better product. I am grateful, too, for Laura Perna who
helped shape the first draft of this project. Thanks also go to the Executive Council of
Magis University who allowed me to use their institutional data?without their
contribution, this dissertation would not exist.
I have also made some life-long friends in this doctoral program and consider
myself blessed to be part of such an amazing and supportive cohort. Heather, Kathleen,
and Michelle have generously shared in the ups and downs of graduate studies and of life.
Their presence made me a better student, scholar, professional, and person. I am grateful
iv
for all they have added to my life and look forward to celebrating the many
accomplishments I know will be in their futures.
Next, I owe a deep debt of gratitude to my colleagues. Despite my protestations,
Susan and Donnie encouraged me to enroll in that first course, knowing all the while that
I would find my place and finish the degree. The hard work and dedication of my co-
workers, Maureen and Sara, made it possible for me to succeed as a full-time professional
and a part-time graduate student. I thank Michelle, Xavier, Susanne, and Kim who
helped me keep my sanity and reminded me to appreciate the simple joys of having good
friends. This dissertation is a testament to the community of colleagues and friends who
have graced my life.
I would also like to thank the many students who have shared their journeys with
me. This study was inspired by their stories, and I feel privileged to have been even a
small part of their collegiate experiences. They have taught me so much.
I am also grateful for the support of my family?Dwight, Elodie, Maria, and
Woody?and friends?especially Jehanna and Kristin?whose support and
encouragement never faltered. They patiently listened as I shared each theory and its
contribution to my thesis, and they genuinely shared in my excitement as the results
began to unfold.
Finally, I thank my wonderful husband, Chris, who appointed himself chief cook
and bottle-washer just to lighten my load. His love and support have made this
dissertation possible, and this triumph is as much his as mine. There are no words
graceful enough to express my love and gratitude.
v
TABLE OF CONTENTS
LIST OF TABLES........................................................................................................... x
LIST OF FIGURES ....................................................................................................... xii
CHAPTER I: INTRODUCTION......................................................................................1
Common Assumptions in College Impact Studies .....................................................1
Accounting for Attributes as a Factor of Student Success..........................................3
Theoretical Frameworks: Student Attributes and Involvement..................................6
Self-determination and Self-efficacy Theories .....................................................7
Definition of Key Terms.............................................................................................8
Purpose and Research Questions ..............................................................................10
Research Design........................................................................................................11
Limitations ................................................................................................................12
Significance of Study................................................................................................13
CHAPTER II: REVIEW OF RELATED LITERATURE..............................................16
Introduction...............................................................................................................16
Understanding College Outcomes ............................................................................17
The Ecology of Person-Environment Interaction ...............................................17
Notable Ecological Theories.........................................................................18
Barker?s Behavior Setting Theory ..........................................................18
Need x Press Culture Theory ..................................................................19
Holland?s Personality Theory .................................................................19
Studying College Impact.....................................................................................20
The I-E-O Model: Astin?s Assessment Methodology ..................................20
The I-E-O Model.....................................................................................21
Benefits of the I-E-O Model Compared to Other Assessment
Strategies.................................................................................................22
Relation of the I-E-O Model to Proposed Study.....................................23
The First Year of College as a Particular Environment............................................25
First-year Seminars .............................................................................................26
Learning Communities........................................................................................27
Faculty and Peer Interaction ...............................................................................28
Measuring Academic Success in the First Year: Grade Point Average....................29
Student Attributes .....................................................................................................30
Self-concept ........................................................................................................31
Academic Self-concept .................................................................................32
Social Self-concept .......................................................................................33
Related Research...........................................................................................34
Self-determination...............................................................................................37
Self-determination Theory............................................................................39
Basic Psychological Needs .....................................................................41
vi
Self-determination and Basic Psychological Needs within Education...42
Related Research.....................................................................................43
Critique of Self-determination Theory....................................................45
Self-efficacy..................................................................................................46
Related Research.....................................................................................47
Critique of Self-efficacy .........................................................................49
Summary.......................................................................................................50
Academic and Social Involvement .....................................................................51
Astin?s Theory of Involvement.....................................................................52
Related Research.....................................................................................54
Critique of Astin?s Theory of Involvement ............................................57
Student Effort................................................................................................58
Summary.......................................................................................................60
Studying Academic Success in the First Year: Literature Supporting
Variables in the Study...............................................................................................60
Student Background Characteristics...................................................................61
Gender...........................................................................................................62
Race...............................................................................................................62
First-generation Status ..................................................................................63
Prior Achievement ........................................................................................63
Student Attributes ...............................................................................................64
Prior Self-concept .........................................................................................64
Self-efficacy..................................................................................................65
Self-determination.........................................................................................65
First-year Experiences: Aspects of the Environment..........................................66
Academic Involvement .................................................................................66
Effort.............................................................................................................67
Social Involvement .......................................................................................67
Conclusion: Blending Sociological and Psychological Theories to Create a
Better Model .............................................................................................................68
CHAPTER III: METHODOLOGY ................................................................................71
Introduction...............................................................................................................71
Data Analysis............................................................................................................74
Descriptive Analysis ...........................................................................................74
Factor Analysis ...................................................................................................75
Path Analysis ......................................................................................................76
Conceptual Model.....................................................................................................79
Inputs?...............................................................................................................83
Student Attributes at Time 1...............................................................................83
Environments ......................................................................................................84
Student Attributes at Time 2...............................................................................85
Variables? ...............................................................................................................86
Dependent Variable ............................................................................................86
Exogenous Variables ..........................................................................................86
vii
Student Background Characteristics.............................................................86
Endogenous Variables ........................................................................................87
Self-concept (Academic and Social).............................................................87
Self-determination.........................................................................................87
First-year Experience....................................................................................88
Instrumentation .........................................................................................................90
Cooperative Institutional Research Program: Freshman Survey........................90
Student Readiness Assessment ...........................................................................92
Cooperative Institutional Research Program: Your First College Year .............93
Sample.......................................................................................................................95
Delimitations.............................................................................................................96
Limitations ..............................................................................................................101
Summary.................................................................................................................103
CHAPTER IV: RESULTS............................................................................................105
Introduction.............................................................................................................105
Exploratory Factor Analysis ...................................................................................105
Student Attributes at Time 1.............................................................................106
Measures of Self-concept............................................................................106
Self-determination.......................................................................................108
First-year Experience........................................................................................110
Faculty Interaction ......................................................................................110
Academic Engagement ...............................................................................111
Sense of Belonging .....................................................................................112
Student Attributes at Time 2.............................................................................113
Academic and Social Self-concept .............................................................113
Self-determination.......................................................................................115
Summary of Factor Analysis ............................................................................117
Inter-correlations.....................................................................................................117
Summary of Inter-correlations..........................................................................119
Missing Data Analysis ............................................................................................123
Summary of Missing Data Analysis .................................................................129
Addressing the Research Questions........................................................................129
Research Question One: Change in Student Attributes Over Time..................130
Academic Self-concept: Change Over Time ..............................................131
Social Self-concept: Change Over Time.....................................................134
Self-determination: Change Over Time......................................................137
Summary of Research Question One..........................................................140
Research Question Two, Three, and Four: The Direct Effects of Student
Attributes on Academic Success in the First Year of College..........................140
Direct Effects for Student Background Characteristics..............................143
Gender...................................................................................................143
Race.......................................................................................................144
First-generation Status ..........................................................................144
High School Grade Point Average........................................................144
viii
SAT Composite.....................................................................................144
Direct Effects for Student Attributes at Time 1..........................................144
Direct Effects for Academic and Social Involvement ................................145
Direct Effects for Student Attributes at Time 2..........................................146
Final Direct Effects Model..........................................................................146
Summary of Direct Effects .........................................................................148
Research Question Two, Three, and Four: The Indirect Effects of Student
Attributes on Academic Success in the First Year of College..........................149
Indirect Effects Predicting Academic Self-concept at Time 2....................149
Indirect Effects Predicting College GPA....................................................153
Summary of Indirect Effects.......................................................................159
Summary?.............................................................................................................159
CHAPTER V: DISCUSSION.......................................................................................161
Introduction.............................................................................................................161
Summary of Findings Related to Research Questions............................................163
Research Question One: Student Attributes?Change Over Time...................163
Change in Self-concept...............................................................................163
Academic Self-concept .........................................................................163
Social Self-concept ...............................................................................165
Change in Self-determination .....................................................................166
Research Question Two: The Impact of Student Attributes on Involvement
in the First Year of College...............................................................................169
First-year Programs.....................................................................................170
Faculty Interaction ......................................................................................171
Academic Engagement ...............................................................................171
Homework Time-on-task............................................................................172
Participation in Student Clubs and Organizations ......................................173
Sense of Belonging .....................................................................................173
Research Question Three: The Impact of Involvement ....................................174
Research Question Four: The Direct and Indirect Effects of Student
Attributes on Academic Success.......................................................................174
Direct Effects on Academic Success ..........................................................174
Indirect Effects on Academic Success........................................................176
Student Background Characteristics.....................................................176
Student Attributes at Time 1.................................................................178
Academic and Social Involvement .......................................................178
Total Effects on Academic Success............................................................180
Conclusions.............................................................................................................180
Implications.............................................................................................................184
Contributions to Practice and Policy.................................................................184
Prospective Student Education ...................................................................185
Admission Practices....................................................................................186
Setting Expectations for Involvement.........................................................187
First-generation Students ............................................................................188
ix
Policy Recommendation: Rethinking Approaches to Accountability ........189
Recommendations for Future Research............................................................190
Final Thoughts ........................................................................................................195
APPENDIX A. Definitions of Variables in the Study with Associated
Theoretical Concepts ....................................................................................................197
APPENDIX B. 2004 Cooperative Institutional Research Program Student
Information Form, Higher Education Research Institute..............................................202
APPENDIX C. Student Readiness Assessment............................................................207
APPENDIX D. 2005 Your First College Year, Higher Education Research
Institute??? .............................................................................................................210
APPENDIX B. 2004 Cooperative Institutional Research Program Student
Information Form, Higher Education Research Institute, Components of the
Factor Composite: Self-concept....................................................................................215
REFERENCES .............................................................................................................217
x
LIST OF TABLES
Table 1. Components of the Factor Composite: Academic Self-concept
at Time 1??. ..............................................................................................................107
Table 2. Components of the Factor Composite: Social Self-concept at Time 1...........107
Table 3. Components of the Factor Composite: Self-determination at Time 1............109
Table 4. Components of the Factor Composite: Faculty Interaction............................111
Table 5. Recoded Components of the Factor Composite: Academic Engagement ......112
Table 6. Components of the Factor Composite: Sense of Belonging...........................113
Table 7. Components of the Factor Composite: Academic Self-concept
at Time 2??. ..............................................................................................................114
Table 8. Components of the Factor Composite: Social Self-concept at Time 2...........114
Table 9. Components of the Factor Composite: Self-determination at Time 2............116
Table 10. Bivariate Correlations...................................................................................120
Table 11. Distribution of Analytic (N = 202) and Population (N = 750) Samples
For Selected Characteristics of Magis University 2004-05 First-year Students...........124
Table 12. Number and Percentage of Cases from the Analytic Sample (N = 202)
that Are Missing for Each Variable in the Analyses.....................................................125
Table 13. Distribution of Complete Cases (N = 176) and Cases with Missing
Data (N = 26) for Selected Characteristics of Magis University 2004-05
First-year Students ........................................................................................................127
Table 14. Paired Samples Comparison between Pre- and Posttest Unstandardized
Mean Scores of Computed Scales for Student Attributes (N = 202)............................131
Table 15. Coefficients for Student Background Characteristics, Prior Student
Attributes, and First-year Experience Predicting Academic Self-concept at Time 2...133
Table 16. Coefficients for Student Background Characteristics, Prior Student
Attributes, and First-year Experience Predicting Social Self-concept at Time 2??..136
xi
Table 17. Coefficients for Student Background Characteristics, Prior Student
Attributes, and First-year Experience Predicting Self-determination at Time 2? ......138
Table 18. Direct Effect Coefficients for Path Analysis Predicting College GPA
and Academic Self-concept at Time 2..........................................................................141
Table 19. Statistically Significant Indirect Effects Predicting Academic
Self-concept at Time 2 for Path Analysis .....................................................................151
Table 20. Statistically Significant Direct, Indirect, and Total Effects
for Independent Variables in the Path Analysis Predicting Academic
Self-concept at Time 2 for Path Analysis .....................................................................153
Table 21. Statistically Significant Indirect Effects Predicting College GPA
for Path Analysis..........................................................................................................155
Table 22. Statistically SignificantDirect, Indirect, and Total Effects
for Independent Variables in the Path Analysis Predicting College GPA ...................158
xii
LIST OF FIGURES
Figure 1. Revised Conceptual Model: Student Attributes and Academic Success
in the First Year of College.............................................................................................82
Figure 2. Direct Effects for Independent Variables: Student Attributes and
Academic Success in the First Year..............................................................................147
1
CHAPTER I
INTRODUCTION
For over 30 years, higher education researchers have measured the effects of
college attendance on students? intellectual, social and psychological development.
Spurred by increasing scrutiny from state and federal lawmakers, college impact studies
respond to calls for institutional accountability (Pace, 1984; Terenzini, 1994).
Specifically, on-going debates over effective spending of public funds have compelled
leading theorists (Astin, Kuh, Pascarella, and Terenzini, for example) to explain not only
how students benefit from their academic and co-curricular experiences, but also how
colleges and universities are uniquely poised to deliver these benefits. As a result, many
higher education studies focus on the impact of specific programs, facilities, or other
institutionally controlled environmental factors on educational outcomes, such as
retention, achievement, and psychological or cognitive development.
Common Assumptions in College Impact Studies
The most cited model for studying college impact is Astin?s (1993, 1991) Input-
Environment-Outcome (I-E-O) model, which measures the effects of the college
environment on educational outcomes while controlling for students? background
characteristics at the time of entry. The I-E-O model assists the higher education
community in understanding how particular interventions, such as academic advising or
residence hall communities, influence outcomes, such as grades, satisfaction, or retention.
Astin (1993) asserted that the best institutional metaphor for understanding higher
education is the hospital. In this schema, Astin likened college students to patients
2
(input) who undergo some form of medical treatment (environment) in order to overcome
illness (outcome). Similarly, he argued, students enter their college or university of
choice, participate in the educational environment supplied by the institution, and
graduate four years later better than they were when they entered. Although Astin
acknowledged the major difference between college students and hospital patients,
namely college students are not typically ill, he missed an obvious hole in his metaphor.
In Astin?s analogy, patients (the student substitute) are passive recipients of treatment.
Students, unlike patients whose treatments are at the mercy of medical staff, have a
particular influence over their experience such that the extent of their involvement in the
academic and social environments is a function of their own determination. In other
words, as autonomous beings, college students have a certain authorship over their own
intellectual and social experiences?they decide if, when, and how to engage their
educational environment. In fact, Astin?s (1984) own theory of student involvement
acknowledges that the ?amount of physical and psychological energy that the student
devotes to the academic experience? (p. 297) is crucial to understanding how the college
environment affects student development.
Astin?s hospital metaphor for higher education focuses on the institution as
owning primary responsibility for the educational process, which is congruent with the
public?s expectations. Sentiments of students, their parents, and the public-at-large
correspond with Astin?s (1985) assertion that the burden of the educational process falls
on the institution. In fact, an entire cottage industry of ranking colleges and universities
exists based on the belief that institutional resources are the ultimate predictors of
educational quality. Entities, the most prominent of which is U.S. News & World Report,
3
have carved a niche for themselves by proclaiming to measure institutional excellence.
By using institutional resources as proxy measures for educational quality, these rankings
purport to list colleges and universities from best to worst in terms of the education they
deliver to students.
Although these ranking systems claim to determine institutional quality, Astin
(1985) has opposed their resource-dependent approach. Instead, Astin has supported
talent development as a more accurate assessment of educational excellence. According
to Astin, the true measure of quality or excellence lies in the institution?s ability to enrich
students? intellectual development, such that students demonstrate cognitive growth and
improvement over the course of their undergraduate program. Truthfully, every
institution is beholden to make a good-faith effort toward providing the most stimulating
environment possible. However, the conversation about excellence in higher education
and even Astin?s influential talent development overlook an influential variable in the
equation: the attributes students bring to bear over their own collegiate experiences.
Accounting for Attributes as a Factor of Student Success
Measuring environmental factors controlled, manipulated, or planned by the
institution does not account for differences in personal authorship on the student?s part.
Intuitively, participants in higher education believe educational outcomes are directly
related to the effort students expend toward their own college experience. Nevertheless,
college impact studies often cite inability to account for students? personal responsibility
as a serious limitation to our understanding of how students achieve particular outcomes
(Hernandez, Hogan, Hathaway, & Lovell, 1999). The result of this limitation is the
4
unintended impression that higher education is either a process institutions perform on
students or even a brand-name commodity available for student purchase.
In their study of effective educational practices, Kuh, Kinzie, Schuh, and Whitt
(2005) concluded that the best institutions ?induce students to assume responsibility for
their own learning? (p. 167). Furthermore, Kuh et al. suggested that the institution and
the student share responsibility for engagement. In terms of shared responsibility for
students? academic success, the institution invests human and financial resources whereas
the student invests time and energy (Upcraft, Gardner, & Barefoot, 2005). Finally,
experience indicates involvement is an interactive arrangement that depends, at least in
part, on the student and how he or she chooses to approach the collegiate experience.
Consider the following advice Brian Mara?a, a member of the baccalaureate Class
of 2004, gave to first-year students entering his alma mater in fall 2004. In an opinion he
wrote for the summer orientation edition of his college?s student newspaper, Mara?a
(2004) counseled new students:
I have three main pieces of advice for all of you: take charge, get involved and
challenge yourself. I managed to accomplish more than I could have imagined
here. I took some really interesting classes, I organized some pretty big events
and I made some incredible friendships. I attribute a lot of that to the fact that for
the most part, I didn?t sit around waiting for things to happen.
I took charge of my academic life by knowing what courses were required,
and I sought advice when I didn?t know what to do.?Take charge of your
extracurricular life by getting involved. Being involved allowed me to be more
than just a student?it allowed me to explore who I am as a person.? Find what
5
you like and dislike, and as you progress through college, you?ll really be able to
focus on those things about which you are passionate. However, in order to find
your passions, you need to challenge yourself. Stretch your mind.?Get out of
your comfort zone.
I might summarize everything I?ve been trying to say by urging you not to
take your time here for granted. You are entering a time of unparalleled freedom
and opportunity. Don?t throw it away. Use the freedom you have to take
advantage of all the opportunities [this institution] provides to grow intellectually,
socially, emotionally, physically, and spiritually. Take charge. Get involved.
Challenge yourself. (p. 5)
In a few short paragraphs, Mara?a?s wisdom echos the conclusion Pascarella and
Terenzini (2005) drew upon completing their meta-analysis of a decade?s worth of
college impact research: Student involvement promotes growth.
Clearly previous research (Berger & Milem, 1999; Hernandez et al., 1999; Huang
& Chang, 2004; Milem & Berger, 1997) indicated that the institutional environment does
influence intellectual and social development, but the role each student plays in shaping
his or her own educational outcomes is less understood. Stakeholders, such as
policymakers, tuition-paying students and families, employers, and accreditation
agencies, have vested interests in ensuring the efficacy of a college or university?s
educational program. However, using performance-based assessment to judge college
impact may not produce an accurate account of institutional effectiveness. It is
necessary, then, to consider and control for student attributes influencing involvement.
Unless these measures are included in outcomes assessment, the likelihood of achieving a
6
true understanding of institutional impact seems slim (Stage, 1989). Holding institutions
accountable for their educational processes makes sense only if outcomes assessment
controls for, or simultaneously measures, the student?s role in becoming an active learner.
Engagement in the educational enterprise is a behavior that can be expressed as a
function of the interaction between person and environment (Lewin, 1936). Research
focusing on the student?s role in the educational process raises awareness that education
is not a commodity for purchase; rather it is a partnership between student and institution
for which both parties have responsibility. A shift in research focus toward the student?s
responsibility for learning begins to hold students accountable for their educational
investment (Pace, 1984). Therefore, the purpose of this research was two-fold. First, this
study was designed to explore and understand how student attributes and the institutional
environment affect one another in a causal model and how each contributes to academic
success, as measured by cumulative grade point average. Second, this study sought to
understand how student attributes change over time and how the institutional
environment affects that change.
Theoretical Frameworks:
Student Attributes and Involvement
Through decades of research, much is known about the effects of college
attendance on students? cognitive, behavioral and psychological growth (Pascarella &
Terenzini, 1991, 2005). Furthermore, in their exhaustive appraisals of college impact
studies, Pascarella and Terenzini (1991, 2005) found very little disparity between
institutional environments in their ability to affect students. Instead, the real differences
7
in outcome measures were found within each institution and were attributable to
variations in individual students? experiences. In addition, Astin?s theory of involvement
(1984) suggested that learning is directly related to college students? investment of
physical and psychological energy. However, very little is known about the effects of
student attributes, or the dimensions of self-concept, self-determination, and self-efficacy
on student involvement, which, as Astin suggested, influences educational outcomes.
Self-determination and Self-efficacy Theories
Generally, the student?s personal influence over outcomes is an acknowledged
missing link in college impact literature (Hernandez, et al., 1999). To rise above this
shortcoming, Stage (1989) recommended adding to college impact models those
psychological dimensions that represent student development over time. The discipline
of psychology offers several interesting concepts for consideration: self-concept, self-
determination (Deci & Ryan,1985; Ryan & Deci, 2000, 2002), and self-efficacy
(Bandura,1977, 1982, 1997).
Self-concept conveys a student?s estimation of his or her own abilities in
comparison to those of the peer group. Self-concept can be used as a global term, or it
can be divided into discrete, domain-specific areas of interest (Waugh, 2001), such as
artistic ability, math/science ability, or leadership ability. For the purposes of this study,
self-concept refers to academic and social self-concept.
Deci and Ryan (1985) have proposed that a person?s ability to act on his or her
own behalf exists on a motivation continuum ranging from a-motivation, through four
levels of extrinsic motivation, and finally reaching intrinsic motivation. According to
8
Deci and Ryan, a person?s place along the motivation continuum depends on his or her
level of self-determination, which in turn is governed by feelings of competence,
autonomy, and relatedness. Examining self-determination contributes to a better
understanding of what responsibility belongs to the student in achieving educational
outcomes.
Self-efficacy theory (Bandura, 1977, 1982, 1997), on the other hand, suggests that
a person?s expectations for success predict achievement. According to Bandura, people
appraise their own abilities based on four environmental cues: prior performance,
observation of others, verbal persuasion from a trusted other, and their emotional state at
the time of appraisal (i.e., the emotional state affects one?s confidence in his or her
abilities). Together with measures of self-concept and self-determination, self-efficacy
contributes to our understanding of a student?s ability to interact with his or her collegiate
environment and to experience academic success.
Definition of Key Terms
This study relies on concepts and terms that may be unfamiliar to the reader. The
following are definitions of key terms that appear throughout the study.
Academic Self-concept: Students? self-evaluative ratings of their ability in several
academic domains comprise ?academic self-concept.? These ratings are relative to
students? perceptions of peer ability.
9
Academic Success: This term refers to the outcome measure for this study. While
numerous valid constructs exist for academic success in the first year (Upcraft, Gardner,
& Barefoot, 2005), this study focuses on cumulative grade point average at the end of the
first year.
Effort: In his work with the College Student Experiences Questionnaire (CSEQ),
Pace (1984) defined student effort as a quality measure detailing ?how students use the
major resources and opportunities for learning and personal growth that are provided by
the college for that purpose? (p. 10). For the purposes of this study, student effort refers
to time-on-task engagement in the academic environment as well as quality of intellectual
and social involvement.
Involvement/Engagement: Student involvement ?refers to the amount of physical
and psychological energy that the student devotes to the academic experience? (Astin,
1984, p. 518). During the undergraduate experience, the physical and psychological
energy a student expends affects his or her ability to engage in the academic or
intellectual atmosphere, to develop relationships with faculty and peers, and to engage in
the co-curricular life of the institution (Pace, 1984). For the purposes of this study, the
terms ?involvement? and ?engagement? are used interchangeably.
Self-determination: Germane to this study, self-determination refers explicitly to
the theory posited by Deci and Ryan (1985). Specifically, self-determination describes a
person?s ability to act of his or her own volition and intrinsic motivation. Self-
10
determination is related to and encompasses a person?s feelings of autonomy,
competence, and relatedness.
Self-efficacy: Taken from the psychology literature, self-efficacy is one?s self
appraisal of personal mastery and the ability to perform given cognitive and behavioral
functions (Bandura, 1977, 1982).
Student Attributes: ?Student attributes? describes those psychological dimensions
that influence the extent to which a student takes responsibility for how he or she engages
the collegiate experience, academically and socially. For the purposes of this study, the
concepts of self-determination, self-efficacy, and self-concept comprise student
attributes.
Social Self-concept: Similar to academic self-concept, social self-concept refers
to students? self-ratings of their interpersonal skills and abilities, relative to those of their
peers.
Purpose and Research Questions
This study describes higher education as a partnership between student and
institution for which student and institution have responsibility. Therefore, the purposes
of this study are two-fold: (a) to explore how initial student attributes shape student
engagement with the institutional environment and how the institutional environment
influences subsequent student attributes; and (b) to understand how student attributes and
institutional environment contribute to academic success, as measured by grade point
11
average at the end of the first college year. Using path analysis to measure the direct and
indirect effects of student attributes, this study proposes a causal model of academic
success. The following research questions guide the study.
1. What changes in student attributes occur during the first year of college,
and what environmental factors influence these changes?
2. Controlling for student background characteristics, how do initial student
attributes influence academic and social involvement in the first year of
college?
3. Controlling for student background characteristics and student attributes at
Time 1, how do academic and social involvement impact subsequent
student attributes and academic success?
4. After controlling for student background characteristics and academic and
social involvement, what are the direct and indirect effects of student
attributes, as measured by self-determination, self-efficacy, and self-
concept, on the academic success of first-year college students?
5. Which student attributes construct (i.e., self-efficacy or self-determination)
is a better predictor of academic success in the first year?
Research Design
This research utilized longitudinal data for first-year students at a Jesuit-Catholic,
comprehensive university in the mid-Atlantic region. The pretest included data from the
institution?s participation in the Cooperative Institutional Research Program (CIRP) and
responses to an adaptation of Ryan and Deci?s (2002) Basic Psychological Needs Scale
12
(BPNS) to measure self-determination. The institution administered these surveys
simultaneously in September, 2004. The posttest consists of institutional data gathered
simultaneously from Your First College Year (YFCY), the one-year follow-up to CIRP,
and a second administration of the BPNS in April, 2005. Student responses to all four
surveys were matched to create the final longitudinal panel.
Three major advantages suggested situating this study in Jesuit higher education.
First, different environmental factors such as size, location, and control can lead to
variations in student behavior (Baird, 1988; Barker, 1968; Upcraft & Gardner, 1999). By
composing a single-institution study, environmental factors are suppressed so as to place
greater emphasis on the contribution of student attributes toward academic success.
Second, Jesuit higher education has foundational principles that are congruent with Kuh?s
(1991) definition of ?involving colleges.? As this research focuses on the effects of
student attributes on educational outcomes, necessity dictates locating an institution of
the type already acknowledged as having an environment that is effective at promoting
student engagement. Third, the target institution supports 15 of the 20 initiatives and
programs that contribute to excellence in the first year of college (Barefoot, Gardner,
Cutright, Morris, Schroeder, Schwartz, Siegel, & Swing, 2005).
Limitations
Although the design has strengths, it also presents limitations. First, the
particularized locale for the study may not produce widely generalizable results. Single-
institution studies limit generalizability to students attending similar institutions
(Creswell, 2003). Therefore I do not expect the results of this study to represent college
13
students nationwide. In addition, this research examined traditional-aged, first-year
students, so results are further limited to exclude adult students and students in their
sophomore through senior years. Furthermore, participating students at the target
institution did not reflect the racial composition of college students nationwide (at this
particular institution, the Class of 2008 was 87% White), so the ability to generalize
results to first-year students of color at similar institutions is limited at best.
Second, the data gathered in this study are self-reported by students. The
limitations associated with self-reported data, including the possibility of inaccurate self-
assessment, are well documented (Schwarz, 1999). Third, the effects of student attributes
or changes in student attributes may not be apparent within the first-year of college, the
period of time covered by this study (Feldman & Newcomb, 1994; Pascarella &
Terenzini, 1991; Terenzini, 1994). Finally, although path analysis will describe direct
and indirect effects of student attributes within the proposed causal model, path analysis
cannot confirm whether or not the proposed model describing causation is correct (Klem,
1995). Rather, path analysis indicated whether or not the proposed model fit the data
(Klem).
Despite these limitations, this study was worth undertaking because student
attributes are essential components to understanding the process of achieving academic
success, as marked by cumulative grade point average at the end of the first year.
Significance of Study
Although this study can not answer all questions, it has immediate implications
for theory and practice. With regard to theory, the research questions guiding this study
14
are an important first step in explaining the student-environment interaction. This
research contributes to our theoretical and conceptual understanding of higher education
by providing a more complete model for outcomes assessment. Astin (1991) designed
the I-E-O model specifically to investigate the role of educational environments in
producing educational outcomes?but almost to the exclusion of understanding the
student?s responsibility to engage in his or her own educational process. The conceptual
model for this study is an I-E-O model because it accounts for student background
characteristics and environmental interventions. However, guided by Lewin?s (1936)
work to understand behavior as a function of the interaction between a person and his or
her environment, the model adds student attributes as contributing factors in influencing
student learning and development. Furthermore, the literature suggested that interaction
with the environment impacts the person as well (Barker, 1968; Ryan & Deci, 2000).
Given these observations, the conceptual model predicts student attributes to have direct
effects on involvement and indirect effects on academic success, as mediated through the
environment. Moreover, the model predicts academic and social involvement to have
direct effects on student attributes at Time 2 and indirect effects on academic success.
Therefore, the conceptual model for this study represents student attributes twice,
first as a measure of students? entry characteristics, which influence how students interact
with the environment, and second as an intermediate outcome, indicating that
participation in the educational environment may change student attributes over time.
These modifications represent a slight departure from Astin?s original I-E-O, such that
measuring and evaluating the role of student attributes in educational achievement is
tantamount to measuring and evaluating the impact of the environment.
15
This contribution to theory highlights the student?s responsibility to engage the
learning environment and presents a more balanced set of expectations for practitioners
and policymakers to consider. Furthermore, research that explains the student?s role in
academic success will aid current and prospective students and their families to adjust
their approach to higher education. Rather than students and parents relying solely on the
institution to provide an excellent product, they can use this research to understand better
the responsibility for educational excellence borne by the student.
With regard to contributions to practice, this study holds particular significance
for the home university and other institutions like it. Results from this study suggest
improvements for delivery of first-year programs, indicate sub-populations in need of
additional assistance, and recommend educational efforts to inform students of
expectations for engagement. Furthermore, results indicate markers of student attributes
that predict academic success. In this case, there are significant implications for how the
target institution and similar others evaluate prospective students for admission. Rather
than relying so heavily on high school GPA and entrance exam scores, institutions may
consider these indicators relative to the prospective student?s ability to take responsibility
for his or her own learning. Again, shifting the focus to student attributes has
implications for prospective college students and their families with regard to how
students prepare for a college education.
16
CHAPTER II
REVIEW OF RELATED LITERATURE
Introduction
The primary purpose of this study was to understand the contributions of student
attributes toward academic success among first-year college students within a particular
institutional environment. This chapter provides the theoretical framework for the study
and its conceptual model. The following pages review and analyze theoretical and
research literature related to academic success within the context of college impact
studies. This chapter also provides a lengthy discussion of methods for understanding
college outcomes, beginning with an overview of ecological models stemming from
Lewin?s (1936) work on the person-environment interaction. A discussion of Astin?s
(1991) input-environment-outcome (I-E-O) assessment model as the primary model of
college impact (Pascarella & Terenzini, 2005) follows. The I-E-O model is the
