ABSTRACT Title of Document: SHOULD THEY STAY OR SHOULD THEY GO? EXAMINING LEGISLATOR BEHAVIOR ON STATE IMMIGRATION POLICY Stephen C. Yoder, Doctor of Philosophy, 2014 Directed By: Professor James G. Gimpel, Department of Government and Politics Undocumented immigration used to be the concern of only states that shared a southern border with Mexico or that contained traditional immigrant gateway cities. No longer. Immigration legislation made the policy agenda in all 50 states in 2007, with 46 states enacting into law a total of 240 immigration-related bills. These bills reflect state legislators' intent on managing the largely Latino undocumented immigrant population, with some legislators working to enact legislation that either restricts or expands this group's access to state benefits. Understanding the personal-, district-, and state-level influences on state legislator behavior in this policy arena is important for understanding the relationship between entrenched power and the representation of disempowered minority groups in the U.S. federal system. The immigration policy arena heightens the salience of both legislators own racial and ethnic characteristics and those of their constituents, making it ideal for assessing legislators' representational roles. This dissertation builds upon and challenges the scholarly literature in the two separate, but linked, fields of state immigration policy and Latino descriptive representation. Prior scholarship on state immigration policy has focused entirely on state immigration policy adoption, leaving scholars none the wiser of legislators' substantive representation of Latino interests in this context. Additionally, several scholarly works have found that Latino legislators offer descriptive representation to Latino interests? representation based on a common ethnic tie beyond that which can be attributed to constituency and party influences?but these findings have been limited by their analysis of Latino descriptive behavior only in states with large Latino populations. This dissertation's analysis of a new, expansive database of state legislator behavior on state immigration policy at the bill sponsorship, committee referral, and floor voting legislative policymaking stages in both chambers of 49 state legislatures challenges conventional scholarly knowledge of the representational role of legislators on Latino issues. It finds that legislators' substantive and descriptive representation of Latino interests is contextual based on the type of legislation and the stage of the policymaking process. Latino legislators' descriptive representation is further confined to whether the legislation threatens the largely Latino undocumented immigrant community. SHOULD THEY STAY OR SHOULD THEY GO? EXAMINING LEGISLATOR BEHAVIOR ON STATE IMMIGRATION POLICY By Stephen C. Yoder 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 2014 Advisory Committee: Professor James G. Gimpel, Chair Professor Frances E. Lee Professor Irwin L. Morris Professor Stella Rouse Professor Christopher H. Foreman ? Copyright by Stephen C. Yoder 2014 ii Dedication To Alyssa, Nicholas, and AnnaLeigh iii Acknowledgments This research project would never have been started let alone finished without the constant and sincere support of faculty, friends, and family. The faculty and staff of the Department of Government and Politics at the University of Maryland, College Park provided my academic training, mentorship, facilities, and funding. My committee?Jim Gimpel, Frances Lee, Irwin Morris, Stella Rouse, and Christopher Foreman?provided feedback that was thoughtful and substantive from this project?s nascent theorizing through to its final editing. Mike Hanmer patiently answered any and all methodological questions in ways that vastly improving this project's measurement and analysis. Jim Gimpel was unwavering in his support for this project, providing expertise and research help in compiling the massive state legislator and state immigration bill datasets. That help came in the form of fellow graduate student Heather Creek and undergraduate Karina Shklyan, without whom the base dataset would never have been completed. Ann Marie Clark and Cissy Roberts were expert departmental staff guides through rocky administrative shoals. My graduate student cohort at the University of Maryland, specifically Heather Creek, Brittany Bramlett, Molly Inman, Jim Curry, and Jill Curry, deserves utmost praise; their sound ideas improved the final project immeasurably, but it is their friendship and support that proves far more valuable. All errors are mine. The last and largest thanks are due my family. My parents and in-laws always expressed interest in my research or, more accurately, in its progress. My wife, Alyssa, was steadfast in her encouragement and support. My son, Nicholas, and daughter, AnnaLeigh, were born during the writing of this dissertation; they remain the best contributions I could ever make to any field. iv Table of Contents Dedication .......................................................................................................................... ii Acknowledgments ............................................................................................................ iii Table of Contents ............................................................................................................. iv List of Tables ..................................................................................................................... vi List of Figures ................................................................................................................. viii Chapter 1: Understanding State Legislative Behavior on Immigration Policy ...........1 Literature Review .......................................................................................................5 State Immigration Policy ..................................................................................5 Latino Descriptive Representation ....................................................................8 Latino Representation and State Immigration Policy ..............................................16 Research Outline ......................................................................................................18 Methodological Note ................................................................................................21 Chapter 2: Assessing Legislative Representation through State Immigration Policy Bill Sponsorship ...............................................................................................................23 Legislative Sponsorship ............................................................................................25 Personal and District Characteristics ..............................................................28 Political and Ideological Considerations.........................................................32 State-level Influences ......................................................................................34 Control Measures ............................................................................................35 Data and Methodology .............................................................................................36 Dependent Variables........................................................................................37 Independent Variables .....................................................................................38 Results.......................................................................................................................40 Expansive Legislation .....................................................................................42 Restrictive Legislation ....................................................................................45 Discussion and Conclusion ......................................................................................48 Chapter 3?Examining Representation of Latino Interests on Immigration in State Legislative Committees ....................................................................................................55 State Legislative Committees ....................................................................................55 Control Measures ............................................................................................63 Methods and Data ....................................................................................................64 Dependent Variables........................................................................................66 Discussion and Conclusions............................................................................79 Chapter 4: Analyzing Representation of Latino Interests through Legislative Roll- Call Voting on State Immigration Legislation ...............................................................83 Legislative Roll-Call Voting .....................................................................................84 Roll-Call Voting and the Representation of Minority Groups ........................86 Methodological Considerations for Studying Roll-Call Votes ........................88 Primary Hypotheses .................................................................................................90 Control Measures ............................................................................................94 Data and Methodology .............................................................................................96 Dependent Variables........................................................................................96 Independent Variables .....................................................................................96 Results.......................................................................................................................97 Chapter 5: Conclusions ................................................................................................. 113 v Appendices ......................................................................................................................122 References .......................................................................................................................125 vi List of Tables Table 2.1. Cross-tabulations of Individual-level Factors on State Legislative Sponsorship and Passage of Immigration Bills, 2007?????????.42 Table 2.2. Negative Binomial Predictions for Influences on Sponsorship of Immigration Legislation, 2007?????????????????...............??44 Table 2.3. Influences on Immigration Bill Passage (Probit Regression Analysis)???????.46 Table 3.1. Difference of Means between Legislators who Served on Committees Where Expansive Immigration Legislation was Referred, and Those Who Did Not, 2007???????????????????????????71 Table 3.2. Difference of Means between Legislators who Served on Committees Where Restrictive Immigration Legislation was Referred and Those Who Did Not, 2007??????????????????????????....72 Table 3.3. Percentages of Group Legislators on Committees to which State Immigration Policy Bills were Referred, 2007????????????????..73 Table 3.4. Percentages of Each Group of Legislators to which State Immigration Bills were Referred, 2007?????????????.???????????.73 Table 3.5. Average Number of Immigration Bills Referred to Legislative Committee by Legislator Committee Member Characteristic, 2007???.?????74 Table 3.6. Average Percentage of Each Type of Immigration Bill Referred to a Committee in Which a Legislator Served, by Group, 2007????..???????.74 Table 3.7. Influence of Committee Composition on Immigration Bill Committee Referral?????????????????...???????..............??76-77 Table 3.8. Influence of Committee Composition on Immigration Bill Committee Referral in States with Latino Legislators?????????..???????..78-79 vii Table 4.1. Comparison between States that Held and did not Hold Recorded Floor Votes on Expansive or Restrictive Immigration Legislation, 2007????...???..100 Table 4.2. Percentage of Votes on Each Type of Immigration Bill Cast by Legislators in Each Group???????????????????...???....103 Table 4.3. Percentage of Each Group Who Voted for at Least One of Each Type of Immigration Bill, 2007????????????????????????103 Table 4.4. Average Number of Yes Votes by Legislators from Each Grouping on Each Type of Immigration Bill, 2007???????????????????..103 Table 4.5. Average Percentage of Yes Votes Cast for Each Type of Immigration Policy by Each Legislator Group?????????????????????.104 Table 4.6. Influences on State Legislators' Voting Yes on Immigration Bill Floor Votes??????????????????...?????????????..106 Table 5.1. Negative Binomial Predictions for Influence of Legislators' State, District, and Personal Characteristics on Introduction of Expansive and Restrictive Immigration Legislation, 2007, by States used in Prior Research??????..??.121 viii List of Figures Figure 1.1 Total Percentages of Latino State Legislators and of Latinos in U.S. Population, by Year???????????????????????.?.4 1 Chapter 1: Understanding State Legislative Behavior on Immigration Policy In June 2007, immigration reform legislation providing a path to citizenship for some undocumented immigrants stalled in the U.S. Senate, despite strong support from President George W. Bush and members of Congress on both sides of the aisle. But, as Cecilia Mu?oz of the National Council of LaRaza, the nation?s largest Latino civil rights group, noted, ?The issue is not going to evaporate because Congress failed to act.? Indeed, undocumented immigration into the U.S. has received increasing attention at the state and local levels in recent years. An estimated 10.6 million immigrants have come to the U.S. since 2000, a number that represents 27.7% of the total foreign-born population currently in the country. Many immigrants, and particularly Latino immigrants, have set up residence in states that lack the infrastructure to facilitate immigrant incorporation and which are unaccustomed to such large numbers of new populations. Significant undocumented immigration used to be the concern of only those states that shared a southern border with Mexico or that contained traditional immigrant gateway cities like New York or Illinois. No longer. While the enactment of high-profile state immigration legislation in Arizona, Georgia, and Alabama has recently refocused national-level media attention on state immigration policymaking, the recent spike in states' consideration of immigration policies started in 2005.1 Legislators in all 50 states introduced a total of 1,562 bills related to immigration in 2007, the first year in which an immigration bill was introduced in every state legislature (see Appendix A), and almost three times the number introduced in 2006. Forty-six states enacted into law a total of 240 1 The National Conference of State Legislatures (NCSL) finds that 2005 is the year in which state immigration policy increased in number and expanded across numerous states (http://www.ncsl.org/issues- research/immig/immigrant-policy-project-state-legislation-117.aspx). 2 immigration-related bills in 2007, more than triple the 84 enacted in 32 states in 2006 (Hegen 2008). U.S. sub-national governments have gone down this legislative road before, most recently in the 1980s and early 1990s during which time they crafted legislation aimed at sending anti-immigrant, often nativist messages to the largely Latino immigrant populations. Bills introduced in this new wave of state immigration legislation cover a wider gamut of public policy arenas, and of messages sent. Some states continue to carry the banner of restrictive immigration sentiment, passing legislation that restricts undocumented immigrants' rights to receive social welfare services, resident tuition, and enables state police to question resident status. Other states strike a more welcoming tone by passing legislation that expands immigrant rights by offering services in languages other than English, higher education at resident rates, and works to prevent abuse of immigrant populations. These state immigration policies have grave real-world effects on the largely Latino2 undocumented immigrant communities they target; they determine how the estimated 11.6 million undocumented immigrants living in the United States interact with state law enforcement, receive social, health, and educational services, and integrate with the state economy, and can influence undocumented immigrants' decision to migrate to or emigrate from a state (see, for example, PBS 2011). This research builds upon and challenges the scholarly literature in two separate, but linked, fields of study, those of state immigration policy and Latino descriptive representation. The few studies that have examined the large increase in state 2 Immigrants from Mexico, El Salvador, Guatemala, Hondurus, Brazil, and Ecuador were estimated by the Department of Homeland Security to comprise ? of the estimated 11.6 million undocumented immigrants in the U.S. in 2008 (Hoefer, Rytina, and Baker 2009). There is evidence that the primary source of immigration to the U.S. is changing from Latin America to Asia, partially as a result of lower immigration from Latin America due to the economic downturn that hit the United States in 2008. Immigrants from Asia are, as a whole, much better educated, wealthier, and more likely to enter the United States legally than are immigrants from Latin America (Pew Research Center 2012). 3 immigration policy activism that began in 2005 focus exclusively on state policy adoption, either of specific immigration-related policies?in-state tuition policy for undocumented immigrants (Reich and Barth 2010), federal/state cooperative immigration law enforcement policy (Creek and Yoder 2012), and state immigration employment policy (Newman et al. 2012)?or of immigration policy of a specific tone: restrictive or expansive of immigrant rights (Monogan 2010; Sanchez, Sanchez, and Ybarra 2012; Chavez and Provine 2009). This study differs from this existing scholarship by building off of and expanding upon Pearson-Merkowitz and Yoder's (2009) first attempt to analyze state immigration policy using individual legislator-level data. Previous studies have found state citizen ideology (Creek and Yoder 2012; Monogan 2010; Newman et al. 2012) or partisanship (Gulasekaram and Ramakrishnan 2013) or the partisan lean of the legislature or of the governor (see, for example, Sanchez, Sanchez, and Ybarra 2012; Chavez and Provine 2009; Reich and Barth 2010) amongst the most powerful indicators of state adoption of state immigration policy that either expands or restricts immigrant rights. This current research tests such findings at the individual legislator level by examining state legislators' behavior on state immigration policy in both legislative chambers of 49 states. It finds inconsistent evidence of individual partisanship and little evidence of citizen ideological influences on state legislator behavior on immigration policy across the legislative policymaking process when controlling for other state-, district-, and personal-level variables. 4 This study also adds to and challenges our growing knowledge of Latino political power. The past several decades have seen marked increases in the percentage of the Latino population overall, its percentage of the voting electorate, and in the number of government positions held by Latinos at all levels of government. At the state level, Latinos have made impressive gains in state legislatures (see Figure 1.1), even if these gains still lag behind the Latino percentage of the population in every state (see Appendix A). This increase in Latino state legislators, however, does not necessarily translate into representation that has a substantive policy impact. What impact does being Latino have on a legislator's championing of Latino interests? Does it matter that state legislators look like America? These questions are important, because without this substantive tie, growth in the number of Latino legislators provides no significant benefit for Latino interests over that of electing non-Latino legislators. Several scholarly works (Bratton 2002; Martinez-Ebers et al. 2000; Rouse 2013) have found that Latino legislators offer descriptive representation to Latino interests?that is, representation based on a common 1970 1980 1990 2000 2010 0 2 4 6 8 10 12 14 16 18 Figure 1.1: Total Percentages of Latino State Legislators and of Latinos in U.S. Population, by Year % Latino Legislators % Latino US Population Year % L a t i n o s 5 ethnic tie beyond that which can be attributed to constituency and party influences. To test whether Latino legislators descriptively represent Latino interests, controlling for constituency and party influences, I examine whether Latino state legislators' support of state immigration legislation that either expands or restricts the rights of the largely Latino immigrant community in the 48 chambers of the 34 states that had Latino legislators in 2007 differed from that of their non-Latino colleagues. This research finds that Latino legislators' descriptive representation of Latino interests is contextual, based on whether the legislation poses a threat to the largely Latino undocumented immigrant community. I address both of these questions by analyzing a newly created database of state legislator behavior on immigration legislation at the bill sponsorship, committee, and chamber floor vote stages of the legislative policymaking process. Literature Review State Immigration Policy Contrary to popular belief, state immigration policy activism is not a recent phenomenon. Throughout the eighteenth and early nineteenth centuries, individual states and localities primarily determined their own immigration policy until the Supreme Court ruled in Henderson v. Mayor of New York (1875) that immigration policy should be uniform and under federal jurisdiction (Tichenor 2002). Though state governments lack the power of the federal government to grant citizenship or to fully enforce immigration provisions through deportation, they can still enact legislation that impacts immigration within the guidelines set by the federal government. These guidelines are still being contested; the Supreme Court's ruling in Arizona v. United States (2012) preempting many measures included in Arizona's SB1070 was just the courts' most recent reconfirmation of the federal government's preeminent role formulating immigration policy. 6 One common explanation used by state legislators to explain the recent increase in state immigration policy activism is the corresponding lack of comprehensive federal immigration policy. Pennsylvania state representative Daryl Metcalfe, founder of the restrictive-immigration policy advocacy group State Legislators for Legal Immigration (SLLI), explains that ?since Washington D.C. remains AWOL on fulfilling its Constitutional responsibilities to secure our nation?s borders against foreign invaders, it is not only incumbent, but the obligation for state lawmakers to step up and do the job that our federal government refuses to do? (SLLI 2007). Though this perspective slights the historical dynamism of the state/federal relationship on immigration policy,3 most scholars agree that the recent increase in state immigration policy activism is related to state and local jurisdictions' needs associated with the changing character of recent immigration that have gone unaddressed at the federal level.4 Over the past two decades, immigration from Latin America has spread beyond the few traditional gateway cities and border states (Frey 2006; McConnell 2008; Passel 3 States enact immigration legislation both in response to federal action and inaction in the immigration policy arena. California's passage of Proposition 187 in 1994 was the state's way of dealing with the failure of the federal government to address the state's concerns in the Immigration Act of 1990, which eased restrictions on and increased the quota of potential immigrants to the U.S. The overturning of Proposition 187 by the federal courts led California's large congressional delegation to push for immigration reform at the national level. This pressure to nationalize what had been state policy impelled the federal government to act once again (Spiro 1997), eventually leading to the passage of the Illegal Immigration Reform and Immigrant Responsibility Act (IIRIRA) of 1996. IIRIRA included many of the provisions that California and other states with high undocumented immigrant populations wanted; namely, it tightened restrictions on the amount of time an immigrant could stay in the U.S., increased provisions for deportation, and strengthened penalties for undocumented immigrants. No federal immigration policy will address all states' needs, so state immigration policy is likely to continue regardless of federal action on immigration policy. The dynamic interaction between the U.S. sub-national governments and the federal government on immigration policy is also subject to changes in the intergovernmental relationship. Recent increases in state immigration policy activism mirror a larger trend of state conflict with the federal government that began under the George W. Bush administration in several policy areas, including welfare, environmental, employment, and law enforcement policy (Krane 2007). 4 Though the failure of immigration reform at the national level has been heavily cited in the scholarly literature, and jurisdictional need has been supported by scholarly findings (for example, Boushey and Luedtkey 2011; Hopkins 2010), Gulasekaram and Ramakrishnan (2013) provide an alternative take on state and localities' increasing passage of restrictive immigration legislation, arguing that this uptick has been partisan, as Republican legislators see political gain in this issue area. 7 and Cohn 2011), which has created unanticipated policy problems and costs for new immigrant destinations. State and local governments legislate immigration policy in part because they bear the fiscal burden of educating immigrant children (Fix and Passel 1994), providing health care and public welfare programs (Hero and Preuhs 2007), and are the sole providers of integration policy for immigrant populations (Fix et al. 2008). States and localities with rapid increases in ethnic heterogeneity due to the influx of immigrant populations have generally crafted restrictive policy solutions in response to these new populations (Boushey and Luedtkey 2011; Hopkins 2010), while states with a firmly established immigrant population have generally crafted more expansive immigration policy (Boushey and Luedtke 2011). Previous state attempts to restrict undocumented immigrants' access to state social services have been challenged on their constitutionality. The Supreme Court's decision in Plyler v. Doe (1982) finding unconstitutional a 1975 Texas law restricting undocumented children from primary and secondary school education has barred states from regulating immigrant access in this specific policy area, though states are free to regulate undocumented immigrants access to public higher education (Reich and Barth 2010). The current wave of restrictive state immigration legislation undoubtedly follows from California's Proposition 187, which limited undocumented immigrants' access to a wide range of state public services and which a federal judge quickly found an unconstitutional infringement on the federal government's plenary power over immigration policy (Calavita 1996). The political science literature richly details state governments' enactment of restrictive English-only legislation and ballot initiatives through studies of public opinion (see, for example, Hood III, Morris, and Shirkey 1997; Hood III and 8 Morris 2000; Tolbert and Hero 1996; Tolbert and Hero 2001; Pantoja and Segura 2003; Valenty and Sylvia 2004) and state-level policy enactment (Schildkraut 2001; Tatalovich 1995). While the few studies that examine the large increase in state immigration policy activism that began in 2005 focus exclusively on state policy adoption, either of specific immigration-related policies?in-state tuition policy for undocumented immigrants (Reich and Barth 2010), federal/state cooperative immigration law enforcement policy (Creek and Yoder 2012), and state immigration employment policy (Newman et al. 2012)?or of immigration policy of a specific tone: restrictive or expansive of immigrant rights (Monogan 2010; Sanchez, Sanchez, and Ybarra 2012; Chavez and Provine 2009). This study builds off of and expands upon Pearson-Merkowitz and Yoder's (2009) first attempt to analyze state immigration policy using individual legislator-level data. Previous studies found state citizen ideology (Creek and Yoder 2012; Monogan 2010; Newman et al. 2012) or partisanship (Gulasekaram and Ramakrishnan 2013) and the partisan lean of the legislature or of the governor (see, for example, Sanchez, Sanchez, and Ybarra 2012; Chavez and Provine 2009; Reich and Barth 2010) amongst the most powerful indicators of state adoption of state immigration policy that either expands or restricts immigrant rights. This research evaluates whether such findings hold in relation to individual state legislators' behavior on state immigration policy. Latino Descriptive Representation Pitkin's (1967) definitions of descriptive and substantive representation have provided a starting point for empirical tests of the kinds of representation minority groups receive in legislative bodies. Descriptive representatives are those who ?look like? the represented community. For Pitkin (1967), descriptive representatives were but a mirror of the citizenry, while substantive representatives were those who act to reflect the represented 9 community's interests. Pitkin (1967), in positing that descriptive representation matters because of the assumption ?that people's characteristics are a guide to the actions they will take" (89), points to the link recent scholars of legislators' representational roles have made between descriptive and substantive representation. Bratton (2006) finds this link in her study of the representational role of Latino state legislators, noting that Latino legislators (descriptive representatives) were more likely than non-Latino legislators to further Latino interests, thus becoming substantive representatives. Rouse (2013) reflects the current focus of empirical scholarship on this link in writing ?a constituency is substantively represented if a legislator or legislative body advances the interests of that constituency even if the demographic characteristics of the constituency are not reflected? by the member (12). This evolving framework is helpful in clarifying the representational behavior of minority group legislators; their behavior furthering group interests beyond that expected by constituency or party concerns becomes descriptive representation, while minority group interests furthered by a legislator not of that minority group becomes substantive representation. Our working test of the link between descriptive and substantive representation thereby becomes whether a legislator's personal characteristics such as race, ethnicity, or gender influences their behavior on legislation that furthers the interests of their ascriptive group beyond that related to their constituency or party representation. Evidence of such a link proves important for the representation of traditionally numerically underrepresented minority groups. Given this link, if present in every legislative body, then a constant factor in the measurement of the representation of minority group interests becomes whether the percentages of minority group members in 10 legislative bodies reflects that of the electorate. In a practical sense, if minority group descriptive representatives advance minority group interests beyond that of a non- minority group member given the same constituency and party concerns, then the election of descriptive representatives must become the primary concern of minority groups. As Rouse (2013, 10) nicely summarizes, a finding of Latino descriptive representation ?points to the fact that minority groups have been historically underrepresented or misrepresented by non-minorities within the legislative process; that non-minorities cannot or will not advocate in the same way for the policy interests of minorities.? Several scholarly works have found evidence of the descriptive representation of Latino interests by Latino legislators (Bratton 2002; Martinez-Ebers et al. 2000; Rouse 2013), but not uniformly across all stages of the legislative process. Indeed, we should expect different stimuli to influence legislator behavior at different stages of the legislative process. Whereas state legislators at other stages of the policymaking process may be motivated by party influences, legislative sponsorship is an individual act linked in past political science scholarship (see, for example, Bratton 2006; Bratton and Haynie 1999; Barnello and Bratton 2007; but see Bratton and Rouse 2011 on state legislative co- sponsorship behavior) to distinct district- and individual-level pressures. We would expect to see legislators sponsor immigration policy responding to the ideological, racial, ethnic, and socioeconomic consistency of their district as well as that which directly relates to their personal representational wishes. As such, previous research (Bratton 2002; Martinez-Ebers et al. 2000; Rouse 2013) finds that minority legislators provide descriptive representation beyond that called upon by constituent representation to 11 members of their own group through their sponsorship activity. Further, in comparing enacted legislation with the broader population of sponsored legislation, these scholars find that minority group legislators are significantly less likely to have the legislation they sponsor enacted into law than are legislators from other ethnic or racial groups. Committees move legislation from bill sponsorship, an individual act, to the floor vote, a collective act of the entire legislative chamber. In the interim, a small subset of the legislative body, committee members, can have a disproportionate influence over policy formulation through the bill markup and committee bill amendment processes. Committee members therefore have ?disproportionate access into the crucial early stages of decision making and a valuable forum for position taking and credit claiming? (Evans 1999, 610). Committee action is also relatively hidden from constituents when compared to bill introductions, committee hearings, or floor votes. This opacity can provide legislators necessary leeway from constituent concerns to engage in legislative actions that impact policy in more substantive and in less symbolic, position taking ways. Differences in the ethnic and racial composition of legislative committees matters because African-American and Latino legislators' behavior in committees differs from that of their white colleagues on issues relating to their ethnic groups (Gamble 2007; Minta 2009; Ellis and Wilson 2013). Though Rouse (2013) has examined how Latino legislators descriptively represent Latino interests in state legislative committees? finding some evidence of descriptive representation in two of the four state legislative chambers she studied?her focus was on members' formal committee participation. Committee members may well influence committee action beyond that witnessed in a markup or a committee hearing. Legislators unable to attend a formal committee markup 12 may have collaborated on proposed changes with committee colleagues informally, or may have had a legislator act in their steed, both examples of participation that would be missed in measuring individual legislator's formal participation. A broader measure than markup participation, simple committee composition, may better capture this informal influence. There has been little analysis of how the composition of committees, and how equal or unequal committee composition vis-a-vis the composition of the full chamber, can influence policy outcomes. There has likewise been little analysis of the composition of the committees to which bills are referred. Examining the composition of the committees to which state immigration legislation is referred may well prove illustrative of whether minority group legislators are able to descriptively represent their own group or substantively represent other ethnic groups at the committee stage of the legislative policymaking process. Descriptive representatives, both by acting in ways that enhance support for minority group interests and by bringing otherwise absent policy issues to the fore, ?may shape a substantially more favorable policy environment for the representation of minority interests? (Ellis and Wilson 2013, 2). Legislative floor action occupies the last legislative step in the legislative policymaking process, and one that determines whether a bill will advance to the executive for signature into law or veto. Legislator behavior at the roll-call voting stage faces different pressures than those that influence legislator behavior at earlier stages of the policymaking process. Whereas state legislative sponsorship and the composition of legislative committees may be motivated predominantly by district- and personal-level influences, legislative roll-call voting has been linked in past political science scholarship 13 with considerable party pressures (see, for example, Cox and McCubbins 2005; Aldrich 1995; Rohde 1991; but see Krehbiel 1993). Roll-call voting therefore is a policymaking stage that offers a distinct arena in which to analyze legislators' reactions to unique pressures. The legislative floor is the only stage in the policymaking process where all members of the legislative chamber have equal access and ability to formally act on a bill. Whereas analyzing state legislator sponsorship and the composition of committees that oversee state immigration policy provides important clues about the policy preferences of the state legislators with the most interest in the policy, analyzing legislator behavior through roll-call votes is important because doing so provides information on how state legislators as a whole behave in response to policy. The roll-call vote also remains the most visible of the policymaking stages to constituents, who do not as closely follow the sponsorship and committee behavior of their representatives. The transparency of the roll-call vote makes it an ideal policymaking stage for the testing of constituency influence on legislative behavior. Research has found that minority group legislators provide descriptive representation to the minority groups they represent by voting differently at the roll-call stage than do their respective white and male colleagues (Swers 1998; 2002; Norton 1999; Welch 1985; Hogan 2008; Vega and Firestone 1995; Tate 2004; Swain 2006; Haynie 2001; Grose 2005; Barrett 1995; Rocca, Sanchez, and Uscinski 2008; Kerr and Miller 1997; Preuhs and Juenke 2011; Casellas 2010; Rouse 2013). Indeed, recent research has found that party- and district-pressures prove influential in state legislators' roll-call voting behavior on all (Casellas 2011) and on only Latino-interest legislation (Rouse 2013). 