foundation for the conceptual model in this study. In addition, the chapter synthesizes the
theoretical and research literature related to student attributes, including self-concept
(Astin, 1993), and, from the field of psychology, self-determination theory (Deci & Ryan,
1985; Ryan & Deci, 2000, 2002) and self-efficacy (Bandura, 1977, 1982, 1997). Astin?s
(1984, 1999) theory of student involvement and Pace?s (1984) work on student effort
round out the literature review. Finally, the chapter closes with a presentation of
variables related to studying academic success in the first year of college.
17
Understanding College Outcomes
The higher education community has been responding to external pressure to
provide evidence that college provides benefits in proportion with public and private
investments (Terenzini, 1994). Over the years, several genres of research approach or
philosophy have emerged: the ecological-environmental approach, the college impact
approach, and the approach examining the process of students? personal and
psychological development. The spectrum of emphasis has waxed and waned between
the three approaches in a manner reflective of societal needs (Baird, 1988; Feldman &
Newcomb, 1994). The research presented here draws upon aspects of all three
approaches, and the following section provides an overview of two of these approaches
and how they shape the current study.
The Ecology of Person-Environment Interaction
Coyne and Clack (1981) defined the environment as having four components: the
physical environment, or that which people perceive through one or more of the five
senses; the social environment, which describes person-to-person behaviors and
relationships or the characteristics of people within the environment; the institutional
environment, which relates to the policies and procedures governing the environment;
and the ecological environment, which refers to the interaction of the three above-listed
components. It follows then, that ecological approaches to studying student outcomes
begin from the perspective that people and their environments have a reciprocal
relationship. First, environments have the ability to shape the behavior of those who
participate within them; second, people and their behavior shape and re-create the
18
environment (Coyne & Clack). Much of the person-environment interaction literature
stems from the work of Lewin (1936), who first suggested that behavior is a function of
the person-environment interaction. Lewin suggested that the behavior of individuals in
particular environments could be represented mathematically. He offered the following
equation to understand human behavior: B = f(P x E), ?where B, the behavior, is the
function (f) of the organism or person (P) interacting with the environment (E)?
(Banning, 1989, p. 54). The different ecological theories endeavor to determine the
appropriate balance between person and environment: Which is more prominent?the
person or the environment?
Notable Ecological Theories
Walsh (1973, 1978) reviewed several ecological theories that have influenced
studies in higher education. These theories differ in their emphases along the person-
environment continuum, such that some theories preference the person over the
environment while the converse is true for others. Three of the major theories are
reviewed briefly below to provide some background.
Barker?s Behavior Setting Theory. From the sociological perspective, Barker
(1968) emphasized the environment over the person. He suggested that the environment
first selects the people who inhabit it and then shapes their behavior regardless of
individual differences. Barker?s theory was based on three assumptions: first, people
generally adhere to the regulations and norms associated with an environment; second,
the structured pattern of the environment dictates human interaction with one another and
with the environment; and third, by measuring the forces of the environment and the
19
behavior of the people, one may be able to understand how the two interact (Walsh,
1978). Furthermore Barker differentiated between ?undermanned? and ?optimally
manned? settings, where undermanned settings increase the involvement of their
inhabitants by having too few people to manage regular operations. Walsh (1978) notes
that people in smaller, or undermanned, settings tend to be more personally productive
than those in optimally manned settings.
Need x Press Culture Theory. Blending sociological and psychological
perspectives, Stern (1970) theorized behavior as a product of the relationship between the
environment (press) and the person (needs). In Stern?s theory, the person represents
needs, or the tendency toward specific behaviors (Walsh, 1978), that bump into the
demands, or presses, of the environment. Need-Press Theory places equal emphasis on
the person and the environment to describe behavior (Walsh).
Holland?s Personality Theory. Leaning much more heavily on psychology,
Holland?s (1973) Personality Theory suggested that behavior results from an interaction
between the environment and the personalities of those who inhabit it. Holland?s six
personality types (as defined by differences in personal attributes) are related to
behavioral differences and, ultimately for Holland, vocational choice. Furthermore,
people select environments that fit or match their personalities, and a good person-
environment fit leads to greater success.
Although the foregoing ecological theories do not play prominently or contribute
substantially to the current study, it is helpful to keep them in mind as a backdrop for the
current study. The ecological approaches do provide a perspective on the balance
20
between person and environment when attempting to describe or predict behavior.
Furthermore, their emphasis on the person-environment interaction was the basis for
suggesting psychological measures as an important addition to the standard college
impact model.
Studying College Impact
Pascarella and Terenzini (1991, 2005) divided studies of college student change
into two families: developmental and college impact. Whereas developmental theories
and models describe student movement through various stages of growth and maturation,
college impact models ?emphasize change associated with the characteristics of the
institutions students attend?or with the experiences students have while enrolled? (2005,
p. 18). This study builds upon college impact theory by adding psychological measures
to Astin?s Input-Environment-Outcome model as the basis for studying student change.
The I-E-O Model: Astin?s Assessment Methodology
The 1980s and 1990s brought increased scrutiny of higher education (Astin,
1991). State and federal legislators called for increased student assessment as a means to
achieve greater accountability for colleges and universities. The call for accountability
suggested a public distrust for the quality of education in the nation?s colleges and
universities. Astin noted, ?public pressures to use more competency testing or outcomes
assessment reflects a concern about how much students are actually learning in our
colleges and universities? (p. 4). Spurred by increasing public demand for accountability,
Astin proposed a methodology for improving assessment in higher education.
21
With his text, Assessment for Excellence, Astin (1991) aimed to make assessment
in higher education more accurate and more practical to stakeholders. According to
Astin, the outcome of true assessment (which includes not only information gathering
and measurement, but also evaluation of that information) should be improvement of the
educational process. More precisely, Astin proffered that ?the basic purpose of assessing
students is to enhance their educational development?[and] advance the educational
mission of our colleges and universities? (p. 4). To this end, Astin introduced and
explained his successful input-environment-outcome (I-E-O) assessment model, so
practitioners and researchers of higher education might pinpoint with greater accuracy
those environmental factors that lead to better educational quality.
The I-E-O Model. Astin (1991) first developed the input-environment-outcome
model in the early 1960s through his work with John L. Holland. The two studied why
certain colleges and universities graduated more students who pursued doctoral work
than others. In their investigation, Astin (1962) found that characteristics of entering
first-year students were far more important predictors of advanced education than any
institutional environmental factor?a finding contrary to previous research. Three
lessons emerged from Astin?s (1991) early work with Holland: (a) to measure educational
impact, the researcher must evaluate outcomes relative to student input measures; (b) the
researcher must consider all input variables that may influence the outcome; and (c) input
and outcome data are more instructive when considered against elements of the
educational environment (pp. 17-18).
These observations led Astin (1991) to create the input-environment-outcome
model, a ?very simple, yet?powerful framework for the design of assessment activities
22
and for dealing with even the most complex and sophisticated issues in assessment and
evaluation? (p. 16). For the purposes of the I-E-O model, input refers to student
characteristics at time of college entry; environment refers to institutional interventions,
including educational programs and student experiences; and outcome refers to student
achievement, development, or growth (Astin, 1991; Pascarella & Terenzini, 1991, 2005).
An asset of the I-E-O model is the ability to measure the relationship between the
institutional environment and the educational outcome, while controlling for differences
in students? personal qualities (Astin, 1991; Pascarella & Terenzini, 1991). In other
words, using the I-E-O model allows the researcher to account for student background
characteristics that otherwise exert influence on outcomes, thereby isolating the
environmental variables of interest in order to measure their educational impact.
Benefits of the I-E-O Model Compared to Other Assessment Strategies. Since its
introduction, the I-E-O model has been influential in college impact studies (Pascarella &
Terenzini, 1991, 2005), perhaps because the I-E-O model is a more complete assessment
tool than many other methodologies (Astin, 1991). According to Astin, there are four
other typical assessment strategies: (a) outcome-only assessments, such as achievement
tests; (b) environment-outcome assessments, such as grade point averages across inter-
collegiate athletic teams; (c) input-outcome assessments, such as pre- and posttest
designs; (d) and environment- or input-only assessments, such as the annual institutional
rankings sponsored by U.S. News & World Report. Seemingly unbeknownst to many
who rely on these assessment strategies, each of these methodologies has a particular
shortcoming rendering it unfit for accurately measuring and evaluating college impact
(Astin).
23
Compared to these assessment methodologies, the I-E-O model is more complete,
allowing for measurement of the impact of environmental interventions given student
skills and traits prior to interaction with the environment (Astin, 1991). Using the I-E-O
model, researchers can investigate the ?comparative effectiveness of different educational
policies and practices? (Astin, p. 37) and be relatively sure that any variance they detect
is due to the intervention and not to differences in students? background characteristics
(Astin). Upcraft, Chrissman Ishler, and Swing (2005) note that the I-E-O model ?is a
useful tool for identifying and estimating the effects of those college experiences over
which institutions have some programmatic or policy control, such as student
experiences, which can be shaped into educational advantage through an institution?s
programmatic or policy concerns,? (p. 497). According to Pascarella and Terenzini
(1991), Astin?s I-E-O formulation has influenced the development of the dominant
models of college impact and student change: Astin?s Theory of Involvement (1985),
Tinto?s Theory of Student Departure (1993), Pascarella?s General Causal Model for
Assessing the Effects of Differential College Environments on Student Learning and
Cognitive Development (1985), and Weidman?s Model of Undergraduate Socialization
(1989).
Relation of the I-E-O Model to the Current Study. As suggested above, Astin?s I-
E-O model provides a superior methodology for measuring and evaluating college
impact. It seems fitting that any researcher whose objective is to understand which
factors lead to or predict certain educational outcomes would build upon the I-E-O
model. As Astin (1991) stated, the focus of I-E-O-based research is on understanding the
impact of educational interventions on outcomes:
24
Input and outcome refer simply to the state of the person at two different time
points, and environment refers to the intervening experiences. We are particularly
interested in learning about environmental experiences that can be controlled or
changed, since it is these experiences that offer the possibility of improving
outcomes in the future. (p. 22)
Accordingly, Astin?s I-E-O model is the foundation for the conceptual model I
present in this investigation of factors contributing to the academic success of first-year
college students. However, my research diverges from Astin?s at a crucial point. Astin
designed the I-E-O model specifically to investigate the role of educational environments
in producing educational outcomes?but almost to the exclusion of understanding the
student?s responsibility to engage in his or her own educational process. In other words,
Astin?s I-E-O model studies the ability of the environment to involve the student while
nearly disregarding the student as an actor with personal accountability for becoming
involved in the academic environment. As Stage (1989) observed, without the addition
of variables that describe students? psychological make-up, college impact models are
insufficient to fully and reliably explain student outcomes.
Although, the conceptual model for this study of academic success in the first-
year is an I-E-O model in that it accounts for student background characteristics and
environmental interventions, it adds student attributes as contributing factors on par with
the environment. Furthermore, the literature suggests that one?s interactions with the
environment impacts behavior (Barker, 1968; Lewin, 1936) and self-determination (Ryan
& Deci, 2000), which ultimately influences the final achievement outcome. Therefore,
the conceptual model proposed in this study suggests that the environment has direct
25
effects on student attributes and indirect effects on the outcome. This modification
represents a slight departure from Astin?s original I-E-O model, such that measuring and
evaluating the role of student attributes in academic success is tantamount to the
measuring and evaluating the impact of the environment.
The First Year of College as a Particular Environment
Since the early 1980s, educators and researchers have turned their attention to the
first-year of college as a critical time for transition that, if done well, leads to persistence
and academic success (Barefoot, Gardner, Cutright, Morris, Schroeder, Schwartz, Siegel,
& Swing, 2005; Upcraft, Gardner, & Barefoot, 2005; Upcraft & Gardner, 1989). Upcraft,
Gardner, Barefoot, and associates suggested that the first year of college requires a
delicate balance of challenge and support in the forms of challenging educational
experiences coupled with effective programs that support the transition. Together these
institutional efforts foster student growth and lead to success in the first year (Upcraft,
Gardner, & Barefoot, 2005). Furthermore, LaSere Erickson and Strommer (2005)
indicated that upon entering college students are typically extrinsically motivated?
whether by grades, a desire to meet family expectation, and/or career aspirations. A
student-centered learning environment in the first-year, they suggested, can inspire
students to focus instead on the academic process and fall in love with learning. Over
time, then, they become more intrinsically motivated toward their college educations.
Elements of a successful first-year environment include interaction with faculty, first-
year seminars, and learning communities?all of which must be specifically designed to
26
meet the transitional needs of first-year college students. The literature surrounding each
element is discussed below.
First-year Seminars
First-year seminars came into vogue in the 1980s and first emerged as University
101 or First-year Experience (FYE) courses (Upcraft & Gardner, 1989); they continue
today in nearly 90 percent of all colleges and universities (Barefoot, Gardner, et al., 2005;
Upcraft, Gardner, & Barefoot, 2005). Typically these FYE and University 101 seminars
are credit-bearing and designed as extended orientations (Gordon, 1989), which cover
standard topics, including time management, study skills, health and wellness, college
survival, conflict management, and values clarification (Gordon). As the name ?seminar?
suggests, these courses are small?about 15 students each (Barefoot, Gardner, et al.,
2005; Gordon)?and give first-year students an opportunity to engage in a more intimate,
discussion-based learning environment (Stuart Hunter & Linder, 2005) heretofore
reserved for upper-division students enrolled in specialized major-related courses. The
small course size facilitates community-building by orchestrating increased student
interaction with the instructors and with one another (Jewler, 1989). Instructors are a
combination of faculty, student development administrators, and upper-division students
(Gordon; Hunter & Linder), and instructor make-up is dependent upon institutional
objectives and course design. FYE courses are linked to academic success measures,
including grade point average, progress toward degree requirements, first- to second-year
retention, and graduation rates (Barefoot, 1993).
27
Along side FYE courses, or perhaps in lieu of them, some institutions offer
academically-focused first-year seminars. Unlike the extended orientation courses, these
seminars fit into the curricular requirements for degree attainment (Gordon, 1989). Like
the FYE course, the first-year academic seminar facilitates student-faculty and peer
interactions and adjustment to university life. In their review of related research, Fidler
and Stuart Hunter (1989) concluded that first-year seminars enhance student success,
including measures such as grade point average, clarification of academic goals, and self-
concept.
Learning Communities
Another programmatic intervention to enhance the first-year experience is the
learning community. Traditionally, learning communities consist of a semester-long
cluster of two or more academic courses that are linked by theme or program area
(Levine Laufgraben, 2005). By enrolling the same cohort of students, learning
communities encourage peer collaboration and inter-disciplinary application of course
material. Learning communities also encourage faculty collaboration, experimental or
interactive pedagogies, and increased faculty-student interaction (Goodsell Love &
Tokuno, 1999). Learning communities may or may not have residential components in
which cohorts are housed together or clustered courses occur within the residence hall
(Levine Laufgraben). In assessing the literature on outcomes related to learning
communites, Levine Laufgraben concluded that enrolling in a learning community has a
significant positive impact on achievement, including the intellectual and social
development of first-year students.
28
Faculty and Peer Interaction
Researchers agree: students interactions with peers and faculty have significant
effects on academic success (Astin, 1993; Banning, 1989; Pascarella & Terenzini, 1991,
2005; Upcraft, 1989). With regard to interaction with faculty, Astin found a positive
correlation with academic self-concept, while Pascarella and Terenzini?s (2005) meta-
analysis revealed that only particular types of faculty-student interactions yielded positive
results. Specifically, Pascarella and Terenzini?s work suggested that faculty-student
interactions are associated with positive gains in academic self-concept when the
interactions are substantively related to course work, but that these gains are eliminated in
models that control for students? perceptions of other aspects of the institutional
environment. Thus, Pascarella and Terenzini concluded that faculty-student interaction is
embedded in students? perceptions of the environment. Social contact with faculty
outside the classroom is positively correlated with other measures of student success,
such as persistence and institutional commitment (Pascarella & Terenzini).
Specific peer interactions, including those within college residences and student
clubs and organizations, have an effect over student development in the first year
(Upcraft, 1989). The segregated and communal nature (many students of a particular age
group living in close proximity to one another) of student residence halls creates a
particular environmental press on students, such that residents have enormous influence
over one another?s perceptions and behaviors (Upcraft). Upcraft warned, however, of the
potential for first-year residences to have a negative influence on academic and social
engagement. Participation in student clubs and organizations, on the other hand, has been
related to gains in social self-concept (Upcraft). These activities increase the potential for
29
students to develop friendships and other meaningful peer associations as well as to build
upon their social and leadership skills.
Measuring Academic Success in the First Year:
Grade Point Average
In measuring students? academic success, educators often rely on indicators such
as grade point average and persistence through the educational system (Upcraft, Gardner,
& Barefoot, 2005). Upcraft, Gardner, and Barefoot urge researchers to consider an
expanded list of indicators when exploring academic success in the first year of college.
Beyond successful course completion and continued enrollment, Upcraft et al. suggest
growth in the following areas as potential measures of first-year student success:
intellectual and academic competence, interpersonal relationships, identity development,
career choice, health and wellness, spirituality, multicultural awareness, and civic
responsibility. For the purposes of this study, the determinant of academic success in the
first year is cumulative grade point average, which is discussed below.
As stated above, grade point average (GPA) indicates successful course
completion, one of the measures often acknowledged as a standard of academic success
(Astin, 1993; Upcraft, Gardner, & Barefoot, 2005). Although an imperfect measure of
intellectual development due to inconsistent calculation, college grades are strong
predictors of success as measured by persistence, degree attainment, and enrollment in
graduate school (Pascarella & Terenzini, 2005). Whether required for entrance into
academic honor societies, student leadership positions, or deans? lists, grade point
average is a common criterion indicating academic achievement. Not surprisingly, many
30
researchers in higher education have dedicated effort to understanding which individual
and environmental factors predict GPA (Pascarella & Terenzini, 1991, 2005). While
studies consistently demonstrate that high school GPA and college entrance exam scores
are strong predictors of college grades (Astin, 1993), other studies have indicated that
grade point average is correlated with prior self-concept (Astin, 1993; Cokley,
Komarraju, King, Cunningham, & Muhammad, 2003; House, 2000; Marsh, 2003; Yeung,
McInerney, Russell-Bowie, Suliman, Chui, & Lau, 2000).
Student Attributes
Colleges and universities are accountable for how they use their financial and
human resources to provide academic and co-curricular programs and facilities that
promote student learning (Pace, 1984). Lest the academic community and the public
mistake this accountability for a one-way responsibility to provide a service to customer-
students, Pace insisted that students be accountable for the effort they put toward their
own learning. These sentiments are echoed by those who call attention to students?
responsibility for the quality of their own education (Kuh, Gonyea, & Williams, 2005;
Kuh, Kinzie, Schuh, & Whitt, 2005). In other words, student authorship is integral in
measuring how students achieve educational outcomes. Nonetheless, college impact
studies rarely account for students? psychological attributes in their models and
measurements. Instead, researchers have recommended examining the influences of
environmental factors to predict success-related outcomes (Upcraft, 2005). However
accounting for their personal attributes may help differentiate why some students achieve
31
particular outcomes at greater rates than their peers. The following section reviews the
literature pertaining to student self-concept, self-determination, and self-efficacy.
Self-concept
Pascarella and Terenzini?s (2005) meta-analysis of college impact research
revealed that self-concept, although a loosely defined term, generally refers to students?
self perceptions of competence relative to those of their peers. Self-concept, as it appears
in the literature, may be designated as an input measure?prior self-concept?an
outcome, or both. Although most studies use data from the Cooperative Institutional
Research Program (CIRP) to study change in self-concept from first to senior year, or
even post-baccalaureate years (Kezar & Moriarty, 2000; Berger & Milem, 2000;
Pascarella, Smart, Ethington, & Nettles, 1987), Terenzini, Theophilides, and Lorang
(1984) argued for more research devoted to year-to-year growth. The results of Terenzini
et al.?s study indicated distinct change in self-concept occurs each year of college beyond
the cumulative growth over four or more years.
Research also has demonstrated consistently that student involvement is a factor
influencing change in self-concept (Berger & Milem, 2000; House, 2000; Kezar &
Moriarty, 2000; Pascarella, Smart, Ethington, & Nettles, 1987). However, studies
examining gender or racial differences in self-concept have produced mixed results: some
demonstrated significant differences based on race (Astin, 1993; Berger & Milem, 2000;
Cokley, Komarraju, King, Cunningham, & Muhammad, 2003; Kezar & Moriarty, 2000;
Marsh & Yeung, 1998) or gender (Kezar & Moriarty, 2000; Marsh & Yeung, 1998),
32
while another study did not reveal any appreciable difference based on race or gender
(Pascarella, Smart, Ethington, & Nettles, 1987).
Additionally self-concept can take several forms, depending on the focus of study.
Whereas some researchers may study students? general self-concept, which refers to their
over all self perception (Graham & Cockriel, 1996; Waugh, 2001), others may examine
particular domains of self-concept, such as academic or social (Berger & Milem, 2000;
Cokley, Komarraju, King, Cunningham, & Muhammad, 2003; Kezar & Moriarty, 2000;
Pascarella, Smart, Ethington, & Nettles, 1987; Terenzini, Theophilides, & Lorang, 1984;
Yeung, McInerney, Russell-Bowie, Suliman, Chui, & Lau, 2000). As academic self-
concept and social self-concept are treated separately in this study, the following sections
review each construct individually.
Academic Self-concept
In urging higher education researchers and practitioners to include students?
intellectual and academic competence in a definition of first-year student success,
Upcraft, Gardner, and Barefoot (2005) concluded: ?Successful first-year students must
not only get off to a good start academically and learn how to learn, but they must begin
to appreciate what it means to become an educated person? (p. 8, emphasis included). In
many ways, academic self-concept captures aspects of this construct of academic and
intellectual competence. Students? self-perceptions of their academic abilities and
intellectual self-confidence as compared to the abilities of other students ?become more
positive during their college years? (Pascarella & Terenzini, 2005, p. 219). In addition,
Marsh (2003) noted that academic self-concept and achievement reinforce one another,
33
such that prior academic self-concept impacts academic performance, which then affects
subsequent self-concept. This finding provides further evidence that academic self-
concept is central to understanding student success.
The body of academic self-concept literature provides evidence that academic
self-concept is hierarchical, such that students may have domain-specific self-concepts
that are empirically different from perceptions of their overall academic competence
(Waugh, 2001; Yeung, McInerney, Russell-Bowie, Suliman, Chui, & Lau, 2000). Yeung
et al. (2000) found reasonably strong support for a hierarchy of academic self-concept,
such that students maintain a global academic self-concept that is separate from their self-
concepts in language, business, math, and the arts. In addition, Waugh?s (2001) study
revealed that global academic self-concept is comprised of separate scales for students?
perceptions of capability, achievement, and confidence. For the purpose of this study,
academic self-concept is defined globally in relation to students? self-ratings of capability
and confidence.
Social Self-concept
In addition to students? intellectual and academic competence, Upcraft, Gardner,
and Barefoot (2005) suggested that the definition of student success includes establishing
and maintaining interpersonal relationships. Upcraft et al. pointed out: ?All first-year
students must develop the interpersonal skills necessary not only to build supportive
relationships, but also to succeed in their many pursuits after college? (p. 8). Although
objective growth in interpersonal relationships may be difficult to measure empirically,
Pascarella and Terenzini (2005) offered social self-concept as an acceptable substitute.
34
Social self-concept pertains to students? self-ratings of their social skills, including their
ability to develop and maintain same-sex friendships, opposite sex-friendships, self-
confidence, and leadership ability (Pascarella & Terenzini). In their review of social self-
concept studies, Pascarella and Terenzini note that social self-concept may suffer a
decline during a student?s transition to college but usually rebounds in the later years.
Given that the transition to college may take a toll on students? social self-perceptions,
this study will examine psychological and environmental factors that may contribute,
either to this decline or to maintenance and even growth in student?s social self-concept.
Related Research
In an effort to understand how college shapes students? social and personal
development, Graham and Cockriel (1996) studied a sample of 9,348 undergraduates
representing 75 colleges and universities. Using exploratory factor analysis, Graham and
Cockriel assessed data from the College Outcomes Survey sponsored by the American
College Testing Program to determine constructs of personal and social growth during
the undergraduate years. The data demonstrated that students experienced the greatest
amount of personal growth in the areas of academic competence, taking personal
responsibility, general knowledge, goal-setting, intellectual curiosity, and self-
confidence. In addition, student responses indicated that the college environment
contributed to growth in these areas as well as in perseverance and in a willingness to
change. Graham and Cockriel?s study revealed that students experienced less growth in
spirituality, civic responsibility, and understanding others. Likewise, students rated as
35
low the college environment?s contribution to these areas as well as to areas such as
family responsibility, financial management, and participation in political elections.
In a related study, House (2000) investigated the effect of student involvement on
academic self-concept. Analyzing CIRP data for 2,134 first-year students at one
institution, House examined the relationship between students? academic self-concept and
their involvement in academic activities during the previous year. Conducting cross-
tabulations, House demonstrated that students? academic self-concept was significantly,
positively related to the number of hours students spent studying, talking with teachers
outside of class, volunteering, and being active in student clubs and organizations. While
House?s involvement construct consists of students? high school activity, this study lends
credence to involvement?s relationship to subsequent self-concept.
Pascarella, Smart, Ethington, and Nettles (1987) presented a causal model to
understand the development of students? academic and social self-concepts. In addition,
they examined racial and gender differences in those factors influencing self-concept.
Pascarella et al. analyzed data from 4,597 men and women who participated in the 1971
CIRP survey as first-year students and then in the 1980 follow-up. Regression analyses
in the Pascarella et al. study revealed that pre-college self-concept influenced college
behavior and that involvement in college affected subsequent self-concept. In addition,
the research team noted that students who attended small colleges reported higher self-
concepts than their peers who graduated from larger institutions. Pascarella et al.
surmised that students perceive smaller colleges and universities to be more
psychologically manageable, thus resulting in a more positive correlation to self-concept
development. Surprisingly, the results of this study demonstrated no appreciable racial or
36
gender difference in the development of academic or social self-concepts. Pascarella et
al. acknowledged that the data for college involvement relied heavily on long-term
memory and that the effects of racial or gender differences in the collegiate experience
may have been diminished by post-baccalaureate experiences.
In a different study, Cokley et al. (2003) examined racial/ethnic differences in
academic self-concept among 396 African American and 291 White students at multiple
predominately White institutions (PWIs) and historically Black colleges and universities
(HBCUs). Students in the study completed the academic self-concept scale along with a
demographic information survey. Cokley et al. utilized principal component factor
analyses to determine whether or not racial differences existed in academic self-concept
constructs. Results demonstrated significant differences in academic self-concept
between African American students and White students. While White students believed
that effort leads to high grade point average and associated academic performance with
ability, Black students related academic performance to prior academic preparation.
Cokley et al.?s study indicates a need to test separate constructs of academic self-concept
for different racial/ethnic groups.
Terenzini, Theophilides, and Lorang (1984) employed longitudinal data to
uncover the effects of college on student?s academic development as measured by
academic self-concept. Using results from the 1978 CIRP survey as baseline data, the
research team followed students from their first year through their junior year in college.
After the third follow-up, Terenzini et al. collected longitudinal panel data for an analytic
sample of 250 students at one institution. Holding constant students? entry
characteristics, Terenzini et al. determined the collegiate experience had a significant,
37
positive impact on students? perceptions of their academic development. Furthermore,
the research team noted that significant effects of college attendance on academic growth
are not only cumulative, but also year-to-year. In other words, distinct and measurable
academic growth occurred during each year of college attendance. Results indicated
students? high school achievement and involvement were the best predictors of perceived
academic growth in the first year. In subsequent years, academic self-concept emerged
from previous growth as well as academic and social involvement. The Terenzini et al.
study did not examine psychological variables that may impact self-concept or
involvement.
Self-determination
Broadly speaking, self-determination stems from the psychology of human
agency, which acknowledges the self as an important actor within the environment. The
person is seen as an actor who imposes upon, reacts to, and cooperates with his or her
surroundings in such a way as to impact the outcome of interactions with others or with
environmental systems (Little, Hawley, Henrich, & Marsland, 2002). Little and
colleagues defined the ?agentic self? as one who ?is the origin of his or her actions, has
high aspirations, perseveres in the face of obstacles, sees more and varied options for
action, learns from failures, and, overall, has a greater sense of well-being? (p. 390). In
other words, the psychological use of ?human agency? refers to active self-regulation and
decision making as opposed to submissive passivity.
Little et al.?s depiction of the agentic self as one who learns from mistakes is
congruent with Emirbayer and Mische?s (1998) theory of human agency as ?temporally
38
constructed engagement? (p. 970), such that the agentic self accounts for the past, present
and future in his or her actions. In the discipline of psychology, then, personal agency,
then, contains the notion that a person?s past experiences can influence his/her present
actions; likewise, a person can regulate present actions to be in accord with future plans.
Agency suggests an integral relationship with one?s surroundings and the people, places
and events found there (Emirbayer & Mische).
With this information in mind, the student begins to emerge as a key factor in
outcomes related to him or her. In their theoretical piece, Little et al. (2002) contended
that individuals with high levels of personal authorship (i.e., those who recognize
themselves as important actors rather than passive recipients) have a greater tendency
toward goal accomplishment and well-being. Conversely, those who have low levels of
personal motivation to be an actor in their own lives are more likely to avoid pursuing
their goals, have lower personal standards, and report lower levels of satisfaction (Little
et al.).
In an effort to clarify and integrate the psychological approach to human agency,
Little et al. (2002) reviewed three theories that help to explain the person-as-actor
throughout the life span: Deci and Ryan?s (1985; Ryan & Deci, 2000, 2002) self-
determination theory (SDT), resource-control theory (Hawley, 1999), and action-control
theory (Little, 1998). Little et al. suggested that Deci and Ryan?s self-determination
theory (SDT) provides an important framework for understanding personal locus of
control in late adolescence and adulthood. Specifically, Little et al. identified the basic
psychological needs that under gird SDT, namely autonomy, relatedness and competence,
as central to the author-self. Moreover, the authors indicated that the strength of SDT lies
39
in its identification of these basic psychological needs as common to all people,
regardless of race, culture, age, or gender. Deci and Ryan?s theory is discussed in detail
below.
Self-determination Theory
Deci and Ryan (1985; Ryan & Deci, 2000, 2002) have posited that human beings
have an innate desire for integration; that is, humans need and want to be authentic and
experience themselves as their own locus of control. Accordingly, Deci and Ryan have
suggested people have a need to be free from control and to experience their own human
authorship, a synonym for their own self-determination. Above all, self-determination
theory (SDT; Deci & Ryan; Ryan & Deci, 2000, 2002) is a theory about people?s
motivation to act on their own behalf. In a sense, SDT is a theory about why people do
the things they do (Deci, 1995).
As stated by SDT, motivation exists on a continuum, from amotivation, or
complete passivity, through stages of extrinsic motivation, and finally to intrinsic
motivation. Ryan and Deci (2002) theorize that intrinsic motivation, or the urge to act
out of sincere interest or inherent satisfaction, is the condition of being fully self-
determined, or experiencing the self as the locus of control (Ryan & Deci, 2000, 2002).
Someone who is intrinsically motivated is free from external controls and assumed to be
acting in accord with his or her true, authentic self. Intrinsic motivation, then, is the
hallmark of the integrated author-self. Extrinsically motivated behaviors, on the other
hand, are in response to some type of control or regulation that exists, wholly or in part,
outside the actor (Deci & Ryan, 1985; Ryan & Deci, 2000, 2002). Ryan and Deci (2000,
40
2002) have hypothesized four types of extrinsic motivation on their continuum: external
regulation, action motivated by foreseeable positive or negative consequences; introjected
regulation, action to increase self-esteem or to avoid guilt; identified regulation, action
stimulated by personal identification and perceived as autonomous; and integrated
regulation, action to obtain personally important goals or to harmonize with personally
important values. Although integrated regulation is autonomous action, Ryan and Deci
have separated it from intrinsic motivation due to its utilitarian nature.
According to Deci and Ryan?s taxonomy of motivation, the origin of a person?s
motivation (its energy source and direction) indicates the level of autonomy, or self-
determination (Ryan & Deci, 2000, 2002). Furthermore, intrinsic motivation and higher-
end extrinsic motivation lead to greater engagement, including creative problem solving,
increases in self-worth, and more efficient learning (Deci, Ryan, & Williams, 1996; Deci,
Vallerand, Pelletier, & Ryan, 1991). However, motivation is also a by-product of one?s
interaction with the environment. SDT proposes that motivation takes place within a
social context, and responds to the extent to which one?s needs for autonomy,
competence and relatedness are either frustrated or supported by the environment (Ryan
& Deci, 2000, 2002). In their most recent discussions of SDT, Ryan and Deci (2002)
suggested that autonomy, competence and relatedness are basic to human functioning and
have referred to them as Basic Psychological Needs (BPN). According to Ryan and Deci
(2002), these BPN are the foundation for self-determination and predicting optimal
human development. Therefore, these basic psychological needs must be considered in
the equation describing personal attributes that influence success.
41
Basic psychological needs. As basic psychological needs, autonomy,
competence, and relatedness are central components of SDT. Moreover, Ryan and Deci
(2002) identified these BPN as universally human, common to all people regardless of
culture, race/ethnicity, gender, or age. Each of the basic psychological needs is
delineated below.
As discussed above, autonomy is the experience of self-governance or an internal
locus of control. The need for autonomy is intimately connected with one?s self-
determination and motivation: ?When autonomous, individuals experience their behavior
as an expression of the self, such that, even when actions are influenced by outside
sources, the actors concur with those influences, feeling both initiative and value with
regard to them? (Ryan & Deci, 2002, p. 8). The need for competence, on the other hand,
does not refer to accumulated skill or knowledge, but to the desire to engage and
accomplish challenging activities commensurate with capabilities (Deci & Ryan, 1985).
The need for competence provides ?energy for learning? (Deci & Ryan, p. 27), such that
one continuously seeks new opportunities to increase skills and abilities (Ryan & Deci).
Finally, the need for relatedness, refers to the basic human need for connectedness
and a sense of belonging (Ryan & Deci, 2002). Unlike autonomy and competence, the
need for relatedness pertains more to the process of internalizing values and goals than it
does to intrinsic motivation (Ryan & Deci). Whereas the needs for autonomy and
competence drive self-determination, the need for relatedness, or the experience of being
in communion with others, inspires movement through the phases of extrinsic motivation.
In other words, relatedness actually fosters autonomy in that feeling supported by and
42
connected to significant others leads to greater security and self-worth (Deci, Ryan, &
Williams, 1996).
Self-determination and basic psychological needs within education. As the
proposed research pertains to the self-determination of college students, it is important to
note SDT?s alignment with theories about student development. All three basic
psychological needs proffered by SDT are harmonious with classic college student
development theory. Chickering and Reisser (1993) put forward a framework for college
student development based on a series of seven vectors, including establishing identity,
developing purpose, and managing emotions. Chickering and Reisser used vectors, as
opposed to stages, to indicate location and direction of each developmental task. Three
of these vectors correspond to the three basic psychological needs of autonomy,
competence and relatedness (respectively, moving through autonomy, developing
competence and developing mature relationships). A fourth vector, developing integrity,
is the result of self-determination, according to Deci and Ryan (1985). Chickering and
Reisser?s classic text provides a theory of student development but does not suggest
discrete variables to measure student progression along each vector. However, the
convergence between these four vectors and the Deci and Ryan?s basic psychological
needs suggests a justification for studying SDT in the context of education and college
student development.
In a departure from student change literature, particularly college impact research,
self-determination theory reframes the role of the environment. Rather than viewing the
environment as shaping behavior, SDT considers the environment as a social context
having the ability to foster or hinder growth (Ryan & Powelson, 1991). According to
43
SDT, the ideal environment will satisfy the basic psychological needs, which leads to
increased motivation and, ultimately, greater educational outcomes (Deci, Ryan, &
Williams, 1996; Deci, Vallerand, Pelletier, & Ryan, 1991; Miserandino, 1996).
Furthermore, when students? basic psychological needs are met in the educational
environment, students begin to value learning and academic success, thereby becoming
truly engaged (Deci, Vallerand, Pelletier, & Ryan, 1991).
According to Ryan and Deci (2002), human well-being is rooted in experiencing
autonomy, competence, and relatedness. These basic psychological needs are essential
for understanding motivation (Deci, Ryan, & Williams, 1996) in that intrinsic motivation
relies on a healthy psyche, which stems from meeting these needs. (Deci, Ryan, &
Williams; Ryan & Deci). When the social or educational context satisfies the basic
psychological needs, students are poised for greater learning and growth (Ryan &
Powelson, 1991). The following section reviews recent studies exploring self-
determination, including the basic psychological needs, within education.
Related research. In a survey of self-determination research, Reeve (2002)
affirmed the use of Deci and Ryan?s theory in the educational setting. According to
Reeve?s literature review, K-12 students with greater self-determination have higher
academic achievement, higher rates of retention, and greater engagement in their
educational processes. In light of these results Reeve concluded, ?The quality of a
student?s motivation explains part of why he or she achieves highly, enjoys school,
prefers optimal challenges, and generates creative products? (p. 183).
Consistent with Reeve?s analysis, Miserandino (1996) found that among 77
elementary school children, perceived competence and autonomy significantly positively
44
predicted student engagement and grades in math and social studies. The results of her
regression analyses indicated that perceived autonomy is a significant positive predictor
of grades in reading, language arts, and spelling. Similarly, Connell and Wellborn (1991)
studied 1,487 third through tenth graders in three separate age-related samples. Using
path analysis, the team demonstrated a significant positive correlation between perceived
autonomy and academic achievement as well as teacher-rated engagement. In addition,
results of the study indicated that perceived competence is significantly positively related
to student engagement, as measured by dedicated time and interest level. The
correlations between feelings of relatedness and engagement or performance were not
significant. Neither the Miserandino nor the Connell and Wellborn study indicated that
the researchers controlled for background characteristics, including prior achievement.
Conversely, in studying 4,537 high school students (male: 2,280; female: 2245),
Vallerand, Fortier, and Guay (1997) found that students who experienced low autonomy
support also reported experiencing low levels of autonomy, competence, and relatedness.
Using structural equation modeling, the research team observed that these students also
reported significantly lower levels of self-determined motivation and were at significantly
higher risk for high school dropout. Vallerand et al.?s study indicated gender differences
such that female students reported significantly higher self-determined behavior and
lower rates of dropout.
While the majority of research on self-determination and BPN concentrates on the
K-12 population, a few studies indicate that self-determination is a strong predictor of
academic outcomes for older adolescents and working adults as well. Investigating 834
college students in the Canadian province of Quebec, Guay, Sen?cal, Gauthier and Fernet
45
(2003) employed structural equation modeling and found that among college students,
self-determination was a strong predictor of career choice, after controlling for gender.
In a study of volunteerism among 121 college students and 227 adult workers, Gagn?
(2003) found that self-determination, specifically the basic psychological needs, had a
significant positive association with volunteer persistence.
While these studies and those listed above further the literature about the
relationship between self-determination and particular outcomes, not much is known
about self-determination with respect to involvement, academic achievement or self-
concept in higher education among college students. Furthermore, none of these studies
investigates which aspects of the university environment cultivate or thwart self-
determination and the basic psychological needs.
Critique of self-determination theory. While current research tends to support
SDT as a valid framework for understanding personal locus of control in adulthood, the
theory is not without its detractors. In his essay, ?Self-Determination: The Tyranny of
Freedom,? Schwartz (2000) warned one to evaluate SDT with a critical eye. According