14 Recent studies have examined the influence of legislators' ethnicity (Rocca, Sanchez, and Uscinski 2008; Kerr and Miller 1997; Preuhs and Juenke 2011; Casellas 2010; Rouse 2013) on their roll-call voting behavior. Though both Rouse (2013) and Casellas (2010) argue that such studies are important to understanding the descriptive and substantive representation of Latinos by their elected legislators, only Rouse (2013) has so far focused exclusively on the roll-call voting behavior of legislators on legislation that disproportionately affects Latinos. Her study of Latino descriptive representation compares Latino legislators' roll-call voting behavior on both general legislation and on that which she classifies as Latino interest legislation, finding little in the way of descriptive representation by Latino legislators for Latino interests at the roll-call voting stage of the policymaking process. State legislatures represent an ideal setting in which to study minority group descriptive representation. The 50 state legislatures differ widely in the ascriptive characteristics of their legislators, in the demographics of the districts they represent, and in the types of bills introduced and advanced through the legislative policymaking process. Political theorists at least as early as James Madison argued that state governments are closer to and legislate and implement policies that have a larger impact on the daily lives of their citizens. Studying immigration policy in state legislatures enables investigation of how variation in Latinos' legislative presence impacts Latino interests at each stage of the legislative policymaking process. If Latinos offer descriptive representation to Latino interests, then state immigration policymaking in states where Latinos comprise a higher percentage of their state legislature, like in California, New Mexico, and Texas, should differ from that in states that have much lower percentages of 15 Latino state legislators, like Pennsylvania, Oklahoma, and Wisconsin. Two works5 have focused specifically on Latino descriptive representation in state legislatures. Bratton (2006) finds that descriptive representation among Latino legislators is particularly activated by immigration legislation, and that Latino legislators are more likely to introduce and pursue legislation beneficial to immigrant populations. Rouse's (2013) is the only study to examine the representative behavior of Latino state legislators across multiple stages of the legislative policymaking process. Her analysis finds that Latino legislators' provision of descriptive representation on Latino interest issues is conditional upon the contexts of the legislative policy stage and upon the state legislature. Her study is also unique in that it works to separate out the behavior of Latino legislators on specifically Latino interest legislation. These studies of Latino descriptive representation in state legislatures have so far largely mirrored the map of traditional immigrant receiving states. They have focused analysis on the heavily Latino states of California, Florida, Illinois, and Texas (Bratton 2006) and Arizona, California, Colorado, Illinois, New Mexico, and Texas (Rouse 2013). Increasingly, scholarship will be called upon to match the real-world spread of recent significant immigration streams beyond these gateway states to determine whether Latino interests are being represented in the states unaccustomed to providing services to new immigrant communities. Latino state legislators in these new immigrant destination states likely labor within a different context than those in states with a large Latino base of support in the public and with an ample and supportive Latino legislative caucus. Latino legislators in states with comparatively fewer Latinos and Latino legislators may 5 Casellas (2011) focuses on Latino legislators at both the congressional and the state levels, but looks across all legislation rather than at that pertaining only to Latino interests. 16 be less willing to sponsor and work to further Latino interest legislation that has little support amongst the public or other legislators. Likewise, Latino legislators in these states may be more willing to attend to budgetary or political concerns in introducing or working to enact legislation that restricts undocumented immigrants? access to state services. My research addresses Latino legislators? descriptive representation of Latino interests by utilizing a unique database of state legislative behavior on state immigration policy through each stage of the legislative policymaking process. In separating out the influence of legislators' Latino ethnicity on their legislative behavior, I account for political, district, and other personal-level influences on their behavior. By pooling the actions of Latino state legislators on state immigration policy across the 50 chambers of the 366 state legislatures in which Latino legislators serve and focusing at each stage of the legislative policymaking process, I am able to draw conclusions from my findings on Latino legislators' descriptive representation that are, in some ways, more generalizable to the behavior of Latino legislators than are those of previous studies.7 Latino Representation and State Immigration Policy Immigration policy is an ideal arena in which to study Latino descriptive representation as past research has found that immigration policy is an issue area to which Latinos pay disproportionate attention. Latinos consistently rank immigration higher as an important issue (see, for example, Carroll 2007a) and are more likely to favor either expanding 6 The total number of states and chambers with Latinos serving is 37 and 51, respectively. Nebraska had one serving Latino legislator in its unicameral state legislature in 2007. However, it is excluded due to its nonpartisan legislature. 7 Prior studies that have focused on Latino descriptive representation in only a few legislatures (Bratton 2006; Casellas 2011; Rouse (2013) have examined a much broader range of policy than does this study. Bratton (2006) and Rouse (2013) both examine policies that they believe appeal particularly to Latinos, while Casellas looks at all policy areas. These studies also examine Latino legislators' behavior across multiple years, while this dissertation only looks at Latino legislators actions on state immigration policy in 2007. 17 immigration or keeping it at the same levels (see, for example, Carroll 2007b; de le Garza and DeSipio 1993; Cain and Kiewiet 1987) than are members of other racial or ethnic groups. While some work (Rouse 2013) argues that immigration policy is not as high a priority amongst the Latino community as are policy issues like education or health care, the large-scale mobilization of Latinos in mass public protests intermittently since 2006 in response to the passage of restrictive state immigration policies like Arizona's SB1070 and of federal movement on immigration policy shows that policy to restrict immigration and immigrant rights motivates Latinos to political action. Immigration policy, therefore, is the type of policy about which ?attentive publics,? those most affected by the law, would be concerned (Arnold 1992), which translates into a policy area in which we can expect Latino legislators to offer clear descriptive representation to Latinos. As this dissertation explicitly argues, the nature of the immigration policy under legislative consideration matters in terms of the legislative response of Latino legislators; restrictive immigration policy threatens the Latino community, leading Latino legislators to explicitly represent the broader Latino community on those policies. The immigration policy arena also naturally raises the salience of specific ethnic and racial conflicts, so that studying immigration policy also yields important insights on legislator behavior in response to legislators' racial and ethnic personal characteristics and those of their constituencies. Indeed, several political science works have studied racial and ethnic conflict through the lens of state immigration policy (Tatalovich 1995; Hood III, Morris, and Shirkey 1997; Hood III and Morris 2000; Tolbert and Hero 1996; Tolbert and Hero 2001; Valenty and Sylvia 2004). Using immigration policy to study racial and ethnic relations allows testing of racial threat and group coalition theories to frame the 18 reactions of African-Americans to Latino interests. At the heart of these analyses is whether members of other social groups provide substantive representation to Latinos. Two other important considerations play into the choice of state immigration policy as a vehicle to study Latino descriptive representation. First, these policies are wide-spread enough?state immigration policies were introduced in 98 of the 99 state legislative chambers, and reached a floor vote in 45 state legislatures in 2007, and in the 48 chambers in the 34 states in which Latino legislators served?to allow for a pooled cross-sectional sample that is representative of the full population of Latino state legislators. This allows for the drawing of conclusions on Latino legislators' representational behavior on state immigration policy that are more generalizable across all Latino legislators than are previous studies that only examined Latino legislators' behavior in only a few states (but see note 7). Second, the nature of state immigration policy and the hypothesized Latino reaction to it is uniquely dichotomous; Latino legislators are expected to support legislation that expands immigrant rights and to withhold their support of legislation that restricts immigrant rights. The interaction between the dichotomy of state immigration legislation and of Latino legislators' hypothesized behavior in reaction to it allows for multiple tests of Latino legislators' descriptive representational behavior. It also provides a richer view of Latino-interest legislation than that previously studied (Rouse 2013), recognizing that such legislation can be favorable or inimical to Latino interests, and that Latino legislator behavior may differ accordingly. Research Outline I analyze the full universe of state legislators to draw generalizable conclusions about the behavior of state legislators on state immigration policy and about the quality of Latino 19 legislators' representation. To do so, I first created a dataset of all state legislators serving in 2007 in both chambers of 49 state legislatures.8 This dataset contains characteristics for each legislator at the personal- (gender, race, ethnicity, party, service in party leadership, chamber of service), district- (district percentage Latino, African-American, below the poverty line, population change from 2000-2006, and the legislator's percentage Republican vote in their last election), and state-levels (unemployment rate, citizen ideology, measures of legislative professionalism, and whether located on the U.S. foreign border, percentage Latino legislators in each chamber of the state legislature). I next paired these data with information on the state immigration legislation introduced in each state in 2007, compiled by the Migration Policy Institute and New York University?s School of Law (2007) and published as "State Responses to Immigration: A Database of All State Legislation." I utilized their dichotomous coding of state immigration legislation that either expands or that restricts immigrant rights. Doing so required researching and coding whether each of the 7,300+ state legislators introduced, served on a committee that oversaw, or cast a vote on the legislative floor on one of the 576 state immigration measures that fit into this dichotomous classification (313 that expand and 263 that restrict immigrant rights). I analyzed this newly created dataset using Ordinary Least Squares, probit, and negative binomial regression models to assess the influences on state legislators' behavior on state immigration policy in general and, more specifically, whether Latino legislators provide more support for measures that expand immigrant rights, and less support for measures that restrict immigrant rights, than do non-Latino legislators when controlling for the predominant personal-, district-, and state-level factors shown to influence legislative behavior. 8 See note 6. 20 Chapter 2 focuses on state legislators' agenda-setting behavior as measured through their sponsorship of state immigration legislation that expands or restricts the rights of immigrants. Its findings indicate that state-level factors such as state unemployment and whether a state resides on the U.S. border, and district-level influences like the percentages of Latinos and of population change in a legislator's district explain the sponsorship actions of state legislators on state immigration policy. Party affiliation, when controlling for the effects of all other variables, has limited substantive effect on legislators' sponsorship of state immigration policy. No evidence of descriptive representation by Latino legislators of Latino interests is shown at this agenda-setting stage of the policymaking process. Chapter 3 assesses the composition of the legislative committees to which state immigration policy was referred in 2007. Studies of representation in state legislative committees have focused on the influence minority group legislators have through participation in committee markups or on the representativeness of the committees as a whole as a measure of the appointment power of political leadership. By analyzing the composition of legislative committees to which state immigration policy is referred, this chapter assesses for the first time how minority group legislator representation in committees might influence substantive policy outcomes. The findings indicate that Latino legislators are no more or less likely to serve on committees referred immigration policy than are non-Latino legislators. Chapter 4 analyzes state legislator roll-call voting in the 33 states that held roll-call votes on expansive or restrictive immigration legislation in 2007. It finds that Latino legislators descriptively represent Latino interests through their roll-call votes on restrictive immigration legislation, when Latino interests are most threatened. Chapter 5 summarizes 21 the results of the empirical analyses, placing the behavior of state legislators and of Latino state legislators within the framework of the extant literatures on state immigration policy adoption and on Latino descriptive representation. In doing so, it lays out a theory of issue salience for Latino legislators' descriptive representation. Carefully tracing the path of these uniquely dichotomous state immigration bills through the legislative process in each state provides a nuanced assessment of where and how state legislators? actions matter to policy outcomes. Electing members of one party as proxies to advance Latinos interests in the state immigration policy realm may not be an effective strategy, nor is electing Latino legislators an effective way to expand the rights of the largely Latino undocumented immigrant community. When looking across all state legislators, the Latino descriptive representation found at the bill introduction stage in prior research becomes increasing conditional on the potential effect of the legislation on the Latino community. Electing Latino representatives to state legislative positions seems to benefit Latinos most when the rights of the largely Latino immigrant community is under greatest threat in the form of restrictive immigration legislation. These findings drive further investigation into the impact that legislators' social identity has on their representational roles. Methodological Note Most dissertations on state and local politics include a case study chapter. This one does not. Case studies are appropriate to give an in-depth example of the typical way a political process works given a small number of cases, which is the approach taken by most such research in this subfield. This research is unique among dissertations in the state and local politics subfield in the breadth of states included in its quantitative analysis; conducting a case study is less illustrative in this case as there is no one state 22 that exemplifies the way that state legislators more broadly and Latino legislators more specifically make policy decisions. Future work beyond the scope of this research might do well to conduct multiple case studies across a range of states judged typical of state immigration policymaking and of Latino legislator behavior. 23 Chapter 2: Assessing Legislative Representation through State Immigration Policy Bill Sponsorship Despite the recent high-profile state immigration enactments, and most notably Arizona's enactment of SB1070, only a few studies have analyzed this most recent wave of state immigration policymaking. These few extant comparative state studies have almost exclusively examined state immigration policymaking through a policy adoption framework (Chavez and Provine 2009; Monogan 2009; Newman et al. 2010; Newton 2012; Reich and Barth 2010; Sanchez, Sanchez, and Ybarra 2012). These help develop our understanding of the preconditions for state immigration policy enactment, but they do little to advance our knowledge of the motivations of individual legislators to support immigration policy or of the policymaking process in the state immigration policy realm. Studying individual legislator behavior in the immigration policymaking realm provides several opportunities not available when studying state immigration policy adoption. This chapter's focus on individual legislator sponsorship behavior extends the range of immigration policy studied from the hundreds of immigration bills that states enact, to the more than a thousand that legislators have introduced onto the legislative policy agenda. This shift in emphasis provides a more robust picture of the internal legislative policymaking process on immigration legislation in the U.S. Focusing on the actions of individual legislators also allows analysis of the effects of legislators' personal characteristics?legislators' ethnicity, race, and gender?on their descriptive and substantive representation of the predominantly Latino immigrant population, of legislators' individual representation of their district, and of legislators collective representation of state interests. Finally, the immigration policy arena naturally heightens 24 the salience of specific ethnic and racial characteristics,9 so that the study of state legislator sponsorship behavior on immigration policy also yields important insight on legislator behavior in response to racial and ethnic constituencies. Such analyses are not possible when analyzing state policy adoption. Sponsoring legislation is the necessary first step toward enacting legislation, and agenda setting affects other stages of the policymaking process by defining the range of alternative policy outcomes. Legislator sponsorship behavior faces different pressures than those that influence legislator behavior at later stages of the policymaking process. Whereas state legislators at other stages of the policy process may be motivated by party influences, legislative sponsorship is an individual act linked in past political science scholarship (see, for example, Bratton 2006; Bratton and Haynie 1999; Barnello and Bratton 2007; but see Bratton and Rouse 2011 on state legislative cosponsorship behavior) to distinct district- and individual-level pressures. The uneven dispersal of immigration nationwide (Frey 2006; McConnell 2008; Passel and Cohn 2011) has made immigration policy more salient in specific states and in specific districts. Legislators respond to the interaction between their state and districts' economic and demographic abilities to handle the strength of the immigration flow with targeted legislative policy solutions. While prior research has found that enacted state public policy does ultimately represent a state's collective ideology (Erikson, Wright, and McIver 1995), it is individual legislators, held accountable by district electoral politics, who place immigration policy on the legislative agenda. While a robust research tradition theorizes that legislators respond primarily to re-elective motives (Mayhew 1974), others 9 Several political science works study racial and ethnic conflict through the lens of state immigration policy (see, for example, Tatalovich 1995; Hood III, Morris, and Shirkey 1997; Hood III and Morris 2000; Tolbert and Hero 1996; Tolbert and Hero 2001; Valenty and Sylvia 2004). 25 (Canes-Wrone, Brady, and Cogan 2002; Carson, Lebo, Koger, and Young 2010) note that members often take actions in loyalty to their party that contradict this seemingly unified focus on reelection, and that these actions can have negative electoral outcomes. While studies of legislative sponsorship acknowledge both of these motivations, several have found that state legislators also sponsor legislation to descriptively and substantively represent the group interests of specific gender, racial, and ethnic groups (Bratton 2006; Bratton and Haynie 1999; Barnello and Bratton 2007), at times independent of constituent or party pressure. The extent to which legislators also represent the collective needs of their state, rather than their individual re-elective needs and that of a specific race, ethnicity, or gender remains an open question. The immigration policy response to these pressures has been two types of immigration legislation: that which expands or that which restricts government protection and services to undocumented immigrants. Extrapolating from the findings of the state legislative sponsorship and the state immigration policy adoption literature, we can expect that the ethnic and ideological composition of a legislator's constituency as well as personal-level factors like a legislator's gender, ethnicity, and race will impact state legislator sponsorship of state immigration policy. This study is novel from previous studies in both its substantive focus and in the data it employs. It is the first work of state legislative sponsorship to examine the state immigration policy context, to analyze the full population of all state legislators, and to examine state legislators' collective representation of state needs. Legislative Sponsorship Though most of the scholarly work on public policymaking and legislator behavior focuses on the end of the policymaking process?floor voting or bill enactment? scholars have good reason to analyze bill sponsorship as a measure of legislator intent 26 and agenda setting. Bill sponsorships are plentiful enough in any given policy area and legislative year to allow for the use of more robust methods than that which can be confidently employed for statistical models of a much more limited number of states or bill enactments. Bill sponsorships are also less beholden10 than are bill enactments to the many legislature-specific variations in institutional rules and norms inherent in the U.S. federalist system. Finally, much previous work has focused on the critical mass a minority group needs within a legislature to enact policies that substantively address minority group concerns (Kanter 1977; Dahlerup 1988). As Rouse (2013) rightly notes, studying agenda setting through bill sponsorship pulls the discussion from what minority groups can accomplish as a critical legislative mass, to what individual legislators with certain characteristics can accomplish as critical actors. Bill sponsorship, argue Burstein, Bauldry, and Froese (2005), more accurately reflects the issues on the legislative policy agenda than does a tally of enacted legislation. Agenda setting affects other stages of the policymaking process by defining the range of alternative policy outcomes. Sponsoring legislation is the necessary first step toward enacting legislation, but legislators may behave differently in response to different pressures when they sponsor legislation than when they vote on the floor or in committee. Talbert and Potoski (2002) find that sponsored legislation reflects multiple policy issue dimensions, but that floor voting reflects only one: ideology/party. Sponsoring a bill is an action that is individual, purposive, and direct in that it sends a clear message to a legislator?s constituents (Schiller 1995). Such clarity is in 10 This is not to say that bill sponsorship is immune to the many state-level variations in institutional rules and behavior. While most states allow but single sponsors of legislation, many allow multiple sponsorship, others limit sponsorship, and others allow committee sponsorship. To the author?s knowledge, no current, comparative list of sponsorship rules and behavior currently exists for all 99 chambers in the 50 state legislatures. 27 contrast to legislative action in committee or on the floor, where party cues often hold sway. Though an individual act, bill sponsorship is not without cost. A sponsoring legislator and their staff pay the opportunity cost of drafting sponsored legislation, as well as any potential political costs associated with position taking (Schiller 1995), which are high enough that a legislator would not undertake sponsorship without a substantial interest in the issue. In their drive to satisfy their primary re-elective motive (Mayhew 1974), legislators sponsor bills to claim credit by advertising to constituents and the media that they took a position on a bill of importance to their district (Fenno 1995). These behaviors match well with theories of legislator behavior as being focused primarily on winning reelection, and legislator sponsorship behavior should then correlate with the specific ideological, racial, ethnic, and socioeconomic characteristics of their district. Schiller (1995) finds that individual- and constituency-level characteristics influence sponsorship behavior. Bratton and Haynie (1999) divide these pressures further into constituency representation and descriptive representation models. The legislator in the constituency representation model reacts solely to district and party needs in sponsoring legislation; the legislator in the descriptive representation model sponsors legislation that responds to the needs of the individual legislator's gender, ethnic, or racial group. Bill sponsorship has received the most scholarly attention in this latter context, particularly in examining whether the representation of women (see, for example, Thomas 1991; Saint-Germain 1989; Vega and Firestone 1995; Swers 2002) and of ethnic or racial groups (Bratton and Haynie 1999; Bratton 2006) in legislatures has been substantive or merely symbolic. The bill sponsorship stage is where the tie between 28 descriptive and substantive representation is the most distinct (Bratton and Haynie 1999; Swers 2002; Bratton 2005). This research has found distinct gender differences in the number and topic of legislation sponsored by state legislators, with female legislators more likely than male legislators to sponsor legislation related to women, children, and the family (Thomas 1991; Bratton and Haynie 1999; and Swers 2002; Reingold 2000), whereas male legislators will sponsor legislation on these same issues given the right interaction of personal-level characteristics like their race, education, age, and family circumstances (Barnello and Bratton 2007). Legislators' race and ethnicity also influence which type of legislation they sponsor, with Latino legislators more likely to introduce legislation beneficial to immigrant populations (Bratton 2002), and with African-American legislators more likely to sponsor bills that disproportionately impact African Americans (Tate 2004). Only a few studies have analyzed the influence of institutional-, district-, and personal-level factors together on state legislators' sponsorship behavior (Hogan, Kromer, and Wrzenski 2006; Kromer 2008). These studies' findings support a rational choice perspective of state legislator sponsorship behavior in finding that legislators sponsor legislation in response to constituency concerns. Based on this understanding of the sponsorship and state immigration policy scholarship, I expect state legislative sponsorship of immigration bills to be a function of legislators' district constituency pressures and personal characteristics. Several political science theories offer testable hypotheses for understanding the relationship between individual- and district-level characteristics and state immigration bill sponsorship. Personal and District Characteristics The heart of the re-election motive (Mayhew 1974), the demographic and political 29 composition of a legislator's district should be directly tied to their legislative action. One theory that ties district demographics to legislator behavior is group conflict theory, which holds that a rapid increase in an ethnically or racially distinct, politically disempowered population provokes majority-group hostility (Williams 1947; Blalock 1967). As the size or percentage of the minority group increases, so too do negative majority-group opinion and the number of restrictive laws targeted at the minority group. Attaching the struggle for economic or political power by an entrenched majority group and an infringing minority group yields realistic group conflict theory (Bobo 1999), which can be readily applied to the immigration policy context. Natives polled in the 2000 General Social Survey entwine their negative perceptions of immigration with the economy, with a majority of respondents reporting immigration as responsible for increased unemployment rates (Rumbaut and Ewing 2007). Common to the federal-level debate over immigration reform is the mantra that immigrants, particularly undocumented immigrants, undermine wages in certain sectors and take jobs that otherwise would go to native-born Americans (Newton 2008), and public opinion toward immigration has been found to reflect concerns about the national economy and about taxes (Citrin et al. 1997; Burns and Gimpel 2000; Lee and Ottati 2002; Lee Ottati and Hussain 2001). Economic status also interacts with public opinion on immigration, with those who are better off having fewer concerns about immigration and those who are economically vulnerable having greater concerns over increased immigration (Burns and Gimpel 2000; Citrin et al. 1997). Immigrants may instigate a threat response in tough economic times in particular. These opinions may translate into immigration policy; if the local population is having a hard time securing work, then newcomers who also need jobs 30 are unlikely to be met with open arms. Additionally, race can factor into economic threat, as jobs filled by immigrants are not evenly distributed throughout the American economy, but cluster in low-skilled positions, a disproportionate percentage of which are held by low-income African Americans (U.S. Census Bureau 2007). Race and ethnicity are also a factor in that the composition of the recent wave of immigration is largely Latino, a stark difference from previous, primarily white immigrant waves. This ethnic identifier may stoke hostility and prejudice among natives that increase with the size of the immigrant group (Alba, Rumbaut and Marotz 2005; Chandler and Tsai 2001; Rocha and Espino 2009). Rapid shifts in a district's demographic composition might also signal a disproportionate increase in immigrant influx, as well as demographic instability. Hopkins (2010) finds that rapid increases in a locale's ethnic diversity increases natives' hostile political attitudes about immigrants, and the chances of that locale passing restrictive immigration legislation. Debate does exist, however, over whether African Americans see arriving Latino immigrants as a racial or economic threat or as a potential minority political coalition partner (see, for example, Kaufmann 2003) to counter majority white political power. Some research on Latino representation (see Bratton 2002; 2006) indicates that at times Latinos find substantive representation in African-American legislators. In particular, since many anti-immigrant initiatives parallel anti-civil rights legislation of the past, minority legislators have been known to bind together against potentially discriminatory attacks from white legislators. African-American and Latino groups also share similar economic positions in American society, with the poverty rates of both groups almost three times that of non-Hispanic whites, and with levels of the insured and average level 31 of education much more similar between African-Americans and Latinos than between either of these groups and whites (U.S. Census Bureau 2007). Such shared group characteristics may lead to shared type and direction of policymaking in African- American legislators' substantive representation of Latino interests, particularly when applied to policy that attempts to expand immigrants' access in these policy areas. Analysis of the type of immigration legislation sponsored by African-American legislators and by those representing districts with large percentages of African Americans may also point us toward which opposing theory?group conflict or group coalition?best encompasses African-American legislator behavior on state immigration legislation that largely impacts Latinos. Numerous studies have found that a legislator's gender, race, and ethnicity impact their legislative behavior, with many finding that these personal characteristics go beyond simple representation of constituency to provide descriptive representation of a group (Haynie 2001; Swers 2002). Bratton (2006) finds that descriptive representation among Latino legislators is particularly activated by immigration legislation, and that Latino legislators are more likely to introduce and pursue legislation beneficial to immigrant populations (Bratton 2002; Martinez-Ebers et al. 2000). Legislators are also known to rely on the expertise of their colleagues in taking voting cues (Kingdon 1989); Latino legislators may therefore be accorded a degree of deference as experts on immigration issues, leading to higher rates of bill passage of immigration bills sponsored by Latino legislators. There may also be a constituency tie in. Latinos consistently rank immigration higher as an important issue than do members of other racial and ethnic groups (see, for example, Carroll 2007a), and Latinos are more likely to favor either expanding 32 immigration or keeping it at the same levels than are other racial or ethnic groups (see, for example, Carroll 2007b; de le Garza and DeSipio, 1994; Cain and Kiewiet 1987). Immigration policy, therefore, is the type of policy about which ?attentive publics,? those most affected by the law, would be concerned (Arnold 1992). Finally, individual legislator's gender influences the number and topic of legislation sponsored (Thomas 1991; Bratton and Haynie 1999; Swers 2002). Policy issues related to women, children, and the family are most likely to be associated with the Democratic Party or a liberal ideology (Reingold 2000), female legislators tend to be more liberal than male legislators on social welfare issues (Barrett 1995; Poggione 2004), and on issues related to African-American interests (Bratton and Haynie 1999). This greater female policy liberalism may also extend to social welfare issues relating to immigration and to the Latino minority group. However, considerable research contradicts these expectations specifically on women's attitudes on immigration policy, finding that women are generally more favorable to restrictive immigration policy than are men (Citrin et al. 1997; Espenshade and Calhoun 1993; Citrin, Reingold, and Green 1990; Burns and Gimpel 2000). This discrepancy in the research is one important reason to control for the effects of gender on legislators' actions on immigration policy. Political and Ideological Considerations State legislator sponsorship behavior on immigration policy also ties into political and ideological divisions. Gimpel and Edwards (1998) argue that the 1990s immigration debate in Congress centered on ideological divisions over redistribution, and that as partisanship has become stronger and more reflective of ideology, attitudes toward immigration likewise have become more partisan. This polarization has at its roots 33 Latinos? propensity to politically incorporate in support of the Democratic Party, which, in turn, has positioned itself as the supporter of less-restrictive immigration policy and has advocated more on issues Latinos find salient than have Republicans (Lewis-Beck, Jacoby, Norpoth, and Weisberg 2008). The Republican Party has been tied to restrictive positions on immigration policy since California Republican Governor Pete Wilson?s decision to publicly support Proposition 187 (Reimers 1999; Wroe 2008). Though some members of the Republican Party, notably George W. Bush in his time as governor of Texas and as president, have tried to change the party?s trajectory on immigration, many Republican lawmakers actively espouse restrictive immigration policy rhetoric. The growing Latino demographic provides conflicting impetus for both parties to act on immigration policy. Democrats rationally act to welcome this wave of overwhelmingly Latino immigration in hopes of securing a partisan advantage, while Republicans act to pass increasingly restrictive policy in hopes of stopping or reversing this immigration flow and the growth of their opponents? partisan advantage. Partisan electoral competition may also play a role in the type of immigration legislation sponsored. Members representing competitive districts may be unwilling to take strong stands on contentious issues like immigration that might alienate swing voters, while legislators from electorally safe districts may have more freedom in their position taking.11 The ideology of a state's citizenship influences the ideology of its state legislators and thus of its policy outcomes (Erikson, Wright, and McIver 1994), with anti-immigrant attitudes associated with a politically conservative ideology (Chandler and Tsai 2001). 11 Due to the potential endogeneity of having a variable at the district-level of data representing percent of the vote that went to the Republican candidate and having a variable at the personal-level of data representing a the party of the legislator (correlation of these variables is 0.75), I have run separate models that include only one of the variables. Neither the coefficients nor the levels of significance change in any statistically significant way across all of the variables in the model. 34 Citizen ideology has also been explicitly linked to enactment of state immigration policies, with more conservative states passing more restrictive state immigration policy (Chavez and Provine 2009; Citrin et al. 1990; Hero and Preuhs 2007) and more liberal states more expansive immigration policy (Chavez and Provine 2009; Monogan 2010). State-level Influences In addition to a state's unemployment rate and citizen ideology, a few other influences at the state level have been theorized to impact state legislative sponsorship. Geography should play a role in the type of immigration policy a state takes up. States on the U.S. foreign border face greater fiscal strain as they work to handle larger initial immigrant flows than states in the U.S. interior. Border states differ in their constituent makeup and in how they choose to deal with increased immigration. States like California, which has a more citizenry that is ideologically liberal, has tended to introduce more measures aimed at expanding immigrant rights in this most recent wave of state immigration policymaking than have, for example, states like Arizona, which are more ideologically conservative and which has introduced a greater number of restrictive immigration bills (see Appendix 1). Increased immigration borne of a state's positioning on the U.S. border can be handled very differently across states. We can expect that these states will take up more of both types of immigration policy than states on the interior. Variables specific to the legislature itself have been shown to influence state legislative sponsorship behavior. Legislators in more-professionalized state legislatures, broadly defined as state legislatures that have unlimited sessions, excellent staff resources, and provide their members pay enough to make public service their full-time gig (Squire 2007), behave differently than do legislators in statehouses without these resources. Monagan (2010) finds that more-professionalized legislatures enact more 35 legislation that expands the rights of immigrants than do less-professionalized legislatures, and surmises that the increased level of careerism in more professionalized legislatures (Squire 1988) results in politicians who have a longer view of the increasing Latino demographic?s potential effects on the legislator's political career. Legislators serving in more-professionalized legislatures also tend to produce more legislation, likely a result of the greater resources provided, than do their peers in less-professionalized legislatures (Kousser 2004). Related, legislative rules of sponsorship are not uniform across all states, with 21 states having limits on the number of bills a legislator can introduce (NCSL 1996). In addition to formal limits, legislators may abide by informal norms unique to each legislature that govern sponsorship activity. A careful read of the prior literature on state immigration policy and on legislator descriptive representation at the bill sponsorship stage yields the following primary hypotheses. Hypothesis 1: African-American state legislators will introduce more legislation that expands immigrant rights, and less legislation that restricts immigrant rights, than will non-African-American legislators. Hypothesis 2: Latino legislators will introduce more legislation that becomes law that expands immigrant rights and introduced fewer that contracts immigrant rights than will non-Latino legislators. Hypothesis 3: Legislators representing districts with rapid population increases will introduce more restrictive immigration bills and fewer expansive immigration bills than will legislators representing districts with less growth or a net exodus of residents. Hypothesis 4: Republican legislators will introduce more restrictive and fewer expansive immigration policies than will Democratic legislators. Control Measures A number of control variables are also included to account for legislator characteristics extemporaneous from this study?s focus on state legislators? descriptive and substantive 36 representation of Latino interests through immigration policy at the bill sponsorship stage of the policymaking process. Briefly, gender may also play a role as female legislators tend to be more liberal than male legislators on social welfare issues (Barrett 1995; Poggione 2004) and on issues related to African-American interests (Bratton and Haynie 1999). Female legislators may therefore offer substantive representation to Latino interests on immigration policy by sponsoring and working to enact more expansive and fewer restrictive state immigration bills than their male counterparts. However, extant research has found that women tend to favor more restrictive immigration policy than do men (Citrin et al. 1997; Espenshade and Calhoun 1993; Citrin, Reingold, and Green 1990; Burns and Gimpel 2000), making gender an important control to include in the model. Immigration is also an economic issue, with legislators from states with higher unemployment rates and higher levels of district poverty expected to sponsor and work to enact more restrictive immigration policy to prevent influx of additional surplus labor than will legislators from states with lower levels of unemployment or representing districts with lower levels of poverty. Legislators residing in states on the U.S. border are expected to sponsor and work to enact more measures of both types of immigration legislation than are legislators in states in the interior. Legislators who work in more professionalized legislatures and those with a history of enacting more bills are expected to introduce and work to enact more immigration legislation of both types than are legislators who work in less professionalized legislatures or those that have a history of enacting fewer bills. Whether a legislator is in their party's leadership is a control variable without a directional hypothesis. Data and Methodology This study analyzes legislator sponsorship activity on both restrictive and expansive 37 immigration legislation across 49 states. Immigration legislation was introduced in all 50 states, and in 98 of the 99 state legislative chambers in 2007.12 The 1,562 state immigration policy bills introduced in state legislatures in 2007 are proof of the diversity of policy areas immigration touches. Some state bills, like Hawaii's SR13, which declares the state's opposition to 2005's Real ID Act, are purely symbolic in that their enactment will not affect a state policy change. Others, like Montana's SB214, which, in part, clarifies for Workers' Compensation Act and Unemployment Insurance Law that "The term 'employment' does not include: . . . service performed by an alien as identified in 8 U.S.C. 1101," specifically exclude undocumented immigrants from receiving benefits in currently existing law. Others still, like Washington's SCR8404, which provides employment and English-language training to immigrants, work to integrate immigrants into communities. MPI divides state immigration policy into two main groups: those that expand (N=313) and those that restrict (N=263) immigrant rights.13 The data for this project are compiled from multiple sources. Data on bill introductions were obtained from the Migration Policy Institute and New York University?s School of Law's (2007) "State Responses to Immigration: A Database of All State Legislation." This data set includes all immigration-related bills and resolutions introduced in state legislatures in 2007 (see Appendix A).14 Dependent Variables Sponsorship information for each bill was compiled from state legislative websites to create both count and dichotomous variables for each legislative introduction of 12 The Alaska Senate was the only state legislative chamber in which an immigration-related bill was not introduced. 13 Twenty-four bills (2.3% of the total number of bills introduced that fit this dichotomous coding) were omnibus bills coded as both expanding and contracting immigrant rights. Sponsorship totals in the analysis differ due to several states that allow multiple sponsors (not just cosponsors) of bills. 14 Information about the methodology used to find and evaluate each piece of legislation is available at: www.migrationpolicy.org/pubs/2007methodology.pdf. 38 restrictive and of expansive immigration legislation.15 I distinguish between restrictive and expansive immigration legislation to see if their sponsorship arises out of different influences. The main legislator dataset contains 7,232 state legislators and 576 state immigration bill introductions.16 I also analyze a subset of this main dataset to more narrowly examine Latino legislators' sponsorship behavior on state immigration policy; this data subset examines the sponsorship behavior of the 4,524 legislators serving in the 48 chambers of the 34 states containing Latino state legislators on 514 state immigration bill introductions (286 expanding immigrant rights, and 228 contracting immigrant rights). Independent Variables Several independent variables test for individual-, district-, and state-level influences on state legislative sponsorship of immigration legislation. Individual-level variables, including each legislator's party, leadership position, chamber, race, and gender, were gathered from state legislature websites. Latino ethnicity was derived from the Directory of Latino Elected Officials, 2007, available from the National Association of Latino Elected Officials (NALEO) Educational Fund. District-level demographic data, including percentages of Latinos, African Americans, households below the poverty level, the rate of each district's population change from 2000-2006, and the percentage vote for the 15 Heather Creek and Karina Shklyan were instrumental coding help for this laborious task, funded, in part, by Dr. James Gimpel. 16 The number of state legislators in this dataset (N=7,232) differs from the number of state legislative seats in the U.S. (N=7,382) for at least three reasons. First, Nebraska's 49 legislators are not included in the dataset because the Nebraska legislature is non-partisan. Second, the numbers differ because some of the state legislative seats were vacant throughout 2007, which reduces the number of seats in the dataset, while for some seats there are data for both the original holder and their replacement, which increases the total number of legislators in this dataset. Third, not all data are available for all legislators, so those legislators for which election data were not available for the most recent election, or whom were independents or members of third parties were not included in the analysis. 39 Republican candidate in each district's most recent state legislative election,17 were all compiled from the 2006 Almanac of State Legislative Elections (Lilley et al. 2007). State- level variables include a state's unemployment rate, degree of state legislative professionalization (Squire 2007),18 citizen ideology (Berry et al. 2010), and whether a state sat on the U.S. foreign border (dichotomous). Descriptive statistics and coding information for each variable are included in Appendix B. Four different multivariate models were run on the effects of state-, district-, and individual-level characteristics on state legislative sponsorship of immigration legislation in 2007. The unit of analysis is the legislator. Since the dependent variables testing legislators' descriptive representation bill sponsorship activity are strongly right-skewed count variables that denote the number of each type of immigration bill a legislator sponsored, I chose a negative binomial regression analysis. Negative binomial regression analysis is the standard method used in the study of state legislative sponsorship (see Appendix C for a summary of the scholarship on state legislative sponsorship). Each model utilizes a different dependent variable. The models in Table 2.3 examine the number of expansive and restrictive immigration bills sponsored by legislators in all state legislative chambers19 and by legislators in state legislative chambers with Latino legislators. Table 2.3 displays the results of probit regression models measuring whether legislators' actions on immigration legislation goes beyond the symbolic stage 17 In some states, coding the Republican percentage of the vote for the most recent election proved problematic due to the existence of multimember districts. In these states, representatives often win elections with (sometimes) less than 30 percent of the vote because of the multimember electoral system. To account for this problem, I tallied the winning percentage of the district if both members were from the same party. If the district was split between candidates from different parties, the senator?s winning margin for the district in which the delegate serves was used. 18 The Squire Index is a widely used measure of state legislative professionalism; I used the data from the 2003 compilation of this index found in Squire (2007) for this project. 19 See note 10. 40 represented by sponsoring a bill, to substantive action to enact the bill into law. These models focus on whether the bills that each legislator sponsors pass the legislative chamber of which they are a member. Because the data for all of the models are gathered for individual- and also for state-level variables, the distribution of state-level variables is not independent among legislators, but rather clusters. In light of the clustered nature of the data used in this study, the models are run with clustered (robust) standard errors.20 Results Table 2.1 contains the percentages of each type of immigration legislation sponsored by Latino, African-American, female, Republican, and state Senators and the percentages of bills sponsored by each group that passed the chamber in which they were introduced in all state legislative chambers and in those state legislative chambers in which Latino legislators serve; the results of these bivariate cross tabulations are similar both for the full dataset and for the subset specifically examining Latino legislator behavior. These data provide initial support for a number of the hypotheses offered. As hypothesized, Latino and female legislators sponsor expansive immigration legislation at rates greater than, and restrictive immigration legislation at rates lower than, their respective percentages in the state legislatures. Latino and female legislators also see a disproportionate percentage of the expansive immigration bills that they sponsor pass, hinting that these groups work to move their representation of Latino group interests beyond the symbolic bill sponsorship stage and into substantive public policy. African- 20 Primo, Jacobsmeier, and Milyo (2007) state that ?clustering arises because the attributes of states in which individuals reside do not vary across individuals within each state? (448). For example, in this chapter's analysis the state unemployment rate, the dichotomous border variable, the degree of legislative professionalism, and the total bill introduction in a state legislature for 2007 for Oregon are all constant across all legislators from Oregon; this violates the assumption that the errors are independently distributed and that their variance is constant. Models with aggregate- and individual-level data must use appropriate statistical techniques to account for this violation of the statistical methods assumptions. Not doing so can lead to an overstating of the statistical significance of coefficient estimates, and particularly the coefficients for states. 41 American legislators likewise sponsor expansive immigration legislation at numbers greater than, and restrictive immigration legislation at rates much lower than, their percentages in state legislatures. These results lend some evidence to group coalition theory, which posits that African-American legislators will provide substantive representation to Latinos, while also putting a damper on group or economic threat theories, which hold that African-American elites will act against Latino group interests to the benefit of their own group. African-American and female legislators also see a disproportionate percentage of the restrictive immigration legislation that they sponsor pass, indicating that these legislator groups exercise considerable legislative sway on this policy area. As hypothesized, sponsorship of immigration legislation shows broad partisan underpinnings, with Republicans sponsoring restrictive immigration legislation at much higher, and expansive immigration legislation at lower, rates than their percentage composition in state legislatures, yet passing fewer of the expansive and restrictive bills that they introduce than their Democratic colleagues, indicating perhaps that Republican lawmakers are more likely to introduce legislation to take a position, rather than to affect public policy. While bivariate cross tabulations can provide a broad overview of the data, it is necessary to control for the effects of multiple covariates in order to draw more confident conclusions. Table 2.2 displays the results of negative binomial regression analyses of the individual-, state-, and district-level influences on legislator sponsorship of expansive and restrictive immigration policy. Table 2.3 displays the results of probit regression analyses of these influences on the success legislators have in passing the immigration legislation they sponsor. 42 Table 2.1. Cross-tabulations of Individual-level Factors on State Legislative Sponsorship and Passage of Immigration Bills, 2007 Legislators in All State Legislative Chambers Independent Variable % of Sample Expansive Restrictive (N=7,232) Sponsored Passed Sponsored Passed (N=724) (N=197) (N=487) (N=53) % % % % Latino 3.2 9.7 7.6 1.5 0.0 African American 8.3 10.7 5.1 2.8 9.4 Female 23.5 32.2 37.6 16.4 26.4 Senate 26.1 32.4 38.6 34.1 35.9 Republican 45.6 31.8 24.9 75.1 58.5 Legislators in Chambers w/ Latino Legislators Independent Variable % of Sample Expansive Restrictive (N=4,623) Sponsored Passed Sponsored Passed (N=371) (N=122) (N=274) (N=37) % % % % Latino 5.1 13.5 12.3 2.2 0.0 African American 8.5 12.1 8.2 3.3 10.8 Female 25.3 32.9 43.4 18.6 29.7 Senate 16.5 26.7 38.5 21.5 16.2 Republican 45.0 32.1 23.8 72.3 59.5 Expansive Legislation Female legislators sponsor greater numbers of expansive immigration legislation, and are more successful at turning their sponsorship of expansive immigration measures into passed legislation, than male legislators. Female legislators thereby offer substantive representation to the largely Latino immigrant communities for which the majority of expansive immigration policy is written. Surprisingly, neither a legislator's party nor racial nor ethnic affiliations influence their sponsorship of expansive immigration legislation. Republican and African-American legislators are both less likely to succeed in passing their sponsored expansive immigration bills. For Republican members, a lack of their party's support coupled with an unwillingness of the Democratic Party to pass 43 legislation sponsored by the opposing party may stymie the passage of expansive immigration measures. African-American legislators have historically found greater difficulty in having the bills they sponsor get enacted than have non-African American legislators (Bratton and Haynie 1999). Of the district-level characteristics, legislators representing districts with high percentages of Latinos sponsored many more expansive immigration bills than did legislators representing districts with low percentages of Latinos. Legislators respond to district reelection concerns when sponsoring expansive immigrant policies. Members representing poor districts are less likely to see their sponsored expansive immigration legislation pass than are members representing wealthy districts. Such a finding, coupled with a similar finding for passage of sponsored restrictive immigration legislation, could signify a class bias in district representation, with members representing poor districts having less success passing any of their sponsored legislation than those representing wealthy districts. Variables related to legislators' substantive representation of Latino interests are those that have substantive impact on legislators' sponsorship of expansive immigration legislation in legislative chambers in which Latino legislators serve. Legislators representing the poorest districts and those with the highest percentages of Latino and African-American constituents were most likely to sponsor expansive immigration legislation. Legislators representing districts with the highest percentages of 44 Independent Variables Coefficient (Rb. SE) Predicted Counts 1 Coefficient (Rb. SE) Predicted Counts 1 Coefficient (Rb. SE) Predicted Counts 1 Coefficient (Rb. SE) Predicted Counts 1 -0.63* -0.39 -0.18 -0.06 0.21 0.07 0.36 0.13 (0.29) (0.31) (0.20) (0.22) 0.02 0.16 0.02 0.11 0.01 0.04 0.01 0.04 (0.01) (0.02) (0.01) (0.02) 0.03* 0.41 0.02 0.14 -0.03 -0.11 -0.03 -0.08 (0.02) (0.02) (0.02) (0.02) -0.70* -0.08 -0.25 -0.02 -0.60 -0.04 -0.36 -0.02 (0.38) (0.41) (0.37) (0.42) 0.00 0.19 0.00 0.02 0.00** 0.46 0.00 0.10 (0.00) (0.00) (0.00) (0.00) -0.00 -0.02 -0.001 -0.01 -0.01* -0.04 -0.01 -0.03 (0.00) (0.003) (0.00) (0.003) -0.02 -0.09 -0.04*** -0.11 -0.03 -0.08 -0.02 -0.04 (0.02) (0.01) (0.02) (0.02) 0.01 0.24 0.03*** 0.54 0.00 0.12 0.01 0.12 (0.01) (0.01) (0.01) (0.01) 0.03*** 1.28 0.03*** 0.98 0.01 0.08 0.01 0.06 (0.01) (0.01) (0.01) (0.01) 0.01 0.53 -0.01 -0.06 0.02* 1.34 0.01 0.26 (0.01) (0.01) (0.01) (0.01) -0.27 -0.03 0.04 0.00 1.45*** 0.10 1.50*** 0.10 (0.25) (0.22) (0.21) (0.21) 0.27* 0.03 0.47** 0.04 0.22 0.02 0.41 0.03 (0.17) (0.19) (0.16) (0.21) 0.15 0.02 0.15 0.01 -0.05 -0.00 0.06 0.00 (0.15) (0.19) (0.27) (0.37) 0.34*** 0.04 0.31* 0.03 -0.07 -0.00 -0.01 0.00 (0.12) (0.16) (0.20) (0.21) -0.24 -0.02 -0.50* -0.03 -0.69 -0.04 -1.13* -0.04 (0.28) (0.24) (0.53) (0.56) 0.26 0.04 0.39 0.04 -0.48 -0.02 -0.53 -0.02 (0.44) (0.39) (0.67) (0.61) -1.85 -3.65*** -4.12*** -4.91*** (1.19) (1.32) (0.93) (1.11) 3.34 3.06 3.89 4.54 (1.63) (1.67) (1.77) (2.42) N 4623 4623 Model chi-square 179.00 161.63 154.02 149.03 P = 0.0000 0.0000 0.0000 0.0000 Table 2.2. Negative Binomial Predictions for Influences on Sponsorship of Immigration Legislation, 2007 1 Predicted counts calculated using the observed values approach (Hanmer and Kalkan 2013). Notes: All p values are one-tailed except for those representing whether a legislator is African American and the percentage African-American of their district due to the contrasting hypotheses offered. * ? 0.05; ** ? 0.01; *** ? 0.001. State and Legislature Characteristics District Characteristics Legislator Characteristics Latino Constant alpha 7232 7232 Republican Senator Leadership Female African American % Republican Vote, Most Recent Election % Below Poverty Level % African-American % Latino Population Change, '00-'06 Unemployment Rate Citizen Ideology Legislative Professionalism On the US Border Total Bill Introductions Model 1: Expansive Legislation Model 2: Restrictive Legislation Legislators in All States Legislators in Chambers w/ Latino Legislators All State Legislators Legislators in Chambers w/ Latino Legislators 45 Latinos and African-Americans were also more likely to have the expansive immigration legislation they sponsored pass their chambers, 13% and 37% more likely, respectively. However, legislators representing the poorest districts in chambers with Latino legislators were 7% less likely to have this same legislation pass their chamber. Legislators from states with high unemployment rates sponsor an average of 0.39 fewer and are 18% less likely to pass expansive immigration bills than are legislators from states with low unemployment rates. This supports the economic threat hypothesis, which states that legislators from states with high unemployment act to preserve the state's rare jobs for natives. Also as hypothesized, legislators from states with a more liberal citizenry sponsor greater numbers of expansive immigration legislation and are more likely to have their sponsored legislation pass than are legislators from states with a more conservative citizenry. Against hypothesized expectations, legislators from border states sponsored fewer expansive immigration bills and had fewer of their sponsored bills expansive immigration bills pass than did legislators from interior states. States on the U.S. border have long had to handle the expense of large immigrant streams without federal recompense; it is not surprising that these legislators are less likely to introduce and see through to passage legislation that welcomes additional immigration. State senators and those from more professionalized legislatures also sponsored a greater number of expansive immigration bills than did legislators from states with less professionalized legislatures. Restrictive Legislation There are clear differences between what influences state legislative sponsorship of expansive and of restrictive immigration policy, indicating that these types of 46 immigration policy are not just two sides of the same coin. As hypothesized, Republican state legislators sponsor a greater number of restrictive immigration bills than do Independent Variables Coefficient (Rob. SE) Predicted Prob. 2 Coefficient (Rob. SE) Predicted Prob. 2 Coefficient (Rob. SE) Predicted Prob. 2 Coefficient (Rob. SE) Predicted Prob. 2 -0.63** -0.18 -0.17 -0.02 0.13* 0.02 0.23** 0.01 (0.21) (0.16) (0.09) (0.10) 0.02** 0.07 0.02 0.03 -0.02** -0.03 -0.03*** -0.03 (0.01) (0.01) (0.01) (0.01) 0.00 0.05 0.004 0.01 -0.01 -0.00 0.01 0.004 (0.02) (0.01) (0.01) (0.01) -0.56** -0.03 -0.29 -0.01 -0.22 -0.00 -0.34 -0.003 (0.29) (0.22) (0.23) (0.29) 0.00 0.06 -0.00 0.00 -0.00 -0.01 -0.0001 0.01 (0.00) (0.00) (0.00) (0.00) -0.00 -0.02 -0.002 -0.01 -0.00 -0.00 -0.002 0.002 (0.00) (0.002) (0.00) (0.003) -0.03* -0.05 -0.06*** -0.07 -0.03*** -0.02 -0.03*** -0.01 (0.02) (0.01) (0.01) (0.01) 0.01 0.15 0.02*** 0.37 -0.01 -0.01 -0.01 -0.004 (0.01) (0.003) (0.01) (0.01) 0.01 0.09 0.01* 0.13 0.01** 0.09 0.01 -0.01 (0.01) (0.01) (0.01) (0.01) -0.00 0.01 -0.01 -0.02 -0.00 -0.00 -0.01 -0.01 (0.01) (0.01) (0.01) (0.01) -0.23* -0.01 -0.11 0.00 0.05 0.00 0.14 0.001 (0.17) (0.09) (0.20) (0.30) 0.37*** 0.02 0.63*** 0.03 0.20 0.01 0.23 0.003 (0.12) (0.09) (0.17) (0.20) 0.09 0.01 0.15 0.01 -0.26 -0.00 -0.08 0.00 (0.13) (0.17) (0.35) (0.36) 0.26*** 0.02 0.29*** 0.01 0.10 0.00 0.13 0.001 (0.08) (0.12) (0.12) (0.16) -0.82*** -0.03 -0.95*** -0.01 0.50* 0.02 0.64* 0.01 (0.13) (0.10) (0.25) (0.29) 0.38 0.03 0.38 0.02 N/A? N/A 1 N/A 1 N/A 1 (0.30) (0.25) -0.25 -1.80*** -1.38** -1.12 (0.87) (0.65) (0.63) (0.66) N = 7232 4623 7000 4391 Model chi-square = 249.64 644.95 92.06 222.97 P> Chi2 = 0.00 0.0000 0.00 0.0000 Pseudo R2= 0.1962 0.2066 0.0949 0.1411 Table 2.3. Influences on Immigration Bill Passage State and Legislature Characteristics District Characteristics Legislator Characteristics 1. No Latino legislators introduced a restrictive immigration bill that passed the legislature, so this variable was dropped. 