to Schwartz, self-determination, especially when conceived as unfettered freedom, can
lead to a paralyzing and unhealthy focus on the self. Schwartz suggested that placing too
much ?emphasis on individual autonomy and control may be undermining a crucial
vaccine against depression: deep commitment and belonging to social groups and
institutions? (p. 86). Rather, Self-determination Theory?s singular focus on autonomy
and choice may serve to alienate the individual from his or her need for interdependence.
46
Self-efficacy
The field of psychology demonstrates that personal locus of control also is
grounded within self-appraisal of personal mastery and one?s ability to perform cognitive
and behavioral tasks (Bandura, 1977, 1982), or self-efficacy. According to Bandura
(1977, 1982), self-efficacy beliefs are the tools of personal authorship. Self-efficacious
expectations determine the effort a person expends toward an activity as well as how long
he or she will persist and how successful the outcome will be. In the words of Bandura,
?self-referent thought is the mediator between knowledge and action? (1982, p. 122)
because people will attempt activities, whether cognitive or behavioral, only if they
reasonably assess that they are capable of accomplishing the task. Similarly, people will
avoid activities when they suspect that their efforts will not be successful. Self-efficacy
is an important factor in student effort because it highlights the importance of a student?s
beliefs about his or her own ability to succeed as a pre-cursor to academic and co-
curricular involvement.
Bandura (1977, 1982) identified four personal and environmental cues from
which individuals make assessments about their own self-efficacy. In order of
significance, these cues are: First, people judge their ability to succeed on prior
performance and accomplishment. That is, if one has excelled in a task before, he or she
is likely to assume subsequent success. Second, individuals gather information about the
probability of success vicariously by observing others? performance. Third, verbal
persuasion from a trusted other can boost an individual?s beliefs of self-efficacy. Fourth,
an individual?s emotional state affects his or her ability to assess capabilities, such that
high stress levels can cause over- or under-estimation of expectations for success.
47
Bandura?s work in developing self-efficacy theory has been applied to the areas of effort,
persistence, achievement and satisfaction among college students. Relevant studies
addressing self-efficacy among college students are summarized below.
Related Research
Research on self-efficacy among college students suggests that self-efficacy is a
significant predictor of educational outcomes. In all, these studies indicate that any
measurement of student attributes must account for students? beliefs in their ability to
successfully engage in the academic and social life of their institution. Although similar,
self-efficacy is distinguishable from self-concept on one major point: the basis for self-
assessment. In self-concept, assessment is comparative and measured against the abilities
of one?s peers. In self-efficacy, however, assessment is based on personal prior
experience and the likelihood of future success. For example, Brown, Lent and Larkin
(1989) researched self-efficacy as a predictor of grades and persistence among 105 first-
year and sophomore students enrolled in a career-planning course for engineering majors
at one institution. Brown et al. used two measures of self-efficacy: (a) students? beliefs
that they will successfully complete the educational requirements for the engineering
degree; and (b) their beliefs that they will persist through certain academic milestones
related to their intended major. The results of the study indicated that students?
assessment of their ability to achieve is a strong positive predictor of academic
performance, regardless of academic aptitude. However, self-efficacy beliefs in the area
of persistence moderated the effect of scholastic aptitude on academic performance. In
other words, students with high self-efficacy beliefs but lower scholastic aptitude were
48
more likely to achieve higher academic standards than their peers who had higher
scholastic aptitude but lower self-efficacy beliefs. Brown et al. concluded that self-
efficacy judgments are most helpful if they are slightly exaggerated.
In their meta-analysis of self-efficacy research, Multon, Brown and Lent (1991)
examined the relationships between self-efficacy and academic performance, effort, and
persistence. They found, as Bandura (1977, 1982) has theorized, that self-efficacy
significantly influences students? choice of academic activity, the effort they put toward
those activities, and their persistence. Multon et al. defined persistence as time on task,
number of tasks completed, and number of academic terms completed. Results from this
study demonstrate that self-efficacy accounted for 14% of the variance for academic
performance and 12% of the variance for persistence.
In a departure from viewing the effects of self-efficacy on performance and
persistence, DeWitz and Walsh (2002) explored the relationship between self-efficacy
and college student satisfaction. DeWitz and Walsh surveyed 312 undergraduate students
enrolled in an introductory psychology course at one institution. The results of their
quantitative analysis suggested that students? self-efficacy beliefs about the college
experience in general are the most significant predictor of satisfaction with the college
experience. More specifically, students with high college self-efficacy were found to
lead happier social lives, believe that they were adequately compensated in grades for
their academic work, believe that their grades matched the effort they expended toward
their academics, recognize greater opportunities for involvement, participate more
frequently in campus events, and perceive a higher quality of education.
49
Finally, Leslie, McClure, and Oaxaca (1998) investigated the effects of self-
efficacy beliefs on the likelihood that college women students and students of color
would choose majors in science, math, or engineering. The sample for this study
included 9,628 entering first-year college students. Using items in the 1971 and 1980
Cooperative Institutional Research Program (CIRP) data files, Leslie et al. conducted
factor analyses to derive composite measures of self-efficacy beliefs. These items
included but were not limited to, student self-reports of math and science preparation,
high school rank, high school college-preparatory program, parental occupation in
science or engineering, and parental income. Multinomial logit regression analyses
demonstrated that self-efficacy beliefs accurately predicted selection of science,
engineering or math as the student?s intended major. Moreover, Leslie et al.?s analyses
revealed significant differences by race and gender. Based on self-efficacy beliefs of
being well-prepared in math and science, White men are more likely than women, Asian-,
and African-American students to select a major in math, science, or engineering. There
was no significant difference between White men and Hispanic/Latino men in this regard.
Critique of self-efficacy
Proponents of self-determination theory claim that self-efficacy?s reliance on
competency beliefs is not sufficient for studying motivation and performance (Connell &
Wellborn, 1991; Deci, Ryan, & Williams, 1996; Deci, Vallerand, Pelletier, & Ryan,
1991). These theorists have argued that although competency beliefs are vital to
motivation, motivation cannot be understood without accounting for autonomy (Deci,
Ryan, & Williams, 1991). More specifically, Deci, Vallerand, Pelletier, and Ryan
50
(1991) concluded that motivation research cannot adequately explain important
differences between extrinsically and intrinsically motivated action without considering
autonomy. Connell and Wellborn (1991) suggested that the flaws in self-efficacy stem
from its identification with social psychology, which emphasizes the social context in
which human development takes place. They point out that social theories, such as self-
efficacy, propose a ?unidirectional? relationship running from the social context to the
person. In other words, self-efficacy does not acknowledge ?what the person brings to the
negotiating table other than a history of social interaction? (Connell & Wellborn, 1991, p.
47).
Summary
Self-concept is a measure of self-assessed abilities as compared to one?s peer
group. Previous studies have linked self-concept and achievement, such that prior self-
concept predicts grade point average. Researchers also have determined that student
involvement in the educational environment predicts self-concept. However, previous
research sheds little light on how other student attribute measures work with self-concept
to influence academic success. Marsh (2003) noted: ?implicit in our discussion is the
untested assumption that the effect of prior self-concept on subsequent achievement was
mediated by student characteristics such as increased conscientious effort [and] enhanced
intrinsic motivation?? (p. 12). The current study brings these concepts together to
understand better how student attributes intersect the institutional environment to produce
educational outcomes.
51
The discipline of psychology supports personal locus of control within two
distinct theories: self-determination and self-efficacy. Both approaches explain human
motivation. Whereas self-efficacy focuses on competency beliefs and explores beliefs
about the ability to succeed as a precursor to action, self-determination suggests that
intrinsic motivation springs from a triad of basic psychological needs for autonomy,
competency, and relatedness. Studies exploring educational outcomes, such as
performance and engagement, have supported the validity and explanatory power of both
self-efficacy and self-determination theories. While both theories also have their critics,
they prove equally useful in understanding the underlying constructs of college student
attributes toward achieving academic success.
Academic and Social Involvement
As a theory of college impact, Pascarella and Terenzini (2005) noted that Astin?s
concept of involvement balances the roles of institutional environment and active student
engagement in producing student change. Through the description of his theory, Astin
(1984) clearly equated involvement with student behavior:
I am not denying that motivation is an important aspect of involvement, but rather
I am emphasizing that the behavioral aspects, in my judgment, are critical: It is
not so much what the student thinks or feels, but what the individual does, how he
or she behaves, that defines and identifies involvement (p. 298).
However, several of the questions Astin has raised for consideration refer back to the
psychological determination from which involvement arises. For instance, Astin stated:
?It seems clear that the effectiveness of any attempt to increase student involvement is
52
highly contingent on the student?s perceived locus of control and attributional
inclinations? (p. 307). In other words, Astin himself opened the door for including
psychological measures, such as self-efficacy and self-determination, as important
variables to consider when studying the effects of involvement on academic success.
Elsewhere, Astin (1984) suggested that the educational experience ?must elicit
sufficient student effort? (p. 301), but also that student effort depends in large part on
?how motivated the student is and how much time and energy the student devotes to the
learning process? (p. 301). Hernandez, Hogan, Hathaway and Lovell (1999) echoed
Astin?s sentiments. In their review of involvement-related literature, Hernandez et al.
noted several limitations inherent in the study of student involvement. First, Hernandez
et al. observed that student intentions were not considered in any of the studies they
analyzed. Second, they noted that college impact depends on the quality of effort
students are willing, or able, to put forward: ?Students bear much of the responsibility
for the extent to which involvement makes a difference in their own development and
learning? (p. 195).
Astin?s Theory of Involvement
In an effort to create a cohesive link between disparate theories that explain how
and why college student development occurs, Astin (1984) introduced student
involvement theory. Remarkably straightforward, the theory focuses primarily on student
behavior rather than perceptions or motivations: ?student involvement refers to the
amount of physical and psychological energy that the student devotes to the academic
experience? (p. 297). Astin regarded student involvement theory as easy to understand
53
and compatible with psychological and sociological explanations of college impact. He
suggested that the theory, ?can explain most of the empirical knowledge about
environmental influences on student development that researchers have gained over the
years? (p. 297).
Five postulates comprise the basis for Astin?s theory of involvement: 1)
involvement signifies an investment of psychological and physical energy in the
educational experience; 2) a continuum of involvement exists along which either the
same student invests a range of energy at various times, or different students invest
different amounts of energy toward the same experience; 3) involvement manifests
quantitative and qualitative characteristics, such that physical and psychological energy
are measurable in amount of time and quality of effort; 4) a student?s educational
outcomes are directly proportional to his or her involvement; and 5) the efficacy of higher
education practices and policies is related to their ability to improve student involvement
(1984, p. 298). Furthermore, Astin (1993) has contended that involvement measurements
include those at time of student entry, such as place of residence, prospective major, and
financial aid, as well as those describing student-environment interaction, such as student
involvement with faculty and peers, work, and academic effort.
In many ways, Astin?s explanation of involvement mirrors Kuh?s (2005)
description of student engagement: ?the amount of time and effort students put into their
studies and other educationally purposeful activities,? (p. 87). Although Kuh?s concept
of engagement differs from Astin?s involvement in that it includes a second feature
explicitly focusing on the institutional environment, the language describing both
54
engagement and involvement is quite similar. Therefore, for the purposes of this study,
engagement and involvement are synonymous terms, and I use them interchangeably.
Related Research
In their anthologies of college impact research, Pascarella and Terenzini (1991,
2005) reviewed several hundred studies linking involvement to educational gains, such as
cognitive skills, subject-matter competence, and persistence. Pascarella and Terenzini?s
findings are as Astin (1999) suggested: ?The greater the student?s degree of involvement,
the greater the learning and personal development? (p. 588). Several contributions to the
involvement literature are discussed below.
After noting the similarities between Astin?s (1984) involvement theory and
Tinto?s theory of student departure (1993), Milem and Berger (1997) designed a study to
empirically test these theories? compatibility with one another; that is, they tested
whether or not Astin?s idea of involvement is a precursor to Tinto?s ideas of academic
and social integration, which ultimately lead to a student?s decision to persist or dropout.
Milem and Berger noted that Tinto?s model was primarily perceptual in orientation,
whereas, Astin?s model was primarily behavioral. The researchers hypothesized that
rather than maintain independence from one another, the models of involvement and
student departure actually influenced one another.
Milem and Berger (1997) chose a private, selective, southeastern university for
their research site and culled their data from three separate surveys: the CIRP Student
Information Form administered during fall orientation 1995; and two locally developed
surveys, one of which was administered in October, 1995 and the other in March 1996.
55
Using data from these three surveys, Milem and Berger constructed a longitudinal panel
of 718 first-year students and conducted a multivariate path analysis to test their research
questions.
Among their results, Milem and Berger found that first-year students who were
not academically engaged in the fall semester were also less likely to perceive the
institution as supportive, less likely to be academically engaged in the spring, and less
likely to perceive their peers as supportive. Likewise, students who reported early social
integration also reported academic non-engagement and were not engaged with the
institution. However, social integration was a significant positive predictor of students?
institutional commitment and their intent to enroll the subsequent fall semester. Their
study led Milem and Berger to conclude that early student involvement was deeply
connected to intent to persist: ?Our findings suggest that the extent to which students
become involved during their first 6 to 7 weeks of a semester are significantly related to
whether they are likely to persist at the institution? (p. 398). The data also demonstrated
that early involvement with faculty is a strong predictor of persistence (Milem & Berger,
1997).
In a follow-up to the study described above, Berger and Milem (1999) tested their
revised model of persistence using fall 1996 enrollment data as measures of actual
persistence. With the exception of the added fall 1996 enrollment figures, Berger and
Milem utilized the same longitudinal panel data as Milem and Berger (1997). In
addition, Berger and Milem conducted a path analysis to test the direct and indirect
effects of involvement on persistence, whereas their previous work examined only direct
effects.
56
In sum, Berger and Milem?s 1999 findings indicated that involvement has indirect
effects on persistence as well as direct effects. For example, students who reported
noninvolvement in the fall also were likely to be uninvolved in the spring, not to be
academically integrated, and were likely not to persist. Spring noninvolvement had
similarly negative predictive effects on academic integration and persistence but was also
a significant negative predictor of social integration.
These two studies (Milem & Berger, 1997; Berger & Milem, 1999) confirm that
involvement is a predictor of academic success and that noninvolvement is a predictor of
departure. These inquiries are related to the proposed research in that they indicate that
involvement should be included as a variable in any examination of the factors leading to
success.
Building on previous research about the effects of student involvement on
academic and interpersonal gains, Huang and Chang (2004) studied the effects of
involvement on third-year college students in Taiwan. Specifically, Huang and Chang
were interested in understanding the relationship between academic and social
involvement and identifying the level of maximum involvement at which students cease
to realize academic and social benefits. Using multistage cluster sampling first to control
for institution type and second to control for institution size, the researchers randomly
selected education classes at each institution. The resulting sample included 627 students
representing 14 different institutions.
In order to answer their first research question, Huang and Chang conducted
scatter plot analysis, which indicated that the relationship between academic and co-
curricular involvement is linear and positive. To address their second question, the
57
researchers conducted an analysis of means for students? growth in self-confidence and
cognitive, communication, and interpersonal skills. All growth measures were self-
reported by the student participants on Likert-type scales. Huang and Chang found that
students with high academic and high co-curricular involvement reported the most gains
in cognitive and communication skills, self-confidence, and interpersonal skills. In
contrast, students with low academic and co-curricular involvement reported the least
gains three areas: communication skills, self-confidence, and interpersonal skills. These
students with low-low designations also reported the second-to-lowest gains in cognitive
skills. Through their statistical analyses, Huang and Chang concluded that co-curricular
involvement does not detract from academic focus. Furthermore, their results suggest
that gains in cognitive and communication skills are linked to both academic and co-
curricular involvement, while gains in self-confidence and interpersonal skills are more
strongly connected to co-curricular than academic involvement.
Huang and Chang?s study helps to establish the link between involvement and
academic success. However, because their study included only a comparison of mean
differences and no analyses to determine prediction or even significance of difference,
more research along these lines is necessary. Furthermore, Huang and Chang?s work
examined Taiwanese students, and their findings may not be generalizable to students in
the United States.
Critique of Astin?s Theory of Involvement
Although Astin?s Theory of Involvement is the focus of much research, Pascarella
and Terenzini (1991, 2005) have questioned whether Astin?s concepts actually meet the
58
definition of a theory. Whereas theories generally offer a system of predicting a
phenomenon by identifying specific variables and their relation to one another (Kerlinger
in Pascarella & Terenzini, 2005), Astin?s involvement concepts are a set of general
suppositions. To be confirmed as ?theory,? Pascarella and Terenzini have recommended
Astin?s ideas of involvement to include ?detailed, systemic description of the behaviors
or phenomena being predicted, the variables presumed to influence involvement, the
mechanisms by which those variables relate to and influence one another, or the precise
nature of the process by which growth or change occurs,? (p. 54). This study will not
settle the debate over the status of Astin?s involvement as theory or general principle. It
will, however, build upon previous studies using Astin?s schema for student involvement
in conjunction with his I-E-O model of college impact in order to examine factors
contributing to academic success of first-year students.
Student Effort
The 2004 annual report of the National Survey of Student Engagement stated that
nearly half (44%) of all college students spend 10 hours or less per week studying or
preparing for their academic courses (National Survey of Student Engagement, 2004).
This is a far cry from the 30-40 hours per week of studying most faculty believe is
necessary in order to be an effective student (National Survey of Student Engagement,
2002). In the estimation of higher education scholars, the fraction of actual study time
versus recommended study time is less a function of educational program quality than of
student effort and involvement.
59
Student effort refers to the amount of time and the quality of the mental activity
college students put toward their undergraduate experiences, inside and outside the
classroom (Astin, 1985; Pace, 1984). If institutions are responsible for providing an
environment conducive to learning, then students are responsible for the investment of
time and effort toward their own education.
According to Pace (1984), student effort begins with the student?s awareness of
the educational opportunities, curricular and co-curricular alike, available at his or her
institution. Once a student is knowledgeable about his or her opportunities, the effort he
or she makes toward participation can be measured. The amount and quality of a
student?s involvement with his or her own college process affects his or her ability to
learn; to create and maintain relationships with faculty and peers; and to participate in
academic, social, and developmental activities (Pace, 1984). In other words, student
authorship is the deciding factor between whether or not a student attains a quality
education.
In support of Pace?s views on student effort, Arrison?s (1988) study revealed that
effort is significantly, positively correlated with academic achievement, as measured by
grade point average. Using a mixed-methods research design, including multivariate
regression analyses and qualitative interview techniques, Arrison examined the predictive
ability of students? academic self-confidence on student effort and academic
achievement. Arrison?s sample for quantitative analyses included 418 first-year students
at one institution, who completed the survey for incoming first-year students sponsored
by the Cooperative Institutional Research Program (CIRP). In related analyses, Arrison
explored the relationship between effort and achievement. She found significant, positive
60
correlations between grade point average and 6 of her 17 effort variables, including 3
related to academic, or discipline-related, effort and 3 related to co-curricular and
interpersonal effort.
Summary
Pace?s (1984) work with the quality of student effort and Astin?s (1984) theory of
involvement are representative of Astin?s (1991, 1993) definition of the environment.
Both frameworks are strikingly similar in that they focused on the quantity and quality of
student behavior within the college environment with the understanding that behaviors
are directly related to educational outcomes. The works of Pace and Astin provide
guidance toward operationalizing student behavior, with respect to awareness of
academic and co-curricular opportunities, time spent on social and academic pursuits, and
quality of effort in those activities. Although neither approach accounted for, or
attempted to explain, psychological characteristics that influence students? behavior,
research literature indicated that both frameworks help to explain educationally relevant
outcomes, such as performance, self-concept, and persistence.
Studying Academic Success in the First Year:
Literature Supporting Variables in the Study
Through his extensive work studying first-year students, Upcraft (2005) devised a
thorough model for successful assessment of the first year of college. This section draws
upon relevant literature to justify the inclusion of variables pertinent to studying the
relationship between student attributes and academic success in the first year. As the
conceptual model for this study follows Astin?s (1991) preferred input-environment-
61
outcome (I-E-O) model for assessment, the variables appear in I-E-O order. First, I
discuss variables representing students? background characteristics, including gender,
race, socioeconomic status, and prior achievement. Second are student attribute
variables, including academic and social self-concept, self-efficacy, and self-
determination. In the conceptual model for this study student attribute variables are both
input measures as well as intermediate outcomes. Last, I discuss variables related to the
educational environment and collegiate experiences, including academic and social
involvement.
Student Background Characteristics
According to Upcraft (2005), ?efforts to promote student success, both inside and
outside the classroom should be based on what is known about the first-year cohort,? (p.
475). In other words, student background characteristics provide the baseline for
assessment. In addition, background characteristics often are related to how students
interact with the educational environment and to any subsequent outcomes, so studying
growth or change over the first year would be impossible without controlling for
differences at time of entry (Astin, 1991). Astin (1991) recommended accounting for
different types of input measures, many of which are germane to this study. Among
these are: ?fixed student attributes,? which include demographic characteristics;
?cognitive functioning,? which includes prior achievement measures; and ?self-ratings,?
which, for the purposes of this study, refer to prior self-concept.
62
Gender
With few exceptions (Guay, Sen?cal, Gauthier, & Fernet, 2003; Pascarella, Smart,
Ethington, & Nettles, 1987), recent studies have uncovered significant gender differences
in variables relevant to this study. For example, Astin (1993) found that being female
was a positive predictor of academic achievement, as measured by grade point average.
With regard to the development of self-concept, a number of studies noted differences in
self ratings, such that men reported higher self-confidence (Delaney, 2004) while women
in the Kezar and Moriarty (2000) study reported greater gains in intellectual and social
self-confidence over four years of college. In addition, among high school students,
Marsh and Yeung (1998) found that despite achieving higher grades in English and math,
girls reported lower math self-concepts. Gender differences are also evident in
involvement, such that women report greater levels of social/peer involvement (Berger &
Milem, 1999; Milem & Berger, 1997) but higher levels of academic non-engagement
(Milem & Berger, 1997). Finally, with regard to measures of student attributes, Leslie,
McClure, and Oaxaca (1998) reported that men experienced higher levels of math/science
self-efficacy, while Vallerand, Fortier, and Guay (1997) indicated that high school girls
report higher levels of self-determined motivation leading to lower levels of dropping
out.
Race
Some studies demonstrate racial differences among variables predicting academic
success. In his longitudinal study of college effects, Astin (1993) found that being White
was a positive predictor of grade point average while being Latino/a was a negative
63
predictor. Results from studies examining racial difference in self-concept are mixed:
Pascarella, Smart, Ethington, and Nettles (1987) found no appreciable difference based
on race, while others (Cokely, Komarraju, King, Cunningham, & Muhammad, 2003)
revealed differences in the structure of academic self-concept constructs based on race.
Similarly, the Milem and Berger (1997) study indicated racial differences in the types of
student involvement that predicted persistence, such that activism was a positive
predictor for African American students while social involvement was a positive
predictor for White students. Finally, while few other studies explored racial differences
in constructs of student attributes, Leslie, McClure, and Oaxaca (1998) noted significant
differences in math/science self-efficacy beliefs based on race, such that White students
were nearly twice as likely as African American and Latino students to report better-than-
average preparation in math and science.
First-generation Status
While there is little to no research linking first-generation status to student
attributes, Astin?s (1993) work indicates that peer group SES has strong positive effects
on intellectual development, indicators of academic involvement, and some
environmental characteristics.
Prior Achievement
According to Astin (1993), prior achievement, as measured by high school grades
and college entrance exam scores, is the single greatest predictor of college grade point
64
average. In addition, Marsh (2003) found prior achievement to be a significant, positive
predictor of academic self-concept.
Student Attributes
Intuitively a relationship between student attributes and academic success makes
sense, and this supposition is supported in the literature (Astin, 1993; Kuh, Kinzie,
Schuh, & Whitt, 2005; Pascarella & Terenzini, 2005). Although prior self-concept, self-
efficacy beliefs, and prior measures of self-determination are technically considered
background characteristics, for the purposes of this study, they belong to a separate
variable category for attributes.
Prior Self-concept
Student self-concept at the time of college entry is a strong predictor of
subsequent achievement, as measured by GPA (Astin, 1993; Cokely, Komarraju, King,
Cunningham, & Muhammad, 2003; House, 2000; Marsh, 2003; Marsh & Yeung, 1998;
Yeung, McInerney, Russell-Bowie, Suliman, Chui, & Lau, 2000). In addition, prior self-
concept predicts academic involvement (Astin, 1993) and social involvement (Pascarella,
Smart, Ethington, & Nettles, 1987). Finally, prior self-concept significantly positively
predicts subsequent self-concept (Astin, 1993; Berger & Milem, 2000; Terenzini,
Theophilides, & Lorang, 1984).
65
Self-efficacy
The results of Brown, Lent and Larkin?s (1989) study led them to conclude self-
efficacy is a predictor of academic performance, as measured by grades and persistence,
regardless of prior aptitude levels. In other words, regardless of aptitude, students with
high self-efficacy beliefs performed better than those with lower self-efficacy. In their
meta-analysis of self-efficacy research, Multon, Brown and Lent (1991) arrived at a
similar conclusion: self-efficacy beliefs significantly and positively predict academic
performance. In a different study, DeWitz and Walsh (2002) found students? self-
efficacy beliefs to have a significant, positive relationship to satisfaction with academic
effort and with academic and social involvement.
Self-determination
As with self-efficacy, research suggests that measures of self-determination also
have a significant association with academic performance and effort. Although few
studies examine the relationship between self-determination and academic success among
college students, self-perceptions of autonomy and competence have a significant,
positive correlation to academic achievement and teacher-rated engagement among
elementary and high school students (Connell & Wellborn, 1991; Miserandino, 1996).
Conversely, in a study about predictors of high school dropout, Vallerand, Fortier, and
Guay (1997) found students with low levels of autonomy support, reported lower levels
of autonomy, competence, and relatedness. Consequently, these students also displayed
lower motivation and higher rates of drop out. My study will include measures of
autonomy, competence, and relatedness as recommended by Deci and Ryan (2002).
66
First-year Experiences: Aspects of the Environment
For the purposes of this study, environmental measures, also known as collegiate
experiences, are divided into academic and social involvement. This dichotomy of
involvement is consistent with Astin?s (1984) theory of involvement and with previous
research (Berger & Milem, 1999; Huang & Chang, 2004; Kezar & Moriarty, 2000;
Milem & Berger, 1997). Both are complex constructs comprised of several different
measures.
Academic Involvement
Academic involvement includes items such as faculty involvement (Astin, 1993;
Berger & Milem, 1999; Milem & Berger, 1997; Pascarella, Smart, Ethington, & Nettles,
1987; Pascarella & Terenzini, 2005), in-class educational behaviors (Huang & Chang,
2004) and participation in educational initiatives such as first-year seminars and living-
learning environments (Astin, 1991; Chrissman Ishler & Upcraft, 2005; Pascarella &
Terenzini). This section details each aspect of academic involvement.
Students? interactions with faculty include such measures as of out-of-class
interaction, visits to faculty homes, meeting with faculty during office hours, and sharing
a meal and/or coffee with faculty. Almost without exception, these measures have had
positive effects on students? self-concept (Astin, 1993; Berger & Milem, 2000), academic
achievement (Astin, 1993; Pascarella & Terenzini, 2005) and institutional involvement
(Milem & Berger, 1997). In addition, Huang and Chang?s (2004) study revealed
students? class-related behaviors such as attendance, taking notes, and completing
67
assignments on time had a significantly positive relationship with self-reported
intellectual and cognitive growth.
Although few studies of student participation in educational initiatives have
measured the effects of participation on grade point average, some studies have indicated
that participation in specific programs has positive effects on academic success.
Specifically, enrolling in first-year seminars and living in college residence halls are
linked to persistence (Chrissman Ishler & Upcraft, 2005; Pascarella & Terenzini, 2005)
and participation in living-learning environments is positively related to grade point
average (Chrissman Ishler & Upcraft). In addition, Inkelas and Weisman (2003) found
that participation in living-learning programs is a strong predictor of enjoying
challenging academic pursuits. Finally, residence living in general is positively related to
academic performance (Blimling, 1999) and persistence (Berger, 1997).
Effort
Astin?s (1993) work demonstrates that student effort, as measured by time-on-task
studying, has a positive association with nearly every academic outcome he tested,
including grades, persistence, and self-rated cognitive and social growth. Similarly,
Pace?s (1984) study on quality of student effort suggests a positive correlation between
effort and learning. The effort variable in my study is patterned after Astin?s work.
Social Involvement
Social involvement denotes a wide array of activities, including peer interaction
(Astin, 1993; Berger & Milem, 1999; Milem & Berger, 1997) and participation in social
68
or co-curricular activities (Astin, 1993; Berger & Milem, 1999; Milem & Berger, 1997;
Huang & Chang, 2004). In general, peer interaction measures, such as discussing course
content, participating in study groups, and socializing, have positive effects on social
self-concept and academic development (Astin, 1993), perceptions of peer and
institutional support, and subsequent involvement with faculty (Milem & Berger).
However, the Milem and Berger study revealed that early peer interaction negatively
influenced subsequent social integration. Participation in co-curricular activities produce
mixed results. On the one hand participation in traditional social activities leads to
academic non-engagement (Milem & Berger), and on the other hand, Huang and Chang
(2004) reported that co-curricular involvement has a significant and positive relationship
with self-rated cognitive growth and academic involvement. Astin (1993) reported that
peer interaction is positively correlated with cognitive and affective development.
Conclusion: Blending Sociological and Psychological Theories
to Create a Better Model
Heretofore, this review of literature has provided an in-depth discussion of two
previously disparate approaches to understanding changes in college students. The
college impact models describe how the environment contributes to student outcomes
while controlling for differences in student background characteristics (Astin, 1991,
1993), but overlooks how the person contributes to the environment. This study weds the
sociological perspective of college impact with psychological dimensions usually
reserved for understanding student development over time. In their meta-analysis of
student outcomes research, Pascarella and Terenzini (2005) concluded that the best
69
models combine the sociological orientation with the psychological orientation. This
study blended these two perspectives to create a more complete model.
This study helps to fill gaps in our understanding of how student attributes, such
as self-concept and self-determination, affect educational outcomes. Specifically the
current study addresses the role of student attributes in predicting academic achievement
among first-year college students, net of previously established predictive factors,
including prior self-concept, high school achievement, and involvement. In addition, this
study builds upon our understanding of the role the psychological dimension of student
attributes play in predicting involvement. The literature on self-efficacy (Bandura, 1977,
1982), self-determination (Deci & Ryan, 1985; Ryan & Deci, 2000, 2002), and self-
concept have suggested that involvement is more a function of psychological attributes
than of the environment.
However, Ryan and Deci (2000, 2002) also have indicated that the environment
can influence self-determination. Therefore, this study was designed to help uncover
how the educational experience influences change in student attributes over time.
Although the current study contributes to our understanding of the role and responsibility
of the individual college student in achieving educational outcomes, I note Schwartz?s
(2000) admonishment not to over-emphasize student attributes in such a way as to
disadvantage the role of environmental factors. To address this concern, this study
examined academic and social involvement along side student attributes in order to
determine the relationship between the two.
This study utilized path analysis to determine the predictive ability of the direct
and indirect effects of involvement on academic success as well as the impact of student
70
attributes on involvement and success. The following chapter describes the methodology
for this study, which provides evidence for a more complete model of academic success,
featuring student responsibility for engaging in the academic process.
71
CHAPTER III
METHODOLOGY
Introduction
Through decades of research, much is known about the effects of college
attendance on students? cognitive, behavioral and psychological growth (Pascarella &
Terenzini, 2005). Furthermore, in their exhaustive appraisals of college impact studies,
Pascarella and Terenzini (1991, 2005) found very little disparity between institutional
environments, in terms of size and control, and their ability to affect students. Instead,
the major differences in effects on outcome measures were found within each institution
and were attributable to variations in individual students? experiences, such as
involvement with peers and faculty, residence arrangement, involvement in co-curricular
activities and academic major (Pascarella & Terenzini, 1991, 2005). In addition,
Pascarella and Terenzini (2005) concluded that student attributes, specifically self-
concept, also contributed to educational gains. However, student attributes, or the
dimensions of self-concept, self-determination, and self-efficacy, are rarely studied in
combination with student involvement. Moreover, Astin?s (1984) work on student
involvement suggested that student involvement influences educational outcomes.
The purposes of this study were two-fold: (a) to explore how initial student
attributes shape student engagement with the institutional environment and how the
institutional environment influences subsequent attributes; and (b) to understand how
student attributes and institutional environment contribute to academic success as
measured by grade point average at the end of the first college year. This study utilized a
72
causal model to measure the contributions of student attributes and academic and social
involvement on academic success in the first year. As such, this study examined a
longitudinal panel using path analysis to investigate the direct and indirect effects of
student attributes on cumulative grade point average of first-year, traditional-aged
students attending a four-year religiously-affiliated, residential, comprehensive master?s
institution. The research questions guiding this study are:
1. What changes in student attributes occur during the first year of college,
and what environmental factors influence these changes?
2. Controlling for student background characteristics, how do initial student
attributes influence academic and social involvement in the first year of
college?
3. Controlling for student background characteristics and student attributes at
Time 1, how do academic and social involvement impact subsequent
student attributes and academic success?
4. After controlling for student background characteristics and academic and
social involvement, what are the direct and indirect effects of student
attributes, as measured by self-determination, self-efficacy, and self-
concept, on the academic success of first-year college students?
5. Which student attributes construct (i.e., self-efficacy or self-determination)
is a better predictor of academic success in the first year?
This quantitative study adapted the classic input-environment-output model (I-E-
O; Astin, 1977, 1991), the primary model for college impact studies. The I-E-O model
73
provides a formula for statistically measuring effects of the college environment on
educational outcomes while controlling for students? background characteristics at the
time of entry. According to Astin?s (1991) definition of ?inputs,? self-concept, self-
efficacy, and self-determination could be considered attitudes and beliefs students
possess at time of entry. However, constructs for student attributes first enter the model
as intermediary variables between student input and institutional environment for two
reasons: (a) to separate these constructs from input variables of a demographic nature that
may influence these attributes; and (b) to be consistent with assertions that self-efficacy
(Bandura, 1997) and self-determination (Deci & Ryan, 1985; Ryan & Deci, 2002) are the
characteristics that determine how a person engages his or her environment. The
literature (Bandura; Berger & Milem, 1999; Deci & Ryan; Milem & Berger, 1997; Ryan
& Deci) indicates that engaging with the environment may change the nature of a
person?s self-determination and self-efficacy beliefs. Indeed, Lewin (1936) theorized
about behavior as a function of the interaction between a person and his or her
environment, and some of the work stemming from Lewin?s theory (Holland, 1973; Stern
1970) indicates that the person and the environment are equal players. In keeping with
the recommendations of these theories, student attributes constructs enter the model again
as mediating factors, or intermediate outcomes, between the environment and the
outcomes of interest.
This chapter outlines an appropriate research methodology, beginning with a
description of data analyses suitable for answering the research questions. A description
of the conceptual model and the variables used in the study follows. Given that this study
relied upon analyses of existing institutional data, I next describe the survey instruments,
74
the data collection strategies employed by the target institution, and the sample. Finally,
the chapter concludes with a description of the study?s limitations and delimitations.
Data Analysis
According to Ferguson?s (1959) classic text on the use of statistical analysis in
education and psychology, statistical analysis should flow naturally from the research
hypotheses, such that data interpretation ?conforms to the rules of scientific evidence? (p.
2). Therefore, this study utilized path analysis, the data analysis technique used to
determine the direct and indirect effects of self-determination (Guay, Sen?cal, Gauthier &
Fernet, 2003; Reeve, Nix & Hamm, 2003; Vallerand, Fortier & Guay, 1997) and college
student involvement (Berger & Milem, 1999; Milem & Berger, 1997). Path analysis is a
statistical method rooted in multiple regression analysis, which ?allows a researcher to
test a theory of causal order? (Klem, 1995, p. 65) by examining the direct and indirect
effects of independent variables. The conceptual model tested in this study includes both
exogenous and endogenous variables. Only those variables related to student background
characteristics are considered exogenous. All other variables in the model are
endogenous because their values may be explained by other variables or constructs in the
model (Klem).
Descriptive Analysis
In order to prepare for the path analysis, I followed several steps of preliminary
analyses. First, I conducted exploratory data analysis to describe the characteristics of
the analytic sample (Ferguson, 1959) and to ascertain the nature of relationships between
75
variables of interest. Second, in order to test for response bias, I conducted missing data
analysis for pivotal independent and dependent variables to determine whether the
analytic sample differed significantly from the target sample on any relevant input
measure, such as gender, race, first-generation status, or high school achievement, or on
any other relevant measure, such as self-determination. Once I understood the impact of
missing data due to incomplete cases or attrition over time, I determined whether these
missing data posed any problems to interpreting the results of statistical analyses.
Possible methods for managing missing data include using list- or pairwise deletions or
taking steps to impute missing data from cases with complete data (Kline, 2005).
In addition to exploratory and missing data analyses, I conducted a paired samples
t test to answer the first research question regarding possible change in student attributes
during the first year. In order to determine which environmental factors influenced that
change, I utilized linear regression analyses.
Factor Analysis
Several theoretical concepts within the proposed conceptual model (self-efficacy,
self-determination, academic involvement, and social involvement) linked to multiple
survey items. Rather than entering each survey item into the model separately, I used
exploratory factor analysis both to substantiate the existence of the underlying constructs
and to reduce redundancy within the model. Exploratory factor analysis allows the
researcher to uncover constructs for which there is no prior knowledge (Bryant &
Yarnold, 1995). To ensure the factors accurately represent the data, Bryant and Yarnold
recommended computing at least one goodness-of-fit index in order to measure how well
76
the data fit the hypothesized constructs. Although data reduction techniques reduce
multicollinearity among variables in the model, use of these techniques may increase the
difficulty of interpreting the results of path analysis, which relies on observed data to
determine causal relationships (Klem, 1995): First, factors often consist of unobservable
indicators (Klem, 2000); and second, there is no guarantee that factors identified in the
model translate to constructs in the real world (Kline, 2005).
Path Analysis
Conceptual models stem from the literature, which provides the justification for
variable inclusion and ordering, yet the information gleaned from the model depends
largely upon the researcher?s choice of statistical analysis. Noting the proliferation of
studies relying upon multiple regression analysis, Pascarella and Terenzini (2005)
suggested that regression analysis ?provides limited information about the connections
between and among variables and about the causal patterns in which experiences
combine to shape student change? (p. 636). Instead, Pascarella and Terenzini
recommended the use of causal modeling procedures, including path analysis.
Path analysis is a predictive technique, which allows a researcher to test more
than one dependent variable within a single model (Klem, 1995). Given that the
literature-supported research questions about the direct and indirect effects of student
attributes on academic success and the influence of attributes and involvement on one
another, path analysis is an appropriate methodological choice. Although path analysis