2. See Hanmer and Kalkan (2012). Senator Leadership Female African American Latino Constant % Republican Vote % Below Poverty Level % African-American % Latino Population Change, '00-'06 Republican Unemployment Rate Citizen Ideology Legislative Professionalism On U.S. Border Total Bill Introductions Expansive Legislation Restrictive Legislation All State Legislators Legislators in Chambers w/ Latino Legislators All State Legislators Legislators in Chambers w/ Latino Legislators 47 Democratic state legislators, a finding that well reflects the partisan rhetoric of the immigration policy arena at the national level, and which is itself reflective of the differences in the parties' bases. Partisanship continues to have a substantive impact on restrictive immigration policy sponsorship in chambers with Latino legislators. Republican members sponsor an average of 0.10 more restrictive measures than do Democratic members in legislative chambers in which Latino legislators serve, a number that mirrors the increase in restrictive immigration sponsorship by Republican members serving in all state legislative chambers. African-American state legislators sponsored fewer restrictive immigration bills than did non-African-American state legislators,21 which provides support for the group coalition hypothesis. While Latino immigrant groups do not find substantive representation in African-American legislators for expansive immigration policies, neither do they find that African-American legislators set the agenda against Latino interests. However, when African-American legislators do sponsor restrictive immigration policy, it is more likely to pass than is legislation sponsored by non-African- American legislators. This bucks previous findings that African-American legislators have less success passing sponsored legislation, and suggests that African-American legislators may be given a degree of deference in crafting legislation associated with the policy needs of a similar minority group. As hypothesized, legislators from districts with large population increases from 2000-2006 sponsored much more, an average of 1.34 more, restrictive immigration bills than did legislators from districts with lower population increases or with population declines. In passing restrictive immigration policy, state legislators from high-growth 21 When these results were calculated at the one-tailed level of statistical significance. 48 districts may be attempting to stem this growth or to limit newcomers' access to state benefits. Going against the direction hypothesized, state legislators representing poor and highly Republican districts sponsored fewer restrictive immigration bills than did those from wealthy and heavily Democratic districts. Those representing poor districts are also less likely to pass their sponsored restrictive immigration legislation than are legislators representing wealthy districts. Legislators representing heavily Latino districts are 9% more likely to pass the restrictive immigration legislation that they sponsor than are those who represent less Latino districts. Legislators from states with high unemployment rates are more likely to pass the restrictive immigration legislation that they sponsor than are those from states with low unemployment rates, and those from states with a more liberal citizenry are less likely to pass the restrictive immigration legislation that they sponsor than are those from states with a more conservative citizenry. These findings reflect the difficulty in getting restrictive immigration policy enacted in liberal states and in states with low unemployment. Legislators in legislatures that had more total bill sponsorships also sponsored, as hypothesized, a greater number of restrictive immigration bills than did those serving in legislatures that sponsor fewer bills. Discussion and Conclusion This chapter examines state legislator sponsorship of immigration legislation that either expands or restricts immigrant rights to test the findings of prior literature on state immigration policy adoption and on the descriptive representational roles of Latino legislators. It adds to the literature on state immigration policy by being the first to examine state legislator behavior in this policy realm, rather than that behavior aggregated up to policy adoption at the state level. This shift in emphasis provides deeper 49 understanding of the broader range of state immigration policies that make the legislative agenda, rather than just those that are enacted into policy. Studying agenda setting through state legislator bill sponsorship results in two findings that expand our knowledge of the influences that impact the setting of the state immigration policy agenda. First, legislator sponsorship of immigration bills that expand immigrant rights differs from legislator sponsorship of immigration bills that restrict immigrant rights. This difference is missed by the past research on state immigration legislation that conjoins expansive and restrictive immigration policy into one measure (Monogan 2010), and perhaps as well by work that analyzes Latino descriptive representation by looking at Latino-interest legislation as if it were unidirectional in nature (Rouse 2013). Latino- interest policies, and particularly state immigration policy, can be bifurcated into policies that benefit and those that are detrimental to Latino interests, and not accounting for this nuance may lead to erroneous results. Second, the state-level findings of this study both reinforce and vary from the findings on state immigration policy adoption of previous studies, which found state citizen ideology (Creek and Yoder 2012; Monogan 2010; Newman et al. 2012) or partisanship (Gulasekaram and Ramakrishnan 2013) and the partisan lean of the legislature or of the governor (see, for example, Sanchez, Sanchez, and Ybarra 2012; Chavez and Provine 2009; Reich and Barth 2010) amongst the most powerful indicators of state adoption of state immigration policy that either expands or restricts immigrant rights. This work finds that a legislator's personal partisanship or that of the district they represent is only a statistically and substantively significant influence on state legislator 50 sponsorship behavior on restrictive immigration policy. In setting the policy agenda for or against Latino interests, partisanship matters for the introduction of measures that restrict immigrant rights; at the other stages of the policymaking process examined in later chapters we may well see partisanship begin to influence legislator action on immigration policy in general as we move toward legislative enactment. Focusing on the actions of individual legislators also allows analysis of the effects of personal characteristics on legislator sponsorship of immigration legislation, a focus that yields important new findings on legislators' descriptive and substantive representation of the predominantly Latino immigrant population. State immigration laws are often highly charged pieces of legislation with very real and painful impacts on immigrant populations. They can determine whether undocumented families must uproot from their adopted land and move to less hostile locales, whether families living paycheck to paycheck lose or gain essential state benefits, and whether individuals detained for what had a been a misdemeanor might be deported. Yet, once controlling for all independent variables, Latino legislative sponsorship behavior on state immigration policy of either type does not differ from that of non- Latino legislators. This finding goes against that of prior literature (Bratton 2002; Martinez-Ebers et al. 2000; Rouse 2013), which found significant Latino descriptive representation of Latino interests at the sponsorship stage of the policymaking process. This finding is countered by the strength of the district constituent tie in, legislators' substantive representation of Latino interests; legislators representing districts with large Latino populations sponsor a greater number of expansive immigration bills than do those representing districts with smaller Latino populations. These results 51 indicate that, at least at the agenda-setting stage of the policymaking process, Latinos should not expect Latino legislators to offer a greater degree of representative attention to their interests than non-Latino legislators. The silver lining here is that those who represent districts with large percentages of Latinos are more likely to sponsor and see through to enactment legislation that expands the rights of the largely Latino immigrant community. If these results hold across the policymaking process, then Latinos might be better served pressuring their current legislators to support expansive immigration legislation than in running Latino candidates for state office. These preliminary results, however, may not fully represent the work of Latino legislators to further Latino interests, as bill sponsorship can largely be a symbolic activity in which legislators claim credit for taking a position to superficially satisfy constituent demands. Though Latino legislators may not sponsor a disproportionate number of bills that further Latino interests nor have success in enacting the expansive legislation that they do sponsor, they may well work harder than non-Latino legislators to enact legislation sponsored by other legislators that furthers Latino interests in the committee and floor voting stages of the policymaking process, thus providing Latinos descriptive representation at more substantive stages of the policymaking process. This study's findings also counteract the racial threat hypothesis and provide some support for the group coalition theory, with evidence that African-American state legislators are less likely to sponsor restrictive immigration legislation than are non- African-American legislators, and those legislators who represent districts with large percentages of African Americans are more likely to introduce legislation that expands immigrant rights than are legislators who represent districts with smaller percentages of 52 African Americans. In combination with this finding of substantive representation, legislators who represent largely African-American districts may have greater expertise crafting policy that benefits both their constituents and the Hispanic constituency that is largely similar in many socioeconomic characteristics. This study's expansive range of data?state legislator sponsorship of immigration policies across 49 states?is more geographically representative of state legislators than are previous studies that have utilized a narrower set of states in different policy realms (see Appendix C), meaning that its findings can be more confidently applied to state legislator sponsorship behavior in general.22 This study is the first to analyze state legislator sponsorship behavior across the full population of state legislators. Though not its primary analytical drive, this is the first study of state legislative sponsorship to test the influence of a state-level economic variable, state unemployment rate, on state legislative sponsorship behavior. This relationship was statistically significant for the number of expansive immigration bills introduced, with legislators from states with higher unemployment sponsoring substantively fewer expansive immigration bills than legislators from states with low unemployment, and was determinative of final passage of both types of sponsored immigration policy, with states with higher unemployment passing fewer expansive and a greater number of restrictive immigration bills. As aggregate measures, state-level unemployment rates may not be fully independent of the unemployment rate of each of a state's legislative districts, but substantially variation is likely to occur, particularly in districts that represent urban 22 Prior work that has examined Latino legislators' descriptive behavior (Bratton 2006; Casellas 2011; Rouse 2013) has focused on the broader policy area of Latino-interest legislation, and has looked across several years of data, making those studies, in some ways, more representative of Latino legislators' descriptive representational behavior than this current study. 53 areas. This finding points toward state legislators' sense of collective representation, representation of state needs, rather than just individual representation, representation of district and individual needs, even after controlling for constituent needs. Though future work would do well in going beyond this study's analysis to control for the influence of other state-level factors on state legislator behavior, the findings do justify the need to take into account a wider range of potential influences on state legislator agenda-setting behavior than endogenous institutional factors related to the legislatures and the legislation, or even of exogenous factors related to individual legislators and their districts. This study's findings also support Hopkins' (2010) findings of the effects of rapid population growth on localities' and states' immigration policy outcomes; legislators representing districts with large population increases introduced a substantively greater number of restrictive immigration bills than did legislators from districts with lower population growth or population declines. Legislators in these districts introduce measures aimed at reducing immigrant inflows. Rapid shifts in a district's demographic composition may signal a large immigration influx. Boushey and Luedtke (2011) find a similar relationship at the state level, with states with recent increases in their foreign- born population more likely to pass restrictive immigration policy than states without such increases. Interestingly, this result disappears when looking only state legislator sponsorship behavior in state legislative chambers that contained Latino legislators. Further analysis of the influences on state legislative behavior on state immigration policy is needed at all levels of the policymaking process. The next chapter will examine state legislator behavior on these immigration policies at the committee 54 stage of the policymaking process. 55 Chapter 3?Examining Representation of Latino Interests on Immigration in State Legislative Committees Research on legislative committees in general, and particularly on those in state legislatures, has been neglected in comparison to research on other stages of the policymaking process. This oversight can be attributed to greater access to quantifiable data at the sponsorship, co-sponsorship, and roll-call voting stages, and should not be viewed as a reflection of the comparative substantive importance of these stages in the policymaking process. Indeed, beyond sponsoring a bill, committees are where individual legislators have the greatest opportunity to alter legislation. The committee setting's relative lack of transparency to constituents also means that legislators are more likely at this stage to make substantive changes to legislation, rather than symbolic changes intended to gain constituent notice. Since legislators' personal characteristics impact their behavior on committees (Kathlene 1994; Rouse 2013; Rosenthal 2001), the composition of committees should impact policy outcomes. This chapter analyzes a new, expansive database of state legislative committee bill referrals in both chambers of the 49 states in which immigration bills were referred to a substantive committee, and bill enactments of state immigration policy across both legislative chambers in the 36 states which enacted state immigration policy in 2007 to analyze the composition of the legislative committees to which state immigration policy was referred. It begins the discussion of whether the ethnic and racial composition of committees impacts the substantive policy that a legislature enacts into law, and is the first work to examine state legislative committees across a wide sampling of chambers and states. State Legislative Committees In favoring, particularly, the roll-call voting stage of the policymaking process in their research, political scientists are guilty of allowing their research pursuits to follow the 56 path of more accessible data. Committees move legislation from bill sponsorship, an individual act, to the floor vote, a collective act of the entire legislative chamber. In the interim, a small subset of the legislative body, committee members, can have a disproportionate influence over policy formulation through the bill markup and committee amendment processes. Committee members therefore have ?disproportionate access into the crucial early stages of decision making and a valuable forum for position taking and credit claiming? (Evans 1999, 610). Committee action is also relatively hidden from constituents when compared to bill introductions, committee hearings, or floor votes. This opacity can provide legislators necessary leeway from constituent concerns to engage in legislative actions that impact policy in more substantive and in less symbolic, position taking ways. Committees also serve a gatekeeping role for public policymaking. In this role, the committee stage plays a much larger role in the policymaking process than does the roll-call voting stage. While almost every bill that is introduced in state legislatures is referred to a committee, as few as 20% ever emerge for floor consideration.23 Party control of the policy agenda (Cox and McCubbins 2005) plays a large role in determining which bills emerge from committee as the majority party does not allow bills that will split the party vote or that will be rejected on the floor out of committee or to be scheduled for a floor vote. Bills that have made it through the committee process are also generally given considerable deference by the full chamber, as these bills have undergone detailed and expert review by those members with the most interest and expertise in the subject matter. Almost all of the bills that gain a floor vote in the state legislatures pass. 23 Only one of the 558 restrictive and expansive immigration bills introduced in 2007 was not referred to any committee. Of these bills, 129 (23.1%) received a floor vote in a state legislative chamber. 57 Of the 129 immigration bills that fit in the expanding/contracting dichotomy analyzed in this project and that received recorded floor votes in 2007, only two (1.6%) were rejected.24 When studied, state legislative committees have most often been used as a mechanism to analyze broader models of organizational theory (Hedlund 1984; Hedlund and Freeman 1981) and to test party control theories (Overby and Kazee 2000; Aldrich and Battista 2002) derived from studies of Congress (Krehbiel 1993; Cox and McCubbins 2005). Given their key role in the policymaking process, it is important to understand the composition of legislative committees to understand whether legislators from ethnic or racial groups disproportionately impacted by a group of legislation have the opportunity to influence such legislation in this setting. Most studies of state legislative committee composition focus on member ideology, finding exceedingly rare committees that do not reflect the general ideology of the full state legislature (Overby and Kazee 2000; Overby, Kazee, and Prince 2004). These findings reinforce Hedlund's (1992) that it is member preference, not formal party control, which determines member committee assignment. Further, though the majority party dominates membership in committees, this majority is not disproportionate to the partisan composition of the chamber (Hedlund and Hamm 1996). Only a few studies have examined the composition of legislative committees by legislator gender, race, and ethnicity. Female state legislators are more likely than their male colleagues to serve on health, welfare, education, and family policy committees (Dolan and Ford 1997), while African-American legislators are more likely than 24 Both rejections, MTHB549 and MTHB735, occurred in the Montana House and restricted immigrant rights to obtain a state driver's license. 58 legislators from other ethnic and racial groups to serve on committees that focus on education, welfare, and that address African-American interests (Haynie, 2000). However, Bratton (2006) found that Latinos were no more likely to serve on legislative committees focusing on health, welfare, and education. Differences in the gender, ethnic, and racial composition of legislative committees matters because minority group membership influences legislator committee behavior. Both Kathlene (1994) and Rosenthal (2001) find that male and female committee members and chairs behave differently in committee than do their male counterparts, and that this behavioral difference can impact committee policy outcomes. Gamble (2007) and Minta (2009) study the composition of congressional committees by African- American and African-American and Latino legislators, respectively, finding that their participation differs from that of their White colleagues on issues relating to their ethnic groups. Ellis and Wilson (2013) find as an example of descriptive representation that congressional committees chaired by African-American and Latino members are more likely to hold hearings on minority issues than those chaired by legislators of other groups. Spun out more broadly, committee composition matters for policy outcomes, which are unlikely to better represent minority interests through descriptive representation that does not countermand the institutional levers of power (Tate 2003). Indeed, Griffin and Keane (2009) find that African-American Congress members hold disproportionately fewer positions of power?including membership on powerful committees?than do members of other racial or ethnic groups, and surmise that this leads to unequal African-American political representation. Ellis and Wilson (2013, 2) 59 further note that ?greater descriptive representation in positions with power over legislative agendas may be key to the functional representation of minority interests? and that ?the marginalization of minority groups from agenda-setting processes results in part from descriptive underrepresentation in positions with influence over legislative agendas.? While no work has examined African-American legislators in state legislative committees, Rouse (2013) has examined how Latino legislators descriptively represent Latino interests in state legislative committees. Her work follows Hall's (1998) groundbreaking analysis of Congress members' participation in committees, which uses a Likert scale to measure member committee participation, with member absence receiving a 0 to member engagement in committee markup receiving a 7. Rouse (2013) finds that Latino legislators provide descriptive representation on Latino interest issues, and that legislators of other racial groups provide substantive representation to Latinos through their committee participation in some state House chambers. Though Hall's (1998) method remains the gold standard for studying legislators' committee behavior, only a few studies have employed it due to the time-consuming data-collection process, which can be compounded in studies of state legislative committees where the availability of committee markup data varies by state, by chamber, and even by committee. As such, until state legislatures start to collect and make more readily available committee markup data, this is not a viable method to examine committees in more than a handful of chambers or legislatures. Studies of other stages of the state legislative policymaking process examine the impact of minority group legislators on policy outcome. Prior research has found that 60 female (Thomas 1991; Bratton and Haynie 1999; and Swers 2002; Reingold 2000), African-American (Tate 2004), and Latino (Bratton 2002; Martinez-Ebers et al. 2000) legislators provide descriptive representation beyond that called upon by constituent representation to members of their own group through their sponsorship activity, and in comparing enacted legislation with the broader population of sponsored legislation gauge the substantive impact of legislators' agenda-setting behavior, finding that minority group legislators are significantly less likely to have the legislation they sponsor enacted into law. Research has found that minority group legislators provide descriptive representation to the minority group they represent by voting differently at the roll-call stage than do their respective white and male colleagues (Swers 1998; 2002; Norton 1999; Welch 1985; Barnello 2001; Hogan 2008; Vega and Firestone 1995; Tate 2004; Swain 2006; Haynie 2001; Grose 2005; Barrett 1995; Rocca, Sanchez, and Uscinski 2008; Kerr and Miller 1997; Preuhs and Juenke 2011; Casellas 2010; Rouse 2013). Though a few studies have examined how ethnicity and race impact legislators' committee actions (Rouse 2013; Gamble 2007; Minta 2009), they have examined only a few legislative chambers and have focused solely on members' formal committee participation. However, committee members may well have influence on committee action beyond that witnessed in a markup or a committee hearing. Legislators unable to attend a formal committee markup may have collaborated on proposed changes with committee colleagues informally, or may have had a legislator act in their steed, both examples of participation that would be missed in Hall's (1998) measure. A broader measure than markup participation, simple committee composition, may better capture this informal influence. Studies of committee composition relative to the legislature as a whole have 61 focused mostly on how legislators are appointed to serve on committees and what this means in terms of theories of partisan control of the legislative process, and not on the disproportionate impact that over-representation of specific groups on committees can have on public policy. There has been little analysis of how the composition of committees, and how equal or unequal committee composition vis-?-vis the composition of the full chamber, can influence policy outcomes. Any one legislator likely has more power to influence her colleagues in the more intimate setting of the legislative committee than on the chamber floor. Descriptive representatives, both by acting in ways that enhance support for minority group interests and by bringing otherwise absent policy issues to the fore, ?may shape a substantially more favorable policy environment for the representation of minority interests? (Ellis and Wilson 2013, 2). These descriptive representatives, however, may well be shut out of the committees that oversee legislation that disproportionately impacts their minority community. This chapter examines the state-, district-, and personal-level characteristics of the legislators who comprised the state legislative committees to which expansive and restrictive immigration legislation was referred in 2007 in light of the composition of the state legislatures in general. Results will indicate whether legislators of certain characteristics are over- or underrepresented on the committees to which immigration legislation is referred. The following hypotheses, derived from social science research, focus the paper on the racial and ethnic composition of the state legislative committees that worked on immigration policy in 2007. Primary Hypotheses Numerous studies have found that a legislator's gender, race, and ethnicity impact their 62 legislative behavior, with many finding that these personal characteristics go beyond simple representation of constituency to provide descriptive or substantive representation of a group (Haynie 2001; Swers 2002). Latino legislators are more likely to introduce and pursue legislation beneficial to immigrant populations?to provide descriptive representation to Latino interests?than are non-Latino legislators (Bratton 2002; Martinez-Ebers et al. 2000). Legislators are also known to rely on the expertise of their colleagues in taking voting cues (Kingdon 1989); Latino legislators may therefore be accorded a degree of deference as experts on immigration issues, leading to higher rates of enactment of immigration bills on which Latino legislators work. There may also be a constituency tie in. Latinos consistently rank immigration higher as an important issue than do members of other racial and ethnic groups, and Latinos are more likely to favor either expanding or keeping immigration at the same levels than are other racial or ethnic groups (see, for example, de le Garza 2004). Immigration policy, therefore, is the type of policy about which ?attentive publics,? those most affected by the law?in this case Latino constituents?would be concerned (Arnold 1992), and therefore Latino legislators can be expected to be more likely to serve on committees to which state immigration bills are referred. The largely Latino composition of the recent wave of immigration, a stark difference from previous, primarily white immigrant waves, makes race and ethnicity a factor in legislators' immigration policy responses. The distinct ethnic character of this immigrant wave may stoke native hostility that increases with the size of the immigrant group (Alba, Rumbaut, and Marotz 2005; Rocha and Espino 2009). Rapid shifts in a district's demographic composition may signal a large immigration influx. Hopkins 63 (2010) finds that rapid increases in a locale's ethnic diversity increases natives' hostile political attitudes toward immigrants, and the probability of that locale passing restrictive immigration legislation. Boushey and Luedtke (2011) find a similar relationship at the state level, with states with recent increases in their foreign-born population more likely to pass restrictive immigration policy than states without such increases. Hypothesis 1: Latino legislators will be more likely to serve on committees to which all types of immigration legislation is referred than will non-Latino legislators. Hypothesis 2: Legislators representing districts with a high percentage of population increase will be more likely to serve on committees to which restrictive immigration legislation is referred than will legislators from districts with low rates of population growth. Control Measures A number of control variables are also included to account for legislator characteristics extemporaneous from this chapter?s focus on state legislators? roles descriptively and substantively representing minority groups in committee. Briefly, each legislator's party is coded because immigration is a partisan measure, with Republicans favoring more restrictive and Democrats more expansive immigration policy. Legislators from states with a more conservative citizen ideology are expected to favor more restrictive immigration policies, while those from states with a more liberal citizenry are expected to favor more expansive policies, and those that represent the most partisan Republican districts will oversee more restrictive and less expansive immigration policy directed to their committees than will those from more moderate or from the most partisan Democratic districts. Gender may also play a role, as female legislators tend to be more liberal than male legislators on social welfare issues (Barrett 1995; Poggione 2004) and on issues related to African-American interests (Bratton and Haynie 1999). Female 64 legislators may therefore offer substantive representation of Latino interests on immigration policy by serving on committees that oversee more expansive immigration policy than do their male counterparts. However, research that focuses specifically on public policy attitudes toward immigration finds that women are more supportive of restrictive immigration policy than are men. These conflicting findings make gender an important control to include in these models. Immigration is also an economic issue, with legislators from states with higher unemployment rates and higher levels of district poverty expected to focus on more restrictive immigration policy to prevent more surplus labor than those from states with lower levels of unemployment or representing districts with lower levels of poverty. Legislators residing in states on the U.S. border are expected to consider more of both types of immigration legislation than are legislators in states in the interior. Legislators who work in more professionalized legislatures and those with a history of enacting more bills are expected to consider more immigration legislation of both types than are legislators who work in less professionalized legislatures or those that have a history of enacting fewer bills. Whether a legislator is in their party's leadership is a control variable without a directional hypothesis. Methods and Data The data for this project are compiled from multiple sources. Data on bill introductions were obtained from the Migration Policy Institute and New York University?s School of Law's (2007) "State Responses to Immigration: A Database of All State Legislation." This dataset includes all immigration-related bills and resolutions introduced in state legislatures in 2007.25 This chapter relies on this database's coding of expansive and restrictive immigration legislation across the 49 states that referred such measures to 25 Information about the methodology used to find and evaluate each piece of legislation is available at: www.migrationpolicy.org/pubs/2007methodology.pdf. 65 substantive committees26 in 2007 to analyze state legislators' descriptive and substantive representation of Latino interests.27 Legislative committees' racial or ethnic composition may impact policy outcomes, particularly in a policy area such as immigration that is related to race or ethnicity. Moreover, a committee may shift the bill toward or away from representation of the interests of a minority group when the composition of the committee differs from that of the chamber. In such cases, committee members play an outsized representational role in formulating policy that may differ from the preferences of the full legislative chamber. Policy voted out of committee that has been amended in committee too far from majority legislative preferences can always be defeated on the chamber floor, but at least in the context of state immigration policy in 2007 analyzed here, that is an exceedingly rare outcome. The 558 state immigration policy bills (301 expansive and 257 restrictive immigration bills) that were referred to substantive committees in 49 state legislatures in 2007, and the 88 bills that were enacted into law (60 expansive and 28 restrictive) from legislatures in 36 states are proof of the diversity of policy areas immigration touches. Some state bills, like Hawaii's SR13, which declares the state's opposition to 2005's Real ID Act, are purely symbolic in that their enactment will not affect a state policy change. Others, like Montana's SB214, which, in part, clarifies for Workers' Compensation Act and Unemployment Insurance Law that "The term 'employment' does not include: . . . service performed by an alien as identified in 8 U.S.C. 1101," specifically exclude undocumented immigrants from receiving benefits in currently existing law. Others still, 26 Committee referrals to rules committees were excluded from this analysis because bill referral to this committee was the automatic option in several states, after which it may have been referred to a substantive committee or committees. Including this committee might skew the results. 27 See note 22. 66 like Washington's SCR8404, which provides employment and English-language training to immigrants, work to integrate immigrants into communities. This diversity is also reflected in the types of committees to which these bills were referred, ranging from Appropriations to Health and Human Services to Higher Education to Inland Fisheries and Wildlife committees. Dependent Variables I followed each of the 558 bills that were referred to a committee through the legislative process to all of the committees in both chambers to which each bill was referred, recording which members had a chance to oversee each bill in committee. Due to multiple committee referral, the 313 expansive immigration bills introduced were referred to a total of 467 substantive committees (and average of 1.49 committee referrals per bill), while the 263 introduced restrictive immigration bills were referred to a total of 362 substantive committees (an average of 1.38 committee referrals per bill) (see Appendix A). In total, 2,262 legislators served on substantive committees to which immigration policy was referred in 2007, many serving on multiple committees. Research that has studied Latino representation in state legislator committees (Rouse 2013) has done so by selecting committees that may see Latino interest legislation, no doubt missing some Latino-interest bills referred to other committees. The dependent variable for the probit regression models (Tables 3.8 and 3.9) is whether a restrictive/expansive immigration bill was referred to a committee on which the legislator sits. To explicitly examine whether Latino state legislators were as likely as non-Latino legislators to have immigration legislation of both types referred to a committee on which they serve, I also ran models (Table 9) that examined bill referral to committee in only those legislative chambers that had Latino state legislators. The primary results of interest to this chapter, 67 therefore, represent the likelihood that a state legislator served on a committee to which an immigration bill of either type was referred, controlling for all other covariates. Independent Variables I examine several independent individual-, district-, state-, and institutional-level variables to determine the characteristics of the committees to which state immigration policy was referred. Individual-level variables, including each legislator's party, whether they were in the majority party, leadership position, chamber, race, and gender, were gathered from state legislature websites. Latino ethnicity was derived from the Directory of Latino Elected Officials, 2007, available from the National Association of Latino Elected Officials (NALEO) Educational Fund. District-level demographic data, including percentages of Latinos, African Americans, households below the poverty level, the rate of each district's population change from 2000-2006, and the percentage vote for the Republican candidate in each district's most recent state legislative election,28 were all compiled from the 2006 Almanac of State Legislative Elections (Lilley et al. 2007). State- level variables include a state's unemployment rate, degree of state legislative professionalization (Squire 2007),29 citizen ideology (Berry et al. 2010), percent Latino in each state legislative chamber (NALEO), and whether a state sat on the U.S. foreign border (dichotomous). Two different multivariate models were run on the state-, institutional-, district-, and individual-level characteristics of the committees to which immigration bills of both types were referred in 2007. The primary unit of analysis is the legislator. Because the data for each of the models are gathered for individual- and also for state-level variables, 28 See note 15. 29 See note 16. 68 the distribution of state-level variables is not independent among legislators, but rather clusters. In light of the clustered nature of the data used in this chapter, the models are run with clustered (robust) standard errors.30 Results It is important to examine whether the composition of the committees to which each type of immigration bill was referred mimics that of the composition of the legislatures at large. Committee composition matters here because of the disproportionate impact committee members may have on the legislative outcome of these bills. If the committees to which these bills were referred differed in composition from the broader legislature, then lawmakers that descriptively or substantively represent specific interests may have greater influence on this legislation than does the average legislator. The preliminary findings presented in Table 3.1 are important because they show that the composition of the committees to which these bills are referred is different from the composition of the legislatures at large. Table 3.1 examines the characteristics of legislators who served on committees to which expansive immigration legislation was referred, while Table 3.2 examines the characteristics of legislators who served on committees to which restrictive immigration legislation was referred. There were stark differences between the legislators who served on committees to which both types of immigration legislation was referred. Legislators who served on committees that considered expansive immigration legislation were from states with higher levels of unemployment, more professional legislatures, were more likely to be on the U.S. border, were from districts with higher percentages of African Americans, Latinos, and more recent population change, were more likely to be Democratic, in the majority party, and 30 See note 18. 69 African American or Latino than were those who did not serve on committees to which these bills were directed. Most of these characteristics have been hypothesized as leading to more expansive immigration policy, except for increases in recent population, and from states with high unemployment rates. Bill referral of restrictive state immigration policy (Table 3.2) saw similar differences between the characteristics of those who served on these committees. Legislators serving on committees referred restrictive immigration policy were from more conservative states with higher unemployment rates and more professional legislatures that were on the U.S. border, represented districts with higher levels of poverty, higher percentages of African Americans, Latinos, and more recent population changes, and were more likely to be African American or Latino. Most of these legislator characteristics have been hypothesized as being correlated with legislator behavior furthering restrictive immigration policy, excepting percentages of a legislator's district Latino and whether a legislator is Latino. Overall, 2,262 individual legislators served on committees to which expansive immigration legislation was referred in 2007, and 2,089 served on committees to which restrictive immigration legislation was referred in 2007. Tables 3.3-3.6 display some bivariate descriptive statistics for the immigration bills that were referred to committees in relation to minority group membership on those committees. Overall, expansive immigration legislation was referred to committees that saw an overrepresentation of Latino, African-American, female, and majority party members, and where Republicans were underrepresented (Table 3.3). Latino legislators were similarly overrepresented on committees to which restrictive immigration legislation was referred. African-American 70 and Republican state legislators served on committees to which restrictive immigration legislation was referred in greater proportion to their composition in the legislative chambers that referred such legislation. Overall, 34.4% of all state legislators serving in states where expansive legislation was referred to committee served on a committee to which expansive immigration legislation was referred. 71 Variable All States Served on Committee Bill Referred to Did not Serve on Committee Bill Referred to Difference of Means p value State and Legislature Characteristics Unemployment Rate 4.28 4.35 4.24 0.00 Citizen Ideology 59.25 61.38 58.00 0.00 Legislative Professionalism 19.17 22.81 17.04 0.00 On the US Border 32.37 33.91 31.47 0.05 Total Bill Enactments 697.98 1000.46 520.17 0.00 District Characteristics % Below Poverty 15.12 15.10 15.13 0.86 % African-American 10.48 11.59 9.84 0.00 % Latino 9.00 12.32 7.04 0.00 Population Change, '00-'06 4.92 5.53 4.55 0.00 Legislator Characteristics Republican 44.35 41.65 45.94 0.01 Majority Party 62.13 63.99 61.04 0.02 Female 23.60 24.93 22.82 0.06 African American 7.86 9.50 6.89 0.00 Latino 3.33 5.35 2.16 0.00 N = 6110 2262 3848 Table 3.1. Difference of Means between Legislators who Served on Committees Where Expansive Immigration Legislation was Referred, and Those Who Did Not, 2007 Notes: The difference of means test is two-tailed to determine whether to reject the null hypothesis that there is no difference in the means of the variables. Nebraska is excluded due to its non-partisan legislature. 72 A larger percentage of Latino legislators had a chance to work on both expansive and restrictive immigration in committee than did any other of the groupings studied here, with 54.5% of all Latino state legislators serving on committees to which expansive immigration legislation was referred and 43.2% of all Latino state legislators serving on Variable All States Served on Committee Bill Referred to Did not Serve on Committee Bill Referred to Difference of Means p value State and Legislature Characteristics Unemployment Rate 4.43 4.55 4.37 0.00 Citizen Ideology 57.83 55.48 59.06 0.00 Legislative Professionalism 18.41 19.06 18.07 0.00 On the US Border 35.80 37.21 33.13 0.00 Total Bill Enactments 747.44 843.31 697.20 0.00 District Characteristics % Below Poverty 15.31 15.83 15.04 0.00 % African-American 11.31 13.07 10.38 0.00 % Latino 8.55 10.27 7.65 0.00 Population Change, '00-'06 5.12 5.45 4.95 0.05 Legislator Characteristics Republican 45.08 46.55 44.32 0.10 Majority Party 61.70 61.29 61.91 0.63 Female 23.88 23.22 24.23 0.38 African American 8.36 9.81 7.60 0.00 Latino 3.21 4.40 2.58 0.00 N = 6075 2089 3986 Table 3.2. Difference of Means between Legislators who Served on Committees Where Restrictive Immigration Legislation was Referred, and Those Who Did Not, 2007 Notes: Difference of means test is two-tailed to determine whether we can reject the null hypothesis that there is no difference in the means of the variables. Nebraska is excluded due to its non-partisan legislature. 73 committees to which restrictive immigration legislation was referred. The average legislator saw 1.04 expansive immigration bills referred to a committee on which they served, but Latino legislators saw 1.79 expansive immigration bills referred to committees on which they served. African-American state legislators also see more of both types of bills referred to their committees, perhaps hinting at their service on committees whose jurisdictions are similar to those on which Latino legislators serve. Finally, while the average legislator saw 17.16% of the expansive state immigration policy referred to a committee in the state in which they served referred to a committee on which they served, Latino legislators saw just 16.79% of the expansive % of Legislature % of Sample (N=6,522) Referred % of sample (N=6,548) Referred % of Sample (N=2,262) (N=2,089) % % 3.4 5.4 4.6 3.3 4.4 3.3 8.1 9.5 7.4 8.5 9.8 7.0 23.4 24.9 25.4 23.7 23.2 25.2 27.1 33.6 27.0 26.1 30.4 27.0 44.3 41.7 43.4 44.9 46.6 49.5 62.1 64.0 62.7 61.7 61.3 59.9 100.0 34.4 100.0 100.0 31.9 100.0All Legislators Latino African Am rican Female Senate Republican Majority Party Table 3.3. Percentages of Group Legislators on Committees to which State Immigration Policy Bills were Referred, 2007 Independent Variable Expansive Restrictive N Referred N Referred (N=2,262) (N=2,089) # % # % 222 54.5 213 43.2 531 40.5 559 36.7 564 36.6 485 31.3 759 42.6 634 37.1 940 32.6 970 33.3 1445 35.7 1278 31.9Majority Party Latino African American Female Senate Republican Table 3.4. ercentages of Each Group of Legislators to which State Immigration Bill were Referred, 20 7 Independent Variable Expansive Restrictive 74 immigration legislation referred to committees on which they served. These bivariate comparisons, however, do not control for the impact of covariates. To control for the impact of these covariates, and to assess the impact of any one characteristic, we must use multivariate models. Table 3.7 shows the results of a probit model whose dependent variables are whether a legislator served on a committee to which each type of immigration bill was referred, and Table 3.8 shows the same model but only for the state legislative chambers in which Latino legislators served. The results are largely complementary, and so have been presented together below. All Legislators 1.04 0.77 Republican 0.90 0.80 Majority Party 1.12 0.75 Female 1.17 0.69 Senate 1.14 0.87 # # Latino 1.79 0.82 African American 1.33 1.07 Table 3.5. Average Number of Immigration Bills Referred to Legislative Committee by Legislator Committee Member Characteristic, 2007 Independent Variables Expansive Restrictive Bills Referred Bills Referred All Legislators 17.16 16.82 Republican 17.46 16.73 Majority Party Member 17.39 17.20 Female 17.21 17.51 Senate 25.38 23.60 % % Lat no 16.79 15.89 African American 20.54 18.46 Table 3.6. Average Percentage of Each Type of Immigration Bill Referred to a Committee in Which a Legislator Served, by Group, 2007 Inde en e t Variabl s Expansive Restrictive Bills Referred Bills Referred 75 Legislators from the most liberal states were 9% more likely to have expansive immigration legislation referred to their committees than were members from the most conservative states, states with the most professional legislatures were 24% more likely to see bills referred to their committees, than were those from the least professionalized legislatures, legislators from states that enact the most bills were 52% more likely and 62% more likely in chambers with Latino legislators to have expansive immigration legislation referred to a committee on which they served than were those from states that enact the fewest number of bills, legislators from legislative chambers with the highest percentages of Latinos were 39% more likely overall and 28% more likely in chambers with Latino legislators to see expansive immigration bills referred to a committee on which they served than were legislators from legislative chambers with the lowest percentage of Latinos. In terms of district characteristics, legislators from districts with the highest percentages of African-Americans were 26% more likely to serve on a committee referred expansive immigration legislation in legislative chambers in which Latino legislators served than were legislators representing those districts with the lowest percentages of African Americans in the same legislative chambers, legislators from districts with the highest Republican vote percentage were 3% more likely to see expansive immigration legislation referred to a committee on which they served than were those from districts with the lowest percentage vote for Republican members, legislators from districts with the highest population change were 47% more likely to see expansive immigration legislation referred to a committee on which they served than were legislators from districts with low population change or even population loss. 76 Table 3.7. Influence of Committee Composition on Immigration Bill Committee Referral Expansive Legislation Restrictive Legislation Referred Referred Independent Variables Coefficient (Rob. SE) Pred. Probabilities Coefficient (Rob. SE) Pred. Probabilities State and Legislature Characteristics Unemployment Rate 0.01 0.01 0.17 0.26 (0.10) (0.14) Citizen Ideology 0.00** 0.09 -0.01 -0.23 (0.01) (0.01) Legislative Professionalism 0.01** 0.24 -0.002 -0.03 (0.01) (0.01) On U.S. Border -0.20 -0.08 -0.18 -0.06 (0.31) (0.27) Total Bill Enactments 0.00*** 0.52 0.000008 0.17 (0.00) (0.000008) % Latino in Chamber 0.02** 0.39 0.02** 0.35 (0.01) (0.01) District Characteristics % Republican Vote 0.00** 0.03 -0.002* -0.08 (0.00) (0.001) % Below Poverty Level -0.01 -0.13 -0.004 -0.07 (0.01) (0.01) % African-American 0.00 0.07 0.002 0.07 (0.00) (0.01) % Latino 0.01 0.20 0.003 0.11 (0.00) (0.004) Population Change, '00-'06 0.01* 0.47 -0.002 -0.14 (0.00) (0.01) Legislator Characteristics Republican -0.11* -0.04 0.11* 0.04 (0.05) (0.07) Senator 0.38*** 0.15 0.24* 0.09 (0.10 (0.11) Leadership -0.32*** -0.11 -0.22* -0.08 (0.10 (0.10) In Chamber Majority Party 0.04 0.01 0.02 0.01 (0.03) (0.03) Female 0.06 0.02 0.02 0.01 (0.05) (0.05) African American 0.07 0.03 -0.06 -0.02 (0.11) (0.13) 77 Latino -0.07 -0.02 0.001 0.001 (0.13) (0.15) Constant -1.09* N/A -0.62 N/A (0.50) (0.69) N = 5977 5948 Wald Chi Squared 281.44 78.82 Prob> Chi Squared 0.00 0.00 Pseudo R2= 0.10 0.04 Notes: All p values are one-tailed. Predicted probabilities calculated from raising an independent variables minimum value to its maximum value while holding other independent variables at their mean values. * ? 0.05; ** ? 0.01; *** ? 0.001. In terms of personal legislator characteristics, Republican members were 4% less likely to see expansive immigration legislation referred to a committee on which they served than were Democratic members, senators were 15% more likely to see expansive immigration legislation referred to a committee on which they served, and 23% more likely in chambers in which Latino legislators served, than were members of the lower chamber, and those in leadership were 11% less likely to see expansive immigration legislation referred to committees on which they served than were those not in leadership. None of the measures of legislators' gender, race, or ethnicity proved statistically significant for the referral of expansive immigration legislation to a committee on which these members served. In terms of whether and how many restrictive immigration bills were referred to a committee on which a legislator serves, legislators in chambers with the highest percentages of Latinos were 35% more likely to have restrictive immigration bills referred to committees on which they served than were legislators who serve on committees in chambers with the lowest number of Latino legislators. Legislators from states that enact the most bills were 32% more likely in chambers with Latino legislators to have restrictive immigration legislation referred to a committee on which they served 78 Table 3.8. Influence of Committee Composition on Immigration Bill Committee Referral in States with Latino Legislators Expansive Legislation Restrictive Legislation Referred Referred Independent Variables Coefficient (Rob. SE) Pred. Probabilities Coefficient (Rob. SE) Pred. Probabilities State and Legislature Characteristics Unemployment Rate 0.16 0.19 0.25 0.39 (0.12) (0.16) Citizen Ideology 0.001 0.02 -0.01 -0.22 (0.01) (0.01) Legislative Professionalism -0.0001 -0.02 -0.01 -0.21 (0.01) (0.01) On U.S. Border -0.08 -0.03 -0.26 -0.10 (0.28) (0.34) Total Bill Enactments 0.0001*** 0.62 0.00001* 0.32 (0.00) (0.00) % Latino in Chamber 0.02* 0.28 0.02 0.31 (0.01) (0.01) District Characteristics % Republican Vote 0.002 0.07 -0.003 -0.10 (0.001) (0.00) % Below Poverty Level -0.004 -0.07 0.003 0.04 (0.01) (0.01) % African-American 0.01*** 0.26 -0.004 -0.12 (0.003) (0.01) % Latino 0.003 0.12 0.001 0.05 (0.003) (0.00) Population Change, '00-'06 0.003 0.20 -0.01 -0.26 (0.006) (0.01) Legislator Characteristics Republican -0.07 -0.03 0.15* 0.06 (0.05) (0.07) Senator 0.59*** 0.23 0.28 0.11 (0.16) (0.23) Leadership -0.26 -0.10 -0.22 -0.08 (0.14) (0.12) Female 0.03 0.01 0.01 0.01 (0.06) (0.06) African American -0.07 -0.03 0.03 0.01 (0.12) (0.14) Latino 0.01 0.01 0.01 0.002 (0.13) (0.15) Constant -1.65 -0.85 79 (0.70)* (0.76) N = 4019 4044 Wald Chi Squared 125.09 54.15 Prob> Chi Squared 0.0000 0.0000 Pseudo R2= 0.1301 0.0391 Notes: All p values are one-tailed. Predicted probabilities calculated from raising an independent variables minimum value to its maximum value while holding other independent variables at their mean values. * ? 0.05; ** ? 0.01; *** ? 0.001. than were those from chambers with Latinos serving that enact the fewest number of bills. In terms of district- and personal-level characteristics, legislators in the most Republican districts were 8% less likely to have restrictive immigration policy referred to committees on which they served than were those that represented the most Democratic districts. Republican legislators were 4% more likely to have restrictive immigration legislation referred to committees on which they served?6% more likely in chambers in which Latinos served?than were Democratic legislators, senators were 9% more likely to have restrictive immigration policy referred to committees on which they served than were members of lower chambers, and those in leadership where 8% less likely to have restrictive immigration legislation referred to their committees than were those not in leadership. Discussion and Conclusions Studies of state legislative policymaking have often focused on the full makeup of the legislature in terms of gender, party, ethnicity, race, but those that serve on committees have a disproportionate influence on the small percentage of legislation that advances from the introduction stage to the floor vote. And, at least in the state immigration policy realm, the vast majority of those that advance, are passed in floor vote. Bill referral in the past has been used to assess party power (for example, see Krehbiel 1993). Member 80 participation in committees has been used to gauge legislators' descriptive and substantive representation of minority group interests in committee. However, the ethnic and racial composition of committees should, in theory, impact the policy that moves to the floor and onward to enactment. Thus, studying the ethnic and racial composition of the committees that work on legislation can tell us whether legislators from specific ethnic or racial groups have the opportunity to descriptively and substantively represent their ascriptive group. Past literature has found that legislators who are members of ethnic and racial groups tend to act for those group interests, even controlling for constituency, ideological, and party influences. For these reasons, it is essential to analyze the composition of the committees to which these bills are sent. What does it tell us that the composition of the committees to which immigration bills are referred differs from the legislatures in general? For immigration bills, these findings show that the committees to which expansive bills are referred are more likely to be composed of Democratic legislators from more liberal states serving in more professionalized legislatures containing higher percentages of Latino legislators, and representing districts that had seen large, positive population change. The hypotheses suggest that these referred expansive immigration bills may therefore be pulled in a more liberal, pro-Latino direction than that represented in the state legislatures as a whole. The population change finding suggests that these legislators are under pressure to integrate these new arrivals through expansive immigration policy. Legislators on the committees to which restrictive immigration policy was referred were more likely to be Republicans from legislatures with a high percentage of Latino legislators, who represent more liberal districts, a profile that hints at ideological moderation. 81 This chapter examined the composition of the committees to which state immigration policy is referred. Though overshadowed in the scholarly literature and in terms of public transparency by both the sponsorship and floor voting stages of the legislative policymaking process, what happens in committee largely determines the fate of legislation, particularly in when passage from committee generally assures passage on the chamber floor. Committee members therefore have considerable influence over what type of policy a state enacts. There has so far been little attention to how committee composition influences the outcome of public policy. This chapter initiates such a discussion by comparing the composition of the committees to which state immigration policy is referred in relation to the composition of the chamber at large. In terms of the study of descriptive and substantive representation of Latino interests, the main emphasis of the broader study of which this chapter is a part, committees to which both types of immigration bills are referred are not significantly different from state legislatures as a whole in terms of legislators' personal ethnic or racial characteristics. As committee assignment is largely a reflection of legislator preference, this suggests that Latino legislators are not disproportionately selecting committees to which Latino interest area legislation is referred. Though this seems to fly in the face of previous research that found that minority group legislators are more likely to choose to serve on committees that are likely to oversee legislation that concerns the ethnic, racial, or gender group that they descriptively represent, it must be kept in mind how broadly immigration policy ranges across the legislative spectrum. Rather than selecting specific committees theorized to oversee a large percentage of legislation of disproportionate import to a specific racial, ethnic, or gender group then count the number of legislators 82 with those ascriptive characteristics that serve on those committees in comparison to on other committees, I have followed the legislation itself to determine the characteristics of legislators' committee oversight. Doing so provides a much fuller picture of which legislators had an oversized role in determining the fate of immigration policy that disproportionately impacts Latinos. Surprisingly, expansive immigration policy was not steered toward committees on which Latino state legislators served, a finding that may be taken to reflect that Latino legislators' behavior on expansive immigration policy is mostly symbolic, were such a result found in terms of Latino legislators' introduction of such policy in Chapter 2. This chapter's findings suggest that Latino legislators had as equal an opportunity as did non-Latino legislators to impact state immigration policy in committee, as least when their potential impact is measured in terms of committee composition. Further analysis of whether Latino legislators were as likely as non-Latino legislators to be chairs of these committees or to serve in the majority party in these committees is warranted. This chapter examined the composition of legislative committees to which state immigration legislation was referred in 2007 to draw conclusions about state legislators' descriptive and substantive representation of Latino interests in committee. It is the first scholarship to study such a broad swath of state legislative committees, and to focus on the impact that committee ethnic and racial composition may have on legislative outcomes. The succeeding chapter will examine how state legislators vote on their chamber floors on the legislation that does make it out of committee. 83 Chapter 4: Analyzing Representation of Latino Interests through Legislative Roll-Call Voting on State Immigration Legislation Legislative floor action occupies the last legislative step in the policymaking process, and one that determines whether a bill will advance to the executive for signature into law or veto. Legislator behavior at the roll-call voting stage faces different pressures than those that influence legislator behavior at earlier stages of the policymaking process. Whereas state legislators at the sponsorship stages of the policy process may be motivated predominantly by district- and personal-level influences, legislative roll-call voting is linked in past political science scholarship with legislator reaction to party pressures (see, for example, Cox and McCubbins 2005; Aldrich 1995; Rohde 1991; but see Krehbiel 1993). Roll-call voting therefore is a policymaking stage that offers a distinct arena in which to analyze legislators' reaction to unique pressures. Indeed, recent research has found that party- and district-pressures prove influential in state legislators' roll-call voting behavior on all (Casellas 2011) and on only Latino-interest legislation (Rouse 2013), while finding no influence for personal-level characteristics. The uneven dispersal of immigration nationwide (Frey 2006; McConnell 2008; Passel and Cohn 2011) has made immigration policy more salient in specific states and in specific districts. Legislators respond to the interaction between their state and districts' economic and demographic abilities to handle the strength of the immigration flow with targeted legislative policy solutions. While prior research has found that enacted state public policy does ultimately represent a state's collective ideology (Erikson, Wright, and McIver 1995), it is individual legislators, held accountable by district electoral politics, who vote on state immigration policy. Robust research shows 84 that legislators respond primarily to re-elective motives (Mayhew 1974). While studies of state legislative roll-call voting acknowledge this motivation, many also test for personal- level motivations in legislators' roll-call voting behavior. Findings have been mixed on the influence of state legislators' gender, race, and ethnicity on their voting behavior (Swers 1998; 2002; Norton 1999; Welch 1985; Barnello 2001; Hogan 2008; Vega and Firestone 1995; Tate 2004; Swain 2006; Haynie 2001; Grose 2005; Barrett 1995; Rocca, Sanchez, and Uscinski 2008; Kerr and Miller 1997; Preuhs and Juenke 2011; Casellas 2010; Rouse 2013). The immigration policy response to state-, district-, and personal- level pressures has been two types of immigration legislation: that which expands or that which restricts government protection and services to undocumented immigrants. The state legislative roll-call voting and the state immigration policy adoption literature has regularly tested hypotheses of the effect of the ethnic and ideological composition of a legislator's constituency as well as their personal-level factors like a legislator's gender, ethnicity, and race have on state legislators' roll-call voting behavior on state immigration policy. This study is novel from previous studies in both its substantive focus and in the data it employs. It is the first work of state legislative roll- call voting to examine the state immigration policy context, to state legislator behavior across the full population of state legislators, and to examine state legislators' collective representation of state needs. Legislative Roll-Call Voting Political scientists have conducted extensive research into legislators' roll-call voting behavior. These studies have measured the impact of legislators' ideology, party, and constituency on their roll-call voting in both Congress (see, for example, Kingdon 1989; Poole and Rosenthal 1985; 1991; 1997; Snyder and Groseclose 2000) and state 85 legislatures (Poole and Rosenthal 2001; Wright and Schaffner 2002; Jenkins 2006). More recently, studies at both national and sub-national levels have examined and found significant the influence of legislators' personal characteristics such as gender (Swers 1998; 2002; Norton 1999; Welch 1985; Barnello 2001; Hogan 2008; Vega and Firestone 1995), race (Tate 2004; Swain 2006; Haynie 2001; Grose 2005; Barrett 1995), and ethnicity (Rocca, Sanchez, and Uscinski 2008; Kerr and Miller 1997; Preuhs and Juenke 2011; Casellas 2010; Rouse 2013) on their roll-call voting behavior. Critics of the use of roll-call votes to gauge legislator behavior note that legislators' preferences are constrained to but three choices at the roll-call vote stage?an aye or nay vote or an abstention?and that while these choices may capture the direction of legislators' preferences, they do not capture their intensity (Highton and Rocca 2005). Recent scholarship has branched out to study legislative behavior at the sponsorship (Schiller 1995; Bratton and Haynie 1999; Barnello and Bratton 2007), cosponsorship (Koger 2003; Bratton and Rouse 2011), and committee (Kathlene 1994; Rosenthal 2001; Rouse 2013) stages of the policymaking process to better assess legislator behavior in arenas when they are not presented with the direct yes-or-no question that the roll-call posits. However, focusing on legislators' roll-call voting behavior remains important for three reasons. First, the legislative floor is the only stage in the policymaking process where all members of the legislative chamber have equal access and ability to formally act on a bill. These actions are visible through members' recorded roll-call votes. Whereas analyzing state legislator sponsorship and committee behavior on state immigration policy provides important clues about the policy preferences of the state legislators with 86 the most interest in the immigration policy arena, analyzing legislator behavior on state immigration policy through roll-call votes is important because doing so provides information on how state legislators as a whole behave in response to state immigration policy. Second, the roll-call vote stage of the policymaking process is important both substantively and in terms of substantive representation; legislators' votes at this final legislative stage of the policymaking process determine whether a bill will pass. Thus, roll-call voting provides a window into how legislators represent their district constituents and constituents of specific racial groups in a substantive manner. Third, the roll-call vote remains the most visible of the policymaking stages to constituents, who do not as closely follow the sponsorship and committee behavior of their representatives. The transparency of the roll-call vote makes it an ideal policymaking stage for the testing of constituency influence on legislative behavior. Roll-Call Voting and the Representation of Minority Groups Prior research has found that female (Thomas 1991; Bratton and Haynie 1999; and Swers 2002; Reingold 2000), African-American (Tate 2004), and Latino (Bratton 2002; Martinez-Ebers et al. 2000) legislators do provide descriptive representation to members of their own group through their sponsorship activity and through their vote at the floor vote stage (Swers 1998; 2002; Norton 1999; Welch 1985; Barnello 2001; Hogan 2008; Vega and Firestone 1995; Tate 2004; Swain 2006; Haynie 2001; Grose 2005; Barrett 1995; Rocca, Sanchez, and Uscinski 2008; Kerr and Miller 1997; Preuhs and Juenke 2011; Casellas 2010; Rouse 2013). Additionally, the extant literature has found relatively robust the impact of having Latino constituents on legislators' substantive representation of Latinos (Welch and Hibbing 1984; Casellas 2010). However, analysis of the link between the descriptive representation of Latino interests by Latino legislators finds more 87 mixed results, with Hero and Tolbert (1995), Casellas (2010), and Rouse (2013) finding no difference and Welch and Hibbing (1984) and Kerr and Miller (1997) finding significant difference in the descriptive representation of Latino constituents when represented by a Latino legislator. Though both Rouse (2013) and Casellas (2010) argue that such a study is important to understanding the substantive representation of Latinos by their elected legislators, only Rouse (2013) has so far focused exclusively on the roll-call voting behavior of legislators on legislation that disproportionately affects Latinos. Her study of Latino representation compares Latino legislators' roll-call voting behavior on both general legislation and on that which she classifies as Latino interest legislation, finding little in the way of an effect for ethnicity on Latino voting behavior. This chapter specifically focuses on the roll-call voting behavior of state legislators on restrictive and expansive immigration policy, a policy issue area which disproportionately affects Latinos and upon which Latinos therefore might be expected to prove an "attentive public" (Arnold 1992). In doing so, it looks to add to the debate on both the descriptive representation of Latinos by Latino representatives and the substantive representation of Latinos by legislators of other minority groups in general as studied by Hero and Tolbert (1995), Casellas (2010), Welch and Hibbing (1984) and Kerr and Miller (1997), and on Latino and other minority legislators' representation of Latinos within a specifically Latino-oriented policy arena (Rouse 2013). By analyzing the roll-call voting behavior of legislators across a broader array of states than that previously assessed, the results of this chapter should be more broadly generalizable of legislators' descriptive and substantive 88 representation of Latino policy interests than have been the results of previous studies.31 Methodological Considerations for Studying Roll-Call Votes Conducting a multivariate analysis on state legislators' voting behavior carries with it a number of methodological pitfalls. Most legislative research on legislators' roll-call voting behavior has attempted to measure whether constituents' ideology is reflected in their legislators' roll-call votes and subsequently their states' enacted public policies. Two main studies have examined this linkage through estimation of state government ideology writ large (Wright, Erikson, and McIver 1993; Berry et al. 1998), but this approach tells us little about individual legislators' roll-call behavior. The most prominently used method of examining this linkage through individual legislators' roll-call voting is to use roll-call votes to estimate legislators' ideal policy points. These studies have most often utilized as a dependent variable Poole and Rosenthal's (1985; 1991; 1997) Nominate scores of legislator ideal points. One limitation to these measures is that a legislator's ideal point exists within the legislative context of the chamber in which they serve, and cannot therefore be compared across chambers or legislatures; most studies that utilize these measures therefore examine Congress, though there have been attempts to apply this method to state legislatures (Poole and Rosenthal 2001). Doing so brings several logistical problems to the fore. First, though interesting analysis of state legislators' roll-call voting behavior has occurred in studying one (Gerber and Lewis 2004; Kousser, Lewis, and Masket 2007; Bertelli and Richardson 2008) or from two to 20 state legislatures (Wright and Schaffner 31 Prior studies of Latino legislators' descriptive representation of Latino interests may well be more representative of Latino legislators' behavior in other ways. Casellas (2011) examines Latino legislators' actions on all legislation in several sessions of the U.S. Congress as well in a few statehouses; Bratton (2006) examines a broader array of policy that corresponds to Latino interests than solely immigration policy, and does so in seven statehouses across multiple years; and Rouse (2013) examines a broader swath of Latino-interest legislation across six statehouses over three years. 