does not prove causation, it does provide for increased understanding of a phenomenon
(Klem; Pascarella & Terenzini, 2005)?in this case, the effects of student attributes. Path
77
analytic procedures test the validity or tenability of theory, although more than one model
may be consistent with any theory (Klem; Pedhazur, 1982). Path analysis is a robust test,
requiring a large sample size?200-300 cases are recommended (Klem; Kline, 2005).
As an extension of multiple regression, path analysis maintains the assumptions of
multiple regression: variables in the model are not related to one another, the
measurement of observed variables is accurate, variables are appropriately included in the
model, and no crucial variables are excluded from the model (Asher, 1983; Klem, 1995;
Pedhazur, 1982). Like multiple regression, path analysis also assumes that residuals ?(a)
have a mean of zero; (b) are homoscedastic (i.e., have equal variances at all values of
predictors); (c) are uncorrelated with each other and with the predictors; and (d) are
normally distributed,? (Klem, p. 49). Klem makes explicit the other major assumptions
of path analysis as well. First, path analysis assumes linearity in the model and that all
hypothesized relationships move in one direction. Second, path analysis assumes that
there are no interaction effects between variables.
Path analytic techniques delineate both direct and indirect relationships between
variables (Asher, 1983; Klem, 1995; Pedhazur, 1982). An arrow between two variables
indicates a direct relationship (Klem; Pedhazur) such that the independent variable at the
arrow?s origin predicts, or is hypothesized to predict, the dependent variable at the
arrow?s point. A direct relationship between variables indicates that a change in the
independent variable results in a change in the dependent variable, holding all other
variables constant (Asher, 1983; Sherlin, 2002). An indirect effect indicates that the
relationship between two variables is mediated by one or more additional variables
(Pedhazur). Depending on their placement in the model, independent variables may
78
exhibit direct effects only, indirect effects only, or both direct and indirect effects. The
sum of direct and indirect effects for any given variable is the total effect (Klem). In
addition, spurious effects arise when two endogenous variables have a common cause,
and unanalyzed effects arise when two exogenous variable are correlated with one
another (Klem). The sum of all four effects (direct, indirect, spurious, and unanalyzed) is
the implied correlation between two variables (Klem; Pedhazur), which can be compared
to the observed correlation to determine model plausibility (Klem).
To calculate the direct and indirect effects of complicated path models, Klem
(1995) suggested using software packages, such as LISREL, EQS, or AMOS, specifically
designed to manage the analysis. The software packages provide advantages in
automated calculation of effects and goodness of fit. However, as the model or this study
is exploratory in nature and relatively straightforward, I utilized a succession of multiple
regressions as the preferred path analytic technique. Multiple regression analysis was the
preferred method of both Asher (1983) and Klem in their explanations of path analysis:
?The simplest ways to obtain the path coefficients is to employ ordinary regression
techniques, providing the regression assumptions are met,? (Asher, p. 30).
Direct effects were obtained through a multiple regression analysis for each
endogenous variable. Significant direct effects were noted, and then indirect effects were
calculated as compound paths (Asher, 1983; Pedhazur, 1982). The indirect effect is the
product of direct paths between one independent variable and the dependent variable as
mediated by one or more intervening variables (Pedhazur). For example, while variable
1 (independent) may not have a direct effect on variable 3 (dependent), the two might
have an indirect relationship mediated by a third variable. The indirect effect of variable
79
1 on variable 3 may pass through variable 2 (mediator), which has a direct effect on
variable 3. The indirect effect of variable 1 on variable 3, then, is be the product of direct
effects between variables 1 and 2 and between variables 2 and 3 (IE
31
= DE
21
x DE
32
).
The indirect effects were added to the direct effects to determine the total effect of each
variable in the model.
Conceptual Model
According to Klem (1995), the conceptual model is the explicit representation of
the researcher?s theory about the causal relationship among variables. Theory indicates
which variables a researcher should include in a model for path analysis as well as the
specific ordering of those variables. The previous chapter outlined the theoretical
concepts, which are the foundation for this research. Astin?s (1984) theory of
involvement suggests that the quantity and quality of students? involvement with the
university environment have tremendous impact on educational outcomes. Research
confirms that involvement is related to gains in self-confidence and cognitive
development, communication and interpersonal skills (Huang & Chang, 2004) and
predicts academic and social integration and persistence (Berger & Milem, 1999; Milem
& Berger, 1997). However, why some students get involved and others do not is less
understood. Borrowing from social psychology, the concepts of self-efficacy (Bandura,
1977, 1982, 1997) and self-determination (Deci & Ryan, 1985; Ryan & Deci, 2000,
2002), and self-concept may be just the tools to explain why some students exert personal
agency over their college experience to greater degrees than others. The conceptual
model for this study was supported by these theoretical frameworks and blended the
80
learning from the effects of involvement theory and student attributes on academic
success.
In an earlier version of the conceptual model, self-efficacy and self-determination
were isolated in their own block at Times 1 and 2. Although no temporal difference
exists between these constructs and academic and social self-concepts, self-efficacy and
self-determination appeared separately because they are the primary focus of the
investigation. However, the correlation study described in Chapter Four indicated that
the available measures for self-efficacy were consistently correlated with academic self-
concept, which violated the assumption of no multicollinearity within path analysis.
Therefore, one of these variables had to be dropped from the model. Given that the
factors describing self-efficacy were theoretically weaker and less robust than those
representing academic self-concept, I removed self-efficacy from the model. With self-
efficacy eliminated from the model, academic and social self-concepts rose to the
foreground as measures of student attributes. Self-concept, then, had parity with self-
determination. This change still allowed for an investigation of the direct and indirect
effects of student attributes on academic success, without violating the ?no
multicollinearity? assumption of path analysis. The revised conceptual model (Figure
3.1) demonstrates the new placement of academic and social self-concepts in the block
with self-determination at Times 1 and 2.
In this model, student background characteristics in block one pertain to gender,
race, first-generation status, high school grade point average, and SAT composite. Path
analysis indicated the direct and indirect effects of these variables on those in all
subsequent blocks. In block two, student attributes refer to academic self-concept, social
81
self-concept, and self-determination, and analysis demonstrated their direct and indirect
effects on the first-year experience, student attributes in Time 2, and college GPA. First-
year experience variables in block three include enrollment in a first-year program,
faculty interaction, academic engagement, homework time-on-task, participation in
student clubs and organizations, and sense of belonging. Direct and indirect effects of
variables in block three are measured for student attributes at Time 2 and college GPA.
With the elimination of self-efficacy, student attributes in the fourth block now contain
academic and social self-concept along with self-determination. Path analysis is used to
determine the direct effects of these variables on college GPA, the dependent variable.
At its core, the conceptual model in Figure 3.1 is an input-environment-outcome
(I-E-O; Astin, 1991) assessment model, which flows from left to right as required by path
analysis (Klem, 1995). In some ways, the model is reminiscent of Tinto?s (1993)
Longitudinal Model for Student Departure in that the variables for both models share a
similar sequence. As with Tinto?s model, the conceptual model guiding this study
outlines student background characteristics as a pre-cursor to understanding personal
student attributes that are critical to the collegiate experience. Both models also
emphasize interaction with the environment as a predictor of student change. In Tinto?s
model, this change is described as a chance for academic and social integration followed
by a confirmation or rejection of goal commitments. In the conceptual model for this
study, student change is marked by the development, or growth, of student attributes
82
S
t
u
d
e
n
t
B
a
c
k
g
r
o
u
n
d
C
h
a
r
a
c
t
e
r
i
s
t
i
c
s
:
G
e
n
d
e
r
R
a
c
e
F
i
r
s
t
-
g
e
n
e
r
a
t
i
o
n
H
i
g
h
S
c
h
o
o
l
G
P
A
C
o
m
p
o
s
i
t
e
S
A
T
S
t
u
d
e
n
t
A
t
t
r
i
b
u
t
e
s
:
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
S
o
c
i
a
l
S
e
l
f
-
c
o
n
c
e
p
t
S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
F
i
r
s
t
-
y
e
a
r
E
x
p
e
r
i
e
n
c
e
:
F
i
r
s
t
-
y
e
a
r
P
r
o
g
r
a
m
s
F
a
c
u
l
t
y
I
n
v
o
l
v
e
m
e
n
t
E
n
g
a
g
e
m
e
n
t
H
W
T
i
m
e
-
o
n
-
t
a
s
k
S
t
u
d
e
n
t
C
l
u
b
s
B
e
l
o
n
g
i
n
g
S
t
u
d
e
n
t
A
t
t
r
i
b
u
t
e
s
:
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
S
o
c
i
a
l
S
e
l
f
-
c
o
n
c
e
p
t
S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
A
c
a
d
e
m
i
c
S
u
c
c
e
s
s
:
C
o
l
l
e
g
e
G
P
A
F
i
g
u
r
e
1
.
R
e
v
i
s
e
d
C
o
n
c
e
p
t
u
a
l
M
o
d
e
l
:
S
t
u
d
e
n
t
A
t
t
r
i
b
u
t
e
s
a
n
d
A
c
a
d
e
m
i
c
S
u
c
c
e
s
s
i
n
t
h
e
F
i
r
s
t
Y
e
a
r
o
f
C
o
l
l
e
g
e
I
n
p
u
t
s
A
t
t
r
i
b
u
t
e
s
T
i
m
e
1
E
n
v
i
r
o
n
m
e
n
t
s
A
t
t
r
i
b
u
t
e
s
T
i
m
e
2
O
u
t
c
o
m
e
s
T
i
m
e
83
at Time 2. Finally, both models are longitudinal, expressing some length of time between
student?s input characteristics and the outcome.
The model guiding this study is recursive, such that all causal links are mono-
directional, as recommended by Kline (2005) and Klem (1995). Beginning with the input
block on the far left, the remainder of this section explains the logic behind the model.
Inputs
Input variables are those characteristics a student possesses at time of entry
(Astin, 1991); they are the qualities students bring to their collegiate experience. In
college impact research, the following are widely recognized as typical input variables
(Astin; Pascarella, 1985; Tinto, 1993): family and personal traits, such as gender,
race/ethnicity, and first-generation status; and academic preparation, such as high school
academic program and grade point average and college entrance exam scores. All of
these were classified as input variables in the proposed model.
Student Attributes at Time 1
Although all pre-entry attitudes and beliefs generally belong with inputs, for the
purposes of this research, those measures constituting the self-efficacy and self-
determination constructs are assigned their own block, entitled ?Student Attributes, Time
1.? The ?Time 1? designation indicates that these attribute measures were taken from the
pretest survey data. The nature of the proposed research is to determine what role student
attributes play in academic and social involvement, and therefore, student attributes were
isolated in their own block. The measures for academic and social self-concept were
84
comprised of students? self-assessments of ability. The self-determination variable is a
composite of student self-perceptions of autonomy, competence, and relatedness, as
suggested by Deci and Ryan (1985; Ryan & Deci, 2000, 2002). The literature (Brown,
Lent, & Larkin, 1989; DeWitz & Walsh, 2002; Deci, Ryan, & Williams, 1996; Deci,
Vallerand, Pelletier, & Ryan, 1991; Leslie, McClure, & Oaxaca, 1998; Miserandino,
1996; Multon, Brown, & Lent, 1991) suggests that these measures of student attributes
have implications for how a person engages his or her environment. As such, student
attributes at Time 1 directly precede the variables within the first-year experience block.
Environments
According to Astin (1991), environmental variables represent aspects of
involvement in the educational experience. The environment has academic and social
components (Astin, 1993; Tinto, 1993), and these components have both formal and
informal configurations (Tinto). The literature on student involvement suggests
important environmental variables including campus residence, co-curricular activities,
employment (Astin, 1984); interactions with faculty and peers (Astin, 1999); and
participation in programmatic interventions for the first year, such as first-year seminars
and learning communities (Chrissman Ishler & Upcraft, 2005). Although the literature
suggests a benefit from residential programs (Astin, 1993; Upcraft, 1989) in the first year
of college, this component is not included in the model because of the highly-residential
nature of the target institution.
Involvement is also described as having quantitative and qualitative components
(Astin, 1984), such that student involvement is reflected in the amount of time and energy
85
students devote to educational activities as well as the quality their efforts (Astin, 1984;
Pace, 1984). This aspect of the first-year experience measures how much time college
students devote to the academic enterprise and how well they engage their classes and
coursework. Variables such as number of hours spent on homework each week and the
average number of times students missed or fell asleep in class all describe the quality of
student effort. An effort construct is necessary to the model because as Astin stated,
?The concepts of time-on-task and effort, for example, appear frequently in the literature
as key determinants of a wide range of cognitive learning outcomes? (p. 305).
Student Attributes at Time 2
Lewin?s (1936) work suggests that behavior results from an interaction between a
person and his or her environment. Aside from affecting how a person interacts with his
or her environment, the literature also indicates that interaction with the environment can
affect a person?s self-determination (Ryan & Deci, 2000). Therefore, the model includes
a second measure of the student attributes composite variables. As student attributes at
Time 2 are the posttest measures of student attributes at Time 1, the model acknowledges
the possibility of change over time, indicated either by increasing or decreasing levels of
academic self-concept, social self-concept, and self-determination. If change occurs,
statistical analyses help determine what part of attribute change is due to environmental
involvement factors represented in the model. Time 2 measures were taken from the
posttest survey data.
86
Variables
Using a combination of survey and institutional data, this study identified a
dependent outcome variable describing academic success and two sets of intermediate
outcome variables: those related to academic and social involvement in the first year and
those related to student attributes at Time 2. The dependent outcome variable is discussed
below; all other variables follow in sequential order.
Dependent Variable
Academic Achievement
Academic achievement is measured by cumulative grade point average for the
2004-2005 academic year. Grades are an often-acknowledged standard of academic
success (Astin, 1993; Upcraft, Gardner, & Barefoot, 2005). The institution provided
information for students? grade point averages.
Exogenous Variables
Student Background Characteristics
In following Astin?s (1993) I-E-O model, the hypothesized causal model accounts
and controls for fixed input characteristics. Variables related to student background
represent demographic information, including gender, race, and first-generation college
student status (a composite indicating parents? educational attainment). Input variables
also include measures of aptitude (high school GPA and college entrance exam scores).
All items pertaining to student background characteristics were located in the CIRP
87
Student Information Form, with the exception of the college entrance exam scores, which
were provided by the institution.
Endogenous Variables
Self-concept (Academic and Social)
Academic and social self-concepts are composites that were constructed from
self-assessment items within the CIRP (Higher Education Research Institute [HERI],
2004) and Your First College Year (YFCY, HERI, 2005) surveys. Exploratory factor
analysis with students? self-ratings determined the parameters for academic and social
self-concepts. Items such as academic ability, intellectual self-confidence, writing
ability, and public speaking ability (Berger & Milem, 2000) were expected to load onto
academic self-concept. Additional items such as computer skills, mathematical ability,
and drive to achieve were also tested for inclusion in the academic self-concept construct.
Items such as leadership ability, popularity, social self-confidence, and understanding of
others were expected to comprise the construct for social self-concept (Pascarella, Smart,
Ethington, & Nettles, 1987). Each of the items listed above was measured on a Likert-
type scale, appeared in pretest form on the Cooperative Institutional Research Program?s
(CIRP) Freshman Survey, and had a direct posttest measure in the Your First College
Year follow-up survey (YFCY).
Self-determination
The items in this category refer to the student?s self-determination at Times 1 and
2. Data related to self-determination were derived from the portion of the Student
88
Readiness Assessment (SRA, Magis University, 2004) that incorporated the Basic
Psychological Needs Survey (BPNS, Deci & Ryan, 2000). These items included student
self-assessments of autonomy, competence and relatedness. Self-determination at Time 1
contributed to the understanding of the direct effect of this student attribute on academic
and social involvement during the first year, while the Time 2 construct measured change
over time and the influence of the college environment over student attributes. All
responses for these items were rated on a Likert-type scale.
First-year Experience
Items in this block refer to academic involvement and social involvement in the
first year of college. Academic and social involvement were placed in the same block
because they both describe the environment of the first-year experience. Academic
involvement was comprised of programmatic interventions in which students may be
enrolled, such as first-year seminars or learning communities, as well as a combination of
composite factors derived from variables describing the student?s interaction with the
institution?s intellectual environment. Previous research (Milem & Berger, 1997)
supports developing factors of academic involvement, which include involvement with
faculty and academic engagement. The variables comprising these factors mapped to
items in the YFCY survey and recorded students? self-assessments, frequency and type of
interaction with faculty, and behaviors associated with intellectual engagement.
Responses were recorded on Likert-type scales. Academic involvement also comprises
quality of effort and time-on-task items obtained from the YFCY survey. The student
effort questions pertain to student self-assessments of number of hours spent studying;
89
frequency of submitting marginal work; and frequency of non-engaging behaviors, such
as feeling bored or falling asleep in class. Together, these variables comprise composites
of academic engagement and time-on-task student effort. All student effort questions are
rated on a continuous scale.
Social involvement refers to students? participation in the social environment of
the institution. Like academic involvement, social involvement is a composite factor
derived from survey items. Previous research (Milem & Berger, 1997) supports
developing factors of social involvement that include involvement with peers, sponsored
student clubs and organizations and athletics (varsity, club-level, and intramural) as well
as affiliation with the institution. The variables comprising these factors mapped to items
in the YFCY survey which recorded students? self-appraisal of interactions with peers,
participation in activities, participation in athletics and organizations, success with
building social relationships, and satisfaction with sense of community. Responses were
recorded on Likert-type scales.
In addition to their properties as independent variables affecting educational
outcomes, variables describing the behavioral aspects of students? academic and social
involvement also appear in the proposed model as intermediate outcome variables
influenced by student background characteristics and student attributes at Time 1.
The operative model for this study, as displayed in Figure 3.1, includes four
categories of independent variables affecting, either directly or indirectly, the dependent
variable: academic achievement. A complete list of variables in this study, their
associated theoretical concepts and their correspondence to specific items on the survey
90
instruments can be found in Appendix A. (For the specific survey instruments, see
Appendices B, C, and D.)
Instrumentation
To explore factors which influence achievement and self-concept, Marsh (2003)
recommended a two-wave research design. Therefore, this study drew upon data from
four survey instruments as part of a longitudinal design. These data were collected at a
target institution as part of its on-going assessment activities. Two of the instruments
were paired surveys from a national study of first-year students, with one survey
administered at the beginning of fall semester and the second survey administered near
the end of spring semester. The additional surveys, measuring self-determination, were
administered concurrently with the other two in a pre-/posttest design. The target
institution granted permission to analyze its data from the 2004 Student Information
Form, the 2005 Your First College Year, and Student Readiness Assessment pre- and
posttests gathered from its first-year students during the 2004-2005 academic year. In
addition, the institution agreed to provide institutional data for the first-year cohort as a
comparison against which to evaluate the analytic sample.
Cooperative Institutional Research Program: Freshman Survey
The target institution is a regular participant in the Cooperative Institutional
Research Program (CIRP), which produces the Student Information Form, an annual
study of the nation?s first-time, first-year college students. Sponsored by the Higher
Education Research Institute (HERI) at the University of California-Los Angeles, the
CIRP Freshman Survey records information about student characteristics, goals, attitudes,
91
and expectations (http://www.gseis.ucla.edu/heri/cirp_survey.html). The Freshman
Survey serves two purposes: first, the data it generates stand alone as an annual snapshot
of the nation?s entering college students, which provide a year-to-year comparison for
trend analysis; and second, the Freshman Survey is one component of a series of related
surveys, which provide longitudinal data for college impact studies (Astin, 1993). In
addition to compiling a national profile of incoming first-year students each year, HERI
provides participating institutions with their own students? data for institutional
assessment and research (http://www.gseis.ucla.edu/heri/cirp_po.html). The 2004
instrument (Appendix B) contained 63 items, including 21 institution-generated questions
and took approximately 20 to 30 minutes to complete.
The CIRP Freshman Survey is the nation?s oldest and largest empirical study of
first-year college students, and the work of HERI researchers spans over three decades
(http://www.gseis.ucla.edu/heri/cirp_survey.html; Miller, Bender, & Schuh, 2005;
Upcraft, Gardner, & Barefoot, 2005). Although HERI staff no longer publish their
methods for establishing reliability and validity?including content, predictive, and
construct validity?for the Freshman Survey, the body of knowledge their research
generates provides substantial evidence that the Freshman Survey is both valid and
reliable. Data from the Freshman Survey comprise the annual profile of students reported
in The American Freshman and have provided the data for Alexander Astin?s volumes
Four Critical Years: Effects of College on Beliefs, Attitudes, and Knowledge (1977) and
What Matters in College: Four Critical Years Revisited (1993) as well as hundreds of
journal articles, books, book chapters, and dissertations.
92
The target institution administered the CIRP Freshman Survey on the first day of
new student orientation in September 2004. In groups of 10 to 15, first-year students met
with orientation leaders who read survey instructions, distributed surveys and pencils,
and then collected completed surveys; the orientation leaders returned completed surveys
to their program director. Typically all first-year students, except varsity athletes
participating in fall sports, attend orientation at this institution; of 960 first-time, first-
year students, the response rate for the fall 2004 administration of CIRP was 92% of all
first-year students.
Student Readiness Assessment
In addition to the CIRP Freshman Survey, the orientation leaders distributed
informed consent forms and administered the university?s Student Readiness Assessment
(Appendix C) to their groups of first-year students. The Student Readiness Assessment
(SRA) is a combination of the Basic Psychological Needs Scale (BPNS; Deci & Ryan,
2000) and institution-sponsored administrative data.
Self-determination Theory posits that the basic psychological needs of autonomy,
competence and relatedness must be satisfied in order for an individual to be self-
determined (Deci & Ryan, 2000). The BPNS consists of 21 items on a Likert-type scale
ranging from 1 (not at all true) to 7 (very true), which measure these three psychological
needs central to Self-Determination Theory (SDT; Deci & Ryan, 1985; Ryan & Deci,
2000, 2002). The BPNS contains three subscales, one for each psychological need, with
seven items per subscale, allowing the researcher to obtain separate scores for each need
or to determine general need satisfaction by combining data from the three subscales
93
(http://www.psych.rochester.edu/SDT/measures/needs.html). Deci and Ryan established
validity and reliability for the BPNS in their University of Rochester laboratory, and
researchers have utilized the scale in a wide variety of domains, including education,
health care, athletics, and work. The BPNS is available on-line at:
http://www.psych.rochester.edu/SDT/measures/needs_scl.html.
Institutional researchers obtained permission to utilize the scale in institutional
research from its authors. The BPNS was not modified in terms of language, question
ordering or response scale. However, to answer other institutional questions eight
additional items addressing parental involvement, perceptions of academic preparation,
and place of college residence were added to the instrument. The additional items were
pilot tested with four students at the institution, who provided feedback about survey
length and question content. These additional questions combined with the original
BPNS constitute the SRA, which has a total of 29 items. The SRA took an average of 10
to 15 minutes to complete. The response rate for the pretest version of the survey was
92%, the same as that for CIRP. The posttest response rate was 38%.
Cooperative Institutional Research Program: Your First College Year
Designed by HERI in conjunction with the Policy Center on the First Year of
College, Your First College Year (YFCY) assesses the personal and academic growth of
students during their first year of college (http://www.gseis.ucla.edu/heri/yfcy.html). As
a follow-up to the Freshman Survey, YFCY collects data related to students?
development since their matriculation and allows assessment of institutional programs
and practices (http://www.gseis.ucla.edu/heri/yfcy/survey_instrument.html). Content
94
areas include academic achievement, behavior, persistence, and involvement, many of
which directly correspond to questions on the Freshman Survey. After establishing
validity and reliability for YFCY through pilot tests, HERI began administering YFCY
nationally in 2000.
The 2005 version of Your First College Year (Appendix D) contained 56 items,
including 30 institution-specific questions and took an average of 20 minutes to
complete. Similar to its practices with results from the CIRP Freshman Survey, HERI
provides participating institutions with their students? YFCY data for institutional
assessment and research (http://www.gseis.ucla.edu/heri/yfcy.html). Used in conjunction
with one another, the CIRP and YFCY surveys are powerful tools for assessing
development and change in the first year of college (Delaney, 2004; Upcraft, Gardner, &
Barefoot, 2005). The host institution granted me authorization to analyze YFCY data,
along with SRA data, gathered from its first-year students during the spring of 2005.
Your First College Year is part of the second wave of data collection along with
the second administration of the SRA in April 2005. First-year students received a letter
from the institution?s president about the importance of participating in the second round
of surveys during the first week of April. During the second week of April, students
received the YFCY in their campus mailboxes along with the second issue of the SRA, a
second consent form, and a cover letter from Institutional Research. This mailing was
followed by an e-mail announcement from their class president reminding them to
complete the survey. Over a three-week period, students returned their completed
surveys to advertised drop boxes, during which time reminders about the survey appeared
in the student newspaper, on the campus radio station, in the daily electronic news
95
bulletin, and by email. As an incentive for students to return their surveys, the institution
raffled two $100 cash prizes and one iPod Shuffle? each of three weeks; only students
who had completed their surveys were eligible. In an additional effort to boost the return
rate, the institution sponsored an incentive program for graduate assistants living in the
first-year residential areas, such that graduate assistants whose residential areas had
greater than an 85% return rate were eligible for a cash prize of $100.
The response rate for YFCY and the BPNS posttest was 38%. The institution
attributed participant attrition to the change in method for survey distribution/collection.
Whereas in September the institution employed a ?captured audience? tactic, in April the
institution relied on voluntary participation by mailing the instruments to students along
with a letter requesting their participation.
The institution sent a data file containing SRA pre- and posttest data, cumulative
grade point averages, SAT scores, and fall 2006 enrollment for each student to HERI to
be merged with CIRP and YFCY data. The students? university identification number
was the merge point. Once all data were merged, HERI stripped students? identification
numbers from the file in order to protect students? anonymity before returning the file to
the institution.
Sample
For the purpose of this study, I had access to data, which were part of a
longitudinal study of first-year student outcomes and persistence funded by the Division
of Academic Affairs at a selective and highly residential, comprehensive university in the
mid-Atlantic region of the United States. The target sample was the fall 2004 entering
96
first-year class of 960 students. Consistent with national gender ratios for first-year
students matriculating in fall 2004 (2004 CIRP Institutional Profile), the sample included
52% women students and 48% men, which promotes the generalizability of findings to
first-year students at liberal arts colleges and universities of similar size and selectivity.
All members of the fall 2004 entering first-year class received four survey
instruments, two pretests on the first day of new student orientation and two follow-up
surveys in April 2005. The longitudinal panel for data analysis consists of participants
who reasonably completed all four surveys (N = 202). On account of the number of
dependent and independent variables of interest and participant attrition between the first
survey administration in September 2004 and the second in April 2005, the analytic
sample includes all members of the first-year class who completed each survey
instrument. In order to maintain reasonably high survey response levels, the target
institution employed a process for systematic data collection. Although I could not
control participant attrition as this research involves secondary data analysis, descriptive
analyses allowed me to investigate whether particular types of students were more, or
less, likely to respond to all four surveys.
Delimitations
The target institution, hereafter referred to as ?Magis University,? is a mid-sized
comprehensive university. Magis University is one of 28 Jesuit-Catholic colleges and
universities in the United States. The student body of Magis University is predominantly
residential, of traditional age, and White. This study is situated in Jesuit higher education
97
for four reasons?one of which was related to research design, two were theoretical, and
the fourth reason was practical.
First, at its core, this study was an examination of the interaction of student
attributes and educational environment in the production of educational outcomes.
Different environmental factors such as size, geographic location, control, and selectivity
all affect student success, and variations in environment lead to variations in student
behavior (Baird, 1988; Kuh, 1993; Upcraft & Gardner, 1989). Locating the study in a
single institution highlighted individual student factors by eliminating environmental
differences that might otherwise influence the outcome and blot out any contributions
made by student attributes. Pascarella and Terenzini (1998) suggested that localized,
single-institution studies are more useful in guiding the understanding of how theoretical
concepts manifest themselves on an individual level. The strength of a single-institution
study is in the ability to disentangle the effects of student characteristics from the
environment. Teasing out the effects of student attributes, contributes to a better
understanding of how the individual functions in relation to academic success prior to
building a multi-institutional model. I delimited the study to a particular institution in
order to isolate an educational environment so as to focus data analysis on differences in
student attributes.
Pascarella and Terenzini (2005) provided further justification for choosing to
locate the study at one institution when they concluded:
On just about any outcome, and after taking into account the characteristics of
students enrolled, the dimensions along which American colleges and universities
are typically categorized, ranked, and studied, such as type of control, size, and
98
selectivity, are simply not linked with important differences in net impacts on
student learning, change, or development. Despite structural and organizational
differences, institutions are more alike than different in their effects on
students?. Rather?what happens to students after they enroll at a college or
university has more impact on learning and change than the structural
characteristics of the institution these students attend. (pp. 641-2)
In other words, the net effects of college attendance are similar across all institutions, and
in a study of the student experience, institution type and size do not play a leading role.
Therefore, a study examining the impact of student attributes on academic success at one
institution contributes to the understanding of students, their learning, and their
development.
The second justification for choosing a single-institution study is related to Magis
University?s Jesuit nature. Jesuit-Catholic higher education has a particular educational
philosophy rooted in academic excellence (Biondi, 1989; Donohue, 1963; Flynn 1989;
McInnes, 1989) and dedication to service and formation of the human person?
spiritually, intellectually, socially, and physically (Biondi). According to Biondi, Jesuit
higher education aims to affirm the dignity of the human person in the likeness of God;
develop intellectual discipline, including critical thinking, language and problem-solving
skills; and improve a student?s ability to lead and serve his or her community. Such a
philosophical approach is congruent with the development of the whole student, and it is
consistent with Kuh?s (1991) description of ?involving colleges.?
The first principle of Kuh?s involving colleges is related to institutional mission.
Kuh (1991, 1993) stipulated that involving colleges have clear missions that are
99
understood by all institutional members and that communicate high expectations for
student performance. The mission, then, becomes a contract of sorts between the
institution and its students. As a statement about the institutional values and goals, the
mission is a philosophical representation of the ideal educational experience for its
students; the mission says both ?what the institution should be expected to do, but also . .
. what its students should be expected to do on their own behalf? (Tinto, 1993, p. 145).
Compared to other institution types, institutions with liberal arts-focused missions, such
as Magis University, tend to have challenging academic programs that are intensive in
reading and writing, a greater sense of community, and a higher degree of student-faculty
interaction (Kezar & Kinzie, 2006), and are related to increases in cognitive and personal
development (Kuh, 1993).
As a Jesuit institution, Magis University shares a common foundation in academic
excellence (Biondi, 1989; Donohue, 1963; Flynn, 1989; McInnes, 1989) and a dedication
to service and formation of the human person?spiritually, intellectually, socially and
physically (Biondi, 1989).
The second criterion for involving colleges is a ?human-scale? campus
environment that also uses its location for a positive educational advantage (Kuh, 1991).
In other words, the campus environment feels good and welcoming. Magis University
prides itself on the 14:1 student-faculty ratio and its well-developed first-year programs,
which connect students to their faculty advisors inside and outside the classroom.
Third, involving colleges value student involvement (Kuh, 1991). One of the
hallmarks of Jesuit education is the commitment to development of the whole person, or
cura personalis (Flynn, 1989). Active development requires participation on the part of
100
the institution and on the part of the student. The commitment to cura personalis
requires Jesuit colleges and universities to involve students in their own growth on many
different levels, including intellectual, spiritual and social.
Adherence to the final two values of involving colleges (mission-consistent
policies and practices and faculty and staff who promote a seamless learning
environment) emerges from the Jesuit tradition of care for the whole person. Jesuit
institutions attract faculty, staff and administrators who understand their obligation as
educators ?to stir the minds and feelings of our students to consider how they should live
in our society?? (Biondi, 1989, p. 99). This educational responsibility is as important
outside the classroom as it is inside. Similarly, for Jesuit higher education to be effective,
institutional practices and policies must reflect the philosophy that informs the mission.
In these ways, Jesuit higher education as a whole strives to embody the characteristics of
involving colleges on the basis of its spiritual tradition. For these reasons, Jesuit higher
education was chosen as the context for this study.
The third reason for selecting a single institution in which to study the first year of
college was related to Magis University in particular. Researchers at the Policy Center
on the First Year of College identified 20 college initiatives and programs that
contributed to excellence in the first year (Barefoot et al., 2005); Magis University had
incorporated 15 of these 20 programmatic interventions, including an advising program
designed for the first year, a common text, academic convocations, first-year seminars, a
core curriculum with a liberal arts foundation, strong leadership and orientation
programs, peer mentoring programs, learning communities, and first-year residences. By
delimiting this study to an institution that met criteria for engaging first-year students, the
101
investigation allowed for more specific attention to students? disposition toward academic
success.
The fourth, and final, reason for situating this study in a single Jesuit institution
was practical. Simply stated, the sample for this study was convenient to obtain. As I am
a long-time associate in the Jesuit network of institutions, I have collegial relationships
with administrators at Magis University. On the basis of our association, Magis
University was willing to grant access to its data.
Limitations
Although the particularized locale allows for deeper understanding of how
students interact with a specific educational environment, it also limits the
generalizability of any results to students at this institution, other Catholic institutions, or
possibly at liberal arts colleges. Furthermore, this research examines traditional-aged,
first-year students at one selective, four-year institution with a student body that does not
reflect the racial composition or socioeconomic levels of college students nationwide;
Magis University draws a student population that is 87% White and economically
privileged. Any attempt to generalize the results of this study beyond Catholic colleges
and universities or liberal arts institutions of similar composition could be a problem.
Typically, quantitative analyses of survey data allow generalizations of results to
portions of the larger population (Creswell, 2003). The survey process allows the
researcher to collect a large volume of data in a short amount of time with relative ease.
Additionally, the researcher can re-administer survey instruments to collect longitudinal
data (as is the case in this study design) without over-taxing participants. The survey
102
process does have its disadvantages, however. As Schwarz (1999) warned, self-reported
data may be less reliable if participants interpret questions differently from one another or
from the researcher, if questions require participants to rely on their long-term memory
for behavioral reports, or if the context for attitudinal questions is murky and unclear.
However, I lessened the effects of student self-reporting by using data collected from
solid survey instruments that have been used in multiple studies and were either
nationally normed and piloted (CIRP and YFCY) or developed specifically to test theory
(the BPNS portion of the SRA). 
Also, although path analysis is a powerful statistical method for determining
direct and indirect effects of exogenous and endogenous variables, it will not confirm or
disprove a particular conceptual model; instead path analysis only allows the researcher
to determine whether or not the data fit the proposed model (Klem, 1995). Klem warned
the researcher to consider the possibility that several models, other than the one specified
in the study, may also fit the data. However, the model proposed for this research tests
relationships as suggested by theory. Whether or not another model may fit does not
detract from the usefulness of testing a theoretically supported model.
The final limitations stem from the longitudinal design of this study. First,
although this study was longitudinal by design, data collection only spanned participants?
first year of college. It is possible that the effects of student attributes or environmental
interventions took longer than one academic year to emerge (Feldman & Newcomb,
1994; Pascarella & Terenzini, 1991), or conversely that any observed change in student
attributes is related to maturity (Pascarella & Terenzini). In fact, Terenzini (1994)
warned that it is unreasonable to expect significant change within the first two years of
103
college. Second, although the institution employed participation incentives, respondent
attrition occurred at a rate of 62%. Matching cases in the study?s four-survey design
further reduced the analytic sample to 259 cases, and the analytic sample was reduced
further as 57 cases were deleted for incomplete data (total participant attrition of 78%). It
is reasonable to anticipate that this level of attrition would cause the analytic sample to
differ from the target sample in meaningful ways. In order to address this concern, I
conducted missing data analysis.
Despite the limitations associated with survey research and quantitative
methodologies, the advantages of undertaking this study to address its research questions
outweighed the disadvantages. No single study can fully explain a theory or conceptual
model. However, this research was a worthy pursuit because it suggests answers to
important questions about the effects of student attributes on educational outcomes.
Whereas typical outcomes-based research focuses on environmental interventions at the
institution level, this study contributes to the understanding of which student factors may
cause different students to benefit from the environment at differing rates.
Summary
Higher education pundits (Kuh, Kinzie, Schuh, & Witt, 2005; Pascarella &
Terenzini, 2005; Upcraft, Gardner, & Barefoot, 2005) persuasively reasoned that college
students bear responsibility for involvement in the educational enterprise. Yet college
impact studies typically focus on institutional efficacy to the exclusion of the effects of
student attributes toward achieving measurable educational outcomes. The purpose of
104
this study was to determine the impact of college student attributes relative to academic
success in the first year.
The sample for this study included first-year students at a religiously-affiliated
comprehensive university in the mid-Atlantic region. A longitudinal panel of these
students? responses to a series of two national and two institutional surveys comprised the
dataset. To account for the personal and institutional factors leading to academic success
in the first year of college, this study utilized a path analytic model to answer the research
questions guiding it. Path analysis was an appropriate statistical technique for evaluating
the direct and indirect effects of student attributes and institutional environment on
academic success.
The following chapter presents the results of the descriptive, exploratory factor,
and missing data analyses, as well as the correlation study and the research questions.
105
CHAPTER IV
RESULTS
Introduction
This study investigated how student attributes influence academic success in the
first year of college. Specifically, descriptive analyses were used to explore
characteristics of the analytic sample, and multivariate analytic techniques were used to
understand how student attributes change over time. Finally, path analytic techniques
were used to examine the direct and indirect effects of academic self-concept, social self-
concept, and self-determination on cumulative grade point average. Changes in student
attributes over time were also investigated along with factors that influence that change.
Prior to addressing the research questions, results of the exploratory factor
analyses that created the variables for the model and the bivariate correlation study are
reported. In addition, the final path model is presented, and missing data analysis is
discussed. Finally, findings related to each research question are discussed.
Exploratory Factor Analysis
Principle components factor analysis was used to create factors that describe
student attributes at Times 1 and 2 as well as measures of the college environment. Each
factor is described in the pages that follow.
106
Student Attributes at Time 1
Three composite variables were created to measure student attributes that may
affect college grade point average, either directly or indirectly, at Time 1. These include
academic self-concept, social self-concept, and self-determination.
Measures of Self-concept
Self-concept refers to a student?s self-assessment of ability as compared to his or
her peer group (Pascarella & Terenzini, 2005). For the purposes of this study,
exploratory factor analyses were employed to create measures of academic and social
self-concept using several items on the 2004 Cooperative Institutional Research Program
(CIRP) Freshman Survey. Twenty-four survey items queried students to rate themselves
on various traits related to different aspects of self-concept. All items were measured on
a Likert-type scale (1 = Lowest 10%, 2 = Below Average, 3 = Average, 4 = Above
Average, and 5 = Highest 10%). Of these 24 traits, four individual items loaded onto a
construct describing academic self-concept. These traits (intellectual self-confidence,
academic ability, drive to achieve, and writing ability) captured this comparative
academic self-assessment. Table 1 summarizes the results of the principle components
factor analysis. Factor loadings for ?Academic Self-concept, Time 1? were moderately
high as was the factor reliability (
 = .649). Similarly, 4 of the 24 trait items represent
social self-concept, including self-assessments of compassion, generosity,
cooperativeness, and understanding of others. This factor had a moderately high
reliability (
 = .685) and factor loadings for each item were moderately high to high
107
(Table 2). Appendix E details a complete listing of the 24 trait items and the factors to
which they belong.
Table 1
Components of the Factor Composite: Academic Self-concept at Time 1
Factor Components Loadings
Self-confidence (intellectual) .776
Academic Ability .746
Drive to Achieve .651
Writing Ability .632
Alpha reliability coefficient .649
Source: Analyses of CIRP:04, Magis University
Note. Items asked respondents to: ?Rate yourself on each of the following traits as compared with the
average person your age.?
Table 2
Components of the Factor Composite: Social Self-concept at Time 1
Factor Components Loadings
Compassion .768
Generosity .759
Cooperativeness .672
Understanding of Others .667
Alpha reliability coefficient .685
Source: Analyses of CIRP:04, Magis University
Note: Items asked respondents to: ?Rate yourself on each of the following traits as compared with the