89 2002; Casellas 2011; Rouse 2013; Jenkins 2006; McCarty, Poole, and Rosenthal 2006; Preuhs and Juenke 2011), running models on each individual chamber across more than a handful of state legislative chambers is cumbersome and, in the end, not illustrative of state legislators' roll-call voting behavior in general. Second, gathering data on each roll- call vote for each state legislator in each legislative chamber is an extremely time consuming and arduous task that had not been completed until Wright (2004) did so recently as part of the Representation in America?s Legislatures project. His compilation of state legislators' roll-call voting is a wonderful resource, and one that has formed the basis for important recent work like Preuhs and Juenke's (2011) determination of the ideological distance of Latino legislators from their legislative peers in 20 states, but it only contains data for the 1999-2000 state legislative session. Most recently, Shor, Berry, and McCarty (2010) collected state legislators' roll-call voting data to create a new measure of state legislator ideology that is comparable across chambers and across legislatures, but only for 11 states. They have since extended this measure to include aggregate state ideology scores for all 50 states for the years 1999-2009, but have not yet released the individual legislator ideal scores. Others (Vega and Firestone 1995; Kau, Keenan, and Rubin 1982; Segal, Cameron, and Cover 1992) have operationalized interest group scores of legislators as the dependent variable as a proxy for Congressional legislators' personal preferences. But this measure also has its detractors, who note that interest groups do not score every vote, but only the votes that they are interested in, so the measure is not an accurate portrayal of a legislator's ideology, and that by analyzing the outcome of ideology, the vote itself, that these measures are not independent of the roll-call vote itself, a criticism that also 90 applies to measures like Nominate that are derived from roll-call votes themselves (Jackson and Kingdon 1992). Additionally, using interest group scores at the state level is particularly problematic because interest groups rarely score state legislative roll-call votes. Finally, others (Miller and Stokes 1963; Jenkins 2006; 2008) have surveyed legislators to gauge their ideology, but uneven survey response rates mean that such an index is often not inclusive of all legislators. Each of these studies looks at either one (Vega and Firestone 1995; Kau, Keenan, and Rubin 1982; Segal, Cameron, and Cover 1992; Miller and Stokes 1963) or only a handful (Jenkins 2006; 2008) of legislatures. This chapter is less interested in determining the influence of ideology versus that of party than it is in determining the impact of racial and ethnic personal characteristics on voting behavior. Further, analyzing legislators' ideology becomes a more interesting exercise and yields more useful findings when assessing numerous roll-call votes across multiple policy areas; this study is moored to one policy area, and the ideological dualism inherent in the vote for each type of immigration policy itself may make it a solid proxy for legislator ideology, with more conservative members voting for more restrictive immigration legislation and more liberal members voting for more expansive immigration legislation. Therefore, measures of state legislators' ideology are less applicable to this study than is their actual voting behavior. Primary Hypotheses Debate exists on whether African Americans see Latino immigrants as an economic or political threat or as a potential minority political coalition partner to counter majority white political power (see, for example, Kaufmann 2003). African-American legislators have explained their opposition to anti-immigrant legislation by linking these measures to 91 past legislative attempts to limit African-American civil rights (see, for example, Bacon 2012). African-American and Latino groups also share similar economic positions in American society, with group levels of poverty, education, and of the insured much more similar between African-Americans and Latinos than between either of these groups and whites (U.S. Census Bureau 2007). In certain contexts, African-American legislators have provided Latinos substantive representation (Bratton 2002; 2006). Such shared group characteristics may lead to shared type and direction of policymaking in African- American legislators' substantive representation of Latino interests in immigration policy, an issue-area that particularly impacts Latinos. Numerous studies have found that a legislator's gender, race, and ethnicity impact their legislative behavior, with many finding that these personal characteristics go beyond simple representation of constituency to provide descriptive or substantive representation of a group (Haynie 2001; Swers 2002). Latino legislators are more likely to introduce and pursue legislation beneficial to immigrant populations?to provide descriptive representation to Latino interests?than are non-Latino legislators (Bratton 2002; Martinez-Ebers et al. 2000). Legislators are also known to rely on the expertise of their colleagues in taking voting cues (Kingdon 1989); Latino legislators may therefore be accorded a degree of deference as experts on immigration issues, leading to higher rates of enactment of immigration bills on which Latino legislators work. There may also be a constituency tie in. Latinos consistently rank immigration higher as an important issue than do members of other racial and ethnic groups, and Latinos are more likely to favor either expanding or keeping immigration at the same levels than are other racial or ethnic groups (see, for example, de le Garza 2004). Immigration policy, therefore, is the type of 92 policy about which ?attentive publics,? those most affected by the law?in this case Latino constituents?would be concerned (Arnold 1992). The largely Latino composition of the recent wave of immigration, a stark difference from previous, primarily white immigrant waves, makes race and ethnicity a factor in legislators' immigration policy responses. The distinct ethnic character of this immigrant wave may stoke native hostility that increases with the size of the immigrant group (Alba, Rumbaut, and Marotz 2005; Rocha and Espino 2009). Rapid shifts in a district's demographic composition may signal a large immigration influx. Hopkins (2010) finds that rapid increases in a locale's ethnic diversity increases natives' hostile political attitudes toward immigrants, and the probability of that locale passing restrictive immigration legislation. Boushey and Luedtke (2011) find a similar relationship at the state level, with states with recent increases in their foreign-born population more likely to pass restrictive immigration policy than states without such increases. Party influence plays a much greater role at the roll-call vote stage than at the sponsorship stage of the legislative policymaking process. Whereas sponsorship is seen as largely an individual act independent of party influence (Schiller 1995), legislators rate highly the influence of party on their voting behavior (Kingdon 1989; Ray 1982; Songer et al. 1986). The roll-call voting stage of the policymaking process is the only stage to involve all members of a chamber, and not just those with greater substantive interest in the legislation who would be more likely to act on a piece of legislation at the sponsorship or committee stages. These less-knowledgeable or -interested members may well rely more heavily on party cues as voting heuristics than would their more-engaged peers, who have already likely sorted themselves into the proper party through their 93 policy interest in such a highly ideological policy issue. No work has broadly assessed party voting in state legislatures, instead focusing on party voting in a few state legislatures (Jenkins 2002; Wright and Schaffner 2002; Ray 1982; Witt and Moncrief 1993). However, much recent research has found party to have the largest single influence on the descriptive representation of Latino constituents by Latino legislators (Rouse 2013; Casellas 2011). In this way, Latinos may be gaining representation not directly through Latino legislators, but indirectly through the Democratic Party, or, as Rouse (2013, 118) puts it: "the representation of Latinos appears to take place in a collective manner through legislative parties." Parties actively work to restrict the choices legislators make at the floor vote stage to those that will not embarrass the party brand or politically endanger its individual members (Aldrich 1995; Cox and McCubbins 1994). As such, legislation that advances out of committee to the chamber floor has gained the majority party's imprimatur, and majority party members may be expected to vote for it at higher rates than non-majority party members (Jenkins 2002). Hypothesis 1: African-American state legislators will vote for a greater percentage of legislation that expands immigrant rights, and a lesser percentage of legislation that restricts immigrant rights, than will non-African-American legislators. Hypothesis 2: Latino legislators will vote for a greater percentage of legislation that expands immigrant rights and a lesser percentage of legislation that contracts immigrant rights than will non-Latino legislators. Hypothesis 3: Legislators representing districts with rapid population increases will vote for a greater percentage of restrictive immigration bills and a lesser percentage of expansive immigration bills than will legislators representing districts with less growth or a net exodus of residents. Hypothesis 4: Republican legislators will vote for a greater percentage of state immigration legislation that restricts and a lesser percentage of immigration 94 legislation that expands immigrant rights than will Democratic legislators. Hypothesis 5: Majority-party members will vote for a greater percentage of immigration legislation of both types than will minority-party members. Control Measures A number of control variables are also included to account for legislator characteristics extemporaneous from this chapter?s focus on state legislators? roles descriptively and substantively representing minority groups through their floor voting behavior. Briefly, legislators from states with a more conservative citizen ideology are expected to vote for a higher percentage of restrictive immigration policies, while those from states with a more liberal citizenry are expected to vote for a higher percentage of expansive policies, and those that represent the most partisan Republican districts will vote for a higher percentage of restrictive and a lower percentage of expansive immigration policy than will those from more moderate or from the most partisan Democratic districts. Gender may also play a role, as female legislators tend to be more liberal than male legislators on social welfare issues (Barrett 1995; Poggione 2004) and on issues related to African- American interests (Bratton and Haynie 1999). Finally, while findings have been relatively clear that gender impacts the number and topic of bills legislators sponsor (Thomas 1991; Bratton and Haynie 1999; Swers 2002), it has been more mixed on the impact of gender on legislators' roll-call voting behavior. Some scholars find significant differences between the roll-call voting behavior of male and female legislators on women's issues with female legislators voting more liberal (Swers 1998; 2002; Hogan 2008; Welch 1985; Barrett 1995; Poggione 2004), while others find no such difference (Barnello 2001; McCarty, Poole, and Rosenthal 1997). Policy issues related to women, children, and the family are most likely to be associated with the Democratic Party or a 95 liberal ideology (Reingold 2000), and female legislators have been found to be more liberal than male legislators on social welfare issues (Barrett 1995; Poggione 2004), and on issues related to African-American interests (Bratton and Haynie 1999). This greater female policy liberalism may also extend to female legislators' substantive representation of Latinos on the social welfare issues relating to immigration. Female legislators may therefore offer substantive representation to Latinos on immigration policy by voting to enact a higher percentage of expansive immigration policy than their male counterparts. However, considerable research contradicts these expectations specifically on women's attitudes on immigration policy, finding that women are generally more favorable to restrictive immigration policy than are men (Citrin et al. 1997; Espenshade and Calhoun 1993; Citrin, Reingold, and Green 1990; Burns and Gimpel 2000). This discrepancy in the research is one important reason to control for the effects of gender on legislators' actions on immigration policy. Immigration is also an economic issue, with legislators from states with higher unemployment rates and higher levels of district poverty expected to vote for a higher percentage of restrictive immigration policy to prevent more surplus labor than those from states with lower levels of unemployment or representing districts with lower levels of poverty. Legislators residing in states on the U.S. border are expected to vote for a higher percentage of both types of immigration than are legislators in states in the interior. Legislators who work in more professionalized legislatures and those with a history of enacting more bills are expected to vote for a higher percentage of immigration legislation (Monogan 2010) of both types than are legislators who work in less professionalized legislatures or those that have a history of enacting fewer bills. Whether a legislator is in their party's leadership is a control variable without a 96 directional hypothesis. Data and Methodology This study relies on the MPI/NYU coding of expansive and restrictive immigration legislation across the 40 states that voted on such measures in 2007 to analyze state legislator roll-call voting behavior.32 A total of 129 state immigration policy bills that fit this coding (77 expansive and 52 restrictive immigration bills) received recorded floor votes in state legislatures in 2007. Dependent Variables In light of the current lack of a measure of state legislators' roll-call votes comparable across state legislatures and across state legislative chambers as discussed above, I have operationalized a measure of state legislators' votes for immigration legislation of both types as the dependent variable. The dependent variable for the OLS multivariate regression models displayed in Table 4.6 is legislators' percentage of yes votes on each type of bill. Independent Variables Several independent variables test for individual-, district-, and state- and institutional- level influences on state legislative roll-call voting on immigration legislation. Individual-level variables, including each legislator's party, whether they were in the majority party, leadership position, chamber, race, and gender, were gathered from state legislature websites. Latino ethnicity was derived from the Directory of Latino Elected Officials, 2007, available from the National Association of Latino Elected Officials (NALEO) Educational Fund. District-level demographic data, including percentages of Latinos, African Americans, households below the poverty level, the rate of each district's population change from 2000-2006, and the percentage vote for the Republican candidate 32 See note 22. 97 in each district's most recent state legislative election,33 were all compiled from the 2006 Almanac of State Legislative Elections (Lilley et al. 2007). State-level variables include a state's unemployment rate, degree of state legislative professionalization (Squire 2007),34 citizen ideology (Berry et al. 2010), percent Latino in each state legislative chamber (NALEO), and whether a state sat on the U.S. foreign border (dichotomous). Two different multivariate models were run on the effects of state- and institutional-, district-, and individual-level characteristics on state legislative roll-call voting on immigration legislation in 2007. The primary unit of analysis is the legislator. Both models, displayed in Table 4.6, utilize a different dependent variable to examine the percentage of expansive and restrictive immigration bills for which each legislator voted. Because the data for each of the models are gathered for individual- and also for state- level variables, the distribution of state-level variables is not independent among legislators, but rather clusters. In light of the clustered nature of the data used in this chapter, the models are run with clustered (robust) standard errors.35 Results Two important considerations must be noted when analyzing state legislative roll-call votes. First, the vast majority of bills that make it to the floor pass. Of 345 total immigration bills that made it to the floor in a chamber of a state legislature in 2007, 339 (98.26%) passed the chamber. Of the 296 immigration bills that made it to a recorded floor vote in a chamber in 40 states in 2007,36 only six (2.03%) were defeated. Individual legislator roll-call votes showed little dissent as well; 110 (37.16%) of the 296 33 See note 15. 34 See note 16. 35 See note 18. 36 Immigration bills were voted on in the House chambers of 36 states, and in the Senate chambers of 33 states. 98 immigration bills receiving recorded floor votes garnered unanimous aye votes. This low level of dissent can stem from the power of the majority party to act as a cartel in weeding out unsanctioned legislation in committee, bringing only the bills that they know will pass (Cox and McCubbins 2004). It also speaks to the trust that individual members have in the specialization crystallized in the legislative committee process, wherein committees containing legislators well-versed in the topic of the legislation are given deference by other legislators on the chamber floor (Shepsle and Weingast 1987). Only two of the 129 bills that fit in the expanding/contracting dichotomy analyzed in this project were rejected (1.6% of all of these types of bills receiving floor votes) across all chambers that had recorded floor votes.37 Individual legislator assent was also very high on these bills, with 46 bills expanding immigrant rights passing a chamber unanimously (59.7% of all expanding legislation that received a recorded floor vote), and 17 bills contracting immigrant rights passing a chamber unanimously (32.7% of all contracting legislation that received a recorded floor vote). Of the bills that fit the scope of this study, 77 bills in 25 states expanding and 52 bills in 21 states contracting immigrant rights received a recorded vote on a chamber floor in 34 state legislatures in 2007.38 For a breakdown of the number of bills that were introduced and were enacted into law in each state, see Appendix A. How representative of the average U.S. state were the states where legislators voted on expansive or restrictive immigration policy in 2007? Table 4.1 displays the 37 See note 21. 38 Three bills that received recorded floor votes were omnibus bills that both contracted and expanded immigrant rights in three state legislatures. Texas, Tennessee, and South Carolina held recorded floor votes on omnibus immigration legislation in 2007. 99 descriptive statistics for states that held recorded floor votes on each type of immigration legislation in comparison to the grouping of states that did not hold recorded floor votes on each type of immigration policy for the variables of primary interest in this study. No difference between the means of states that voted on and those that did not vote on expansive immigration legislation achieved standard levels of statistical significance. The difference between those states that voted for restrictive immigration legislation and those that did not is statistically significant at standard accepted social science levels for three variables: population change from 2000-2006, state citizen ideology, and percent of Republican legislators. Comparing the descriptive statistics on these variables does lend support to the hypotheses offered, with states that have a more conservative citizen ideology, more Republican legislators in their state legislatures, and with greater population changes (Newman et al. 2012; Hopkins 2010) all more likely to vote on restrictive immigration measures. However, the point here is not to analyze these descriptive statistics for support for our hypotheses, but rather to determine whether our sample of state legislatures that held recorded floor votes on immigration bills in 2007 is representative of all state legislatures. These findings suggest that we can freely apply our findings of state legislator roll-call voting behavior on expansive immigration legislation to state legislators in general, but that we must be more careful in applying our findings on state legislator roll-call voting behavior on restrictive immigration to all state legislators in general. However, this study does analyze all state legislator roll-call voting on restrictive and expansive immigration legislation, so results from these analyses can readily be taken as representative of legislator voting behavior in this policy realm. Moving to the individual-legislator level, expansive immigration legislation 100 received a floor vote in state legislatures containing 3,194 legislators in 2007; of these legislators, a vast majority, 3,068 (96.06%), voted yes on at least one of these bills. In fact in state House chambers, 80.17% of legislators voted for every one while only 4.41% of state House members did not vote for any of the expansive immigration bills brought to a recorded vote in their chamber, whereas 85.06% of state Senators voted for every one and only 2.91% of state Senators did not vote for any of the expansive immigration bills brought to a recorded vote in their chamber. Personal-level variables impacted legislators' voting actions on state immigration legislation, at least when assessing the relationship between voting for these measures through bivariate analysis as shown in Tables 4.2-4.5. Table 4.2 compares Latino, African-American, female, members in the upper chamber, Republican, and majority- party legislators' votes for expansive and restrictive immigration legislation with these Variable All States Voted On Did Not Vote On Difference of Means p value Voted On Did Not Vote On Difference of Means p value State and Legislature Characteristics Unemployment Rate 4.35 4.24 4.53 0.29 4.27 4.46 0.50 Citizen Ideology 58.42 61.01 55.97 0.27 51.79 63.62 0.01 Legislative Professionalism 0.18 0.17 0.19 0.54 0.17 0.19 0.49 On the US Border 0.30 0.28 0.33 0.69 0.38 0.25 0.34 Total Bill Enactments 681.96 616.52 741.96 0.61 488.19 820.29 0.18 District Characteristics % Below Poverty 12.60 12.72 12.54 0.84 12.20 12.95 0.41 % African-American 10.30 10.23 10.62 0.89 9.05 11.45 0.39 % Latino 9.56 11.90 7.20 0.10 10.71 8.77 0.50 Population Change, '00-'06 5.96 6.81 5.18 0.28 8.40 4.23 0.005 Legislator Characteristics R publican 46.31 44.16 48.56 0.29 53.31 41.07 0.002 Femal 23.27 23.59 22.94 0.76 24.11 22.64 0.49 Af ican American 7.71 7.22 8.22 0.65 6.40 8.70 0.30 Latino 3.58 5.13 1.97 0.13 3.61 3.56 0.98 N = 50 25 24 21 28 Table 4.1. Comparison between States that Held and did not Hold Recorded Floor Votes on Expansive or Restrictive Immigration Legislation, 2007 Expansive Legislation Restrictive Legislation Notes: Difference of means test is two-tailed to determine whether we can reject the null hypothesis that there is no difference in the means of the variables. Nebraska is excluded due to its non-partisan legislature. 101 groups' overall representation in state legislatures. Table 4.3 shows the percentages of each of these groupings who voted for at least one expansive or one restrictive immigration bill that reached the floor in state legislatures in 2007; the high percentages voting for each type of legislation in each group show in part the high level of assent on bills that reach the floor in state legislatures. Table 4.4 shows the average number of votes for each type of immigration bill by members of these groupings, which gives a better indication than does the simple dichotomous variable analyzed in Tables 4.2 and 4.3 of the variability in the intensity of group members' support for these measures. Table 4.5 displays the average percentage of votes for each type of immigration bill by members of these groupings, which better accounts for the opportunities each legislator had to vote on each type of bill than does the simple count variables analyzed in Table 4.4. Assessing these statistics overall, African-American, female, majority party members, and members in the upper chamber all supported expansive immigration legislation at rates greater than the average legislator, while African-American legislators supported restrictive immigration at much lower rates than did the average legislator. Republican legislators supported expansive immigration legislation at rates lower than did Democrats while supporting restrictive immigration legislation at rates higher than did Democrats. Senators, and majority-party and, surprisingly, female legislators generally voted at higher rates for legislation of both types. Perhaps the most interesting finding of these bivariate analyses pertains to Latino support of expansive immigration policy. Latino legislators show high levels of intensity of support for expansive immigration legislation when compared to the average legislator as evident in their high average number of votes for expansive immigration bills (Table 102 4.4), yet the average percentage of votes Latino legislators cast for expansive immigration legislation was lower than that of the average member. Though these simply summary statistics provide greater insight into the voting behavior of these groups on state immigration legislation and generate support for some of the hypotheses posited, multivariate analysis is necessary to control for the effect of each covariant on state legislator voting behavior. The OLS regression model results shown in Table 4.6, which control for the influence of all other variables, offer further support for some of the hypotheses. In this table, I have run two separate models on the voting behavior of state legislators on expansive and restrictive immigration policy: one model on all state legislative chambers that had recorded floor votes on these measures, and one on only those chambers in which Latino state legislators served. These latter models should provide a clearer view of Latino state legislator behavior on immigration policy than would analyzing chambers without Latinos serving therein. Analyzing only those variables that achieved standard social science levels of statistical significance, for expansive immigration legislation, legislative professionalism and legislators' gender proved statistically significant influences on legislators' roll-call votes, but only the coefficient for legislators' gender was in the direction hypothesized. Female legislators' percentage voting for expansive immigration legislation was two percentage points higher than that of their male colleagues. This finding does not prove significant when looking only at state legislative chambers in which Latinos served. In these latter legislative chambers only, Republican legislators were 11 percentage points less likely to vote for expansive immigration measures than were Democratic legislators. The percentage of Latinos in a legislator's 103 district also becomes statistically significant in this model, but the coefficient indicates that this finding has little substantive impact on legislator's voting behavior on expansive immigration policy. (N=7,348) Group % of legislatures 3.16 8.27 23.49 26.24 45.62 61.95 All Legislators 2.50 1.70 Republican 2.21 1.72 Majority Party Member 2.69 1.83 Female 2.69 1.75 Senate 2.39 1.67 # # Latino 4.44 2.40 African American 3.18 1.39 Table 4.4. Average Number of Yes Votes by Legislators from Each Grouping on Each Type of Immigration Bill, 2007 Independent Variables Expansive Restrictive Vote For Vote For Majority Party Member 96.77 82.02 All Legislators 96.06 78.84 Senate 97.01 87.73 Republican 95.13 80.54 African American 97.61 71.95 Female 96.98 79.60 Vote For Vote For % % Latino 94.61 78.89 Majority Party Member 62.77 58.65 Table 4.3. Percentage of Each Group Who Voted for at Least One of Each Type of Immigration Bill, 2007 Independent Variables Expansive Restrictive Senate 30.68 29.32 Republican 43.43 57.02 African American 7.67 7.11 Female 24.11 22.22 (N=3,194) (N=2,336) % % Latino 4.95 3.30 Table 4.2. Percentage of Votes on Each Type of Immigration Bill Cast by Legislators in Each Group Independent Variables Expansive Restrictive Vote For Vote For 104 The variables that proved significant for voting for restrictive immigration legislation differed from those for voting for expansive immigration legislation, suggesting that these policies are not two sides of the same coin. The district characteristic for percentage of Latinos and the personal legislator characteristics of party leadership, majority party membership, gender, and Latino ethnicity achieved standard social science levels of statistical significance. Of these, the coefficients for gender and district percentage Latino were in the opposite direction of that hypothesized, with female state legislators voting for a higher percentage of restrictive immigration legislation than did their male colleagues, a finding that does not hold when only examining chambers in which Latino legislators served, and with those representing districts with a higher percentage of Latinos also more likely to vote for restrictive immigration legislation than their peers who represent districts with a lower percentage of Latinos. Membership in the majority party increased the vote for restrictive immigration bills as hypothesized, and by a substantively significant 14 percentage?15 percentage points in legislative chambers in which Latinos served?over that of minority party members. Most important for this study, Latino legislators were a substantively significant 15 percentage points less likely, as hypothesized, to vote for restrictive immigration policy than were non-Latino Majority Party Member 92.62 66.14 All Legislators 90.38 59.87 Senate 92.21 61.29 Republican 86.47 62.06 African American 93.72 55.87 Female 91.22 60.99 Vote For Vote For % % Latino 90.13 42.90 Table 4.5. Average Percentage of Yes Votes Cast for Each Type of Immigration Policy by Each Legislator Group Independent Variables Expansive Restrictive 105 legislators, or 11 percentage points less when only looking at legislative chambers in which Latino legislators served. Discussion and Conclusions Legislator roll-call voting on immigration bills that expand immigrant rights differs from legislator roll-call voting on immigration bills that restrict immigrant rights. This difference is an important one, and one that is missed by the past research on state immigration legislation that conjoins expansive and restrictive immigration policy into one measure (Monogan 2010). These policies are not simply two sides of the same coin, but rather different coins altogether, with some legislators voting for some bills that restrict and some bills that expand immigrant rights. There are two explanations for this, one based on the legislation, and one on individual legislators. First, this finding may belie that some of these policies are not quite as clearly drawn between expansion and restriction of immigrant rights as their coding typology might indicate. Second, legislative behavior on immigration policy may be a great deal more complex than simply drawn typologies and hypotheses can capture. This complexity may be contingent upon policy area, with legislators holding, perhaps, a moral belief in the right of immigrants to a health care delivery system comparable to that offered U.S. citizens, but also that undocumented immigrants are in the country illegally and should be targeted by state law enforcement. More research into the passage of both types of state immigration policy within various policy realms is needed. The lack of a number of findings of this study when compared to prior literature on state immigration policy is also important. State ideology influences state policy adoption of both restrictive and expansive immigration policy (Chavez and Provine 106 Independent Variables All Chambers Only Chambers w/ Latino Legislators All Chambers Only Chambers w/ Latino Legislators Unemployment Rate 0.00 0.04 0.11 0.09 (0.01) (0.03) (0.11) (0.12) Citizen Ideology -0.00 -0.003 -0.01 -0.0002 (0.00) (0.00) (0.00) (0.01) Legislative Professionalism -0.01*** -0.01*** -0.01 -0.02 (0.00) (0.00) (0.01) (0.01) On U.S. Border -0.05 -0.03 -0.08 -0.02 (0.05) (0.06) (0.13) (0.14) Total Bill Enactments 0.00** -0.00002* -0.00 -0.00001 (0.00) (0.00) (0.00) (0.00) % Latino in Chamber -0.00 -0.001 -0.00 -0.002 (0.00) (0.00) (0.02) (0.02) % Republican Vote 0.00 0.001 -0.00 -0.001 (0.00) (0.00) (0.00) (0.00) % Below Poverty Level 0.00 0.0001 -0.00 -0.01 (0.00) (0.00) (0.00) (0.00) % African-American -0.00 0.001 -0.00 -0.003 (0.00) (0.00) (0.00) (0.00) % Latino 0.00 0.001* 0.01** 0.004 (0.00) (0.00) (0.00) (0.00) Population Change, '00-'06 0.00 0.001 0.00 -0.0001 (0.00) (0.00) (0.00) (0.00) Republican -0.07 -0.11* 0.02 0.04 (0.04) (0.06) (0.04) (0.05) Senator 0.03 0.04 -0.03 0.11 (0.01) (0.02) (0.10) (0.13) Leadership -0.01 -0.03 0.06* 0.04 (0.02) (0.03) (0.03) -0.03 In Chamber Majority Party 0.03 0.05 0.14*** 0.15*** (0.01) (0.04) (0.03) (0.04) Female 0.02* 0.01 0.04* -0.04 (0.01) (0.01) (0.02) (0.02) African American 0.01 -0.02 0.07 0.09 (0.02) (0.03) (0.08) (0.09) Latino 0.01 -0.0001 -0.15** -0.11* (0.05) (0.05) (0.06) (0.06) Constant 0.98*** 1.02*** 0.57 0.61 (0.06) (0.14) (0.41) (0.37) N = 3325 2346 2642 2149 R 2 = 0.1092 0.1391 0.1221 0.1528 Table 4.6. Influences on State Legislators' Voting Yes on Immigration Bill Floor Votes Expansive Legislation Restrictive Legislation Coefficient (Rob. SE) Coefficient (Rob. SE) Notes: All p values are one-tailed except for those for variables representing whether a legislator is African American and the percentage African American of their district due to the contrasting hypotheses offered for these variables. State and Legislature Characteristics District Characteristics Legislator Characteristics 107 2009), but does not impact legislators' roll-call voting on either type of immigration policy. The size of a state's Latino population plays a predictive role in state enactment of both kinds of immigration policy (Sanchez et al. 2012) or of only expansive immigration policy (Chavez and Provine 2009), but the percentage of Latinos in a state legislative district is only influential on roll-call voting on restrictive immigration legislation, and here higher percentages of Latinos increase legislators' percentage vote for restrictive immigration legislation. Critical mass theory does not seem to factor in on the voting decisions of legislators on immigration policy as there was no significance for the percentage of Latino legislators in a state legislature for voting for either expansive or restrictive immigration policy, but this finding does give further evidence that Latino legislators are not the only legislators who substantively legislate in this policy arena that disproportionately impacts the Latino community. Additionally, this study finds no support at the roll-call voting stage for Hopkins' (2010) findings of the effects of rapid population growth on localities' immigration policy outcomes. Legislators from districts with large population increases did not vote differently on immigration legislation than did their colleagues in districts with a lesser increase or a decrease in population over several years. The differences in these findings may seem somewhat surprising at first because, due to the incredibly high percentage of legislative passage of immigration bills that reach the roll-call vote stage, the current study's breadth of data is much more reflective of the range of bills studied through the policy adoption lens than would be examinations of state immigration policy at earlier stages of the policymaking process. Difference between these two populations does exist however, first because many immigration laws 108 were passed in only one legislative chamber or were vetoed by the governor and therefore were not enacted into law, and second because not every bill that passed a legislative chamber received a recorded roll-call vote. Of the 299 total state immigration bills that were passed into law, 256 (85.62%) received recorded roll-call votes. This study's main difference, however, is not in its examination of a slightly larger population of legislation than that studied in the state immigration policy adoption literature, but in its focus on a different level of analysis, individual legislators' actions, rather than the aggregation of those legislators' actions to the state level. In doing so, the current study provides a more robust picture of the state immigration policymaking process by allowing analysis of the effects of personal characteristics on legislator roll-call voting on immigration legislation. This focus yields important new findings on legislators' descriptive and substantive representation of the predominantly Latino immigrant population. State immigration laws are often highly charged pieces of legislation with very real and painful impacts on immigrant populations. They can determine whether undocumented families must uproot from their adopted land and move to less hostile locales, whether families living paycheck to paycheck lose or gain essential state benefits, and whether individuals detained for what had a been a misdemeanor might be deported. One of, if not the largest fear among the undocumented community is of families being separated due to deportation, a fear 15-year old undocumented immigrant Juan Flores notes: "Oklahoma is closing all the doors because we are illegal. Work for my parents brings fear. They will not even let me go outside anymore. They are so scared someone is going to come take me? (Walker 2007). The finding that Latino legislators provide descriptive representation to Latinos through their dissent on restrictive immigration 109 policy is best understood in this context. Latino legislators descriptively represent Latino interests when those interests are most salient, that is, when they are threatened, but not, as the finding of lack of descriptive representation of Latino interests on expansive immigration legislation shows, in a subarena in which those interests are not threatened by rather proactively expanding. Communities under policy threat have been known to stimulate legislative policy response, as have smokers threatened by anti-smoking legislation and cigarette taxes (Green and Gerkin 1989) and gun owners threatened by gun control legislation (Wolpert and Gimpel 1998). This chapter's finding of Latino legislators' descriptive representation of Latino interests does counter the findings of both Rouse (2013) and Casellas (2011), whose studies of roll-call voting behavior found that Latino legislators were no more likely to vote for Latino interests than were non-Latino members. To understand this discrepancy of findings, it is important to note how these previous studies differed methodologically from the current study. The breadth of the issue areas addressed is one important difference between this and previous studies of Latino state legislators' roll-call voting behavior. Casellas (2011) examines all legislation that received a roll-call vote, while Rouse (2013) studies the difference between Latino legislators' roll-call vote on all legislation and on Latino- interest issues, a policy area that includes immigration policy, but also legislation "aimed at improving healthcare services, bills addressing discrimination, more general education policies, and economic improvement initiatives for impoverished areas" (111). The current study focuses solely on state immigration legislation, and its finding of Latino legislators' descriptive representation of Latino interests emerges more specifically in the 110 policy subarena of restrictive immigration legislation. The narrowness of the policy subarena in which this finding occurs, and that it is not paired with a similar finding on expansive immigration policy, may also hold the key to resolving the discrepancy of this finding with that of the extant literature. Work that analyzes Latino descriptive representation by looking at Latino-interest legislation as if it were unidirectional in nature (Rouse 2013) may also miss the nuanced nature of this finding. Latino-interest policies, and particularly state immigration policy, can be bifurcated into policies that benefit and those that are detrimental to Latino interests, and not accounting for this nuance may lead to erroneous results of Latino descriptive representation. The findings from this chapter indicate that Latino descriptive representation does occur, but only when Latino interests are most salient, when Latino interests are threatened. Rouse's (2013) work also assessed a broader range of policy than does this dissertation; it may well be that these findings only hold for immigration policy, and disappear when other Latino-interest areas are examined. The other important methodological distinction between the present study and prior studies lies in the included range of state legislatures. The present study analyzes roll-call voting on restrictive immigration legislation in every state that saw recorded roll- call votes on this type of legislation in 2007, a total of 21 states (in 19 House chambers and in 18 Senate chambers). Rouse (2013) examined the recorded roll-call votes of state legislatures in both chambers of five states chosen both for their high percentage of Latino legislators and their high percentage of Latinos in the general population (Arizona, California, Florida, Texas, and New Mexico). Casellas (2011) restricts his study of Latino roll-call voting behavior in state legislatures to the Texas House and both chambers of the 111 Colorado and New Jersey legislatures. The discrepancy in these findings of Latino legislators' descriptive representation of Latino interests might lie in states with a high percentage of Latinos in the legislature who may not feel the need to offer direct descriptive representation to Latino constituents because, as Rouse notes, those interests are already being addressed in the legislature through the actions of the Democratic Party; however, judging from the insignificant results related to the percentage of Latino legislators in state legislatures and for the party variable for the restrictive immigration legislation model, this does not appear to be the case. Rather, it may be that Latino legislators in states with high percentages of Latino legislators that are not included in Rouse (2013) and Casellas' (2011) studies are descriptively representing Latino interests. The current study's expansive range of data is more representative of state legislators' roll-call voting behavior than are previous studies that have utilized a narrower set of states in different policy realms, meaning that its findings can be more confidently applied to state legislator roll-call voting behavior in general. This study analyzed state legislator roll-call voting in the 33 states that held roll- call votes on expansive or restrictive immigration legislation in 2007. While finding mixed results for female legislators' substantive representation of Latino interests, contrary to previous literature (Rouse 2013; Casellas 2011), it finds both little role for the influence of party on legislators' roll-call voting in the immigration policy realm and that Latino legislators descriptively represent Latino interests through their roll-call votes on restrictive immigration legislation when Latino interests are threatened. The finding of the influence of legislators' personal characteristics is important in our understanding of the representative tie between citizen, legislator, and public policy. In the state 112 immigration policy realm, at least, Latino interests are best served through the election of Latino legislators to state legislatures. The roll-call voting stage is, however, but one of many in the public policymaking process, and critics note that while studying roll-call voting provides a unique stage to capture the direction of legislators' policy preferences, it does not capture its intensity as do other legislative stages. Many scholars (Rouse 2013; Bratton 2002; Bratton and Rouse 2012) have also argued that other legislative policymaking stages are better reflection of legislators' descriptive representation of minority groups. To fully understanding the stories of state immigration policy and Latino representation in state legislatures, in addition to the roll-call voting stage one must also consider the sponsorship and committee stages of the policymaking process. 113 Chapter 5: Conclusions Legislators in all 50 states introduced a total of 1,562 bills related to immigration in 2007, the first year in which an immigration bill was introduced in every state legislature (see Appendix 1), and almost three times the number introduced in 2006. Forty-six states enacted into law a total of 240 immigration-related bills in 2007, more than triple the 84 enacted in 32 states in 2006 (Hegen 2008). U.S. sub-national governments have gone down this legislative road before, most recently in the 1980s and early 1990s during which time they crafted legislation aimed at sending anti-immigrant, often nativist messages to the largely Latino immigrant populations. These state immigration policies have grave real-world effects on the largely Latino undocumented immigrant communities they target; they determine how the estimated 11.6 million undocumented immigrants living in the United States interact with state law enforcement, receive social, health, and educational services, and integrate with the state economy, and can influence undocumented immigrants' decision to migrate to or emigrate from a state (see, for example, PBS 2011). This increase in state immigration policy is occurring in parallel to the marked increases in the percentage of the Latino population overall, its percentage of the voting electorate, and in the number of government positions held by Latinos at all levels of government over the past several decades. At the state level, Latinos have made impressive gains in state legislatures, even if these gains still lag behind the Latino percentage of the population in every state (see Appendix 1). This increase in Latino state legislators, however, does not necessarily translate into representation that has a substantive policy impact. By examining how Latino legislators represent Latino interests through state immigration policy, this dissertation addresses important questions of 114 representation in American democracy. Do ethnic personal characteristics and shared experiences influence how legislators behave within a policy area that disproportionately impacts members of that ethnicity? These questions are important, because without this descriptive representational tie, growth in the number of Latino legislators provides no significant benefit for Latino interests over that of electing non-Latino legislators. Several scholarly works (Bratton 2002; Martinez-Ebers et al. 2000; Rouse 2013) have found that Latino legislators offer descriptive representation to Latino interests?that is, representation based on a common ethnic tie beyond that which can be attributed to constituency and party influences. This dissertation examines state legislator behavior on immigration legislation that either expands or restricts immigrant rights to test the findings of prior literature on state immigration policy adoption and on the descriptive representational roles of Latino legislators. Previous state-level studies of state immigration policy have found state citizen ideology (Creek and Yoder 2012; Monogan 2010; Newman et al. 2012), partisanship (Gulasekaram and Ramakrishnan 2013), and the partisan lean of the legislature or of the governor (see, for example, Sanchez, Sanchez, and Ybarra 2012; Chavez and Provine 2009; Reich and Barth 2010) amongst the most powerful indicators of state immigration policy adoption. In contrast to the findings of the state immigration policy adoption literature, partisanship plays a real, but inconsistent role contextual to the legislative policy stage and the direction of the policy under consideration in determining legislator behavior on immigration policy. Republican legislators tend to use the bill introduction stage to set the policy agenda for restrictive state immigration policy, and are moderately more likely to serve on committees that see fewer expansive immigration 115 bills and that see a larger number of restrictive immigration bills. These findings may well bubble up into the partisanship seen at the state policy adoption stage. State citizen ideology had little impact on state legislator behavior on state immigration policy, but this may be due to the mismatch between the levels of constituency, with each state legislature responsive to their own differing district constituency for re-election, and not the state-level constituency. State district or legislator-level ideology data have been, until recently, nonexistent. Recent work (Shor and McCarty 2013) mapping individual state legislators ideology might well be used in future work to better measure the impact of state legislator ideology on state legislator behavior. In contrast to the findings of the extant literature on descriptive representation, this dissertation finds that Latino legislators' representation of Latino interests is conditional on whether those interests are directly threatened. Whereas Latino legislative sponsorship behavior on state immigration policy of either type does not differ from that of non-Latino legislators, a finding that goes against that of prior literature (Bratton 2002; Martinez-Ebers et al. 2000; Rouse 2013) that found significant Latino descriptive representation of Latino interests at the agenda-setting stage of the legislative policymaking process, their behavior in not voting for such measures on chamber floors speaks to their descriptive representational role. Latino legislators here mirror the ?second face of power? (Bachrach and Baratz 1962) in their descriptive representation of Latino issues by working to stop restrictive immigration policy from being enacted. The differences between this dissertation's findings and that of prior literature may well flow from the differences between the scope of the data they examine. This dissertation is the first analysis of state legislator behavior on state immigration policy 116 and is therefore different from the previous work in this policy realm that has focused on policy adoption at the state level. In terms of its findings on Latino descriptive representation, two major differences emerge. First, this study analyzes a newly created database of state legislator behavior across the legislative policymaking process in 49 states, whereas prior studies focused their attention on only a handful of states selected for their high percentages of Latino legislators, Latino residents, or both. This dissertation's expansive range of data?immigration policies across 49 states?is more representative of state legislators than are previous studies that have utilized a narrower set of states in different policy realms (see Appendix 2), meaning that its findings can be more confidently applied to state legislator sponsorship behavior in general. This study is the first to analyze state legislator sponsorship behavior among all state legislators. Second, this dissertation's focus on state immigration policy differs from prior studies' analysis of a broader range of policy arenas that impact Latinos. The breadth of the issue areas addressed is one important difference between this and previous studies of Latino state legislators' behavior. Rouse (2013) studies the difference between Latino legislators' behavior on all legislation and on legislation related to Latino-interest issues, a policy area that includes immigration policy, but also legislation "aimed at improving healthcare services, bills addressing discrimination, more general education policies, and economic improvement initiatives for impoverished areas" (111). Rouse?s (2013) excellent research design wisely compares state legislator?s representational behavior on those issues thought to be the most salient to the Latino community?Latino-interest issues?with all other legislation. However, though this research design was a major step forward over that which examined Latino legislative behavior only on all legislation 117 (Casellas 2011), it did not examine the potential directional nuances of Latino-interest legislation. In contrast, this dissertation focuses solely on state immigration legislation. This dissertation's comparative narrowness of focus provides considerable payoff in that the standard typology for classifying state immigration policy accounts for its dichotomous nature: policy that either expands or restricts immigrant rights. The bi- directional nature of this policy area allows for greater nuance in examining the descriptive representation of Latino legislators, with descriptive representation showing when Latino legislators take actions that further policy expanding immigrant rights or do not take or block policy that restricts immigrant rights. Previous studies therefore may well overstate Latino descriptive representation on Latino-interest legislation because they do not take into account that some legislation in this arena may be beneficial while other legislation may be inimical to Latino interests. It is also quite possible that these findings are salient for immigration policy, and that by expanding the range of policy that previous scholars have classified as Latino-interest that these findings disappear. This dissertation finds that Latino legislators substantively represent Latino interests on immigration policy only when those interests are most threatened. Communities under policy threat have been known to stimulate legislative policy response, as have smokers threatened by anti-smoking legislation and cigarette taxes (Green and Gerkin 1989) and gun owners threatened by gun control legislation (Wolpert and Gimpel 1998). Issue salience has generally been used to describe the importance that the public pays to an issue-area (Wlezien 2005), and about which ?attentive publics,? those most affected by the law, would be concerned (Arnold 1992). This dissertation's findings of Latino descriptive representation only to block state immigration policy that 118 restricts immigrant rights suggests that these attentive publics form amongst Latino constituents in opposition to passage of such bills, and that they pressure Latino legislators to kill such legislation. Lack of a finding for Latino descriptive representation for state immigration policy that expands immigrant rights suggests that these communities do not form to support?and therefore legislators do not feel pressure to fight for?the expansion of rights for the largely Latino immigrant community. This finding follows that of relatively recent research examining into the differing motivational attributes of threat and of opportunity, which finds threat the more likely to spur political action (Miller and Krosnick 2004). The finding of the importance of issue salience to legislator descriptive representation may well extend beyond representation of the Latino community to a broader discussion of what stokes legislator action on the behalf of an ethnic, racial, or underprivileged constituency, and future work on descriptive representation would do well to parse out the intended legislative impact on the studied ethnic or racial group to determine whether legislators? descriptive representation is brought about through legislative threat. It may be that this is the only policy stimulus that earns a legislator?s response. Though issue salience has largely been examined in the context of the public?s opinion (Wlezien 2005), the representational tie between legislators and their constituency?s attentive publics makes the argument for consideration of issue salience in future models of legislator behavior. This dissertation's findings indicate that future work should pay attention to the potential multi-directionality of purpose in certain policy areas. There may be further directional nuance to policies within policy arenas that can yield greater political insight that would be obtained from examination of these policies as if they were a unidirectional 119 mass. Studying policy areas where there is a clear bi-directionality in purpose allows the researcher to more accurately test theories of representation. This dissertation also calls to attention the potential advantages of assessing state legislator behavior across a larger collection of states than that assessed by prior research in this field. Latino state legislator behavior in largely Latino states, where there are likely other Latino legislators to create a legislative mass capable of moving policy, may well differ from Latino legislators' actions when they have no such network of support. Preliminary analysis of the differences in legislator behavior in only those states examined by prior scholarship indicates that state legislator sponsorship behavior in those states differs from that of the full population of state legislators analyzed in this dissertation (Table 5.1). Analyzing a sample of legislators from fewer states chosen for their large Latino population and the high percentage of state legislators who are Latino might have led to the conclusion that Latino legislators provide descriptive representation to Latino interests on immigration policy that expands immigrant rights, a finding that disappears when analyzing the data used in this dissertation, which is the full population of state legislators. Recent work (for example, Shor and McCarty 2013) has started to improve the quality and quantity of data that we have on state legislatures and legislators across all 50 states. Future work will do well to take advantage of these improving resources to continue this trend toward the inclusion of all state legislators in studies of state legislative behavior. Such work can enhance the theoretical insights advanced by earlier studies by using data from which we can draw insights that are generalizable across all state legislators. Understanding how legislators behave on state immigration policy is important. 120 The states are offering up a variety of policy solutions in their struggle to cope with increasing numbers of largely poor, unskilled immigrants. State-, district-, and personal- level factors impact state legislators? behavior on state immigration policy. While no one factor uniformly influences state legislators? behavior on the two types of state immigration policy examined here, this dissertation finds that while important, legislators? partisanship and states? citizens? ideology are not as universally important in the formation of state immigration policy as the state immigration policy adoption literature finds. Finally, it offers a more nuanced look at Latino descriptive representation, finding that Latino legislators only offer descriptive representation on state immigration policy that restricts immigrant rights. The past several decades have seen marked increases in the percentage of the Latino population overall, its percentage of the voting electorate, and in the number of government positions held by Latinos at all levels of government. This increase in Latino state legislators, however, does not necessarily translate into representation that has a substantive policy impact. In terms of legislation that would expand the rights of the largely Latino immigrant community, Latino legislators offer no significant benefit for Latino interests over that of electing non-Latino legislators. It is only in protecting the already provided rights of this community that Latino legislators make a small difference. Understanding how these influences impact state legislator behavior gives us a more informed picture of how and why representation is provided in the U.S. political system. 121 States Analyzed in Prior Research Yoder (ND) 1 Bratton and Haynie (1999) 2 Barnello and Bratton (2007) 3 Rouse (2013) 4 Yoder (ND) 1 Bratton and Haynie (1999) 2 Barnello and Bratton (2007) 3 Rouse (2013) 4 Independent Variables Unemployment Rate -0.63** -4.75*** -1.06*** 2.80*** 0.21 -11.73*** -0.42* -6.30*** (0.29) (0.16) (0.43) (0.00) (0.20) (1.53) (0.23) (0.55) Citizen Ideology 0.02* -0.06*** 0.04 -0.03*** 0.01 -0.31*** 0.02 -0.10*** (0.01) (0.01) (0.03) (0.01) (0.01) (0.05) (0.03) (0.01) Legislative Professionalism 0.03** -0.09*** 0.04*** -0.05** -0.03 -0.36*** -0.02 0.16*** (0.02) (0.02) (0.02) (0.02) (0.02) (0.12) (0.02) (0.02) On the US Border -0.70** 8.06*** 0.16 0.75*** -0.60 26.16*** 0.93** 1.41*** (0.38) (0.56) (0.51) (0.25) (0.37) (5.51) (0.52) (0.27) Total Bill Introductions 0.00 0.00*** 0.00* 0.00 0.00*** 0.00*** 0.00 0.00*** (0.00) (0.00) (0.00) (0.00) (0.00) '(0.00) (0.00) (0.00) % Republican Vote -0.00 -0.01* -0.00 0.00 -0.01** -0.01 -0.00 0.00 (0.00) (0.00) (0.00) (0.01) (0.00) (0.01) (0.01) (0.01) % Below Poverty Level -0.02 -0.00 -0.04* -0.06** -0.03* -0.02 -0.01 0.02* (0.02) (0.01) (0.02) (0.03) (0.02) (0.03) (0.02) (0.02) % African-American 0.01 -0.01*** 0.01 0.02 0.00 -0.02 0.01 0.01 (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.02) % Latino 0.03*** 0.02 0.00 0.02* 0.01 -0.11* -0.04 -1.25*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.08) (0.03) (0.41) % Population Change, '00-'06 0.01 -0.04*** -0.01 -0.03** 0.02** -0.01 0.01 0.02 (0.01) (0.01) (0.01) (0.01) (0.01) (0.03) (0.01) (0.02) Republican -0.27 0.58*** 0.06 -0.19 1.45*** 1.66*** 1.20*** 1.05** (0.25) (0.17) (0.22) (0.28) (0.21) (0.52) (0.30) (0.52) Senator 0.27** -0.01 -0.06 -0.04 0.22 -0.09 0.15 0.66*** (0.17) (0.21) (0.15) (0.32) (0.16) (0.12) (0.15) (0.18) Leadership 0.15 0.38 0.30 0.61 -0.05 -0.65 -0.15 0.76*** (0.15) (0.32) (0.28) (0.44) (0.27) (0.46) (0.33) (0.29) Female 0.34*** 0.18 0.16 0.03 -0.07 -0.30 -0.03 0.44 (0.12) (0.21) (0.17) (0.29) (0.20) (0.28) (0.23) (0.30) African American -0.24 -0.12 -0.32 0.38 -0.69 -1.18*** -1.00 0.01 (0.28) (0.18) (0.28) (0.43) (0.53) (0.09) (0.61) (1.13) Latino 0.26 0.53* 0.68** 0.61* -0.48 0.17 -0.37 -1.25*** (0.44) (0.34) (0.35) (0.44) (0.67) (0.32) (0.60) -0.41 N = 7232 917 2141 793 7232 917 2141 793 6 Models are estimated controlling for clustering of standard errors by state. State and Legislature Characteristics District Characteristics Legislator Characteristics Expansive Immigration Legislation Restrictive Immigration Legislation Coef. (Robust Std. Error) 6 Table 5.1. Negative Binomial Predictions for Influence of Legislators' State, District, and Personal Characteristics on Introduction of Expansive and Restrictive Immigration Legislation, 2007, by States used in Prior Research 5 Notes: All p-values are one tailed, except for African American and District % African American, which are two tailed due to the contrasting hypotheses offered, and Chamber, Leadership, and the Constant term, for which no directional hypotheses were put forth. * ? .1; **? .05; ***? .01 1 States used in Yoder's (ND) analysis: All states and legislative chambers except for NE, which was excluded due to its nonpartisan legislators. 2 States used in Barnello and Bratton's (1999) analysis: AR, CA, IL, MD, NC, NJ 3 States used in Barnello and Bratton's (2007) analysis: AZ, AR, CA, CO, FL, IL, LA, MD, MI, MS, NJ, NM, NC, SC, TX 4 States used in Rouse's (2013) analysis: AZ, CA, CO, IL, NM, TX 5 Due to the limited number of cases, full models could not be estimated on the states used in Bratton's (2005) [CA, IL, MD] and (2006) [CA, FL, IL, TX] analyses. Coef. (Robust Std. Error) 6 122 Appendices Appendix A. Percent Latino Legislators, Introduced, and Passed Legislation by State and Legislative Typology, 2007 State Expanding Contracting Introduced Passed Introduced Passed Alabama 3.9 0.0 0.0 0.0 1 0 5 0 Alaska 5.5 0.0 0.0 0.0 0 0 1 0 Arizona 29.6 18.9 18.3 20.0 3 0 4 0 Arkansas 6.4 0.0 0.0 0.0 1 0 0 0 California 37.6 22.5 20.5 25.0 18 4 5 0 Colorado 20.7 5.0 4.6 5.7 1 1 5 3 Connecticut 13.4 3.2 4.0 0.0 8 1 5 0 Delaware 8.2 1.6 2.4 0.0 1 1 2 1 Florida 22.5 10.6 11.7 7.5 7 1 3 1 Georgia 8.8 1.3 1.7 0.0 0 0 3 1 Hawaii 8.9 1.3 0.0 4.0 15 7 0 0 Idaho 11.2 1.0 1.4 0.0 2 0 1 1 Illinois 15.8 6.0 6.8 5.1 4 2 1 0 Indiana 6 0.7 1.0 0.0 2 1 5 2 Iowa 5 0.0 0.0 0.0 6 0 2 0 Kansas 10.5 2.4 3.2 0.0 0 0 5 3 Kentucky 3.1 0.0 0.0 0.0 8 1 1 0 Louisiana 4.2 0.0 0.0 0.0 1 1 1 1 Maine 1.3 0.0 0.0 0.0 0 0 2 1 Maryland 8.2 2.1 2.1 2.1 8 4 3 0 Massachusetts 9.6 2.0 2.5 0.0 14 0 1 0 Michigan 4.4 2.0 1.8 2.6 0 0 17 1 Minnesota 4.7 1.5 1.5 1.5 17 3 4 0 Mississippi 2.7 0.0 0.0 0.0 1 0 7 0 Missouri 3.5 0.5 0.6 0.0 4 0 12 0 Montana 2.9 0.7 1.0 0.0 2 1 7 1 Nebraska 9.2 2.0 N/A 2.0 1 0 1 1 Nevada 26.5 4.8 4.8 4.8 5 3 4 2 New Hampshire 2.8 0.5 0.5 0.0 1 1 1 0 New Jersey 17.7 4.9 7.5 0.0 6 0 0 0 New Mexico 46.3 39.3 42.9 33.3 5 1 3 0 New York 17.6 7.6 8.7 6.5 61 3 17 0 North Carolina 8.4 1.2 0.8 2.0 10 2 15 1 North Dakota 2 0.0 0.0 0.0 2 1 0 0 Ohio 3.1 0.0 0.0 0.0 4 0 1 0 Oklahoma 8.9 0.0 0.0 0.0 6 0 10 2 Oregon 11.7 1.1 1.7 0.0 8 2 3 0 Pennsylvania 5.7 0.4 0.5 0.0 2 0 4 0 12.4 2.7 2.7 2.6 9 2 8 0 South Carolina 5.1 0.6 0.8 0.0 7 0 13 1 South Dakota 2.7 0.0 0.0 0.0 2 1 1 0 Tennessee 4.6 0.8 1.0 0.0 9 1 36 0 Texas 37.6 18.9 20.7 16.1 30 7 24 3 Utah 13 1.9 1.3 3.4 1 1 1 0 Vermont 1.5 0.0 0.0 0.0 1 0 2 1 Virginia 7.9 0.7 1.0 0.0 12 5 10 0 Washington 11.2 2.0 2.0 2.0 4 1 5 1 West Virginia 1.2 0.0 0.0 0.0 2 1 1 0 Wisconsin 5.9 0.8 1.0 0.0 1 0 0 0 Wyoming 8.9 2.2 3.3 0.0 0 0 1 0 Total 16.3 3.2 3.8 2.9 313 60 263 28 % Latino by State % Latino State Legislators % Latino Legislators Lower Chamber % Latino Legislators Upper Chamber Rhode Island Source: Laglagaron et al. (2008), U.S. Census, and author's coding. 