average person your age.?
108
Self-determination
The Basic Psychological Needs Scale (Deci & Ryan, 2000) contains 21 items that
may be divided into three sub-scales, one each for measuring autonomy, competence, and
relatedness. These sub-scales may be used independently or averaged together to create
one scale for self-determination (Gagn?, 2003). As this study is concerned with the
larger concept of self-determination, analysis followed Gagn??s example to create one
factor representing self-determination using all 21 items.
Example items included: ?I feel like I can pretty much be myself in daily
situations,? (autonomy); ?Most days I feel a sense of accomplishment from what I do,?
(competence); and ?People are generally very friendly to me,? (relatedness). The items
were measured on a Likert scale (1-2 = ?Not true at all,? 3-5 = ?Somewhat true,? 6-7 =
?Very true?). All 21 items were tested using principle component analysis. However,
unlike the Gagn? (2003) study, only 19 items loaded onto the factor with factor loadings
above .300. The two items with low factor loadings include ?I feel pressured in my life,?
(factor loading = .270); and ?In my daily life, I frequently have to do what I am told,?
(factor loading = -.071). Table 3 reports the factor loadings for the 19 items, which range
from .369 to .702, and the alpha reliability for the factor (
 = .841).
109
Table 3
Components of the Factor Composite: Self-determination at Time 1
Factor Components Loadings
People are generally pretty friendly towards me .702
I feel like I can pretty much be myself in my daily situations .666
Most days I feel a sense of accomplishment from what I do .626
I get along with people I come into contact .618
People I interact with on a daily basis tend to take my feelings
into consideration
.595
I really like the people I interact with .570
I generally feel free to express my ideas and opinions .549
?In my life I do not get much of a chance to show how
capable I am
.538
?The people I interact with regularly do not seem to like me
much
.523
People in my life care about me .501
People I know tell me I am good at what I do .495
I consider the people I regularly interact with to be my friends .487
?I pretty much keep to myself and don?t have a lot of social
contacts
.478
I have been able to learn interesting and new skills recently .474
?I often do not feel very capable .465
?There are not many people that I am close to .415
I feel like I am free to decide for myself how to live my life .411
?There is not much opportunity for me to decide for myself
how to do things in my daily life
.398
?Often I do not feel very competent .369
Alpha reliability coefficient .841
Source: Analyses of SRA:04, Magis University
Note. (?) Indicates item was recoded to achieve metric with consistent direction
110
First-year Experience
Astin (1984) suggested that the measures of the environment include students?
academic and social involvement as well as programmatic interventions. Variables
describing the first-year experience included three composites (faculty interaction,
academic engagement, and sense of belonging), one dummy variable indicating whether
or not a student was enrolled in a special first-year academic program, and two
standardized single measures of time allocation, according to Pace?s (1984) concept of
student effort. These items include average weekly hours for homework time-on-task
and average weekly hours devoted to clubs and organizations. Items used to create these
measures of the college environment appeared on the 2005 Your First College Year
survey. Each factor was derived using exploratory factor analysis.
Faculty Interaction
The items considered for this factor measure other-than-classroom contact
between students and faculty. The questions asked how often respondents interacted with
faculty during office hours and outside class or office hours. Each was measured on a
Likert scale (1 = ?Never,? 2 = ?1 or 2 times per term,? 3 = ?1 or 2 times per month,? 4 =
?Once a week,? 5 = ?2 or 3 times per week,? 6 = ?Daily?). Results of the exploratory
factor analysis are summarized in Table 4. Although factor loadings for each item were
high (.841), the reliability coefficient was moderate (
 = .587).
111
Table 4
Components of the Factor Composite: Faculty Interaction
Factor Components Loadings
Interacted with faculty during office hours .841
Interacted with faculty outside of class or office hours .841
Alpha reliability coefficient .587
Source: Analyses of YFCY:05, Magis University
Note. Items asked respondents: ?Since entering this college, how often you???
Academic Engagement
Astin?s (1984) student involvement theory describes involvement as having
quantitative and qualitative aspects. The single-item variables measuring time devoted to
homework and student clubs and organizations describe the quantitative nature of
involvement, while the composite variable for academic engagement portrays the
qualitative nature. This composite variable also reflects Kuh?s (1991, 2005) work on
student engagement, which delineates behaviors associated with active learning. Items in
this factor described the number of times respondents handed in late homework, skipped
class, were late to class, and turned in sub-par work. Items are scored on a Likert scale
and are recoded such that 1 = ?Frequently,? 2 = ?Occasionally,? 3 = ?Rarely,? and 4 =
?Not at all.? The factor had high reliability (
 = .710) with loadings ranging from .697 to
.769 (Table 5).
112
Table 5
Recoded Components of the Factor Composite: Academic Engagement
Factor Components Loadings
Turned in course assignment(s) late .769
Skipped class .753
Come late to class .707
Turned in course assignments that did not reflect your best
work
.697
Alpha reliability coefficient .710
Source: Analyses of YFCY:05, Magis University
Note. Items asked respondents: ?Since entering this college, indicate how often you??
Note. All items were recoded to accurately reflect factor title.
Sense of Belonging
The factor describing students? sense of belonging captures an important aspect of
Astin?s (1984) social involvement and Tinto?s (1993) social integration. Two items from
the 2005 Your First College Year survey comprise this highly reliable factor (
 = .813)
with loadings of .920 each (Table 6). Both items ask respondents the extent to which
they feel a sense of belonging or affiliation with the campus community and are scored
on a Likert scale (1 = ?Strongly disagree,? 2 = ?Agree,? 3 = ?Disagree,? and 4 =
?Strongly agree?).
113
Table 6
Components of the Factor Composite: Sense of Belonging
Factor Components Loadings
I feel I am a member of this college .920
I feel I have a sense of belonging to this college .920
Alpha reliability coefficient .813
Source: Analyses of YFCY:05, Magis University
Note. Items asked respondents: ?Please indicate the extent to which you agree or disagree with the
following statements.?
Student Attributes at Time 2
The composite variables representing student attributes at Time 2, academic and
social self-concept, self-efficacy, and self-determination, are identical to those of Time 1.
Although the items are identical, the reliability and loadings for each factor are somewhat
different at Time 2.
Academic and Social Self-concept
Items for these scales appear on the 2005 Your First College Year survey. As
with Time 1, these are self-assessment items that are scored on a Likert scale with 1 =
Lowest 10% to 5 = Highest 10%. The factor for academic self-concept at Time 2 has a
high reliability (
 = .717), with loadings between .651 and .813 (Table 7). The social
self-concept factor is highly reliable (
 = .779), with item loadings ranging from .725 to
.795 (Table 8). Reliability for academic and social self concepts increased from Time 1.
114
Table 7
Components of the Factor Composite: Academic Self-concept at Time 2
Factor Components Loadings Time 1
Academic Ability .812
Self-confidence (intellectual) .804
Writing Ability .689
Drive to Achieve .648
Alpha reliability coefficient .717 .649
Source: Analyses of YFCY:05, Magis University
Note. Items asked respondents to: ?Rate yourself on each of the following traits as compared with the
average person your age.?
Table 8
Components of the Factor Composite: Social Self-concept at Time 2
Factor Components Loadings Time 1
Generosity .795
Compassion .792
Understanding of Others .786
Cooperativeness .728
Alpha reliability coefficient .779 .685
Source: Analyses of YFCY:05, Magis University
Note. Items asked respondents to: ?Rate yourself on each of the following traits as compared with the
average person your age.?
115
Self-determination
Table 9 displays the results from the principle components analysis that created
the factor for self-determination at Time 2. As with Time 1, 19 of the 21 potential items
loaded onto the factor with a score higher than .30. The two items not meeting the
requirements of the scale are: ?I feel pressured in my life,? (factor loading = .227); and
?In my daily life, I frequently have to do what I am told,? (factor loading = -.037).
Reliability for the Time 2 scale is high (
 = .871) and equals the reliability for self-
determination at Time 1.
116
Table 9
Components of the Factor Composite: Self-determination at Time 2
Factor Components Loadings Time 1
People I interact with on a daily basis tend to take my feelings
into consideration
.736
I really like the people I interact with .724
I consider the people I regularly interact with to be my friends .690
People are generally pretty friendly towards me .660
Most days I feel a sense of accomplishment from what I do .647
People in my life care about me .638
I get along with people I come into contact .611
?The people I interact with regularly do not seem to like me
much
.599
People I know tell me I am good at what I do .581
I generally feel free to express my ideas and opinions .578
?I often do not feel very capable .562
I feel like I can pretty much be myself in my daily situations .535
?There are not many people that I am close to .528
I have been able to learn interesting and new skills recently .525
?I pretty much keep to myself and don?t have a lot of social
contacts
.509
?Often I do not feel very competent .493
I feel like I am free to decide for myself how to live my life .493
?There is not much opportunity for me to decide for myself
how to do things in my daily life
.466
?In my life I do not get much of a chance to show how
capable I am
.465
Alpha reliability coefficient .871 .871
Source: Analyses of SRA:05, Magis University
Note. (?) Indicates item was recoded to achieve metric with consistent direction
117
Summary of Factor Analyses
Exploratory factor analysis was used to create multi-item composites describing
key constructs in the conceptual model, including academic self-concept, social self-
concept, and self-determination at Times 1 and 2. In addition, exploratory factor analysis
was used to create several of the environmental measures describing academic and social
involvement in the first year of college, including faculty interaction, academic
engagement, and sense of belonging. In each case, the multi-item composite created a
stronger, more reliable and robust measure than would have been available with a single-
item measure.
Inter-correlations
In order to investigate potential covariance among variables selected for the
model, I computed correlations for all variables (Table 10). Generally, variables in the
model were correlated only slightly with one another, with a few notable exceptions.
Typical correlations were as low as r = -.01 for gender and first-generation status or as
high as r = .35 for SAT composite and academic self-concept at Time 1. Although the
correlation between race (?White? is the reference group) and SAT composite (r = -.22)
was relatively low, the significant negative relationship between the two is worth noting.
As one would expect, exceptions to the pattern of generally low correlations
occurred between pre- and posttest measures of the attribute variables. Academic self-
concept at Time 1 was highly correlated with academic self-concept at Time 2 (r = .73).
In addition, the Time 1 measurements of social self-concept and self-determination were
moderately correlated with their Time 2 counterparts (r = .60 and r = .52, respectively).
118
However, the pretests for each student attribute accounted for less than 100% of the
variance at Time 2, indicating that the pretest values were not the sole determinants of the
posttest values. A moderate correlation also existed between sense of belonging and self-
determination at Time 2 (r = .54).
The dependent variable, college grade point average, was correlated slightly with
almost all variables in the model, with two exceptions. As expected, college GPA was
moderately correlated with its pre-cursor, high school GPA (r = .46). College GPA also
was moderately correlated with academic self-concept at Time 2 (r = .44).
Finally, self-efficacy at Time 1 was correlated with academic self concept at
Times 1 (r = .47) and 2 (r = .34). In addition, the Time 2 construct for self-efficacy was
moderately correlated with academic self-concept at Times 1 (r = .45) and 2 (r = .44).
The measure of self-efficacy was consistently correlated with academic self-concept and
was removed from the model because it was less robust than the measure for academic
self-concept. Although several items from the CIRP Student Information Form would
have contributed to a more reliable measure of self-efficacy, these items were not
repeated on the Your First College Year survey. Therefore, it was unfeasible to develop
a strong and reliable measure of self-efficacy at Time 2 that was equivalent to that at
Time 1. Although the literature has supported a vigorous debate over which attribute
construct, self-efficacy or self-determination, is a better predictor of academic success,
this study was not able to contribute to the debate. Hence I abandoned the fifth research
question to determine which of these constructs proves to have a greater ability to predict
the outcome. From this point forward, this study only reports results for research
questions one through four.
119
Summary of Inter-correlations
With a few exceptions, variables were correlated only slightly. Moderate to
strong correlations between pre- and post-test measures were anticipated. Notable
correlations include the negative relationship between race and SAT composite, and the
strong correlation between academic self-concept at Time 2 and college GPA.
Consistent correlations between self-efficacy and academic self-concept posed a
violation of the non-multicollinearity assumption that guides path analytic techniques.
As the measure of self-efficacy was less robust than the measure of academic self-
concept, it was advisable to eliminate the self-efficacy construct from the model
altogether. Altering the model in this way resulted in abandoning the final research
question to determine whether self-efficacy or self-determination is the better predictor of
academic success.
120
T
a
b
l
e
1
0
B
i
v
a
r
i
a
t
e
C
o
r
r
e
l
a
t
i
o
n
s
G
e
n
d
e
r
R
a
c
e
1
s
t
G
e
n
H
S
G
P
A
S
A
T
S
C
1
a
c
a
d
S
C
1
s
o
c
S
E
1
G
e
n
d
e
r
1
R
a
c
e
-
0
.
0
4
1
1
s
t
G
e
n
-
0
.
0
1
0
.
0
3
1
H
S
G
P
A
0
.
0
5
-
0
.
0
1
0
.
0
4
1
S
A
T
-
0
.
1
2
-
0
.
2
2
*
*
0
.
0
7
0
.
2
4
*
*
1
S
C
1
a
c
a
d
-
0
.
2
1
*
*
0
.
0
2
-
0
.
0
6
0
.
3
1
*
*
0
.
3
5
*
*
1
S
C
1
s
o
c
0
.
0
2
-
0
.
1
1
-
0
.
1
0
0
.
0
5
-
0
.
0
7
0
.
2
9
*
*
1
S
E
1
-
0
.
1
2
0
.
1
2
0
.
0
7
0
.
2
1
*
*
0
.
3
5
*
*
0
.
4
7
*
*
0
.
1
6
*
1
S
D
1
-
0
.
0
6
0
.
1
1
-
0
.
0
3
0
.
1
6
*
0
.
0
1
0
.
3
0
*
*
0
.
3
2
*
*
0
.
2
4
*
*
1
y
r
P
r
o
g
0
.
1
3
-
0
.
0
7
0
.
0
6
0
.
0
4
0
.
0
7
0
.
0
8
0
.
0
2
0
.
0
4
F
a
c
u
l
t
y
0
.
0
2
-
0
.
1
2
0
.
0
7
0
.
1
0
-
0
.
1
7
*
0
.
1
0
-
0
.
0
5
0
.
0
4
E
n
g
a
g
e
0
.
0
1
0
.
0
8
-
0
.
0
9
0
.
2
2
*
*
-
0
.
1
1
0
.
1
3
0
.
0
3
0
.
0
6
H
W
t
i
m
e
0
.
1
4
*
-
0
.
0
6
0
.
0
2
0
.
1
9
*
*
0
.
0
5
0
.
0
7
0
.
0
6
0
.
1
3
C
l
u
b
s
-
0
.
2
0
*
*
-
0
.
1
0
0
.
0
5
0
.
0
6
0
.
0
5
0
.
1
8
*
0
.
0
9
-
0
.
0
4
B
e
l
o
n
g
i
n
g
-
0
.
1
7
*
0
.
1
2
-
0
.
1
1
0
.
0
6
0
.
0
5
0
.
1
6
*
0
.
0
4
0
.
1
5
*
S
C
2
a
c
a
d
-
0
.
2
2
*
*
0
.
0
1
-
0
.
0
4
0
.
3
0
*
*
0
.
3
4
*
*
0
.
7
3
*
*
0
.
1
7
*
0
.
3
4
*
*
S
C
2
s
o
c
-
0
.
0
8
-
0
.
0
1
-
0
.
2
1
*
*
-
0
.
1
0
-
0
.
0
4
0
.
2
0
*
*
0
.
6
0
*
*
0
.
1
5
*
S
E
2
-
0
.
0
6
0
.
1
8
*
0
.
0
5
0
.
1
6
*
0
.
3
0
*
*
0
.
4
5
*
*
0
.
1
4
0
.
5
9
*
*
S
D
2
0
.
0
4
0
.
0
4
-
0
.
0
5
0
.
1
1
-
0
.
0
9
0
.
1
3
0
.
1
4
0
.
1
0
U
g
G
P
A
0
.
0
0
0
.
1
0
0
.
0
4
0
.
4
6
*
*
0
.
3
1
*
*
0
.
3
5
*
*
0
.
0
7
0
.
1
2
121
T
a
b
l
e
1
0
(
c
o
n
t
i
n
u
e
d
)
B
i
v
a
r
i
a
t
e
C
o
r
r
e
l
a
t
i
o
n
s
S
D
1
1
y
r
P
r
o
g
F
a
c
u
l
t
y
E
n
g
a
g
e
H
W
t
i
m
e
C
l
u
b
s
B
e
l
o
n
g
i
n
g
S
C
2
a
c
a
d
G
e
n
d
e
r
R
a
c
e
1
s
t
g
e
n
H
S
G
P
A
S
A
T
S
C
1
a
c
a
d
S
C
1
s
o
c
S
E
1
S
D
1
1
1
y
r
P
r
o
g
0
.
0
1
1
F
a
c
u
l
t
y
0
.
0
5
0
.
0
2
1
E
n
g
a
g
e
0
.
2
5
*
*
-
0
.
1
2
-
0
.
0
7
1
H
W
t
i
m
e
0
.
1
3
0
.
0
8
0
.
1
6
*
0
.
1
6
*
1
C
l
u
b
s
0
.
0
7
0
.
3
2
*
*
0
.
0
6
-
0
.
1
4
0
.
0
8
1
B
e
l
o
n
g
i
n
g
0
.
3
2
*
*
0
.
0
4
-
0
.
0
3
0
.
0
4
-
0
.
0
2
0
.
0
6
1
S
C
2
a
c
a
d
0
.
3
0
*
*
0
.
1
5
*
0
.
1
5
*
0
.
1
8
*
*
0
.
0
9
0
.
1
5
*
0
.
1
8
*
1
S
C
2
s
o
c
0
.
2
8
*
*
-
0
.
0
7
0
.
0
4
-
0
.
0
2
0
.
0
8
0
.
0
3
0
.
1
7
*
0
.
2
1
*
*
S
E
2
0
.
2
8
*
*
0
.
0
0
0
.
0
2
0
.
0
8
0
.
1
7
*
-
0
.
0
6
0
.
1
5
*
0
.
4
4
*
*
S
D
2
0
.
5
2
*
*
-
0
.
1
0
0
.
0
9
0
.
1
9
*
*
-
0
.
0
1
-
0
.
0
5
0
.
5
4
*
*
0
.
1
9
*
*
U
g
G
P
A
0
.
1
0
0
.
1
0
-
0
.
0
2
0
.
3
5
*
*
0
.
1
5
*
0
.
0
2
-
0
.
0
2
0
.
4
4
*
*
122
T
a
b
l
e
1
0
(
c
o
n
t
i
n
u
e
d
)
B
i
v
a
r
i
a
t
e
C
o
r
r
e
l
a
t
i
o
n
s
S
C
2
s
o
c
S
E
2
S
D
2
U
g
G
P
A
G
e
n
d
e
r
R
a
c
e
1
s
t
g
e
n
H
S
G
P
A
S
A
T
S
C
1
a
c
a
d
S
C
1
s
o
c
S
E
1
S
D
1
1
y
r
P
r
o
g
F
a
c
u
l
t
y
E
n
g
a
g
e
H
W
t
i
m
e
C
l
u
b
s
B
e
l
o
n
g
i
n
g
S
C
2
a
c
a
d
S
C
2
s
o
c
1
S
E
2
0
.
2
0
*
*
1
S
D
2
0
.
1
5
*
0
.
2
7
*
*
1
U
g
G
P
A
-
0
.
0
8
0
.
2
7
*
*
0
.
0
5
1
S
o
u
r
c
e
:
A
n
a
l
y
s
e
s
o
f
C
I
R
P
:
0
4
;
S
R
A
:
0
4
/
0
5
;
a
n
d
Y
F
C
Y
:
0
5
,
M
a
g
i
s
U
n
i
v
e
r
s
i
t
y
*
p
<
.
0
5
*
*
p
<
.
0
1
*
*
*
p
<
.
0
0
1
123
Missing Data Analysis
Among the limitations listed for this study in Chapter III is the high attrition
between test Times 1 and 2. Although 952 first-year students enrolled in Magis
University during the 2004-2005 academic year completed at least one of the four survey
instruments for this study, reasonably complete data were available for only 202 students
(21.2%). For the purposes of this study, ?reasonably complete? refers to students who
completed all four surveys, skipped very few questions, and missed no more than one
item within each of the composite factors. Given the number of factors in the model,
cases with data missing for more than one item per factor were eliminated from the
analytic sample. Despite the high level of participant attrition, Table 11 indicated that the
analytic sample varies significantly from the population on only the three measures of
gender, high school grade point average, and cumulative college GPA. A chi-square test
revealed that the analytic sample contains significantly more women (71.8%). Likewise,
t tests indicated that the students in the analytic sample had significantly higher high
school GPAs (mean z-score = 0.1) and college GPAs (mean z-score = 0.2) than students
who were not included in the analyses.
124
Table 11
Distribution of Analytic (N = 202) and Missing (N = 750) Samples for Selected
Characteristics of Magis University 2004-05 First-year Students
Characteristics %Analytic % Missing % Population
Sample 21.2 78.8 100
Gender***
Women 71.8 58.2 61.3
Men 28.2 41.8 38.7
Race
White 84.7 87.2 86.7
Non-White 15.3 12.8 13.3
First-generation College Student
Yes 19.2 21.9 21.2
No 80.8 78.1 78.8
Mean z-score: High School GPA* 0.1 0.0 0.0
Mean z-score: SAT-Verbal 0.1 0.0 0.0
Mean z-score: SAT-Math 0.0 0.0 0.0
Mean z-score: Cumulative College GPA*** 0.2 -0.1 0.0
Source: Analyses of CIRP:04; SRA:04/05; and YFCY:05, Magis University
*p<.05 **p<.01 ***p<.001
Of the 202 cases in the analytic sample, 26 (12.9%) were missing data for at least
one variable in the model. Table 12 demonstrates that nine variables have relatively
small amounts of missing data. Of these nine variables, self-determination at Times 1
and 2 have the most missing data (5.0% and 5.9%, respectively). The other seven
variables with missing data are missing 1% or less.
125
Table 12
Number and Percentage of Cases from the Analytic Sample (N = 202) that are Missing
for each of the Variables in the Analyses
Variable Number
Complete
Number
Missing
Percent
Missing
Gender 202 0 0
Race 202 0 0
First-generation College Student 202 0 0
High School GPA 202 0 0
SAT Composite 202 0 0
Academic Self-concept, Time 1 201 1 0.5
Social Self-concept, Time 1 201 1 0.5
Self-determination, Time 1 192 10 5.0
First-year Programs 202 0 0
Faculty Interaction 202 0 0
Engagement 202 0 0
Student Clubs 201 1 0.5
Sense of Belonging 201 1 0.5
Self-determination, Time 2 190 12 5.9
Academic Effort: Homework Time-on-task 201 1 0.5
Cumulative College GPA 202 0 0
Academic Self-concept, Time 2 201 1 0.5
Social Self-concept, Time 2 200 2 1.0
Number of cases with data for all variables 176 26 12.9
Source: Analyses of CIRP:04; SRA:04/05; and YFCY:05, Magis University
Further missing data analysis reveals that these data were missing randomly
(Table 13). That is, no patterns of missing data emerged, and the characteristics of
126
students with missing data match the characteristics of students with complete data. As
cases with missing data did not differ in any significant way from those cases with
complete data, the generalizability of the analytic sample to the population sample is
preserved. For the purposes of this study, the random nature of missing data means that
the analytic sample is representative of the population sample and that findings related to
the analytic sample may be applied confidently to the population sample.
127
Table 13
Distribution of Complete Cases (N = 176) and Cases with Missing Data (N = 26) for
Selected Characteristics of Magis University 2004-05 First-year Students
Characteristics %Complete % Missing % Total
Sample 87.1 12.9 100
Gender
Women 71.6 73.1 71.8
Men 28.4 26.9 28.2
Race
White 86.0 77.4 84.7
Non-White 14.0 22.6 15.3
First-generation College Student
Yes 18.2 23.1 18.8
No 81.8 76.9 81.2
High School Type
Public 54.5 61.5 55.4
Private 45.5 38.5 44.6
Highest Degree Aspiration
Less than Bachelors 2.6 0.0 2.3
Bachelors 7.9 0.0 6.9
More than Bachelors 88.7 100.0 90.2
Other 0.7 0.0 0.6
Institution Choice
1
st
54.6 57.7 55.0
2
nd
29.3 34.6 30.0
3
rd
9.2 3.8 8.5
Less than 3
rd
6.9 3.8 6.5
128
Table 13 (continued)
Characteristics %Complete % Missing % Total
Mean z-score: High School GPA 0.0 0.0 0.0
Mean z-score: SAT- Math 0.0 -0.2 0.0
Mean z-score: SAT- Verbal 0.0 0.0 0.0
Mean z-score: Academic Self-concept Time 1 0.0 -0.3 0.0
Mean z-score: Social Self-concept Time 1 0.0 0.2 0.0
Mean z-score: Self-determination Time 1 0.0 0.2 0.0
Mean z-score: Cumulative College GPA 0.0 0.0 0.0
Source: Analyses of CIRP:04; SRA:04/05; and YFCY:05, Magis University
*p<.05 **p<.01 ***p<.001
Cohen and Cohen (1983) suggest pairwise deletion as a suitable technique for
handling randomly missing data; however, this procedure would have reduced the sample
size for some analyses below the 200 recommended for reliable path analysis (Klem,
1995; Kline, 2005). In order to preserve sample size, I imputed missing data for
continuous variables since no variable had more than 6% missing data. To determine
imputation values I calculated the mean scores by race and gender for each variable with
missing data using one-way analysis of variance. The average value for students of the
same race and gender replaced the missing values for items in the self-concept, self-
determination, and engagement scales and for the variables describing involvement in
student clubs and homework time-on-task.
129
Summary of Missing Data Analysis
Participant attrition reduced the size of the population sample by nearly 80%,
such that the analytic sample retained only 202 of the original 952 cases. Despite this
loss, the analytic sample met the size requirements (200-300 cases, minimum) for path
analysis (Klem, 1995) and differed significantly from the population sample only in
gender, high school grade point average, and college GPA. Of the 202 cases in the
analytic sample, 26 (nearly 13%) were incomplete; and of the nine variables with missing
data, only two were missing more than one percent. Missing data for the analytic sample
were hand-imputed to preserve sample size, and none of the 202 cases in the analytic
sample were missing data for multiple variables.
Addressing the Research Questions
The following sections present results for four of the five research questions
guiding this study. (Note: Research Question Five was eliminated due to
multicollinearity between composites for academic self-concept and self-efficacy.) The
first section addresses research question one through mean comparisons of student
attributes at Times 1 and 2. Statistically significant change is explored next through
multivariate analysis of input characteristics and first-year involvement variables
influencing this change. The next section presents results of the path analysis and
addresses research questions two through four about the direct and indirect effects of
student attributes and first-year involvement on academic success.
130
Research Question One:
Change in Student Attributes Over Time
The first research question is guided by the assertion that interaction with the
college environment can affect student?s self-determination (Deci, Ryan, & Williams,
1996; Deci, Vallerand, Pelletier, & Ryan, 1991; Miserandino, 1996; Ryan & Powelson,
1991) and self-concept (Berger & Milem, 2000; House, 2000; Kezar & Moriarty, 2000;
Pascarella, Smart, Ethington, & Nettles, 1987). The first step is to determine whether or
not change in student attributes takes place during the first year of college. Table 14
demonstrates results from paired-samples t tests, which indicate statistically significant
change does occur over time for all three measures of student attributes (academic self-
concept, social self-concept, and self-determination). Because I measured change in
student attributes between only two points in time, I chose to use the paired samples t test
rather than an ANOVA for repeated measures, which requires data from at least three
time points. Academic self-concept shows statistically significant (p = .05) positive
change between Times 1 and 2 (mean difference = .28; percent change = 1.82) as does
social self-concept (mean difference = .29; percent change = 1.87). Self-determination,
on the other hand, shows statistically significant (p = .001) negative change between
Times 1 and 2 (mean difference = 7.86; percent change = 7.98). Multivariate analyses
exploring the factors influencing these changes are examined in the following section.
131
Table 14
Paired Samples Comparison between Pre- and Posttest Mean Scores of Computed
Scales for Student Attributes (N = 202)
Source: Analyses of CIRP:04; SRA:04/05; and YFCY:05, Magis University
*p<.05 **p<.01 ***p<.001
Academic Self-concept: Change Over Time
Over the course of the first year of college, academic self-concept increased
nearly two percent. Linear regression analyses were conducted to explain what factors
contribute to the statistically significant positive change in academic self-concept over
time. The blocked-entry strategy employed in these analyses is consistent with the
conceptual framework presented in this study. Table 15 demonstrates the change in
unstandardized coefficients as additional blocks are added to the model. While all three
Mean
Mean
Difference
Percent
Change
Academic Self-concept*
Pre-test 15.17
Post-test 15.44 .28 1.82
Social Self-concept*
Pre-test 15.50
Post-test 15.79 .29 1.87
Self-determination***
Pre-test 106.27
Post-test 98.41 -7.86 7.98
132
blocks in the model are statistically significant, the results validate the assertion that
involvement in the collegiate environment positively influences academic self-concept at
Time 2 (Block 3). After controlling for input characteristics, student attributes at Time 1,
and first-year experiences, only composite SAT scores (b = .151), prior academic self-
concept (b = .594), involvement in a first-year program (b = .224), faculty interaction (b
= .121), and academic engagement (b = .122) contribute significantly to an increase in
academic self-concept at Time 2.
Results from the regression analysis indicate the possibility of a suppression
effect for SAT composite. The positive significant effect size shown by SAT composite
(b = .275) is reduced to a non-significant size (b = .099) when academic self-concept is
introduced in block 2. However, as block three was added, the effect size for SAT
composite (b = .151) increased in magnitude and became a positive significant predictor
once again.
133
Table 15
Coefficients for Student Background Characteristics, Prior Student Attributes, and First-
year Experience Predicting Academic Self-concept at Time 2
Blocks
Independent Variable Background
Characteristics
(1)
Prior Student
Attributes
(2)
First-year
Experience
(3)
Gender
Women -.435** -.150 -.198
Men (ref.)
Race
White -.140 -.087 -.105
Student of Color (ref.)
First-generation Status
Yes -.155 -.017 -.030
No (ref.)
z-score H.S. GPA .250*** .073 .037
z-score SAT Composite .275*** .099 .151**
Academic Self-concept 1 .639*** .594***
Social Self-concept 1 -.043 -.014
Self-determination 1 .107* .076
134
Table 15 (continued)
Blocks
Independent Variable Background
Characteristics
(1)
Prior Student
Attributes
(2)
First-year
Experience
(3)
First-year Program
Yes .224*
No (ref.)
Faculty Interaction .121* 
Academic Engagement .122* 
z-score HW Time-on-task -.007
z-score Student Clubs -.015
Sense of Belonging .014
R
2
.211*** .560*** .590***
Change in R
2
.211*** .349*** .030*
Source: Analyses of CIRP:04; SRA:04/05; and YFCY:05, Magis University
*p<.05 **p<.01 ***p<.001
Social Self-concept: Change Over Time
Between Time 1 and Time 2, social self-concept increased nearly two percent.
Table 16 summarizes the results of the regression predicting change in social self-concept
over time. After controlling only for students? background characteristics, first-
generation college students report lower social self-concept (b = -.504) than their peers
who have had at least one parent attend college. After all other variables are added into
the model, first-generation students still report lower social self-concept (b = -.339) than
their peers. When controlling for input characteristics, prior attributes, and first-year
experiences, results indicate that students with higher average high school GPA indicate
135
lower social self-concept (b = -.156). When controlling for input characteristics, prior
attributes, and first-year experience, students with higher social self-concept at Time 1
experience greater social self-concept at Time 2 (b = .564). Factors representing
involvement in the first year (Block 3) did not serve as significant predictors of social
self-concept at Time 2 when controlling for input characteristics and prior attributes.
Results from the regression analysis indicate the possibility of a suppression
effect for high school GPA, which did not have a significant effect when it entered the
model in the first block (b = -.084). However, as blocks two and three were added, the
effect size for high school GPA (b = -.149 and b = -.156, respectively) increased in
magnitude and became a negative significant predictor of social self-concept.
136
Table 16
Coefficients for Student Background Characteristics, Prior Student Attributes, and
First-year Experience Predicting Social Self-concept at Time 2
Blocks
Independent Variable Background
Characteristics
(1)
Prior Student
Attributes
(2)
First-year
Experience
(3)
Gender
Women -.185 -.164 -.148
Men (ref.)
Race
White -.014 .036 .039
Student of Color (ref.)
First-generation Status
Yes -.504** -.352* -.339*
No (ref.)
z-score H.S. GPA -.084 -.149* -.156*
z-score SAT Composite -.019 .012 .023
Academic Self-concept 1 .031 .025
Social Self-concept 1 .548*** .564***
Self-determination 1 .113 .080
137
Table 16 (continued)
Blocks
Independent Variable Background
Characteristics
(1)
Prior Student
Attributes
(2)
First-year
Experience
(3)
First-year Program
Yes -.148
No (ref.)
Faculty Interaction .083
Academic Engagement -.063
z-score HW Time-on-task .079
z-score Student Clubs -.036
Sense of Belonging .111
R
2
.056* .413*** .444***
Change in R
2
.056* .356*** .031
Source: Analyses of CIRP:04; SRA:04/05; and YFCY:05, Magis University
*p<.05 **p<.01 ***p<.001
Self-determination: Change Over Time
Unlike the two measures of self-concept, which increased during the first year of
college, self-determination experienced a significant decline over the first year of college.
Results of linear regression analyses investigating the factors contributing to this decline
appear in Table 17. When controlling only for student background characteristics,
students with higher GPAs in high school (b = .142) report greater self-determination at
Time 2. However, once all other variables in the model were added, high school GPA
did not serve as a significant predictor of self-determination at Time 2. After controlling
for input characteristics, prior student attributes, and first-year experiences, only gender
138
(b = .318), prior self-determination (b = .388), enrollment in first-year programs (b = -
.249), and sense of belonging (b = .463) contribute significantly to self-determination at
Time 2.
Results from the regression analysis indicate the possibility of a suppression
effect for gender, which did not have a significant effect in blocks one (b = .030) or two
(b = .125). However, as block three was added, the effect size for gender (b = .318)
increased in magnitude and became a positive significant predictor of self-determination
at Time 2.
Table 17
Coefficients for Student Background Characteristics, Prior Student Attributes, and First-
year Experience Predicting Self-determination at Time 2
Blocks
Independent Variable Background
Characteristics
(1)
Prior Student
Attributes
(2)
First-year
Experience
(3)
Gender
Women .030 .125 .318*
Men (ref.)
Race
White .102 -.120 -.217
Student of Color (ref.)
First-generation Status
Yes -.111 -.063 .095
No (ref.)
z-score H.S. GPA .142* .054 .040
z-score SAT Composite -.133 -.102 -.052
139
Table 17 (continued)
Blocks
Independent Variable Background
Characteristics
(1)
Prior Student
Attributes
(2)
First-year
Experience
(3)
Academic Self-concept 1 .014 -.014
Social Self-concept 1 -.058 .000
Self-determination 1 .540*** .388***
First-year Program
Yes -.249*
No (ref.)
Faculty Interaction .084
Academic Engagement .075
z-score HW Time-on-task -.088
z-score Student Clubs -.032
Sense of Belonging .463***
R
2
.031 .295*** .503***
Change in R
2
.031 .263*** .208***
Source: Analyses of CIRP:04; SRA:04/05; and YFCY:05, Magis University
*p<.05 **p<.01 ***p<.001
140
Summary of Research Question One
Although academic and social self-concepts increased significantly over students?
first year of college, self-determination decreased significantly. Aspects of the first-year
experience contributing to the increase in academic self-concept were involvement in a
first-year program, faculty interaction, and academic engagement. Results from the
regression analysis predicting social self-concept did not show that any of the
environmental variables served as significant predictors. Enrollment in first-year
programs was a negative predictor for self-determination at Time 2, while sense of
belonging was a positive predictor. Sense of belonging was a stronger predictor than the
pre-test for self-determination at Time 2.
Research Questions Two, Three, and Four: The Direct Effects of Student
Attributes on Academic Success in the First Year of College
This section reports the path analytic results for research questions two through
four. Results are organized by the order in which blocks were entered into the model and
follow Table 18, which reports the direct effects for all variables in the model. Figure 2
describes the direct effects for variables ultimately leading to significant change in
academic success in the first year of college as measured by cumulative grade point
average. (Note: the following variables were dropped from the model in Figure 2
because they did not make significant contributions to the outcome: First-generation
student status, social self-concept at Time 1, homework time-on-task, participation in
student clubs, sense of belonging, and social self-concept and self-determination at Time
2.) Finally, this section reports the indirect effects for variables in the final model.
141
T
a
b
l
e
1
8
.
D
i
r
e
c
t
E
f
f
e
c
t
C
o
e
f
f
i
c
i
e
n
t
s
f
o
r
P
a
t
h
A
n
a
l
y
s
i
s
P
r
e
d
i
c
t
i
n
g
C
o
l
l
e
g
e
G
P
A
a
n
d
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
a
t
T
i
m
e
2
V
a
r
i
a
b
l
e
6
7
8
9
1
0
1
1
1
2
1
3
1
4
1
5
1
6
1
7
1
8
1
.
G
e
n
d
e
r
:
F
e
m
a
l
e
-
.
4
1
8
*
*
.
0
1
8
-
.
1
6
2
.
1
7
1
*
.
0
5
5
.
0
2
8
.
3
2
9
*
-
.
3
9
5
*
-
.
3
0
6
*
-
.
1
9
8
-
.
1
4
8
.
3
1
8
*
.
0
5
6
2
.
R
a
c
e
:
W
h
i
t
e
-
.
1
6
1
-
.
1
6
5
.
3
9
7
*
-
.
0
8
8
-
.
0
9
2
.
3
6
0
-
.
0
5
1
-
.
2
1
0
.
0
9
7
-
.
1
0
5
.
0
3
9
-
.
2
1
7
.
1
8
5
3
.
1
s
t
G
e
n
e
r
a
t
i
o
n
S
t
u
d
e
n
t
-
.
2
1
4
-
.
2
4
3
-
.
1
1
1
.
0
8
6
.
2
3
1
-
.
2
2
1
.
0
4
1
.
1
6
8
-
.
2
8
0
-
.
0
3
0
-
.
3
3
9
*
.
0
9
5
.
0
6
9
4
.
?
H
i
g
h
S
c
h
o
o
l
G
P
A
.
2
5
4
*
*
*
.
0
7
1
.
1
6
5
*
-
.
0
0
4
.
1
0
3
.
2
1
0
*
*
.
1
5
1
*
.
0
3
0
.
0
0
5
.
0
3
7
-
.
1
5
6
*
.
0
4
0
.
2
5
6
*
*
*
5
.
?
S
A
T
C
o
m
p
o
s
i
t
e
.
2
8
4
*
*
*
-
.
0
5
8
-
.
0
7
2
.
0
3
5
-
.
2
7
0
*
*
*
-
.
2
2
4
*
*
.
0
2
7
-
.
0
0
5
.
0
0
2
.
1
5
1
*
*
.
0
2
3
-
.
0
5
2
.
1
5
9
*
6
.
?
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1
.
0
4
6
.
2
0
1
*
.
0
9
5
.
0
1
4
.
1
2
2
.
0
4
4
.
5
9
4
*
*
*
.
0
2
5
-
.
0
1
4
-
.
0
0
1
7
.
?
S
o
c
i
a
l
S
e
l
f
-
c
o
n
c
e
p
t
1
-
.
0
0
1
-
.
1
3
8
-
.
0
8
1
.
0
1
2
.
0
5
6
-
.
0
8
1
-
.
0
1
4
.
5
6
4
*
*
*
.
0
0
0
.
1
2
1
8
.
?
S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1
.
0
0
2
.
0
2
8
.
1
9
4
*
*
.
1
0
8
.
0
0
9
.
3
1
3
*
*
*
.
0
7
6
.
0
8
0
.
3
8
8
*
*
*
-
.
1
0
6
9
.
F
i
r
s
t
-
y
e
a
r
P
r
o
g
r
a
m
s
.
2
2
4
*
-
.
1
4
8
-
.
2
4
9
*
.
1
2
9
1
0
.
?
F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n
.
1
2
1
*
.
0
8
3
.
0
8
4
-
.
0
3
2
1
1
.
?
A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t
.
1
2
2
*
-
.
0
6
3
.
0
7
5
.
2
6
2
*
*
*
1
2
.
?
H
o
m
e
w
o
r
k
T
i
m
e
-
o
n
-
t
a
s
k
-
.
0
0
7
.
0
7
9
-
.
0
8
8
.
0
3
9
1
3
.
?
C
l
u
b
s
&
O
r
g
a
n
i
z
a
t
i
o
n
s
-
.
0
1
8
-
.
0
3
6
-
.
0
3
2
-
.
0
2
6
1
4
.
?
S
e
n
s
e
o
f
B
e
l
o
n
g
i
n
g
.
0
1
4
.
1
1
1
.
4
6
3
*
*
*
-
.
0
5
8
142
T
a
b
l
e
1
8
c
o
n
t
i
n
u
e
d
.
S
o
u
r
c
e
:
A
n
a
l
y
s
e
s
o
f
C
I
R
P
:
0
4
;
S
R
A
:
0
4
/
0
5
;
a
n
d
Y
F
C
Y
:
0
5
,
M
a
g
i
s
U
n
i
v
e
r
s
i
t
y
*
p
<
.
0
5
*
*
p
<
.
0
1
*
*
*
p
<
.
0
0
1
?
z
-
s
c
o
r
e
d
(
s
t
a
n
d
a
r
d
i
z
e
d
)
v
a
r
i
a
b
l
e
V
a
r
i
a
b
l
e
6
7
8
9
1
0
1
1
1
2
1
3
1
4
1
5
1
6
1
7
1
8
1
5
.
?
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
.
3
0
1
*
*
*
1
6
.
?
S
o
c
i
a
l
S
e
l
f
-
c
o
n
c
e
p
t
2
-
.
1
2
7
1
7
.
?
S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
2
.
0
2
3
1
8
.
?
C
o
l
l
e
g
e
G
P
A
R
2
.
2
1
8
*
*
*
.
0
2
2
.
0
5
1
.
0
3
9
.
0
9
5
*
.
1
5
1
*
*
*
.
0
6
7
.
0
7
2
.
1
4
3
*
*
*
.
5
9
0
*
*
*
.
4
4
4
*
*
*
.
5
0
3
*
*
*
.
4
2
4
*
*
*
143
Direct Effects for Student Background Characteristics
Table 18 describes the direct effects of student background characteristics
(gender, race, first-generation status, high school grade point average, and SAT
composite) on student attributes at Time 1, academic and social involvement in the first-
year, student attributes at Time 2, and cumulative GPA at the end of the first year of
college. The results in Table 18 demonstrate that the predicted paths for student
background characteristics explained nearly 22% of the variance in academic self-
concept at Time 1, but explained very little of the variance in social self-concept and self-
determination at Time 1 (R
2
= .022 and R
2
= .051, respectively). Of these background
characteristics, only gender, race, high school GPA, and SAT composite remain in the
final model. The direct effects for all background characteristics are discussed in turn
below.
Gender. Gender had a significant direct effect on academic self-concept in that
women have lower academic self-concept than men (b = -.418) upon entering college.
However, gender had no effect on social self-concept or self-determination. Gender was
a significant predictor of four variables describing academic and social involvement in
the first year. Specifically, women were more likely to be enrolled in first-year academic
programs (b = .171) and to spend more time on their homework (b = .329) than their male
counterparts. However, women were not as involved socially, being less likely to be
involved in student clubs (b = -.395) and to report a lower sense of belonging to the
campus community (b = -.306). Finally, although gender did not predict self-
determination at Time 1, it did predict self-determination at Time 2, with women
demonstrating greater self-determination (b = .318) than men.
144
Race. Race had a direct effect on only self-determination at Time 1. White
students in the analytic sample tended to report greater self-determination at Time 1 (b =
.397) than students of color.
First-generation Status. Within the conceptual model, first-generation students
differed from other students in only one respect: they reported lower social self-concept
at Time 2 (b = -.339).
High School Grade Point Average. High school GPA positively predicted several
variables in the model. Students who reported higher high school GPA were more likely
to report greater academic self-concept at Time 1 (b = .254), greater self-determination at
Time 1 (b = .165), greater academic engagement (b = .210), and to spend more time
doing homework (b = .151). However, students who had higher GPAs in high school
were more likely to report lower social self-concept (b = -.156) at the end of the first
year. Finally, as one would expect, students with higher GPAs in high school tended to
have higher GPAs in college (b = .256).
SAT Composite. The composite score for SAT college entrance exams positively
predicted academic self-concept at times one (b = .284) and two (b = .151) and college
GPA (b = .159). However, SAT composite negatively predicted two measures of
academic involvement. Students with higher SAT scores were less likely to interact with
faculty (b = -.270) and to be less academically engaged (b = -.224).
Direct Effects for Student Attributes at Time 1
With respect to academic and social involvement in the first year of college, only
academic self-concept and self-determination had any influence (Table 18). Academic
145
self-concept at Time 1 was a positive predictor of faculty interaction (b = .201). Self-
determination at Time 1 positively predicted academic engagement (b = .194) and a
student?s sense of belonging (b = .313). Social self-concept at Time 1 had no impact on
the first-year experience. As Table 18 demonstrates, the paths predicting faculty
interaction (R
2
= .095), academic engagement (R
2
= .151), and sense of belonging (R
2
=
.143) explain a significant amount of the variance for each. As was expected, student
attributes at Time 1 positively predicted their counterparts at Time 2. Academic-self
concept at Time 1 predicted academic self-concept at Time 2 (b = .594); social self-
concept at Time 1 predicted social self-concept at Time 2 (b = .564); and self-
determination at Time 1 predicted self-determination at Time 2 (b = .388).
Direct Effects for Academic and Social Involvement
Four of the first-year experience variables describing academic and social
involvement had significant influence over student attributes at Time 2 (Table 18).
Enrollment in a first-year program positively affected students? subsequent academic
self-concept (b = .224) but negatively influenced self-determination at Time 2 (b = -
.249). Faculty interaction (b = .121) and academic engagement (pink lines; b = .122)
positively influenced academic self-concept at Time 2. In addition, sense of belonging
positively influenced self-determination at Time 2 (b = .463). None of the first-year
experience variables had a significant effect on social self-concept, and academic
engagement was the only first-year experience variable with a significant direct effect on
college grade point average (b = .262). The paths predicting academic self-concept at
Time 2 (R
2
= .590), social self-concept at Time 2 (R
2
= .444), and self-determination at
146
Time 2 (R
2
= .503) explained a substantial percentage of the variance for each of these
student attribute variables at Time 2 (Table 18).
Direct Effects for Student Attributes at Time 2
Although three paths were predicted, only academic self-concept at Time 2 is a
significant predictor of college GPA (b = .301) (see pink lines in Figure 4). Table 18
demonstrates that the full model explains 42 % of college GPA (R
2
= .42). The following
sections explore significant indirect and total effects explaining college GPA and
academic self-concept at Time 2, which is directly linked to college GPA.
Final Direct Effects Model
Figure 2 describes the direct effects for variables ultimately leading to significant
change in academic success in the first year of college as measured by cumulative grade
point average. (Note: the following variables were dropped from the model in Figure 2
because they did not make significant direct or indirect contributions to the outcome:
First-generation student status, social self-concept at Time 1, homework time-on-task,
participation in student clubs, sense of belonging, and social self-concept and self-
determination at Time 2.) Consistent with path analytic models, from left to right Figure
2 demonstrates first the correlations between remaining exogenous input variables
comprising students? background characteristics. The highest correlation was the small
positive relationship between high school GPA and SAT composite (r = .24). Following
the correlation notations, Figure 2 depicts those variables that affect college GPA, either
147
F
i
g
u
r
e
2
.
D
i
r
e
c
t
E
f
f
e
c
t
s
f
o
r
I
n
d
e
p
e
n
d
e
n
t
V
a
r
i
a
b
l
e
s
:
S
t
u
d
e
n
t
A
t
t
r
i
b
u
t
e
s
a
n
d
A
c
a
d
e
m
i
c
S
u
c
c
e
s
s
i
n
t
h
e
F
i
r
s
t
Y
e
a
r
o
f
C
o
l
l
e
g
e
I
n
p
u
t
s
S
t
u
d
e
n
t
A
t
t
r
i
b
u
t
e
s
1
F
i
r
s
t
-
y
e
a
r
E
x
p
e
r
i
e
n
c
e
S
t
u
d
e
n
t
A
t
t
r
i
b
u
t
e
s
2
O
u
t
c
o
m
e
T
i
m
e
F
i
r
s
t
-
y
e
a
r
P
r
o
g
r
a
m
F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n
G
e
n
d
e
r
R
a
c
e
H
i
g
h
S
c
h
o
o
l
G
P
A
S
A
T
C
o
m
p
o
s
i
t
e
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
E
n
g
a
g
e
m
e
n
t
C
o
l
l
e
g
e
G
P
A
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
.
3
0
1
.
2
2
4
.
1
2
1 .
1
2
2
.
2
6
2
.
5
9
4
.
2
0
1
.
1
9
4
.
1
7
1
-
.
4
1
8
.
3
9
7
.
2
5
4
.
1
6
5
.
2
1
0
.
2
5
6
.
2
8
4
-
.
2
2
4
.
1
5
9
.
1
5
1
-
.
2
7
0
-
.
1
2
-
.
0
4
-
.
2
2
-
.
0
1
.
2
4
.
0
5
148
directly or indirectly. Direct effects between variables in the final model are discussed
here from left to right. Indirect effects are discussed in the next section.
Of the background characteristics only high school GPA and SAT composite had
direct positive effects on college grade point average (b = .256 and b = .159,
respectively). Of the student attributes at Time 1, neither academic self-concept nor self-
determination had direct effects on college GPA. Of the three remaining involvement
measures, only academic engagement was a predictor of college grade point average (b =
.262). Finally, academic self-concept at Time 2 was a significant positive predictor of
college GPA (b = .301).
Summary of Direct Effects
Initial student attributes predicted some aspects of academic and social
involvement in the first year of college. Specifically, academic self-concept at Time 1
predicted interaction with faculty, although lower SAT composite was a stronger
predictor of faculty interaction. Self-determination at Time 1 predicted both academic
engagement and sense of belonging, although student background characteristics aided in
the prediction of these variables as well. Social self-concept did not predict any of the
variables describing the first-year experience.
Academic and social involvement did have some effect on student attributes at
Time 2. Specifically, academic self-concept at Time 2 was positively predicted by
enrollment in first-year programs, faculty interaction, and academic engagement.
Enrollment in first-year programs was a negative predictor for self-determination at Time
2, while sense of belonging was a very strong positive predictor for the same attribute.
149
None of the measures for academic or social involvement were significant predictors of
social self-concept at Time 2. No involvement measures, with the exception of academic
engagement, significantly predicted college grade point average. The findings regarding
academic and social involvement are noteworthy specifically because they suggest that
effects of most involvement behaviors are mediated through the student rather than
directly linked to the outcome.
Research Questions Two, Three, and Four: The Indirect Effects of Student
Attributes on Academic Success in the First Year of College
The relationships between variables in any model often are complicated. Path
analytic models examine the more hidden indirect effects as well as the more obvious
direct effects. Although only four variables in the final path model had direct effects on
grade point average at the end of the first year of college, the others are related indirectly
through their relationships with academic self-concept at Time 2. This section reports the
results of these indirect effects. Because this study focused on the contribution of student
attributes toward academic success, this section also reports the indirect effects on
academic self-concept at Time 2, which itself was a significant positive predictor of
college grade point average.
Indirect Effects Predicting Academic Self-concept at Time 2
As with college GPA, the indirect effects for academic self-concept at Time 2 are
many but of small magnitude (Table 19). Being female had a small negative indirect
effect on academic self-concept at Time 2 (b = -.221), while the effect of being White
150
was quite small (b = .009). High school GPA (b = .183) and SAT composite (b = .114)
had the greatest indirect effects, albeit still small. Finally academic self-concept (b =
.024) and self-determination at Time 1 (b = .023) also had small positive indirect effects
on academic self-concept at Time 2. Table 20 shows the total effects, comprised of
significant direct and indirect effects, for variables predicting academic self-concept at
Time 2.
151
T
a
b
l
e
1
9
S
t
a
t
i
s
t
i
c
a
l
l
y
S
i
g
n
i
f
i
c
a
n
t
I
n
d
i
r
e
c
t
E
f
f
e
c
t
s
P
r
e
d
i
c
t
i
n
g
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
a
t
T
i
m
e
2
f
o
r
P
a
t
h
A
n
a
l
y
s
i
s
I
n
d
i
r
e
c
t
P
a
t
h
E
s
t
.
1
.
G
e
n
d
e
r
:
F
e
m
a
l
e
G
e
n
d
e
r