123 Variables Mean Std. Dev. Min. Max. Mean Std. Dev. Min. Max. 0.10 0.43 0 7 0.11 0.45 0 7 0.07 0.33 0 9 0.07 0.36 0 6 0.07 0.26 0 1 0.03 0.16 0 1 0.05 0.23 0 1 0.01 0.09 0 1 4.36 0.87 2.60 7.10 4.30 0.84 2.6 7.1 57.34 15.55 22.73 88.17 58.31 15.06 23.52 88.17 18.40 12.01 2.70 62.60 20.03 12.86 2.7 62.6 0.34 0.47 0.00 1.00 0.39 0.49 0 1 3171.39 2893.66 486.00 16131.00 3762.83 3348.59 486 16131 45.97 30.06 0.00 100.00 45.86 30.81 0 100 15.27 7.30 1.50 52.10 14.11 7.16 1.5 87.3 10.91 17.02 0.00 92.20 10.85 16.47 0.0 92.2 8.71 13.46 0.00 98.00 11.59 15.80 0.3 98.0 5.23 9.95 -85.00 157.00 6.52 9.40 -17 157 Republican 0.46 0.50 0.00 1.00 0.45 0.50 0 1 Senator 0.26 0.44 0.00 1.00 0.16 0.37 0 1 Leadership 0.05 0.22 0.00 1.00 0.05 0.21 0 1 Female 0.24 0.42 0.00 1.00 0.25 0.43 0 1 African American 0.08 0.28 0.00 1.00 0.08 0.29 0 1 Latino 0.03 0.18 0.00 1.00 0.05 0.22 0 1 N 7232 4623 1=Senator; 0=House member. 1=leadership; 0=not. 1=female; 0=male. 1=African American; 0=not. 1=Hispanic; 0=not. (NALEO 2007). Coding Info Notes: All variables coded by author unless otherwise noted. * (Lilley et al. 2007). Count. # of expansive bills a legislator sponsored. Count. # of restrictive bills a legislator sponsored. 1=sponsored and was enacted; 0=did not sponsor or was not enacted District % Latino District Population Change, 2000- 2006 District % Hispanic, 2006.* District % pop. change, 2000-2006.* 1=Republican; 0=not Republican. District % Republican Vote, Most Recent Election District % Below Poverty Level District % African-American % Republican vote in district in latest state legislative election for that seat. See note 9.* District % below poverty, 2006.* District % African American, 2006.* State Legislative Professionalism Border State Total Bill Introductions Each state's legislative professionalism (Squire 2007). 1=on US border; 0=not. # of bills and resolutions introduced in regular or special session in 2007 in each state (Book of States). Independent Variables State Unemployment Rate State Citizen Ideology Each state's unemployment rate. U.S. BLS. NOMINATE state citizen ideology. (Berry et al. 2012). # Introduced Contracting Legislation Introduced Enacted Expanding Legislation Introduced Enacted Contracting Legislation 1=sponsored and was enacted; 0=did not sponsor or was not enacted Legislators in All State Legislative Chambers Legislators in Chambers w/ Latino Legislators Dependent Variables # Introduced Expanding Legislation Appendix B. Descriptive Statistics and Coding Information for Legislator Introduction of State Immigration Bills, 2007 124 Article Summary Method States Examined Bratton and Haynie (1999) Negative binomial regression AR, CA, IL, MD, NC, NJ Bratton (2005) Negative binomial regression CA, IL, MD Bratton (2006) Negative binomial regression CA, FL, IL, TX Barnello and Bratton (2007) Negative binomial regression AZ, AR, CA, CO, FL, IL, LA, MD, MI, MS, NJ, NM, NC, SC, TX Rouse (2013), Chapter 3 Negative binomial regression AZ, CA, CO, IL, NM, TX Appendix C. List of Prominent Articles on State Legislative Sponsorship Investigates the agenda-setting behavior of female and black state legislators, contrasting their success with that of white men in passing legislation on ?black interest? and ?women's interest? measures. Uses a six-state, three-year sample, to test a descriptive representation model in which group members (blacks and women) represent group interests above and beyond the extent motivated by constituency and party pressures. Find that blacks and women share a set of distinctive policy interests. Analyzes data from three state legislatures in four years, finding that even in extremely skewed state legislatures, women are generally more active than men in sponsoring legislation that focuses on women?s interests and that women are generally as successful as men in passing the legislation that they sponsor. Finds that Latino representatives in three of the four states studied sponsor more measures focusing on Latino interests, and that Latino representatives are more likely to serve on committees that are relatively likely to meet those interests. Also finds that Latino representatives are as successful as other legislators in passing the bills that they introduce. Explores the conditions under which men are likely to sponsor legislation in policy areas involving women?s issues and children?s issues in the upper and lower chambers of 15 state legislatures in 2001. Differences in sponsorship are relatively marked in the sponsorship of legislation that focuses on reproduction or other health issues particularly relevant to women. Focuses on ethnic differences in agenda setting; specifically, the factors that influence the sponsorship activity of Latino legislators. Examines the effects of institutional characteristics (specifically, the percentage of Latinos in the chamber) on the link between descriptive and substantive representation by examining the total number of Latino interest bills introduced in both chambers of six state legislatures for three years in four policy areas (specific Latino interests, education, health, and welfare policy). 125 References Alba, Richard, Rub?n G. Rumbaut, and Karen Marotz. 2005. ?A Distorted Nation: Perceptions of Racial/Ethnic Group Sizes and Attitudes toward Immigrants and Other Minorities.? Social Forces 84(2): 901?919. Aldrich, John H. 1995. Why Parties? The Origin and Transformation of Political Parties in America. Chicago: University of Chicago Press. Arnold, R. Douglas. 1992. The Logic of Congressional Action. New Haven: Yale University Press. Bachrach, Peter, and Morton S. Baratz. 1962. ?Two Faces of Power.? American Political Science Review 56 (4): 947?952. Bacon, David. 2012. ?How Mississippi?s Black/Brown Strategy Beat the South?s Anti- Immigrant Wave.? The Nation. http://www.thenation.com/article/167465/how- mississippis-blackbrown-strategy-beat-souths-anti-immigrant-wave# (May 11, 2013). Barnello, Michelle A., and Kathleen A. Bratton. 2007. ?Bridging the Gender Gap In Bill Sponsorship.? Legislative Studies Quarterly 32(3): 449?474. Barnello, Michelle A. 2001. ?Gender and Roll Call Voting in the New York State Assembly.? Women & Politics 20(4): 77?94. Barrett, Edith J. 1995. ?The Policy Priorities of African American Women in State Legislatures.? Legislative Studies Quarterly 20(2): 223?247. Berry, William D. et al. 1998. ?Measuring Citizen and Government Ideology in the American States, 1960?93.? American Journal of Political Science 42(1): 327?348. Berry, William D. et al. 2010. ?Measuring Citizen and Government Ideology in the U.S. States: A Re-appraisal.? State Politics & Policy Quarterly 10(2): 117?135. Bertelli, Anthony, and Lilliard E. Richardson. 2008. ?Ideological Extremism and Electoral Design: Multimember versus Single Member Districts.? Public Choice 137: 347?368. Blalock, Hubert M. 1967. Toward a Theory of Minority Group Relations. New York: Capricorn. Bobo, Lawrence D. 1999. ?Prejudice as Group Position: Microfoundations of a Sociological Approach to Racism and Race Relations.? Journal of Social Issues 55(3): 445?472. Boushey, Graeme, and Adam Luedtke. 2011. ?Immigrants Across the U.S. Federal Laboratory: Explaining State-Level Innovation in Immigration Policy.? State Politics & 126 Policy Quarterly 11(4): 390?414. Bratton, Kathleen A. 2002. ?The Effect of Legislative Diversity on Agenda-Setting: Evidence from Six State Legislatures.? American Politics Research 30(2): 115?142. Bratton, Kathleen A. 2006. ?The Behavior and Success of Latino Legislators: Evidence from the States.? Social Science Quarterly 87(s1): 1136?1157. Bratton, Kathleen A., and Kerry L. Haynie. 1999. ?Agenda Setting and Legislative Success in State Legislatures: The Effects of Gender and Race.? Journal of Politics 61(3): 658?679. Bratton, Kathleen A, and Stella M Rouse. 2011. ?Networks in the Legislative Arena: How Group Dynamics Affect Cosponsorship.? Legislative Studies Quarterly 36(3): 423?460. Burns, Peter, and James G. Gimpel. 2000. ?Economic Insecurity, Prejudicial Stereotypes, and Public Opinion on Immigration Policy.? Political Science Quarterly 115(2): 201? 225. Burstein, Paul, Shawn Bauldry, and Paul Froese. 2005. ?Bill Sponsorship and Congressional Support for Policy Proposals, from Introduction to Enactment or Disappearance.? Political Research Quarterly 58(2): 295?302. Calavita, Kitty. 1996. ?The New Politics of Immigration: ?Balanced-Budget Conservatism? and the Symbolism of Proposition 187.? Social Problems 43(3): 284?305. Cain, Bruce E., and D. Roderick Kiewiet. 1987. ?Latinos and the 1984 Election: A Comparative Perspective.? In Ignored Voices: Public Opinion Polls and the Latino Community, ed. Rodolfo O De la Garza. Austin, TX: Center for Mexican American Studies. Canes-Wrone, Brandice, David W. Brady, and John F. Cogan. 2002. ?Out of Step, Out of Office: Electoral Accountability and House Members? Voting.? American Political Science Review 96(1): 127?140. Carroll, Joseph. 2007a. About One in Three Hispanics Say Immigration is Nation?s Top Problem. Gallup News Service. http://www.gallup.com/poll/28165/about-one-three- hispanics-say-immigration-nations-top-problem.aspx. Carroll, Joseph. 2007b. Hispanics? Views on Immigration Unchanged from Last Year. Gallup News Service. http://www.gallup.com/poll/28003/hispanics-views-immigration- unchanged-from-last-year.aspx. Carson, Jamie L., Gregory Koger, Matthew J. Lebo, and Everett Young. 2010. ?The Electoral Costs of Party Loyalty in Congress.? American Journal of Political Science 54(3): 598?616. 127 Casellas, Jason Paul. 2010. Latino Representation in State Houses and Congress. Cambridge: Cambridge University Press. Chandler, Charles R., and Yung-mei Tsai. 2001. ?Social Factors Influencing Immigration Attitudes: An Analysis of Data from the General Social Survey.? Social Science Journal 38(2): 177?188. Chavez, J. M., and D. M. Provine. 2009. ?Race and the Response of State Legislatures to Unauthorized Immigrants.? ANNALS of the American Academy of Political and Social Science 623(1): 78?92. Childs, Sarah, and Mona Lena Krook. 2008. ?Critical Mass Theory and Women?s Political Representation.? Political Studies 56: 725?736. Citrin, Jack, Donald Green, Christopher Muste, and Cara Wong. 1997. ?Public Opinion toward Immigration Reform: The Role of Economic Motivations.? Journal of Politics 59(3): 858?881. Citrin, Jack, Beth Reingold, and Donald Green. 1990. ?American Identity and the Politics of Ethnic Change.? Journal of Politics 52(4): 1124?1154. Cox, Gary W., and Mathew D. McCubbins. 2005. Setting the Agenda: Responsible Party Government in the U.S. House of Representatives. Cambridge: Cambridge University Press. Creek, Heather M., and Stephen Yoder. 2012. ?With a Little Help from Our Feds: Understanding State/Federal Cooperation on Immigration Enforcement.? Policy Studies Journal 40(4): 674?697. Dahlerup, Drude. 1988. ?From a Small to a Large Minority: Women in Scandinavian Politics.? Scandinavian Political Studies 11(4): 275?298. De la Garza, Rodolfo O, and Louis DeSipio. 1993. ?Save the Baby, Change the Bathwater, and Scrub the Tub: Latino Electoral Participation after Seventeen Years of Voting Rights Act Coverage.? Texas Law Review 71(7): 1479?1540. Dolan, Kathleen, and Lynne E. Ford. 1997. ?Change and Continuity among Women State Legislators: Evidence from Three Decades.? Political Research Quarterly 50(1): 137? 151. Ellis, William Curtis, and Walter Clark Wilson. 2013, forthcoming. ?Minority Chairs and Congressional Attention to Minority Issues: The Effect of Descriptive Representation in Positions of Institutional Power.? Social Science Quarterly. Erikson, Robert S., Gerald C. Wright, and John P. McIver. 1994. Statehouse Democracy: 128 Public Opinion and the American States. Cambridge: Cambridge University Press. Espenshade, Thomas J., and Charles A. Calhoun. 1990. ?An Analysis of Public Opinion toward Undocumented Immigration.? Population Research and Policy Review 13 (February):189?224. Esses, Victoria M et al. 2001. ?The Immigration Dilemma: The Role of Perceived Group Competition, Ethnic Prejudice, and National Identity.? Journal of Social Issues 57(3): 389?412. Estrada, Ismael, and Keith Oppenheim. 2007. ?Oklahoma targets illegal immigrants with tough new law.? CNN. http://articles.cnn.com/2007-11- 02/us/oklahoma.immigration_1_illegal-immigrants-immigration-law-immigration- status?_s=PM:US (Accessed January 6, 2013). Evans, C. Lawrence. 1999. ?Legislative Structure: Rules, Precedents, and Jurisdictions.? Legislative Studies Quarterly 24(4): 605?642. Fenno, Richard F. 1973. Congressmen in Committees. Berkeley: Institute of Government Studies Press. Fix, Michael E., and Jeffrey S. Passel. 1994. Immigration and Immigrants: Setting the Record Straight. Washington, D.C.: The Urban Institute. Fix, Michael E. et al. 2008. Los Angeles on the Leading Edge: Immigration Integration Indicators and their Policy Implications. Washington, D.C.: Migration Policy Institute. Frey, W. H. 2006. Diversity Spreads Out: Metropolitan Shifts in Hispanic, Asian, and Black Populations Since 2000. Washington, D.C.: Center on Urban Metropolitan Policy, Brookings Institution. Gamble, Katrina L. 2007. ?Black Political Representation: An Examination of Legislative Activity within U.S. House Committees.? Legislative Studies Quarterly 32(3): 421?447. Gerber, Elisabeth R, and Jeffrey B. Lewis. 2004. ?Beyond the Median: Voter Preferences, District Heterogeneity, and Political Representation.? Journal of Political Economy 112: 1364?1383. Gimpel, James G., and James R. Edwards. 1998. The Congressional Politics of Immigration Reform. Boston: Longman. Green, Donald Philip, and Ann Elizabeth Gerken. 1989. ?Self-Interest and Public Opinion toward Smoking Restrictions and Cigarette Taxes.? Public Opinion Quarterly 53(1): 1?16. Griffin, J. D., and M. Keane. 2009. ?Are African Americans Effectively Represented in 129 Congress?? Political Research Quarterly 64(1): 145?156. Grose, Christian R. 2005. ?Disentangling Constituency and Legislator Effects in Legislative Representation: Black Legislators or Black Districts?? Social Science Quarterly 86(2): 427?443. Gulasekaram, Pratheepan, and S. Karthick Ramakrishnan. 2013, forthcoming. ?Immigration Federalism: A Reappraisal.? New York University Law Review. Hall, Richard L. 1998. Participation in Congress. New Haven: Yale University Press. Hanmer, Michael J., and Kerem Ozan Kalkan. 2013. ?Behind the Curve: Clarifying the Best Approach to Calculating Predicted Probabilities and Marginal Effects from Limited Dependent Variable Models.? American Journal of Political Science 57 (1): 263?277. Haynie, Kerry L. 2001. African American Legislators in the American States. New York: Columbia University Press. Hedlund, Ronald D. 1984. ?Organizational Attributes of Legislatures: Structure, Rules, Norms, Resources.? Legislative Studies Quarterly 9(1): 51?121. Hedlund, Ronald D., and Patricia K. Freeman. 1981. ?A Strategy for Measuring the Performance of Legislatures in Processing Decisions.? Legislative Studies Quarterly 6(1): 87?113. Hedlund, Ronald D., and Samuel C. Patterson. 1992. ?The Electoral Antecedents of State Legislative Committee Assignments.? Legislative Studies Quarterly 17(4): 539?559. Hedlund, Ronald D, and Keith E Hamm. 1996. ?Political Parties as Vehicles for Organizing U.S. State Legislative Committees.? Legislative Studies Quarterly 21(3): 383?408. Hegen, Dirk. 2008. 2007 Enacted State Legislation Related to Immigrants and Immigration. Washington, D.C.: National Conference of State Legislatures. Hero, Rodney E., and Robert R. Preuhs. 2007. ?Immigration and the Evolving American Welfare State: Examining Policies in the U.S. States.? American Journal of Political Science 51(3): 498?517. Hero, Rodney E., and Caroline J. Tolbert. 1995. ?Latinos and Substantive Representation in the U.S. House of Representatives: Direct, Indirect, or Nonexistent?? American Journal of Political Science 39(3): 640?652. Highton, Benjamin, and Michael S. Rocca. 2005. ?Beyond the Roll-Call Arena: The Determinants of Position Taking in Congress.? Political Research Quarterly 58(2): 303? 316. 130 Hoefer, Michael, Nancy Rytina, and Bryan C. Baker. 2009. Population Estimates: Estimates of the Unauthorized Immigrant Population Residing in the United States: January 2008. Washington, D.C.: Department of Homeland Security, Office of Immigration Statistics. http://ipv6.dhs.gov/xlibrary/assets/statistics/publications/ILL_PE_2005.pdf (Accessed October 28, 2012). Hogan, Robert E. 2008. ?Sex and the Statehouse: The Effects of Gender on Legislative Roll-Call Voting.? Social Science Quarterly 89(4): 955?968. Hood III, M.V., Irwin L. Morris, and Kurt A. Shirkey. 1997. ??!Quedate o Vente!?: Uncovering the Determinants of Hispanic Public Opinion toward Immigration.? Political Research Quarterly 50(3): 627?647. Hood III, M.V., and Irwin L Morris. 2000. ?Brother, Can You Spare a Dime? Racial/Ethnic Context and the Anglo Vote on Proposition 187.? Social Science Quarterly 81(1): 194?206. Hopkins, Daniel J. 2010. ?Politicized Places: Explaining Where and When Immigrants Provoke Local Opposition.? American Political Science Review 104(1): 40?60. Jackson, John E., and John W. Kingdon. 1992. ?Ideology, Interest Group Scores, and Legislative Votes.? American Journal of Political Science 36: 805?823. Jenkins, S. 2002. ?Party Voting in State Legislatures.? Spectrum: Journal of State Government 75(4): 10?13. Jenkins, Shannon. 2006. ?The Impact of Party and Ideology on Roll-Call Voting in State Legislatures.? Legislative Studies Quarterly 31(2): 235?257. Jenkins, Shannon. 2008. ?Party Influence on Roll Call Voting: A View from the U.S. States.? State Politics & Policy Quarterly 8(3): 239?262. Kanter, Rosabeth Moss. 1977. ?Some Effects of Proportions on Group Life: Skewed Sex Ratios and Responses to Token Women.? American Journal of Sociology 82(5): 965?990. Kathlene, Lyn. 1994. ?Power and Influence in State Legislative Policymaking: The Interaction of Gender and Position in Committee Hearing Debates.? American Political Science Review 88(3): 560?576. Kau, James B., Donald Keenan, and Paul H. Rubin. 1982. ?A General Equilibrium Model of Congressional Voting.? Quarterly Journal of Economics 93: 271?93. Kaufmann, Karen M. 2003. ?Cracks in the Rainbow: Group Commonality as a Basis for Latino and African-American Political Coalitions.? Political Research Quarterly 56(2): 131 199?210. Kerr, Brinck, and Will Miller. 1997. ?Latino Representation, It?s Direct and Indirect.? American Journal of Political Science 41(3): 1066?1071. Kingdon, John W. 1989. Congressmen?s Voting Decisions. Third Edition. Ann Arbor: University of Michigan Press. Koger, Gregory. 2003. ?Position Taking and Cosponsorship in the U.S. House.? Legislative Studies Quarterly 28(2): 225?246. Kousser, Thad. 2004. Term Limits and the Dismantling of State Legislative Professionalism. Boston: Cambridge University Press. Kousser, Thad, Jeffrey B. Lewis, and Seth Masket. 2007. ?Ideological Adaptation? The Survival Instinct of Threatened Legislators.? Journal of Politics 69: 828?843. Krane, Dale. 2007. ?The Middle Tier in American Federalism: State Government Policy Activism During the Bush Presidency.? Publius: The Journal of Federalism 37(3): 453? 477. Krehbiel, Keith. 1993. ?Where?s the Party?? British Journal of Political Science 23(2): 235?266. Kromer, M.K. 2008. ?Legislative Participation in the American States: Determinants and Consequences.? Ph.D. Dissertation. Louisiana State University. http://etd.lsu.edu/docs/available/etd-07092008-211200/unrestricted/kromerdiss.pdf. Laglagaron, Laureen, Cristina M. Rodriguez, Alexa Silver, and Sirithon Thanasombat. 2008. Regulating Immigration at the State Level: Highlights from the Database of 2007 State Immigration Legislation and the Methodology. Washington, D.C.: Migration Policy Institute. www.migrationpolicy.org/pubs/2007methodology.pdf. Lee, Yueh-Ting, Victor Ottati, and Imtiaz Hussain. 2001. ?Attitudes toward ?Illegal? Immigration into the United States: California Proposition 187.? Hispanic Journal of Behavioral Sciences 23(4): 430?443. Lee, Yueh-Ting, and Victor Ottati. 2002. ?Attitudes toward U.S. Immigration Policy: The Roles of In-Group-Out-Group Bias, Economic Concern, and Obedience to Law.? Journal of Social Psychology 142: 617?634. Lewis-Beck, Michael et al. 2008. The American Voter Revisited. Ann Arbor: University of Michigan Press. Lilley III, William, Laurence J. Defranco, and Mark F. Bernstein. 2007. Almanac of State Legislative Elections. 3rd ed. Washington, D.C.: CQ Press. 132 Martinez-Ebers, Valerie et al. 2000. ?Latino Interests in Education, Health, and Criminal Justice Policy.? PS: Political Science and Politics 33(3): 547?554. Mayhew, David R. 1974. Congress: The Electoral Connection. New Haven: Yale University Press. McCarty, Nolan, Keith T. Poole, and Howard Rosenthal. 2006. Polarized America: The Dance of Ideology and Unequal Riches. Cambridge, MA: MIT Press. McConnell, Eileen Diaz. 2008. ?The U.S. Destinations of Contemporary Mexican Immigrants.? International Migration Review 42(4): 767?802. Migration Policy Institute and New York University School of Law. 2007. ?State Responses to Immigration: A Database of All State Legislation.? http://www.migrationpolicy.org/pubs/2007methodology.pdf (Accessed March 7, 2009). Miller, Joanne M., and Jon A. Krosnick. 2004. ?Threat as a Motivator of Political Activism: A Field Experiment.? Political Psychology 25(4): 507-523. Miller, Warren E., and Donald E. Stokes. 1963. ?Constituency Influence in Congress.? American Political Science Review 57(1): 45?56. Minta, Michael D. 2009. ?Legislative Oversight and the Substantive Representation of Black and Latino Interests in Congress.? Legislative Studies Quarterly 34(2): 193?218. Monogan, James E. 2010. ?The Long-term Consequences of Immigration Politics.? Ph.D. dissertation, University of North Carolina, Chapel Hill. http://search.proquest.com.proxy- um.researchport.umd.edu/pqdtft/docview/305218045/abstract/137BD12AD934B9CE696 /46?accountid=14696 (Accessed July 5, 2012). National Conference of State Legislatures. 1996. ?Bills and Bill Processing.? In Inside the Legislative Process, Washington, D.C.: National Conference of State Legislatures. http://www.ncsl.org/documents/legismgt/ILP/96Tab3Pt1.pdf. National Conference of State Legislatures. 2012. ?Percentages of Women in State Legislatures.? Washington, D.C.: National Conference of State Legislatures. http://www.ncsl.org/legislatures-elections/wln/women-in-state-legislatures-2012.aspx (Accessed October 28, 2012). Newman, Benjamin J. et al. 2012. ?Immigration Crackdown in the American Workplace: Explaining Variation in E-Verify Policy Adoption across the U.S. States.? State Politics & Policy Quarterly 12(2): 160?182. Newton, Lina. 2008. Illegal, Alien, or Immigrant: The Politics of Immigration Reform. New York: NYU Press. 133 Newton, Lina. 2012. ?Policy Innovation or Vertical Integration? A View of Immigration Federalism from the States.? Law & Policy 34(2): 113?137. Norton, Noelle H. 1999. ?Uncovering the Dimensionality of Gender Voting in Congress.? Legislative Studies Quarterly 24(1): 65?86. Overby, L. Marvin, Thomas A. Kazee, and David W. Prince. 2004. ?Committee Outliers in State Legislatures.? Legislative Studies Quarterly 29(1): 81?107. Overby, L. Marvin, Thomas A. Kazee. 2003. ?Outlying Committees in the Statehouse: An Examination of the Prevalence of Committee Outliers in State Legislatures.? Journal of Politics 62(3): 701?728. Pantoja, Adrian D., and Gary M. Segura. 2003. ?Fear and Loathing in California: Contextual Threat and Political Sophistication among Latino Voters.? Political Behavior 25(3): 265?286. Passel, Jeffrey S., and D?Vera Cohn. 2011. Unauthorized Immigrant Population: National and State Trends, 2010. Washington, D.C.: Pew Hispanic Center. http://njdac.org/blog/wp-content/plugins/downloads- manager/upload/2010%20undocumented%20trends%20by%20state.pdf (Accessed October 25, 2012). PBS. 2011. ?Alabama?s Immigration Law: Assessing the Economic, Social Impact.? Newshour. http://www.pbs.org/newshour/bb/business/july-dec11/alimmigration_10- 13.html (Accessed October 19, 2012). Pearson-Merkowitz, Shanna, and Stephen Yoder. 2009. ?Legislative Minutemen: The Politics of Immigration Policy in the US States.? Presented at the Spring Conference of the Graduate Student at the University of Maryland, College Park, Maryland. May 8. Pew Research Center. 2012. The Rise of Asian Americans. Washington, D.C.: Pew Research Center. http://www.pewsocialtrends.org/files/2012/06/SDT-The-Rise-of-Asian- Americans-Full-Report.pdf (Accessed October 29, 2012). Pitkin, Hanna F. 1972. The Concept of Representation. Berkeley: University of California Press. Poggione, Sarah. 2004. ?Exploring Gender Differences in State Legislators? Policy Preferences.? Political Research Quarterly 57(2): 305?314. Poole, Keith T., and Howard Rosenthal. 1985. ?A Spatial Model for Legislative Roll Call Analysis.? American Journal of Political Science 29: 357?384. Poole, Keith T., and Howard Rosenthal. 1991. ?Patterns of Congressional Voting.? American Journal of Political Science 35(1): 228?278. 134 Poole, Keith T., and Howard Rosenthal. 1997. Congress: A Political-Economic History of Roll Call Voting. New York: Oxford University Press. Poole, Keith T., and Howard Rosenthal. 2001. ?D-NOMINATE after 10 Years: A Comparative Update to Congress: A Political-Economic History of Roll-Call Voting.? Legislative Studies Quarterly 26: 5?29. Preuhs, R. R., and E. G. Juenke. 2011. ?Latino U.S. State Legislators in the 1990s: Majority-Minority Districts, Minority Incorporation, and Institutional Position.? State Politics & Policy Quarterly 11(1): 48?75. Primo, David M., Matthew L. Jacobsmeier, and Jeffrey Milyo. 2007. ?Estimating the Impact of State Policies and Institutions with Mixed-Level Data.? State Politics & Policy Quarterly 7(4): 446?459. Ray, David. 1982. ?The Sources of Voting Cues in Three State Legislatures.? Journal of Politics 44(4): 1074?1087. Reich, Gary, and Jay Barth. 2010. ?Educating Citizens or Defying Federal Authority? A Comparative Study of In-State Tuition for Undocumented Students.? Policy Studies Journal 38(3): 419?445. Reimers, David. 1999. Unwelcome Strangers: American Identity and the Turn Against Immigration. New York: Columbia University Press. Reingold, Beth. 2000. Representing Women?: Sex, Gender, and Legislative Behavior in Arizona and California. Chapel Hill: University of North Carolina Press. Rocca, M. S., G. R. Sanchez, and J. Uscinski. 2008. ?Personal Attributes and Latino Voting Behavior in Congress.? Social Science Quarterly 89(2): 392?405. Rocha, Rene R., and Rodolfo Espino. 2009. ?Segregation, Immigration, and Latino Participation in Ethnic Politics.? American Politics Research 38 (4): 614?635. Rohde, David. 1991. Parties and Leaders in the Postreform House. Chicago: University of Chicago Press. Rosenthal, Cindy Simon. 2001. ?Gender Styles in State Legislative Committees.? Women & Politics 21(2): 21?45. Rouse, Stella M. 2013. Latinos in the Legislative Process: Interests and Influence. Cambridge: Cambridge University Press. Rumbaut, R. G., and W. A. Ewing. 2007. The Myth of Immigrant Criminality and the Paradox of Assimilation: Incarceration Rates among Native and Foreign-born Men. 135 Washington, D.C.: Immigration Policy Center, American Immigration Law Foundation. Saint-Germain, Michelle A. 1989. ?Does Their Difference Make a Difference? The Impact of Women on Public Policy in Arizona Legislature.? Paper presented at the Annual Meeting of the American Political Science Association, Atlanta, GA. http://www.eric.ed.gov/ERICWebPortal/contentdelivery/servlet/ERICServlet?accno=ED3 14324 (Accessed April 29, 2009). Sanchez, Gabriel R., Lisa Sanchez, and Vickie Ybarra. 2012. ?Filling the Void: Factors Leading to Punitive Immigration Policy Across the American States.? Paper presented at the Annual Meeting of the Western Political Science Association, Portland, OR. http://wpsa.research.pdx.edu/meet/2012/sanchezgabriel.pdf. Schildkraut, Deborah J. 2001. ?Official-English and the States: Influences on Declaring English the Official Language in the United States.? Political Research Quarterly 54(2): 445?457. Schiller, Wendy J. 1995. ?Senators as Political Entrepreneurs: Using Bill Sponsorship to Shape Legislative Agendas.? American Journal of Political Science 39(1): 186?203. Segal, Jeffrey A., Charles M. Cameron, and Albert D. Cover. 1992. ?A Spatial Model of Roll Call Voting: Senators, Constituents, Presidents, and Interest Groups in Supreme Court Confirmations.? American Journal of Political Science 36: 96?121. Shepsle, Kenneth A., and Barry R. Weingast. 1987. ?The Institutional Foundations of Committee Power.? American Political Science Review 81(1): 85?104. Shor, Boris, Christopher Berry, and Nolan McCarty. 2010. ?A Bridge to Somewhere: Mapping State and Congressional Ideology on a Cross-institutional Common Space.? Legislative Studies Quarterly 35(3): 417?448. Snyder, James M., and Tim Groseclose. 2000. ?Estimating Party Influence in Congressional Roll-Call Voting.? American Journal of Political Science 44(2): 193?211. Songer, Donald R. et al. 1986. ?The Influence of Issues on Choice of Voting Cues Utilized by State Legislators.? Western Political Quarterly 39(1): 118?125. Spiro, Peter J. 1997. ?Learning to Live with Immigration Federalism.? Connecticut Law Review 29: 1627?1646. Squire, Peverill. 1988. ?Career Opportunities and Membership Stability in Legislatures.? Legislative Studies Quarterly 13(1): 65?82. Squire, Peverill. 2007. ?Measuring State Legislative Professionalism: The Squire Index Revisited.? State Politics & Policy Quarterly 7(2): 211?227. 136 State Legislators for Legal Immigration. 2007a. ?State Legislators for Legal Immigration Personally Deliver Illegal Alien Invasion Halting Marching Orders on Pearl Harbor Day.? http://www.statelegislatorsforlegalimmigration.com/NewsItem.aspx?NewsID=5643. State Legislators for Legal Immigration. 2007b. ?Virginia?s Sen. Cuccinelli Joins State Legislators for Legal Immigration; Denounces Federal Amnesty Bill.? http://www.statelegislatorsforlegalimmigration.com/NewsItem.aspx?NewsID=5673. Swain, Carol M. 2006. Black Faces, Black Interests: The Representation of African Americans in Congress. Lanham, MD: University Press of America. Swers, Michele L. 1998. ?Are Women More Likely to Vote for Women?s Issue Bills than Their Male Colleagues?? Legislative Studies Quarterly 23(3): 435?448. Swers, Michele L. 2002. The Difference Women Make: The Policy Impact of Women in Congress. Chicago: University of Chicago Press. Talbert, Jeffery C., and Matthew Potoski. 2002. ?Setting the Legislative Agenda: The Dimensional Structure of Bill Cosponsoring and Floor Voting.? Journal of Politics 64(3): 864?891. Tatalovich, Raymond. 1995. ?Voting on Official English Language Referenda in Five States: What Kind of Backlash against Spanish-Speakers?? Language Problems and Language Planning 19(1): 47?59. Tate, Katherine. 2004. Black Faces in the Mirror: African Americans and Their Representatives in the U.S. Congress. Princeton: Princeton University Press. Thomas, Sue. 1991. ?The Impact of Women on State Legislative Policies.? Journal of Politics 53(4): 958?976. Tichenor, Daniel J. 2002. Dividing Lines: The Politics of Immigration Control in America. Princeton: Princeton University Press. Tolbert, Caroline J., and Rodney E. Hero. 1996. ?Race/Ethnicity and Direct Democracy: An Analysis of California?s Illegal Immigration Initiative.? Journal of Politics 58(3): 806?818. Tolbert, Caroline J., John A. Grummel, and Daniel A. Smith. 2001. ?The Effects of Ballot Initiatives on Voter Turnout in the American States.? American Politics Research 29(6): 625?648. U.S. Census Bureau. 2007. Income, Poverty, and Health Insurance Coverage in the United States: 2006. Washington, D.C.: Government Printing Office. http://www.census.gov/prod/2007pubs/p60-233.pdf (Accessed October 26, 2012). 137 Valenty, Linda O, and Ronald D Sylvia. 2004. ?Thresholds for Tolerance: The Impact of Racial and Ethnic Population Composition on the Vote for California Propositions 187 and 209.? Social Science Journal 41(3): 433?446. Vega, Arturo, and Juanita M. Firestone. 1995. ?The Effects of Gender on Congressional Behavior and the Substantive Representation of Women.? Legislative Studies Quarterly 20(2): 213?222. Walker, Devona. 2007. ?Families say they worry they will be torn apart.? NewsOK. http://newsok.com/families-say-they-worry-they-will-be-torn-apart/article/3178078 (Accessed January 6, 2013). Welch, Susan. 1985. ?Are Women More Liberal than Men in the U.S. Congress?? Legislative Studies Quarterly 10(1): 125?134. Welch, S., and J. R. Hibbing. 1984. ?Hispanic Representation in the US Congress.? Social Science Quarterly 65(2): 328?335. Williams, Robin Murphy. 1947. The Reduction of Intergroup Tensions: A Survey of Research on Problems of Ethnic, Racial, and Religious Group Relations. New York: Social Science Research Council. Witt, Stephanie L., and Gary Moncrief. 1993. ?Religion and Roll Call Voting in Idaho.? American Politics Research 21(1): 140?149. Wlezien, Christopher. 2005. ?On the Salience of Political Issues: The Problem with ?Most Important Problem.?? Electoral Studies 24: 555-579. Wolpert, Robin M., and James G. Gimpel. 1998. ?Self-Interest, Symbolic Politics, and Public Attitudes toward Gun Control.? Political Behavior 20(3): 241?262. Wright, Gerald C, and Brian F Schaffner. 2002. ?The Influence of Party: Evidence from the State Legislatures.? American Political Science Review 96(2): 367?379. Wright, Gerald. 2004. Representation in America?s Legislatures. Indiana University: National Science Foundation Grant. http://www.indiana.edu/~ral/data_9900.html. Wroe, Andrew. 2008. The Republican Party and Immigration Politics: From Proposition 187 to George W. Bush. Boston: Palgrave Macmillan.