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
-
.
2
4
8
G
e
n
d
e
r

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
-
.
0
1
0
G
e
n
d
e
r

F
i
r
s
t
-
y
e
a
r
P
r
o
g
r
a
m
s

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
.
0
3
7
T
o
t
a
l
-
.
2
2
1
2
.
R
a
c
e
:
W
h
i
t
e
R
a
c
e

S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
.
0
0
9
3
.
H
i
g
h
S
c
h
o
o
l
G
P
A
H
i
g
h
S
c
h
o
o
l
G
P
A

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
.
1
4
8
H
i
g
h
S
c
h
o
o
l
G
P
A

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
.
0
0
6
H
i
g
h
S
c
h
o
o
l
G
P
A

S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
.
0
0
4
H
i
g
h
S
c
h
o
o
l
G
P
A

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
.
0
2
5
T
o
t
a
l
.
1
8
3
152
T
a
b
l
e
1
9
c
o
n
t
i
n
u
e
d
S
o
u
r
c
e
:
A
n
a
l
y
s
e
s
o
f
C
I
R
P
:
0
4
;
S
R
A
:
0
4
/
0
5
;
a
n
d
Y
F
C
Y
:
0
5
,
M
a
g
i
s
U
n
i
v
e
r
s
i
t
y
*
p
<
.
0
5
*
*
p
<
.
0
1
*
*
*
p
<
.
0
0
1
I
n
d
i
r
e
c
t
P
a
t
h
E
s
t
.
4
.
S
A
T
C
o
m
p
o
s
i
t
e
S
A
T
C
o
m
p
o
s
i
t
e

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
.
1
6
5
S
A
T
C
o
m
p
o
s
i
t
e

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
.
0
0
7
S
A
T
C
o
m
p
o
s
i
t
e

F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
-
.
0
3
2
S
A
T
C
o
m
p
o
s
i
t
e

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
-
.
0
2
6
T
o
t
a
l
.
1
1
4
5
.
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
.
0
2
4
6
.
S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1
S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2
.
0
2
3
153
Table 20
Statistically Significant Direct, Indirect, and Total Effects for Independent Variables in
the Path Analysis Predicting Academic Self-concept at Time 2
Source: Analyses of CIRP:04; SRA:04/05; and YFCY:05, Magis University
*p<.05 **p<.01 ***p<.001
Indirect Effects Predicting College GPA
Table 21 lists the numerous significant indirect effects for college GPA, all of
which are small in magnitude. Overall, being female (b = -.064) had a negative indirect
effect on college GPA, while being White (b = .023) had a positive indirect effect. High
Variable Total Direct
Effect
Total Indirect
Effect
Total Effect
1. Gender: Female -.221 -.221
2. Race: White .010 .010
3. 1
st
Generation Student
4. High School GPA .183 .183
5. SAT Composite .151** .114 .265
6. Academic Self-concept 1 .594*** .024 .618
7. Social Self-concept 1
8. Self-determination 1 .023 .023
9. First-year Programs .224* .224
10. Faculty Interaction .121* .121
11. Academic Engagement .122* .122
12. Homework Time-on-task
13. Clubs & Organizations
14. Sense of Belonging
154
school GPA (b = .116) and SAT composite scores (b = .020) both had small, positive
indirect effects. The positive indirect effect of academic self-concept at Time 1 (b =
.178) is three times greater than that of self-determination at Time 1 (b = .056). Of the
six variables describing academic and social involvement in the first year, only three have
an indirect effect on college GPA, all of which were positive. These included enrollment
in a first-year academic program (b = .064), interaction with faculty outside the
classroom (b = .035), and academic engagement (b = .035). Table 22 lists the total
effects, comprised of significant direct and indirect effects, for variables predicting
college GPA.
155
T
a
b
l
e
2
1
S
t
a
t
i
s
t
i
c
a
l
l
y
S
i
g
n
i
f
i
c
a
n
t
I
n
d
i
r
e
c
t
E
f
f
e
c
t
s
P
r
e
d
i
c
t
i
n
g
C
o
l
l
e
g
e
G
P
A
f
o
r
P
a
t
h
A
n
a
l
y
s
i
s
I
n
d
i
r
e
c
t
P
a
t
h
E
s
t
.
1
.
G
e
n
d
e
r
:
F
e
m
a
l
e
G
e
n
d
e
r

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
-
.
0
7
2
G
e
n
d
e
r

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
-
.
0
0
3
G
e
n
d
e
r

F
i
r
s
t
-
y
e
a
r
P
r
o
g
r
a
m
s

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
1
1
T
o
t
a
l
-
.
0
6
4
2
.
R
a
c
e
:
W
h
i
t
e
R
a
c
e

S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
0
3
R
a
c
e

S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

C
o
l
l
e
g
e
G
P
A
.
0
2
0
T
o
t
a
l
.
0
2
3
3
.
H
i
g
h
S
c
h
o
o
l
G
P
A
H
i
g
h
S
c
h
o
o
l
G
P
A

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
4
3
H
i
g
h
S
c
h
o
o
l
G
P
A

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
0
2
H
i
g
h
S
c
h
o
o
l
G
P
A

S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
0
1
H
i
g
h
S
c
h
o
o
l
G
P
A

S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

C
o
l
l
e
g
e
G
P
A
.
0
0
8
H
i
g
h
S
c
h
o
o
l
G
P
A

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
0
7
H
i
g
h
S
c
h
o
o
l
G
P
A

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

C
o
l
l
e
g
e
G
P
A
.
0
5
5
T
o
t
a
l
.
1
1
6
156
T
a
b
l
e
2
1
c
o
n
t
i
n
u
e
d
I
n
d
i
r
e
c
t
P
a
t
h
E
s
t
.
4
.
S
A
T
C
o
m
p
o
s
i
t
e
S
A
T
C
o
m
p
o
s
i
t
e

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
4
8
S
A
T
C
o
m
p
o
s
i
t
e

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
0
2
S
A
T
C
o
m
p
o
s
i
t
e

F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
-
.
0
0
9
S
A
T
C
o
m
p
o
s
i
t
e

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
-
.
0
0
8
S
A
T
C
o
m
p
o
s
i
t
e

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

C
o
l
l
e
g
e
G
P
A
-
.
0
5
7
S
A
T
C
o
m
p
o
s
i
t
e

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
4
4
T
o
t
a
l
.
0
2
0
5
.
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
1
7
1
A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
1

F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
0
7
T
o
t
a
l
.
1
7
8
6
.
S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1
S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
0
7
S
e
l
f
-
d
e
t
e
r
m
i
n
a
t
i
o
n
1

A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

C
o
l
l
e
g
e
G
P
A
.
0
4
9
T
o
t
a
l
.
0
5
6
157
T
a
b
l
e
2
1
c
o
n
t
i
n
u
e
d
S
o
u
r
c
e
:
A
n
a
l
y
s
e
s
o
f
C
I
R
P
:
0
4
;
S
R
A
:
0
4
/
0
5
;
a
n
d
Y
F
C
Y
:
0
5
,
M
a
g
i
s
U
n
i
v
e
r
s
i
t
y
*
p
<
.
0
5
*
*
p
<
.
0
1
*
*
*
p
<
.
0
0
1
I
n
d
i
r
e
c
t
P
a
t
h
E
s
t
.
7
.
F
i
r
s
t
-
y
e
a
r
P
r
o
g
r
a
m
s
F
i
r
s
t
-
y
e
a
r
P
r
o
g
r
a
m
s

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
6
4
8
.
F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n
F
a
c
u
l
t
y
I
n
t
e
r
a
c
t
i
o
n

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
3
5
9
.
A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t
A
c
a
d
e
m
i
c
E
n
g
a
g
e
m
e
n
t

A
c
a
d
e
m
i
c
S
e
l
f
-
c
o
n
c
e
p
t
2

C
o
l
l
e
g
e
G
P
A
.
0
3
5
158
Table 22
Statistically Significant Direct, Indirect, and Total Effects for Independent Variables in
the Path Analysis Predicting College GPA
Source: Analyses of CIRP:04; SRA:04/05; and YFCY:05, Magis University
*p<.05 **p<.01 ***p<.001
Variable Total Direct
Effect
Total Indirect
Effect
Total Effect
1. Gender: Female -.064 -.064
2. Race: White .023 .023
3. 1
st
Generation Student
4. High School GPA .256*** .116 .372
5. SAT Composite .159* .020 .179
6. Academic Self-concept 1 .178 .178
7. Social Self-concept 1
8. Self-determination 1 .056 .056
9. First-year Programs .064 .064
10. Faculty Interaction .035 .035
11. Academic Engagement .262*** .035 .297
12. Homework Time-on-task
13. Clubs & Organizations
14. Sense of Belonging
15. Academic Self-concept 2 .301*** .301
16. Social Self-concept 2
17. Self-determination 2
159
Summary of Indirect Effects
With regard to direct effects for academic success in the first year of college, only
academic self-concept at Time 2 was a significant contributor. However, both academic
self-concept and self-determination at Time 1 had significant positive indirect effects on
college grade point average. Similarly, these time-one attributes contributed indirectly to
academic self-concept at Time 2, which had a direct effect on the dependent variable.
Many variables contributed indirectly to students? grade point averages in the first year,
indicating a more complicated picture of academic success.
Summary
This study used descriptive, multivariate, and path analytic techniques to answer
research questions related to the impact of student attributes on academic success in the
first year of college. Exploratory factor analyses were used to create composite variables
describing student attributes at Times 1 and 2 as well as aspects of academic and social
involvement. Missing data analyses were conducted to investigate possible problems
related to the high rate of participant attrition. Although the analytic sample was nearly
80% smaller than the original population sample, results from missing data analyses
indicated that the analytic sample did not differ significantly from the population sample.
Therefore results from the multivariate and path analytic techniques are generalizable to
the population sample. Paired-samples t tests and multivariate linear regression were
used to explore change in student attributes over time. Finally, path analysis revealed the
direct and indirect effects of student attributes on academic success, as measured by
college grade point average, in the first year of college. Four variables in the final model
160
had significant direct effects on college GPA, while all other variables, including
measures of student involvement, had indirect effects as mediated through academic self-
concept at Time 2. The results presented in this chapter are discussed in the following
chapter along with implications for future research and practice.
161
CHAPTER V
DISCUSSION
Introduction
This study investigated the relationship between student self-responsibility and
academic success in the first year of college. By incorporating psychological constructs
in an I-E-O college impact model, this study explored how student attributes influence
academic outcomes beyond what is already explained by aspects of the environment.
Specifically, this study analyzed the direct and indirect effects of academic self-concept,
social self-concept, and self-determination on cumulative grade point average. In
addition, this study examined the university environment and the influence it exerts over
change in and development of student attributes during the first collegiate year. Data for
this inquiry were supplied by Magis University, a comprehensive Jesuit institution in the
mid-Atlantic region. Magis University supplied institutional data for the entering Class
of 2008, as well as access to these students? responses to the 2004 Cooperative
Institutional Research Program?s (CIRP) Student Information Form for first-year
students, the 2005 follow-up Your First College Year, and the pre- and posttest versions
of the 2004-2005 Student Readiness Assessment. These data were used to explore the
following research questions:
1. What changes in student attributes occur during the first year of college,
and what environmental factors influence these changes?
2. Controlling for student background characteristics, how do initial student
attributes influence academic and social involvement in the first year of
college?
162
3. Controlling for student background characteristics and student attributes at
Time 1, how do academic and social involvement impact subsequent
student attributes and academic success?
4. After controlling for student background characteristics and academic and
social involvement, what are the direct and indirect effects of student
attributes, as measured by self-determination, self-efficacy and effort, on
the academic success of first-year college students?
(Note: due to consistent correlations between self-efficacy and academic self-concept,
self-efficacy was eliminated from the model as the less robust construct of the two.
Therefore, I was not able to pursue Research Question Five comparing self-efficacy to
self-determination, and the discussion in this chapter pertains only to Research Questions
One through Four.)
This chapter reviews the findings presented in Chapter IV, draws appropriate
conclusions, and offers implications for theory, practice, and future research. The leading
sections summarize the analyses for each research question and the test model, which was
proposed and tested in this study. The second section discusses the conclusions that may
be inferred from these results, and the last sections consider the contributions to
scholarship and practice as well as directions for future study. The chapter closes with
final thoughts on the meaning of this study.
163
Summary of Findings Related to Research Questions
Research Question One: Student Attributes?Change over Time
Several meta-analyses of student change related to college attendance revealed
that it may be unreasonable to expect noticeable change in student attributes within one
year of college (Feldman & Newcomb, 1994; Pascarella & Terenzini, 1991, 2005).
However, the results of this study indicated that students at Magis University did
experience significant change in self-concept and self-determination during their first
year of college. Further analyses indicated those factors that contributed to these
changes. The findings for Research Question One about change in student attributes over
time are discussed below.
Change in Self-concept
Self-concept refers to students? self-appraisal of abilities as compared to their peer
group. Based on their findings that self-concept changes each year of college beyond the
cumulative growth over four years, Terenzini, Theophilides, and Lorang (1984)
suggested the need for additional research investigating year-to-year differences. This
study confirms the assertions of the Terenzini et al. study and advances the understanding
of change over time.
Academic self-concept. This study demonstrated that students? academic self-
concept experienced small but significant increases over the first year of college. Several
constructs in the model contribute to this change in some interesting ways. The
discussion begins with the effects of student background characteristics in predicting
164
academic self-concept at Time 2. Although initially female students have significantly
lower academic self-concept than their male counterparts, the effects of gender on
academic self-concept at the end of the first year are non-existent. After adding prior
student attributes and first-year experiences to the model, gender no longer played a
significant role in determining academic self-concept at Time 2. This finding confirms
the Pascarella, Smart, Ethington, and Nettles (1987) study which did not reveal any
difference in self-concept based on race or gender.
An alternate explanation for the change in gender?s effect on academic self-
concept at Time 2 is the possibility a suppression effect. Specifically, introducing
academic self-concept at Time 1 to the model in Block 2 may mask the effect of gender
on subsequent academic self-concept (Cohen, Cohen, West, & Aiken, 2003). In other
words, the relationship between gender and prior academic self-concept may be such that
the presence of prior academic self-concept in the model dominates and obscures the true
effect of gender on later academic self-concept.
High school grade point average saw a similar decline in its predictive ability.
When considering student background characteristics alone, high school GPA was a
positive predictor of academic self-concept at Time 2. However, the effects of this
measure dissipated once other variables were added to the model. The composite score
for SAT was the only background characteristic that retained a significant effect on
academic self-concept at Time 2, which is not surprising since students who have tested
well in the past may appraise their abilities at a higher level than their peers, regardless of
the experiences they have.
165
As expected, prior academic self-concept is the strongest predictor of academic
self-concept at Time 2. The results of the regression analysis confirmed that the pre- and
posttest measures of academic self-concept are highly related. However, the analysis also
revealed that involvement with aspects of the first-year experience, such as enrolling in a
first-year program, interacting with faculty, and exhibiting engagement behaviors, have
positive effects on academic self-concept over and beyond prior self-concept. Self-
determination at Time 1 proved to be a significant positive predictor of academic self-
concept at Time 2 when prior student attributes were added to the model. However, the
effects of self-determination were eliminated once the involvement variables were added
to the model. These findings are consistent with previous research (Berger & Milem,
2000; House, 2000; Pascarella, Smart, Ethington, & Nettles, 1987), which demonstrated
that student involvement is a strong predictor of change in self-concept.
Social self-concept. As with academic self-concept, the pretest for social self-
concept was the strongest predictor for the posttest measure. Despite the strength of this
relationship, however, two background characteristics were significant predictors as well.
First-generation status and high school grade point average both had negative effects on
social self-concept at Time 2, even after all other variables were added to the model.
The effects of high school GPA became significant only after prior social self-concept
was added to the model, suggesting that high achieving students may be more
academically focused and less likely to see themselves as socially confident from day
one.
The effects of first-generation status are curious. Although the negative effect of
first-generation status on social self-concept at Time 2 is mitigated slightly by prior
166
student attributes and first-year experiences, the negative relationship between the two
remains strong throughout the model. A possible explanation for this finding could be
that first-generation students have limited financial resources, which could diminish their
ability to participate in social activities with their peers (Berger & Milem, 1999).
However, further investigations revealed that first-generation status was not correlated
with student-reported parental income (albeit an unreliable measure). Other explanations
include the possibilities that first-generation students may be more academically focused
and spend less time socializing with their peers and/or that they enter college with less
cultural capital and understand the benefits of social involvement to a lesser degree
(Lubrano, 2004; Pascarella, Pierson, Wolniak, & Terenzini, 2004). Either of these
explanations may lead first-generation students to appraise their social ability at a lower
level than their peers who have a family history of college attendance. Finally, this
finding is a curious one because social self-concept at Time 2 is the only place where
generational status makes any difference in the model. The relationship between first-
generation status and other variables in the model is worthy of further investigation.
Change in Self-determination
The explanation of change in self-determination over time is less straightforward
than that for self-concept. First, the most interesting and unexpected effects appeared in
the last block of the model, and second, unlike either measure of self-concept, self-
determination decreased significantly over the first year of college. The early positive
effect of high school grade point average was mitigated by subsequent variables in the
model, which is understandable considering the relatively high correlation between the
167
Time 1 and Time 2 measures for self-determination. The initially strong positive effect
of the pretest measure for self-determination, however, was diminished by variables
describing first-year involvement in the third and final block in the model. In fact, the
effect of the pretest measure for self-determination was second in size, with the effect of
sense of belonging having the largest effect. Although not measured explicitly, this
finding may indicate a relationship between sense of belonging and the subscale of self-
determination which measures relatedness.
The effect of gender, although not significant in earlier models, became
significant in the third block such that women have higher levels of self-determination at
Time 2 than their male counterparts. Interestingly, the predictive effect of gender in the
third block was almost equal that of self-determination at Time 1. As with academic self-
concept, this finding may point to a possible suppression effect. In this case, the sharp
rise in the effect of gender in Block 3 may indicate a potential interaction between gender
and another variable in the model (Cohen, Cohen, West, & Aiken, 2003).
Finally, enrollment in a first-year academic program was a significant negative
predictor of self-determination at Time 2. Given the goals of most first-year academic
programs?to facilitate meaningful connection between new students and faculty, to ease
the transition to college, and to provide a challenging academic introduction to college
accompanied by appropriate support systems?this finding is particularly surprising.
?Enrollment in first-year programs? was a composite variable created to capture all
students who participated in first-year seminars or living-learning communities.
However, even in its disaggregated form, there were no apparent relationships that would
168
have described or explained the negative relationship between first-year program
enrollment and self-determination at Time 2.
Instead, explanation may lie in one of the considerations of self-determination
theory itself. As Ryan and Powelson (1991) asserted, the environment is a social context
with the ability to hinder or foster growth. This viewpoint is not unlike higher
education?s ecological theories, which propose that behavior is explained through the
person-environment interaction (e.g. Barker, 1968; Stern, 1970). Perhaps the key to
understanding why enrolling in first-year programs negatively predicts self-determination
at Time 2 is in the environment. Simply said, these findings suggest that some aspect of
these first-year programs negatively impacts the development of autonomy, competence,
and/or relatedness.
Similarly, a closer look at the environment might help explain why self-
determination decreases over time while self-concept tends to increase. Previous
researchers have linked change in self-determination directly to environmental
interventions amongst K-12 students (Connell & Wellborn, 1991; Miserandino, 1996;
Vallerand, Fortier, & Guay, 1997). Specifically, students who experienced low support
for the development of their autonomy reported lower subsequent levels of autonomy,
competence, and relatedness. The initial findings from this study suggest a need to
understand why students who enroll in specialized programs have lower self-
determination at the end of the first-year than their counterparts who were not similarly
enrolled?especially when the programs are intended to aid students and their
development. Perhaps the balance of challenge and support that is cited as a key
component to beneficial first-year programs (Barefoot, et al., 2005; Upcraft, Gardner, &
169
Barefoot, 2005; Upcraft & Gardner, 1989) is off kilter such that students in these
programs are receiving too much support and not enough challenge.
Other explanations for the significant decrease in self-determination for
participants in first-year programs include the possibility that students referenced their
high school relationships and experiences when responding to the pretest. The familiarity
of the high school environment and an established peer group may have boosted students?
confidence in their autonomy, competence, and/or relatedness; whereas responses to the
posttest may refer to students? college experiences and relationships, which are newer
and, therefore, possibly less comfortable and secure. Finally, it is possible that students
participating in first-year programs have a better understanding or higher expectations of
the college experience and may assess themselves against these expectations, resulting in
lower self-appraisals than their non-participating peers (K. Inkelas, personal
communication, May 16, 2006).
Research Question Two: The Impact of Student Attributes on
Involvement in the First Year of College
Within the context of Astin?s (1991) input-environment-outcome model for
assessing college impact, the environment refers to any and all institutional interventions,
including student experiences and educational programs. This definition of the
environment is harmonious with Astin?s (1984) theory of student involvement, which
describes student experiences in terms of psychological and physical energy toward the
educational enterprise. Included in Astin?s understanding of involvement are time and
quality of effort and measures of the student-environment interaction, including
170
interaction with peers and faculty and program participation. For the purposes of this
study, the environmental measures were limited to students? involvement during the first
year of college. These measures included enrollment in first-year programs, interactions
with faculty outside the classroom, behaviors of academic engagement, homework time-
on-task, participation in student clubs and organizations, and the development of college
affiliation, or a sense of belonging. The second research question investigated the
influence of student attributes on academic and social involvement. The following
section reviews the findings for each involvement measure in the first year of college.
First-year Programs
The student attributes tested in the model had no significant effect on first-year
program enrollment. Although one might expect prior academic self-concept to play a
role in students deciding to enroll in a first-year seminar or social self-concept to be a
contributing factor toward joining a living-learning community, none of these
psychological attributes seem to have made a difference. In fact, only gender proved to
be a significant predictor of first-year programs, in that women were more likely to be
enrolled than men. This is not surprising given that women have greater representation
than men within the Magis University?s student population. One would expect
significantly higher female enrollment in an environment with more women than men
students.
171
Faculty Interaction
Of the student attributes at Time 1, only academic self-concept had any effect on
student-faculty interaction outside the classroom. Prior academic self-concept was a
positive predictor of faculty interaction, such that students with higher academic self-
concept reported greater contact with faculty outside of classroom or office hours. This
finding is consistent with that of House (2000), who examined the relationship between
academic self-concept and academic involvement. House?s study demonstrated that
academic self-concept at the beginning of the first year of college was positively related
to time spent with faculty in high school.
Other than prior academic self-concept, only SAT composite was a significant
predictor of faculty interaction. In the case of SAT composite, the relationship to faculty
interaction was negative, meaning that students with higher SAT scores reported less out-
of-class contact with their faculty during the first year of college. This finding is similar
to that of Berger and Milem (1999), who reported that students who were high achieving
in high school (as represented by grade point average) had fewer interactions with faculty
outside of class. Berger and Milem surmised that first-year students might be more likely
to see faculty outside of class or office hours if they are experiencing academic difficulty.
Along these lines, students whose college entrance exams scores were high might not
have perceived connecting with faculty as a need or benefit.
Academic Engagement
Prior self-determination had a significant positive effect on behaviors associated
with academic engagement. This finding is consistent with Reeve?s (2002) analysis of
172
several studies involving K-12 students. Reeve concluded that, across the board, greater
self-determination leads to higher levels of engagement in the educational process.
Studying the K-12 population, Connell and Wellborn (1991) found that students who
reported higher levels of self-determination were also rated as more highly engaged by
their teachers. The current study extended this understanding by reproducing these
findings in the college setting.
Not surprisingly, high school grade point average was a positive predictor of
engagement as well. It follows logically that students who are high achieving in high
school would continue their academic habits in college. Unfortunately, the same logic
did not hold true for students with high SAT composite scores. Unlike high school GPA,
composite SAT was a significant negative predictor of engagement. This finding is
notable, if only to draw attention to the fact that testing ability did not translate into
behaviors associated with academic engagement.
Homework Time-on-task
Of the background characteristics, only gender and high school grade point
average were significantly associated with the number of hours students spent doing
homework. Women students devoted significantly more time to homework than their
male counterparts. Similarly, students who were high achieving in high school spent
more time on their studies than their peers. None of the student attributes at Time 1 had a
significant effect on homework time-on-task. This finding was surprising considering
previous research linking academic self-concept (House, 2000) and self-determination
(Connell & Wellborn, 1991) with hours spent studying.
173
Participation in Student Clubs and Organizations
Student attributes at Time 1 were not significant predictors of student
participation in clubs and organizations during the first year of college. In fact, only
gender had a significant effect on participation, such that men reported greater
involvement in student clubs than did women. These findings seem counterintuitive
given the greater enrollment in first-year programs by women students and the
expectations that social self-concept and self-determination would influence participation
in student clubs. Perhaps these findings simply suggest that the end of the first year is too
early to measure involvement in student clubs at Magis University.
Sense of Belonging
Self-determination at Time 1 is a significant positive predictor of developing a
sense of belonging in the first-year of college. Although previous research does not help
in explaining this relationship, this finding is not all that surprising. The measure of self-
determination used for this study accounted for students? senses of autonomy,
competence, and relatedness. It is likely that students who scored higher on the self-
determination pretest would also be more apt to experience affiliation with the institution.
The surprising predictor of sense of belonging was gender. Women students,
although they constituted the majority of respondents and the majority of students at
Magis University, felt a lower level of belonging to the institution than their male
counterparts. Although not tested in this study, this discrepancy in sense of belonging
could be related to the previous finding that women students in the first year of college do
not participate in student organizations at the same rate as men.
174
Research Question Three: The Impact of Involvement
The direct effects of involvement measures on student attributes at Time 2 were
enumerated in the discussion of the first research question. Likewise, the discussion of
involvement?s impact on academic success is better understood in context. Therefore, the
discussion of the effects of involvement on academic success is withheld here in favor of
the discussion of research question four.
Research Question Four: The Direct and Indirect Effects of
Student Attributes on Academic Success
The final research question concerned the direct and indirect effects of variables
in the model on college grade point average. Several constructs have direct positive
effects, but a more interesting story can be found in the indirect effects. These findings
are discussed below.
Direct Effects on Academic Success
Of the three student attribute measures at Time 2, only academic self-concept had
a significant direct effect on academic success as measured by cumulative college grade
point average. Although much of the literature treats posttest measures of academic self-
concept as an outcome (Graham & Cockriel, 1996; House, 2000; Waugh, 2001), this
finding is consistent with at least one previous study asserting the effects of academic
self-concept on performance (Marsh, 2003). In addition, academic self-concept at Time
2 has the greatest direct effect on college GPA, followed closely by academic
175
engagement and high school GPA, which are followed by composite SAT score. All of
these variables were positive predictors of college GPA.
The finding that academic engagement positively predicted college GPA even
when accounting for student attributes is consistent with Kuh?s work on student
engagement (2005; Kuh, Kinzie, Schuh, & Whitt, 2005; National Study of Student
Engagement 2002, 2004). Kuh?s research focuses on institutional practices that increase
students? engagement behaviors; however, the findings in this study suggest that student
attributes are at least partially responsible for the students? dispositions toward
engagement, regardless of institutional interventions. In fact, given the findings of this
study, it is entirely plausible to argue that academic engagement is the behavioral
manifestation of students? self-responsibility. From this perspective, institutions are to be
held accountable for providing environments that challenge students in their pursuit of
higher education, but students are to be held accountable for the extent to which they
engage. Contrary to some of Kuh?s work on student engagement, then, the behavior is
more a function of the person than of the environment (Lewin, 1936).
In addition, this study?s finding that prior achievement was a positive predictor of
college grades is consistent with previous research (Astin, 1993; Marsh 2003). However,
in his comparison of students? pre- and post-college survey responses, Astin asserted that
prior achievement was the single greatest predictor of college grade point average. This
study?s findings were slightly different. By including student attributes as an
intermediate outcome, this study found that academic self-concept at Time 2 was a
slightly better predictor of academic success than prior achievement. This suggests that
some of the effects of prior achievement are mediated by academic self-concept.
176
Indirect Effects on Academic Success
Although academic self-concept at Time 2 had the greatest direct effect on
academic success, many other variables in the model passed through this posttest
measure to have significant indirect effects. Beginning with student background
characteristics, indirect effects are discussed below.
Student background characteristics. In the model specified for this study, gender
had three indirect paths to college GPA. The first two indirect paths passed first through
academic self-concept at Time 1, for which being female was a negative predictor. Thus,
two of gender?s indirect paths to college grade point average favor men, which raises
some concerns about how young women develop their academic self-concepts. The third
path passed first through first-year programs, which enroll more women than men. This
path to college GPA was a positive predictor for women.
The race variable had two indirect paths to college GPA, both of which passed
through self-determination at Time 1. As White students tended to score higher on self-
determination, being White was ultimately a positive predictor of college grades at the
end of the first year. Race was not a significant factor anywhere else in the model,
which raises several red flags. This finding may signify either that the instrument
measuring self-determination does not capture this construct for students of color, or that
there were some environmental circumstances?perhaps the effects of being a student of
color on an overwhelmingly majority campus?that led to differences in scores. More
likely, however, is the explanation that the analytic sample contained such a small
proportion of students of color (so small, in fact, I was unable to tease out separate
racial/ethnic groups) that the data lacked enough power to draw any substantial
177
conclusions with regard to race. This finding may also indicate a need for further
research, given the constraints presented by the particular student population in this study
and the inability to disaggregate the different racial/ethnic groups.
Aside from their direct effects on college grades, high school GPA and the
composite SAT scores had several indirect effects as well. High school grades were
positive predictors for academic self-concept, self-determination, and academic
engagement behaviors. This finding furthers Astin?s (1993) assertion that prior
performance is a robust measure of academic success in college and suggests that
students? academic behaviors tend to carry over from high school into college. The
composite scores for college entrance exams painted a slightly different picture, however.
Although SAT composite was a positive predictor for academic self-concept at Times 1
and 2, which led to a positive indirect path to college grades, it was a negative predictor
of academic involvement. The negative paths through involvement nearly canceled out
the positive effects of SAT composite on student attributes, which led to a nearly
negligible total indirect path for SAT scores. This finding suggests that students who test
well think highly of their abilities and tend to feel more authorship for their experience
but do not necessarily know how to translate this self-confidence into positive behaviors.
Perhaps this finding is cause for concern considering the emphasis college admission
teams, high school counselors, and families place on achieving high scores on college
entrance exams. The other possibility to consider is that the academic program is not
challenging enough to engage students who test well. In this case, the finding becomes a
concern for academic administrators and those faculty who educate first-year students at
Magis University.
178
Student attributes at Time 1. The pretest measures for academic self-concept and
self-determination both had indirect paths to college GPA; social self-concept did not.
Academic self-concept passed through the posttest measure as well as faculty interaction,
while self-determination passed only through academic engagement. The student
attribute measures at Time 1 enhance what previous research suggests about students and
their paths to success in college (Astin, 1993; Kuh, 2005; Kuh, Kinzie, Schuh, & Whitt,
2005). It is important to note that the effects of student attributes are apparent during the
first year of college, and it would be interesting to learn if these effects hold stable,
increase, or decrease over the college experience. These findings underscore the need to
account for student disposition toward the academic enterprise when attempting to
understand how and why students achieve certain outcomes.
Academic and social involvement. Three of the involvement measures had
indirect effects on academic success at the end of the first year of college: enrollment in
first-year programs, interaction with faculty outside the classroom, and behaviors related
to academic engagement. All three passed through academic self-concept at Time 2 on
their way to impacting college GPA. These findings confirm Astin?s (1984) and Kuh?s
(2005; Kuh, Kinzie, Schuh, & Whitt, 2005; National Study of Student Engagement 2002,
2004) assertions that involvement and engagement have positive effects on achievement.
These findings also suggest that the psychological and physical energy students devote to
their college experiences contribute to their personal growth and development, which
can, in turn, influence measures of success. Where these findings differ from previous
work by Astin and Kuh is in the discovery that involvement behaviors?with the
exception of academic engagement, which was discussed earlier as a manifestation of
179
student responsibility?do not exhibit direct effects on college GPA. Astin?s I-E-O
model (1991) contains a direct link between measures of the environment and the
outcome. Instead, the findings of this study support the notion that environmental effects
are mediated through the student before exerting any influence over the outcome. This
point is a bit of a departure from previous research.
Homework time-on-task had no indirect effects on college grades. Taken together
with the fact that homework effort also had no direct effect on college GPA, this finding
raises some concerns. Immediately, this finding suggests that first-year students at Magis
University see no tangible return on their investment of time spent studying. Given the
data, it is possible that those students who devote more hours to studying each week than
their peers do not achieve at higher rates. It is likely that students who spend more time
on their homework could become discouraged that their study habits produce results no
different from their peers who do not work as hard. However, it remains to be seen
whether or not this discouragement would negatively affect their future academic
involvement.
On another note, the range of hours devoted to homework at Magis University
were consistent with the number of hours reported in the National Study of Student
Engagement (2002), that is to say students spent on average fewer than 10 hours per
week studying (M = 5.5, SD = 1.2). However, the average grade point average for
students at the end of their first year at Magis University was relatively high (M = 3.2, SD
= 0.46). This suggests at least some level of grade inflation exists at Magis University,
which may distort the picture of which factors contribute to academic success and which
do not.
180
Total Effects on Academic Success
As expected from previous research (Astin, 1993), prior achievement (by way of
high school grades and college entrance exam scores) are positive predictors of academic
success in the first year of college. Measures of academic involvement also have a
positive predictive ability, as previous research would suggest (Astin, 1993; Pascarella &
Terenzini, 2005). Interestingly, though, above and beyond these typical measures,
student attributes also contribute to academic success in the first year. Again, this finding
suggests that adding student attributes creates a more complete model of student success.
Conclusions
In recent decades, higher education has responded to increasing scrutiny by
endeavoring to express the benefits students receive by attending college (Pace, 1984;
Terenzini, 1994). Calls for greater accountability have resulted in research to
demonstrate how an undergraduate education is a value-added experience worthy of
federal, state, and family investments. Outcomes-based assessment purports to evaluate
program and institutional quality. While these assessment efforts underscore institutional
responsibility to create and maintain effective programs, they often do not include a key
factor to success: the student and his or her responsibility toward earning a quality
education. Even current efforts to measure students? psychological and physical
investments in the undergraduate experience (e.g., the National Survey of Student
Engagement) consider student effort an outcome measuring the institution?s ability to
engage students.
181
The purpose of this study, then, was to begin reframing responsibility for higher
education by demonstrating one way to account for student disposition toward academic
success. By examining student attributes and their contribution to academic and social
involvement, and subsequently to achievement, this study describes higher education as a
partnership between student and institution for which student and institution have
responsibility. The findings of the study indicate the shared responsibility between
student and institution. From these findings, at least four conclusions can be drawn: (a)
accounting for incoming students? attributes contributes to an understanding of academic
success; (b) measurable change in student attributes occurs over the space of one year; (c)
academic engagement can be interpreted as the behavioral manifestation of students?
collegiate responsibilities; and (d) the classic input-environment-outcome model (Astin,
1991) is enhanced by incorporating the development of student attributes as predictors.
First, personal attributes, such as self-concept and self-determination, provide an
avenue for describing the student?s disposition toward the educational process. Although
the attributes tested in this study are not exhaustive, they at least begin to describe the
psychological characteristics that influence success in college. Rather than focus on the
institution?s responsibility to engage students (Kuh, Kinzie, Schuh, & Whitt, 2005), this
study demonstrates that academic and social involvement are products, at least in part, of
matriculating students? academic self-concept and self-determination. Furthermore,
together these attributes help explain success in the first year of college beyond what can
be explained by prior achievement. The results of this study begin to illustrate the
student as actor with responsibility for becoming engaged.
182
Second, this study revealed that student attributes undergo measurable change
over one year. Although some of this change may be a result of natural maturation
(Pascarella & Terenzini, 1991), results clearly indicated that at least part of the change is
attributable to students? academic and social involvement. In this way, this study
substantiates previous research on the impact of academic and social involvement on
students? personal development (Astin, 1994; Berger & Milem, 1999; Huang & Chang,
2004) and affirms the benefits of college attendance. For example, change in academic
self-concept over the first year of college was influenced by involvement such as
enrolling in a first-year program and interacting with faculty. Likewise, growth in self-
determination was influenced by developing a sense of belonging and affiliation with the
institution.
In addition, the results of this study further research addressing self-determination
in education, which heretofore has focused on K-12 environments. Previous researchers
have examined the effects of the K-12 classroom environment on the development of
self-determination (Connell & Wellborn, 1991; Miserandino, 1996). This study extends
this research to the undergraduate environment and confirms that interactions with the
environment influence subsequent self-determination. Finally, while the model for this
study included the possibility of change in student attributes over time, a healthy
skepticism questioned whether or not one academic year was long enough to observe
appreciable change. The results indicated, however, that an academic year is long
enough to manifest personal development, confirming Terenzini, Theophilides, and
Lorang?s (1984) assertion that distinct change in student attributes occurs each year.
183
Third, the results of this study suggest a departure from the prevailing
interpretation of academic engagement. Currently the National Survey of Student
Engagement measures engagement behaviors as an effect of the institutional
environment. This line of research examines how aspects of the environment create
atmospheres that promote student engagement. Although this research is a noble and
worthy undertaking, it unintentionally de-emphasizes the student?s role in becoming
engaged. This study?s findings indicate that engagement is, at least, just as much a
function of the student as it is of the environment. In other words, this study
demonstrates that student attributes predict academic engagement. In this vein, it is
necessary to reframe an understanding of academic success as a shared responsibility
between the student and the environment.
Finally, the results of the study confirm the inclination to incorporate students?
psychological attributes in order to build a better college impact model (Pascarella &
Terenzini, 2005; Stage, 1989). Although ecological perspectives of higher education
(Barker, 1968; Lewin, 1936) stress the person-environment interaction as the catalyst for
behavior or outcomes, college impact models rarely account for the person beyond what
are typically listed as ?student background characteristics.? Theories explaining student
involvement (Astin,1984), involving colleges (Kuh, 1991) and student engagement (Kuh,
2005) all attempt to describe how students interact with the college environment.
However, none of these accounts for why students become engaged academically or
involved socially. Instead these theories privilege the environment as the subject of
action and relegate students to the role of indirect object. In measuring college impact, it
is at least equally important to understand the student as the primary actor in his or her
184
educational endeavors. By incorporating student attributes into existing models for
student involvement, this study sought to bridge the gap between the psychological and
sociological phenomena that produce college outcomes. Rather than viewing these two
aspects as wholly separate, the model tested in this study blended the psychology of
student attributes with the sociology of the college environment to produce a better
model. Results from the study indicate that accounting for student attributes contributes
to our understanding of why students become academically and socially involved.
Furthermore, while student involvement predicts at least some portion of
academic success, alone it does not tell the whole story. Instead, the psychological
dimensions of student attributes greatly enhance an understanding of the path to academic
success. The effects of the environment on the outcome were indirect rather than direct.
In other words, the environmental effects on academic success were mediated through the
student (represented by academic self-concept at Time 2) rather than directly linked.
Implications
The results of this study suggest implications for practice, policy, and theory. The
following section outlines contributions to practice and policy and then offers
recommendations for future research.
Contributions to Practice and Policy
This study highlights the role students play in their own academic success,
including the responsibility students bear for becoming engaged in their own educational
process. This approach to understanding academic success suggests at least one policy
185
implication and four recommendations for practice. Given the particularity of locating
the study in a single institution, the recommendations for practice are tailored to meet the
needs at Magis University. Certainly, the findings of this study may resonate with the
climate of student achievement at other institutions, in which case faculty and
administrators at those institutions also may find use for the following recommendations.
Directions for practice fall into four areas: (a) education for prospective college students
about the student?s role in achieving academic success; (b) admission practices; (c)
setting expectations for students and how they engage their own learning; and (d)
attention to first-generation students. These recommendations are outlined below,
followed by the policy implication that has emerged from this study.
Prospective Student Education
Magis University, like many other institutions, produces a broad array of
marketing materials for prospective students. Colorful, eye-catching brochures describe
every aspect of the university from residential living and the fitness center to the full
listing of academic majors. Although intended to inform prospective students who are
preparing to choose a college community to join, these materials inadvertently may send
a message that is counter-intuitive. Higher education theorists (Frank, 2001; Winston &
Zimmerman, 2000) describe this phenomenon of amenities-first-marketing as the great
?arms race? of higher education. In an effort to attract students to choose Magis,
administrators may be feeding into a growing consumer mentality in which prospective
students and their parents shop for the best amenities at the greatest value rather than
looking for an institution in which the student can flourish intellectually.
186
In order to counter-balance the marketing materials highlighting student services
and facilities, the admission staff may consider developing equally captivating tools,
from additional brochures to specially designed programs for prospective students and
high school guidance counselors, that describe the student?s responsibility to become
academically and socially involved. These developments would capitalize on the
findings presented in this study which suggest that students who assume responsibility for
their experience and exhibit involvement behaviors are more likely to succeed
academically. As a Jesuit institution, Magis University is committed to developing the
student as a whole person. Certainly well-placed messages about the student?s role in
achieving academic success would affirm the institution?s guiding philosophy. In this
way, Magis would send a clear message about the partnership it enters with each student
who matriculates?a partnership in which the institution promises to provide challenging
academic and social opportunities in a supportive environment and the student promises
to engage. This type of communication to prospective students prepares them to
understand themselves as authors of their own experience rather than consumers of goods
and services.
Admission Practices
Beyond early communication to prospective students about how to engage the
Magis University environment, the findings of this study suggest a need for altering some
admission practices. In weighing prospective student profiles, Magis University?like
many other institutions across the country?considers college entrance exam scores in the
admission decision. Although the results of the study indicated that the SAT composite
187
score predicts at least some portion of college grade point average in the first year, it is
less predictive than high school grades. Furthermore, findings demonstrated that SAT
composite is a negative predictor of academic involvement, including engagement
behavior, which was a positive predictor of academic success. Although it is unrealistic
to imagine Magis University eliminating college entrance exam scores from the student
profile, it is reasonable to suggest that admission professionals reconsider the weight they
assign to these scores. In addition to adjusting the prominence of entrance exam scores in
the admission decision, Magis University may consider methods for ascertaining
prospective students? academic self-concept and self-determination, both of which
predicted different aspects of students? academic and social engagement in their first year
of college. Given the results of this study, these shifts in practice have the potential to
yield a student body that is better disposed to meet the demands of higher education.
Setting Expectations for Involvement
Admission practices aside, Magis University would do well to communicate high
expectations for students? academic and social involvement. The findings of this study
suggest that of all the involvement measures, only behaviors depicting academic
engagement are directly linked to academic success in a significant manner. Therefore,
Magis University could consider focusing on delineating and setting expectations for
students? active engagement in their learning processes.
Setting expectations begins the process of reframing accountability for academic
success to include student responsibility for engaging the academic and social
environment. One approach to communicating expectations is developing rubrics for
188
academic and social involvement. Currently, rubrics are offered as an alternative method
for assessing students? academic progress (Anderson, 1998). A rubric involves a defined
set of competencies or goals as well as the dimensions or criteria for achieving them. An
involvement rubric, then, might indicate ?engaged learning? as a desirable competency
and then list a set of behaviors associated with academic engagement, such as completing
reading assignments before class and contributing to classroom discussions. In addition,
a rubric also might delineate the spectrum of progress. For instance, in setting an
expectation for social involvement, a rubric might define behaviors from beginning
involvement, which might involve participating in a student club, attending a retreat, or
serving the greater community, all the way to exemplary involvement, which might
involve taking a leadership role in an organization or activity. Armed with a set of
rubrics outlining desired academic and social development outcomes, students would be
in a better position to see themselves as authors of their own experience. Although this
recommendation for practice does not address initial student attributes that contribute to
academic success, it is an environmental intervention with the potential for influencing
growth and positive change in the psychological dimensions that predict success.
First-generation Students
Although this was not a study focused on first-generation students, one finding
stood out as deserving attention. Results from the causal model indicated that first-
generation status had a negative influence over social self-concept at the end of the first
year of college. Whereas first-generation status did not predict any other variable in the
model (i.e., first-generation students appear to have similar experiences as those students
189
who have a family history of college attendance), it did negatively predict social self-
concept. This finding may indicate a social atmosphere or environment that favors in
some way one sub-population over another. A needs or climate assessment for first-
generation students may help Magis University understand the needs of a sub-population
that otherwise may remain hidden by the student culture.
Policy Recommendation: Rethinking Approaches to Accountability
The results of this study provide a new perspective on accountability for academic
success in the first year of college. Specifically, findings indicated that the student brings
his or her own psychological attributes to bear on the educational process, and these
dispositions can be used to predict achievement. Given these results, it seems imperative
to reframe the federal- and state-level conversations around accountability?particularly
where performance-based funding is concerned. Performance-based funding is a system
of ?allocating resources to institutions based on the extent to which they achieve
previously established goals, objectives, and outcomes? (Layzell, 1999, p. 233).
Among other indicators, Layzell (1999) listed inputs and outcomes, such as SAT
scores for the entering class and retention rates, as well as ?customer? needs, such as
student satisfaction and employment rates for graduating seniors, as measures states
commonly used to evaluate institutional effectiveness. In this way, state policymakers
monitor institutional progress toward explicit goals and then often use the information to
determine funding levels. Although the performance-based funding approach ensures
institutional accountability to state legislators and the tax-paying public, they mistakenly
place the burden of responsibility for student success squarely on the institutions
190
themselves. Instead, the results of this study indicate that students, by their psychological
dispositions, are at least partly responsible for their own success. Whereas the institution
is responsible for providing an engaging environment, students are responsible for how
and when they choose to engage?academically and socially. Furthermore, performance
indicators addressing student satisfaction without the caveat for student responsibility
reinforces the notion of student-as-consumer of an educational product. Inadvertently,
state monitoring procedures may counteract efforts to engage students as learners who
earn rather than consume a higher education.
Given these remarks, the policy recommendation stemming from this study
focuses on reframing the monitoring systems that hold institutions accountable to the
public. Although this is not a recommendation to eliminate performance-based funding,
it is a proposal to address the philosophical underpinnings that guide the choice of
monitoring criteria. Understanding the shared responsibility between institution and
student for academic success may not eliminate criteria such as effective use of human
and other resources or impact on state economy. However, balancing institutional and
student responsibility for learning may alter the weight given to student satisfaction and
similar criteria.
Recommendations for Future Research
No one study can answer all questions about a given topic. Although this study
was a good first step in understanding how student attributes contribute to academic
success in the first year of college, it was just that?a first step. Additional research is
191
needed to address some of the questions raised by the current study. This section
describes five areas for investigation in the future.
First, the model in this study produced interesting results. However, further
testing and tweaking of the model is required. This study examined only three non-
cognitive predictors of success?academic self-concept, social self-concept, and self-
determination as measured by basic psychological needs. What other student attributes
would contribute to the model? Additional elements of the psychosocial dimension, such
as self-understanding, intellectual orientation, self-esteem, or alternative measures of
autonomy (Pascarella & Terenzini, 2005) may contribute to the model just as well as, or
perhaps even better than, the three student attributes that were included. In addition, this
model tested only one aspect of student success: grade point average at the end of the first
year of college. Other outcomes that may suit the model include, first-to-second-year
retention, successful course completion, or major declaration (Upcraft, Gardner, &
Barefoot, 2005). Moreover, the model could be extended to capture academic success
over four years of college, in which case the dependent variables would reflect senior-
year data, such as graduation rates, cumulative undergraduate grade point average,
employment rates, or graduate school acceptance rates.
With regard to the model itself, future investigators may see a need to adjust the
placement of the variable describing enrollment in first-year programs. Although this
study included first-year programs as a measure of involvement (Astin, 1984), other
researchers (Fidler & Stuart Hunter,1989; Stuart Hunter & Linder, 2005) make
convincing arguments in another direction. Specifically, their research indicates that
enrolling in first-year programs leads to greater academic and social involvement.
192
Therefore, a theory supports enrollment in specific programmatic interventions as an
intermediate outcome with direct influence on engagement. For the purposes of testing
programmatic efficacy, future researchers may modify the model presented in the current
study by placing the programmatic variable between those variables describing student
attributes and those describing involvement.
A second direction for future research is greater examination of the influence of
race/ethnicity. The analytic sample for the current study was not diverse enough to draw
substantial inferences with respect to the role of race/ethnicity in the model. Because of
the small percentage of non-White students in the sample, the current study can say
nothing about how well the model predicts academic success in the first year with regard
to students of color. Furthermore, the students of color represented in the current analytic
sample could not be disaggregated by racial or ethnic group, further limiting any
understanding of how well the model works. Future researchers will want to test the
model at institutions with greater diversity so as to provide more robust comparison
groups by race.
The third recommendation for additional study is to develop a better
understanding of student attributes that relate to success. For example, this study utilized
an instrument measuring basic psychological needs (Deci & Ryan, 2000) to capture the
construct of self-determination. Future researchers will want to investigate different
instruments that measure self-determination and compare their results to those presented
in this study. In addition, self-efficacy was dropped from the model despite the
theoretical foundations that support its inclusion. Further investigation may uncover a
more robust posttest measurement for self-efficacy, in which case I would recommend
193
reintroducing self-efficacy to the model. Future research should build upon the current
study and begin to unfold the complexity surrounding student attributes that impact
success. If results prove to be consistent, I recommend advocating for the inclusion of
items measuring these attributes in survey instruments designed to measure college
impact. The findings of this study indicate that national surveys, such as those sponsored
by the Higher Education Research Institute and the National Survey of Student
Engagement, would provide more complete data for outcomes assessment if they also
included reliable measures of student attributes that predict success. Furthermore, these
findings indicate a possible amendment of I-E-O models in which academic success is
the identified outcome. Future testing of these models should include student attributes
in the intermediate outcome position between measures of the environment and the
outcome.
The fourth recommendation for future research is to develop a multi-level model
that will contribute to an understanding of the relative impact of student attributes and
institutional environment on student success. As a first step, the current study focused on
understanding the contribution of student attributes toward academic success. As a result
the study was delimited to a single institution in order to isolate environmental
differences that may also influence success. Future researchers may choose to advance
the current study by investigating the effects of student attributes within several different
institutional types. A study along these lines could advance knowledge by describing
how different environments influence the development of student attributes, whether
different environments attract students with particular psychological traits, and how
194
student attributes influence academic and social involvement within different
environments.
A fifth direction for future research addresses the implications of methodological
choice. As a quantitative investigation, the current study accesses predictors of success
only on the macro level. In order to understand the particulars of student attributes and
their impact on success, qualitative analysis is a necessary companion to the current
study. A qualitative methodology, such as case study, would provide insight into the
experiences of students and those university educators who interact with them. By
describing their own stories, students and educators may shed light onto students?
processes of developing self-concept or self-determination. What programs, persons, or
experiences promote growth? In what ways do students feel responsible for their own
education, or conversely, in what respects do they hold the institution responsible? Other
questions may describe the characteristics or experiences typical of students with high
versus low student attribute levels. This level of analysis cannot be achieved through
quantitative methods, yet this type of contribution is necessary to build our understanding
of students and their paths to success.
195
Final Thoughts
The current climate of assessment in higher education is designed to express how
students benefit from college attendance and how colleges and universities are uniquely
poised to provide these benefits. The current study contributes to assessment philosophy
by demonstrating that student attributes predict academic success beyond what can be
explained by prior achievement and involvement. Specifically, the findings suggested
that prior measures of academic self-concept and self-determination directly influence
students? patterns of academic and social involvement, and thereby indirectly influence
achievement. Furthermore, study results demonstrated that the later measure of academic
self-concept had the greatest direct effect on college grade point average and was a better
predictor than prior achievement. This finding is particularly notable for two reasons: (a)
inserting student attributes as an intermediate outcome between the environment and the
outcome of interest adds value to the traditional I-E-O model (Astin, 1991); and (b) the
effects of the environment, which are emphasized by leading higher education
researchers (Astin, 1993; Kuh, 1991; 2005; Kuh, Kinzie, Schuh, & Whitt, 2005), are
largely mediated through the student. These results confirm previous admonitions to wed
the psychological orientation of student development theory to the sociological
orientation of college impact in order to build more effective models (Pascarella &
Terenzini, 2005; Stage, 1989).
In addition, this study creates implications for practice, policy, and future
research. This study is an important first step in understanding how student attributes
contribute to academic success in the first year of college. Further research is necessary
to understand the different psychological dimensions that contribute to success as well as
196
the multiple measures of success they might predict. Additionally researchers who build
multi-level studies based on the current research will be able to determine the relative
impact of student attributes and institutional environment on student success.
With regard to policy and practice, this study?s focus on the effect of student
attributes on academic success indicates a need to reframe the discussions of
accountability for student success. Educational efforts on the part of institutions will
improve students? awareness of their role in the educational process?that while the
institution provides an engaging environment, students are responsible for their own
engagement. Likewise, this study suggests a need to educate policymakers on
appropriate criteria for funding initiatives. Although policymakers are duty-bound to
hold institutions accountable, current assessment practices unintentionally may be
creating a climate where the public is encouraged to see students as consumers of an
educational good rather than learners with authorship over the extent to which they
engage. The results of this study underscore the need to view responsibility for student
success as shared between the institution and the student. Students who see themselves
as authors of their college experience are more likely to engage in meaningful ways and
subsequently be in a better position to succeed.
197
Appendix A
Definitions of Variables in the Study with Associated Theoretical Concepts
Theoretical
Concept Variable Definition
CIRP
Student background
characteristics
Gender Dichotomous variable (0 =
male; 1 = female)
Q #1
Race/ethnicity Composite of dichotomous
variable to identify
race/ethnicity (0/1):
White (reference) and
Students of color comprised
of multiracial, African
American, Native American,
Asian/Pacific American,
Latino.
Q #25
First-generation
student status:
Mother?s/Father?s
education
Composite of categorical
variables (0 = college
degree; 1 = no more than
high school; 2 = greater than
high school but less than
college; 3 = greater than
college)
Q #28
High school GPA Continuous variable from
student self-reported data.
The grade scale ranges from
1 to 7, with 1 = 1.0 GPA or
less and 7 = 4.0 or greater.
Q #7
College entrance
exam score
Z-score of composite
continuous variable adding
math and verbal SAT results
Inst. Data
198
Theoretical
Concept Variable Definition
CIRP
Student Attributes at
Time 1
Academic self-
concept
A continuous interval
variable on a Likert-type
scale based on student self-
reported data (1 = lowest
10%; 2 = below average; 3 =
average; 4 = above average;
5 = highest 10%), including
academic ability, computer
skills, mathematical ability,
public speaking ability,
intellectual self-confidence,
and writing ability.
Factor
analysis
of items
in Q #31
Social self-concept A continuous interval
variable on a Likert-type
scale based on student self-
reported data (1 = lowest
10%; 2 = below average; 3 =
average; 4 = above average;
5 = highest 10%), including
kindness, leadership ability,
social self-confidence, and
understanding of others.
Factor
analysis
of items
in Q #31
SRA1
Self-determination A continuous interval
variable on a 7-point Likert-
type scale based on student
self-reported data (1-2 = not
at all true; 3-5 = somewhat
true; 6-7 = very true)
Factor
analysis
of Q #s 1
through
21
199
Theoretical
Concept Variable Definition
YFCY
First-year Experience First-year programs A categorical variable based
on student self-reported data
(0 = no seminar; 1 =
extended orientation
seminar, 2 = academic or
pre-professional seminar, 3
= other)
An item
in Q # 17
Faculty Interaction A continuous interval
variable on a Likert-type
scale based on student self-
reported data (1 = not at all;
2 = rarely; 3 = occasionally;
4 = frequently)
Factor
analysis
of items
in Q # 5
Academic
engagement
A continuous interval
variable on a Likert-type
scale based on student self-
reported data (1 = not at all;
2 = rarely; 3 = occasionally;
4 = frequently)
Factor
analysis
of items
in Q #s
12, 18
Homework time-on-
task
A continuous variable based
on student self-reported data
(1 = none; 2 = less than 1
hour; 3 = 1-2; 4 = 3-5; 5 = 6-
10; 6 = 11-15; 7 = 16-20; 8 =
21-30; 9 = over 30)
An item
in Q #21
Student Clubs A continuous variable based
on student self-reported data
(1 = none; 2 = less than 1
hour; 3 = 1-2; 4 = 3-5; 5 = 6-
10; 6 = 11-15; 7 = 16-20; 8 =
21-30; 9 = over 30)
An item
in Q #21
200
Theoretical
Concept Variable Definition
YFCY
First-year Experience Sense of Belonging A continuous interval
variable on a Likert-type
scale based on student self-
reported data (1 = strongly
agree; 2 = agree; 3 =
disagree; 4 = strongly
disagree)
Factor
analysis
of items
in Q # 15
Student Attributes at
Time 2
Academic self-
concept
A continuous interval
variable on a Likert-type
scale based on student self-
reported data (1 = lowest
10%; 2 = below average; 3 =
average; 4 = above average;
5 = highest 10%), including
academic ability, computer
skills, mathematical ability,
public speaking ability,
intellectual self-confidence,
and writing ability.
Factor
analysis
of items
in Q #9
Social self-concept A continuous interval
variable on a Likert-type
scale based on student self-
reported data (1 = lowest
10%; 2 = below average; 3 =
average; 4 = above average;
5 = highest 10%), including
leadership ability, popularity,
social self-confidence, and
understanding of others.
Factor
analysis
of items
in Q #9
201
Theoretical
Concept Variable Definition
SRA
Student Attributes at
Time 2
Self-determination A continuous interval
variable on a 7-point Likert-
type scale based on student
self-reported data (1-2 = not
at all true; 3-5 = somewhat
true; 6-7 = very true)
Factor
analysis
of Q #s 1
through
21
Academic Success Cumulative GPA Continuous variable. The
grade scale ranges from 1 to
7, with 1 = 1.0 GPA or less
and 7 = 4.0 or greater.
Inst. data
202
APPENDIX B
2004 Cooperative Institutional Research Program Student Information Form
Higher Education Research Institute
203
204
205
206
207
APPENDIX C
Student Readiness Assessment
208
209
210
APPENDIX D
2005 Your First College Year
Higher Education Research Institute
211
212
213
214
215
APPENDIX E
2004 Cooperative Institutional Research Program Student Information Form
Higher Education Research Institute
Components of the Factor Composite: Self-concept
Factor Components Loadings
Academic Self-concept
Self-confidence (intellectual) .776
Academic Ability .746
Drive to Achieve .651
Writing Ability .632
Social Self-concept
Compassion .768
Generosity .759
Cooperativeness .672
Understanding of Others .667
Self-esteem
Self-understanding .783
Emotional Health .676
Self-confidence (social) .625
Forgiveness .522
Leadership Self-concept
Leadership Ability .747
Creativity .644
Public Speaking Ability .589
Courage .554
216
Source: Analyses of CIRP:04, Magis University
Note. Items asked respondents to: ?Rate yourself on each of the following traits as compared with the
average person your age.?
Factor Components Loadings
Spirituality
Religiousness .916
Spirituality .903
217
REFERENCES
Anderson, R. S. (1998). Why talk about different ways to grade? The shift from
traditional assessment to alternative assessment. New Directions for Teaching and
Learning, 1998, (74), 5-16.
Arrison, E. G. (1998). Academic self-confidence as a predictor of first year college
student quality of effort and achievement. Dissertation Abstracts International, 59
(3-A), 0742. (UMI No. 9826138)
Asher, H. B. (1983). Quantitative Applications in the Social Sciences: Vol. 3. Causal
modeling. (2
nd
ed.). Beverly Hills, CA: Sage.
Astin, A. W. (1962). Influences on the student?s motivation to seek advanced training:
Another look. Journal of Educational Psychology, 53, 303-309.
Astin, A. W. (1977). Four critical years. Effects of college on beliefs, attitudes, and
knowledge. San Francisco: Jossey-Bass Publishers.
Astin, A. W. (1984). Student involvement: A developmental theory for higher education.
Journal of College Student Personnel, 25, (4), 297-308.
Astin, A. W. (1985). Achieving educational excellence: A critical assessment of priorities
and practices in higher education. San Francisco: Jossey-Bass Publishers.
Astin, A. W. (1991). Assessment for excellence: The philosophy and practice of
assessment and evaluation in higher education. New York: American Council on
Education/Macmillan.
Astin, A. W. (1993). What matters in college? Four critical years revisited. San
Francisco: Jossey-Bass Publishers.
218
Astin, A. W. (1999). Involvement in Learning revisited: Lessons we have learned.
Journal of College Student Personnel, 40, (5), 587-598.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.
Psychological Review, 84, (2), 191-215.
Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist,
37, (2), 122-147.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W. H. Freeman.
Banning, J. H. (1989). Impact of college environments on freshman students. In M. L
Upcraft; J. N. Gardner, & Associates (Eds.). The Freshman Year Experience (pp.
53-62). San Francisco: Jossey-Bass Publishers.
Barefoot, B. O. (1993). Exploring the evidence: Reporting outcomes of freshman
seminars. Columbia: University of South Carolina, National Resource Center for
the Freshman Year Experience.
Barefoot, B. O., Gardner, J. N., Cutright, M., Morris, L. V., Schroeder, C. C., Schwartz,
S. W., Siegel, M. J., & Swing, R. L. (2005). Achieving and sustaining institutional
excellence for the first year of college. San Francisco: Jossey-Bass.
Berger, J. B. (1997). Students? sense of community in residence halls, social integration,
and first-year persistence. Journal of College Student Development 38, (5), 441-
453.
Berger, J. B. & Milem, J. F. (1999). The role of student involvement and perceptions of
integration in a causal model of student persistence. Research in Higher
Education, 40, (6), 641-664.
219
Berger, J. B. & Milem, J. F. (2000). Exploring the impact of historically Black colleges in
promoting the development of undergraduates? self-concept. Journal of College
Student Development 41, (4), 381-394.
Biondi, L. (1989). Educational aims of the liberal arts curriculum: Contextual education.
In R. E. Bonachea (Ed.), Jesuit Higher Education, 94-104. Pittsburgh, PA:
Duquesne University Press.
Blimling, G. S. (1999). A meta-analysis of the influence of college residence halls on
academic performance. Journal of College Student Development, 40, (5), 551-
561.
Brown, S. D., Lent, R. W., & Larkin, K. C. (1989). Self-efficacy as a moderator of
scholastic aptitude?Academic performance relationships. Journal of Vocational
Behavior, 35, (1), 64-75.
Bryant, F. B., & Yarnold, P. R. (1995). Principle-components analysis and exploratory
and confirmatory factor analysis. In L. G. Grimm & P. R. Yarnold (Eds.),
Reading and understanding multivariate statistics (pp. 99-136). Washington, DC:
American Psychological Association.
Chickering, A. W., & Reisser, L. (1993). Education and Identity (2
nd
Ed.). San Francisco:
Jossey-Bass Publishers.
Chrissman Ishler, J. L., & Upcraft, M. L. (2005). The keys to first-year student
persistence. In M. Upcraft; J. N. Gardner, & B. O. Barefoot (Eds.). Challenging &
Supporting the First-year Student: A Handbook for Improving the First Year of
College (pp. 27-46). San Francisco: Jossey-Bass Publishers.
220
Cohen, J., & Cohen, P. (1983). Missing data. In J. Cohen & P. Cohen (Eds.), Applied
multivariate regression/Correlation analyses for the behavioral sciences (2
nd
ed.,
pp. 275-300). Hillsdale, NJ: Lawrence Erlbaum.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple
regression/correlation analysis for the behavioral sciences (3
rd
ed.). Mahwah, NJ:
Lawrence Erlbaum Associates.
Cokley, K., Komarraju, M., King, A., Cunningham, D., & Muhammad, G. (2003). Ethnic
differences in the measurement of academic self-concept in a sample of African
American and European American college students. Educational and
Psychological Measurement, 63, (4), 707-722.
Connell, J. P., & Wellborn, J. G. (1991). Competence, autonomy, and relatedness: A
motivational analysis of self-system processes. In M. R. Gunnar & L. A. Sroufe
(Eds.), The Minnesota Symposia on Child Development: Vol. 23. Self Processes
and Development (pp. 43-77).
Creswell, J. W. (2003). Research design: Qualitative, quantitative, and mixed methods
approaches (2
nd
ed.). Thousand Oaks, CA: Sage Publications.
Deci, E. L. (1995). Why we do what we do: Understanding self-motivation. New
York: Penguin Books.
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human
behavior. New York: Plenum.
Deci, E. L., & Ryan, R. M. (2004). Basic Psychological Needs Survey [Survey
instrument]. Available from University of Rochester Web site,
http://www.psychology.rochester.edu/SDT.
221
Deci, E. L., Ryan, R. M., & Williams, G. C. (1996). Need satisfaction and the self-
regulation of learning. Learning & Individual Differences, 8, (3), 165-183.
Deci, E. L., Vallerand, R. J., Pelletier, L. G., & Ryan, R. M. (1991). Motivation and
education: The self-determination perspective. Educational Psychology, 26, (3 &
4), 325-346.
Delaney, A. M. (2004, November). Discovering what makes a difference in the first year
in college: Insights to enlighten planning. Paper presented at the meeting of the
Conference of the North East Association for Institutional Research, Portsmouth,
NH.
DeWitz, S. J. & Walsh, W. B. (2002, August). Self-efficacy and college student
satisfaction. Journal of Career Assessment, 10, (3), 315-326.
Donohue, J. W. (1963). Jesuit Education: An essay on the foundations of its idea. New
York: Fordham University Press.
Emirbayer, M. & Mische, A. (1998). What is agency? The American Journal of
Sociology, 103 (4), 962-1023.
Feldman, K. A., & Newcomb, T. M. (1994). The impact of college on students. New
Brunswick, NJ: Transaction.
Ferguson, G. A. (1959). Statistical analysis in psychology and education. New York:
McGraw-Hill.
Fidler, P. P., & Stuart Hunter, M. (1989). How seminars enhance student success. In M. L
Upcraft; J. N. Gardner, & Associates (Eds.). The Freshman Year Experience (pp.
216-237). San Francisco: Jossey-Bass Publishers.
222
Flynn, J. E. (1989, March). Sharing the gift: Academic excellence and Ignatian
spirituality: An invitation to partnership in American Jesuit higher education.
(Dissertation). Newton Centre, Massachusetts: Andover Newton Theological
School.
Frank, R. H. (2001). Higher education: The ultimate winner-take-all market. In M.
Devlin, & J. Meyerson (Eds.) Forum Futures: Exploring the Future of Higher
Education (pp. 3-12). San Francisco: Jossey-Bass Publishers.
Gagn?, M. (2003). The role of autonomy support and autonomy orientation in prosocial
behavior engagement. Motivation and Emotion, 27, (3), 199-223.
Goodsell Love, A., & Tokuno K. A. (1999). Learning communities models. In J. H.
Levine (Ed.), Learning communities: New structures, new partnerships for
learning. (pp. 9-17). Columbia, SC: University of South Carolina, National
Resource Center for The First-Year Experience and Students in Transition.
Gordon, V. P. (1989). Origins and purposes of the freshman seminar. In M. L
Upcraft; J. N. Gardner, & Associates (Eds.). The Freshman Year Experience (pp.
183-197). San Francisco: Jossey-Bass Publishers.
Graham, S., & Cockriel, I. (1996). Indexes to assess social and personal development and
the impact of college. College Student Journal, 30, (4), 502-515.
Guay, F., Sen?cal, C., Gauthier, L., & Fernet, C. (2003). Predicting career indecision: A
self-determination theory perspective. Journal of Counseling Psychology, 50, (2),
165-177.
Hawley, P. H. (1999). The ontogenesis of social dominance: A strategy-based
evolutionary perspective. Developmental Review, 19, 97-132.
223
Hernandez, K., Hogan, S., Hathaway, C. & Lovell, C. D. (1999). Analysis of the
literature on the impact of student involvement on student development and
learning: More questions than answers?. NASPA Journal, 36, (3), 184-197.
Higher Education Research Institute. (2004). Student Information Form [Survey
instrument]. Available from HERI Web site, http://www.gseis.ucla.edu/heri.
Higher Education Research Institute. (2005). Your First College Year [Survey
instrument]. Available from HERI Web site, http://www.gseis.ucla.edu/heri.
House, J. D. (2000). The effect of student involvement on the development of academic
self-concept. The Journal of Social Psychology, 140, (2), 261-263.
Huang, Y. & Chang, S. (2004). Academic and cocurricular involvement: Their
relationship and the best combinations for student growth. Journal of College
Student Development, 45, (4), 391-406.
Inkelas, K. K., & Weisman, J. L. (2003). Different by design: An examination of student
outcomes among participants in three types of living-learning programs. Journal
of College Student Development, 44, (3), 335-368.
Jewler, A. J. (1989). Elements of an effective seminar: The University 101 program. In
M. L. Upcraft; J. N. Gardner, & Associates (Eds.). The Freshman Year
Experience (pp. 198-215). San Francisco: Jossey-Bass Publishers.
Kezar, A., & Kinzie, J. (2006). Examining the ways institutions create student
engagement: The role of mission. Journal of College Student Development, 45,
(2), 149-171).
224
Kezar, A., & Moriarty, D. (2000). Expanding our understanding of student leadership
development: A study exploring gender and ethnic identity. Journal of College
Student Development, 41, (1), 55-69.
Klem, L. (1995). Path analysis. In L. G. Grimm & P. R. Yarnold (Eds.), Reading and
understanding multivariate statistics (pp. 65-98). Washington, DC: American
Psychological Association.
Klem, L. (2000). Structural equation modeling. In L. G. Grimm & P. R. Yarnold (Eds.),
Reading and understanding more multivariate statistics (pp. 227-259).
Washington,vDC: American Psychological Association.
Kline, R. B. (2005). Principles and practice of structural equation modeling (2
nd
ed.)New
York: The Guilford Press.
Kuh, G. D. (1991). Characteristics of involving colleges. In G. D. Kuh & J. H. Schuh
(Eds.). The Role and Contribution of Student Affairs in Involving Colleges (pp.
11-29). Washington, DC: National Association of Student Personnel
Administrators.
Kuh, G. D. (1993). In their own words: What students learn outside the classroom.
American Educational Research Journal, 30, (2), 277-304.
Kuh, G. D. (2005). Student engagement in the first year of college. In M. L Upcraft, J. N.
Gardner, & B. O. Barefoot (Eds.). Challenging & Supporting the First-year
Student: A Handbook for Improving the First Year of College (pp. 86-107). San
Francisco: Jossey-Bass Publishers.
225
Kuh, G. D., Gonyea, R. M., & Williams, J. M. (2005). What students expect from college
and what they get. In T. E. Miller, B. E. Bender, & J. H. Schuh (Eds.). Promoting
Reasonable Expectations: Aligning Student and Institutional Views of the College
Experience (pp. 34-64). San Francisco: Jossey-Bass Publishers.
Kuh, G. D., Kinzie, J., Schuh, J. H., Whitt, E. J. & Associates. (2005). Student success in
college: Creating conditions that matter. San Francisco: Jossey-Bass Publishers.
LaSere Ericson, B., & Strommer, D. W. (2005). Inside the first-year classroom:
Challenges and constraints. In M. L Upcraft; J. N. Gardner, & B. O. Barefoot
(Eds.). Challenging & Supporting the First-year Student: A Handbook for
Improving the First Year of College (pp. 241-256). San Francisco: Jossey-Bass
Publishers.
Layzell, D. T. (1999). Linking performance to funding outcomes at the state level for
public institutions of higher education: Past, present, and future. Research in
Higher Education, 40, (2), 233-246.
Leslie, L. L., McClure, G. T., & Oaxaca, R. L. (1998). Women and minorities in science
and engineering: A life sequence analysis. The Journal of Higher Education, 69,
239-276.
Levine Laufgraben, J. (2005). Learning communities. In M. Upcraft; J. N. Gardner, & B.
O. Barefoot (Eds.). Challenging & Supporting the First-year Student: A
Handbook for Improving the First Year of College (pp. 371-387). San Francisco:
Jossey-Bass Publishers.
Lewin, R. (1936). Principles of topological psychology (F. Heider & G. M. Heider,
Trans.). New York: McGraw-Hill.
226
Little, T. D. (1998). Sociocultural influences on the development of children?s action-
control beliefs. In J. Heckhausen & C. S. Dweck (Eds.), Motivation and self-
regulation across the life span (pp. 281-315). New York: Cambridge University
Press.
Little, T. D., Hawley, P. H., Heinrich, C. C., & Marsland, K. W. (2002). Three views of
the agentic self: A developmental synthesis. In E. L. Deci & R. M. Ryan (Eds.)
Handbook of Self-determination Research (pp. 3-33). Rochester, NY: The
University of Rochester Press.
Lubrano, A. (2004). Limbo: Blue-collar roots, white-collar dreams. Hoboken, NJ: John
Wiley & Sons.
Mara?a, B. (June-July, 2004). Living at Loyola: Find your passions. The
Greyhound, 77 (26), p. 5.
Marsh, H. W. (2003, November). A reciprocal effects model of the causal ordering of
academic self-concept and achievement. Paper presented at the meeting of New
Zealand Association for Research in Education/Australian Association for
Research in Education, Auckland, New Zealand.
Marsh, H. W., & Yeung, A. S. (1998). Longitudinal structural equation models of
academic self-concept and achievement: Gender differences in the development
of math and English constructs. American Educational Research Journal, 35, (4),
705-738.
McInnes, W. C. (1989). The current state of the Jesuit philosophy of education. In R. E.
Bonachea (Ed.), Jesuit Higher Education, 26-45. Pittsburgh, PA: Duquesne
University Press.
227
Milem, J. F., & Berger, J. B. (1997). A modified model of college student persistence:
Exploring the relationship between Astin?s Theory of Involvement and Tinto?s
Theory of Student Departure. Journal of College Student Development, 38, (4),
387-400.
Miller, T. E., Bender, B. E., Schuh, J. H., & Associates. (2005). Promoting reasonable
expectations: Aligning student and institutional views of the college experience.
San Francisco: Josse-Bass Publishers.
Miserandino, M. (1996). Children who do well in school: Individual differences in
perceived competence and autonomy in above-average children. Journal of
Educational Psychology, 88, (2), 203-214.
Multon, K. D., Brown, S. D. & Lent, R. W. (1991, January). Relation of self-efficacy
beliefs to academic outcomes: A meta-analytic investigation. Journal of
Counseling Psychology, 38, (1), 30-38.
National Survey of Student Engagement. (2002). Student engagement: Pathways to
collegiate success (2002 Annual Survey Results). Bloomington, IN: Author.
National Survey of Student Engagement. (2004). Student engagement: Pathways to
collegiate success (2004 Annual Survey Results). Bloomington, IN: Author.
Pace, C. R. (1984). Measuring the quality of college student experiences: An account of
the development and use of the College Student Experiences Questionnaire, 4-22.
Los Angeles: Higher Education Research Institute, Graduate School of Education,
University of California, Los Angeles.
228
Pascarella, E. T. (1985). College environmental influences on learning and cognitive
development: A critical review and synthesis. In J. Smart (Ed.), Higher
education: Handbook of theory and research (Vol. 1). New York: Agathon.
Pascarella, E. T., Pierson, C. T., Wolniak, G. C., & Terenzini, P. T. (2004). First-
generation college students: Additional evidence on college experiences and
outcomes. The Journal of Higher Education, 75, (3), 249-284.
Pascarella, E. T., Smart, J. C., Ethington, C. A., & Nettles, M. T. (1987). The influence of
college on self-concept: A consideration of race and gender differences. American
Educational Research Journal, 24, (1), 49-77.
Pascarella, E. T. & Terenzini, P. T. (1991). How college affects students. San Francisco:
Jossey-Bass Publishers
Pascarella, E. T. & Terenzini, P. T. (1998). Studying college students in the 21
st
century:
Meeting new challenges. Review of Higher Education, 21, (2), 151-165.
Pascarella, E. T. & Terenzini, P. T. (2005). How college affects students: A third decade
of research. San Francisco: Jossey-Bass Publishers
Pedhazur, E. J. (1982). Multiple regression in behavioral research. New York: CBS
College Publishing.
Reeve, J. (2002). Self-determination theory applied to educational settings. In E. L. Deci
& R. M. Ryan (Eds.) Handbook of Self-determination Research (pp. 183-203).
Rochester, NY: The University of Rochester Press.
Reeve, J., Nix, G., & Hamm, D. (2003). Testing models of the experience of self-
determination in intrinsic motivation and the conundrum of choice. Journal of
Educational Psychology, 95, (2), 375-392.
229
Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of
intrinsic motivation, social development, and well-being. American Psychologist,
55, (1), 68-78.
Ryan, R. M., & Deci, E. L. (2002). Overview of self-determination theory: An
organismic dialectical perspective. In E. L. Deci & R. M. Ryan (Eds.) Handbook
of Self-determination Research (pp. 3-33). Rochester, NY: The University of
Rochester Press.
Ryan, R. M., & Powelson, C. L. (1991). Autonomy and relatedness as fundamental to
motivation and education. Journal of Experimental Education, 60, (1), 49-66.
Schwarz, N. (1999). Self-reports: How the questions shape the answers. American
Psychologist, 54, (2), 93-105.
Schwartz, B. (2000). Self-determination: The tyranny of freedom. American
Psychologist, 55, (1), 79-88.
Sherlin, Jr., J. H. (2002). Understanding the system persistence of first-generation
students through path modeling. Unpublished doctoral dissertation, University of
Maryland, College Park.
Stewart Hunter, M., & Linder, C. W. (2005). First-year seminars. In M. L Upcraft; J. N.
Gardner, & B. O. Barefoot (Eds.). Challenging & Supporting the First-year
Student: A Handbook for Improving the First Year of College (pp. 275-291). San
Francisco: Jossey-Bass Publishers.
Terenzini, P. T., Theophilides, C., & Lorang, W. G. (1984). Influences on students?
perceptions of their academic skill development during college. Journal of Higher
Education, 55, (5), 621-636.
230
Tinto, V. (1993). Leaving college: Rethinking the causes and cures of student attrition.
(2
nd
ed.). Chicago: The University of Chicago Press.
Upcraft, M. L. (1989). Residence halls and campus activities. In M. L. Upcraft; J. N.
Gardner, & Associates (Eds.). The Freshman Year Experience (pp. 142-155). San
Francisco: Jossey-Bass Publishers.
Upcraft, M. L. (2005). Assessing the first college year. In M. L Upcraft; J. N. Gardner, &
B. O. Barefoot (Eds.). Challenging & Supporting the First-year Student: A
Handbook for Improving the First Year of College (pp. 469-485). San Francisco:
Jossey-Bass Publishers.
Upcraft, M. L., Chrissman Ishler, J. L., & Swing, R. L. (2005). A beginner?s guide for
assessing the first college year. In M. L Upcraft; J. N. Gardner, & B. O. Barefoot
(Eds.). Challenging & Supporting the First-year Student: A Handbook for
Improving the First Year of College (pp. 486-500). San Francisco: Jossey-Bass
Publishers.
Upcraft, M. L., & Gardner, J. N. (1989). A comprehensive approach to enhancing
freshman success. In M. L. Upcraft; J. N.Gardner, & Associates (Eds.). The
Freshman Year Experience (pp. 1-12). San Francisco: Jossey-Bass Publishers.
Upcraft, M. L., Gardner, J. N., & Barefoot, B. O. (2005). The first year of college
revisited. In M. L Upcraft; J. N. Gardner, & B. O. Barefoot (Eds.). Challenging &
Supporting the First-year Student: A Handbook for Improving the First Year of
College (pp. 486-500). San Francisco: Jossey-Bass Publishers.
231
Vallerand, R. J., Fortier, M. S., & Guay, F. (1997). Self-determination and persistence in
a real life setting: Toward a motivational model of high school dropout. Journal
of Personality and Social Psychology, 72, (5), 1161-1176.
Waugh, R. F. (2001). Measuring ideal and real self-concept on the same scale, based on a
multifaceted, hierarchical model of self-concept. Educational and Psychological
Measurement, 61, (1), 85-101.
Weidman, J. (1989). Undergraduate socialization: A conceptual approach. In J. Smart
(Ed.), Higher education: Handbook of theory and research (Vol. 5). New York:
Agathon.
Winston, G.C., & Zimmerman, D. J. (July/August, 2000). Where is aggressive price
competition taking higher education? Change, 32, (4), 10-19.
Yeung, A. S., McInerney, D. M., Russell-Bowie, D., Suliman, R., Chui, H., & Lau, I. C.
(2000). Where is the hierarchy of academic self-concept? Educational
Psychology, 92, (3), 556-567.