University of Southern California Dissertations and Theses

The changing state of tolerance and tolerance in changing states

(USC Thesis Other)

The changing state of tolerance and tolerance in changing states

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content The Changing State of Tolerance and Tolerance in Changing States by Youssef Chouhoud A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Political Science and International Relations) May 2019 Dedication To my parents, Elsayed and Amani. Each day I appreciate more and more the sacrifices you made to get me where I am. I wish we had had more time together. To the brave women and men of Egypt that inspired me to take this path in life. I hope one day we can make the better society you dreamed of a reality. i Acknowledgements It has been a long, winding road. I would not have been able to traverse it without a small army of masterful guides and gracious allies helping me along the way. Inside the academy, my gratitude goes first and foremost to my co-chairs, Laurie Brand and Gerardo Munck. From literally the first day I set foot on USC’s campus, Laurie has been a constant source of wisdom and compassion. Her breadth and depth of knowledge about the Middle East was inspiring, as was the personal respect she afforded the region and its people. For six years, Laurie was my anchor. I could not have asked for a more exemplary advisor. Gerry was no less a source of inspiration and support. His insights on democracy and democratization will forever influence the way I think about these richest of concepts. When I was fortunate enough to be part of an inaugural cohort of teaching fellows at USC, Gerry was an invaluable mentor who devoted countless hours helping me improve my course on the Arab Uprisings. I am very lucky to have someone as thoughtful as him in my corner. My committee also included two titans in their field. Many in academia can claim that they are an authority on this or that niche topic. What makes Dennis Chong stand out is that he can rightfully claim authoritative knowledge on “how people think about things.” That standing can also make him a bit intimidating at ii times. Whenever I got over my intimidation, however, I was always richly rewarded with probing critiques that immensely improved my work. Beyond that, there were numerous times when we would just talk about life and other mundane things and it meant more to me than any research feedback. And then there is Sherman Jackson. For my money, Dr. J is the top scholar of Islam in the nation. He was someone I looked up to long before we met, and he one of those rare heroes whom you admire all the more once you get to know them better. Our conversations in his office are some of my most treasured memories from USC–they were always intellectually challenging and spiritually rejuvenating. I was truly blessed to have the opportunity to learn from him. I owe a debt of gratitude to many faculty members beyond my dissertation com- mittee. In particular, Morris Levy, Anthony Kammas, Jeb Barnes, and Christian Grose each continually offered me invaluable council and friendship during my time at USC. Whenever I hear horror stories from other programs, I realize how good I had it as a member of the POIR community. This community, of course, extends beyond the faculty. Cathy Ballard, Veridiana Chavarin, Aurora Ramirez, and Linda Cole each helped me navigate university bu- reaucracy on a regular basis and (on more than one occasion) saved me when I was in a jam (usually of my own making). But more than that, they always just made my day brighter. I cannot thank them enough. The PhD grind would not have been doable were it was not for my amazing colleagues. My deepest gratitude to Joey, Juve, Justin, Simon, Sara, NdZ, Brian, Parker, Rod, Dave, Cyrus, Mark, Jordan, Kelly, Jennifer, Guez, and Suzie. You all made grad school a blast when it can often be a bummer. Outside of the academy, my family was my rock. My stepmom, Hala Eladawi, and my sister, Joudi, always offered words of encouragement and kept me connected to my homeland. Love you both. Stateside, having my sister, Rabab, and her husband, Aly, along with their two beautiful daughters, Taliyah and Alyah, close to home made the last three years in LA so much more enjoyable. Having Rabab at graduation made me feel like my mom and dad were there, too. Love you lots. iii Last, to my wife and daughters, I am at a loss for the right words, but I will do my best. Dahlia, Hana, and Zeyna light up my life. They remind me each day why I keep doing what I am doing. I cannot wait to see what amazing things they will all do when they grow up, God willing. As for Suzanne, none of this–literally not any of it–would have been possible without her. She did not just support me, she was my foundation. She did not simply sacrifice for my pursuit, she did so without ever complaining. From working a full time job, to raising the kids, to putting up with me and all my idiosyncrasies, she deserves more praise than I can possibly give her in this space. I love you sweetheart. iv Contents Dedication i Acknowledgements ii List of Figures viii List of Tables x Abstract xi 1 Introduction 1 1.1 The Value and Conceptualization of Tolerance . . . . . . . . . . . . . 2 1.2 Research on Political Tolerance . . . . . . . . . . . . . . . . . . . . . 5 1.2.1 The Stouffer Agenda . . . . . . . . . . . . . . . . . . . . . . . 5 1.2.2 Muslims and Democratic Norms . . . . . . . . . . . . . . . . . 7 1.3 Overview of The Dissertation . . . . . . . . . . . . . . . . . . . . . . 11 2 Are Minorities (In)Tolerant Like Everyone Else?: American Mus- lims and the Enhanced Role of Personal Threat 15 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2.2 Determinants of Political Tolerance . . . . . . . . . . . . . . . . . . . 19 2.2.1 Threat Perception . . . . . . . . . . . . . . . . . . . . . . . . 20 2.2.2 Minority Status . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2.2.3 Religiosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 v 2.3 Hypotheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.4 Data and Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.4.1 Dependent Variable . . . . . . . . . . . . . . . . . . . . . . . . 27 2.4.2 Key Independent Variables and Controls . . . . . . . . . . . . 28 2.4.3 Experimental Analysis . . . . . . . . . . . . . . . . . . . . . . 30 2.5 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 33 2.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Appendices 52 2.A Sampling Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . 52 2.B Question Wording . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 3 (When) Are Muslims (In)Tolerant?: The Conditional Expression of Democratic Norms in Tunisia and Egypt 55 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 3.2 Political Tolerance: An Overview . . . . . . . . . . . . . . . . . . . . 58 3.2.1 Comparative Studies of Political Tolerance . . . . . . . . . . . 58 3.2.2 Political Tolerance in the MENA Region . . . . . . . . . . . . 60 3.3 Political Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 3.4 Hypotheses and Model Specification . . . . . . . . . . . . . . . . . . . 64 3.4.1 Hypotheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 3.4.2 The Data and the Dependent Variable . . . . . . . . . . . . . 69 3.4.3 Additional Controls . . . . . . . . . . . . . . . . . . . . . . . . 72 3.5 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 73 3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Appendices 83 3.A Additional Tables and Figures . . . . . . . . . . . . . . . . . . . . . . 83 3.B Question Wording . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 4 Tolerance and the Prospects for Tunisian Democracy 85 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 vi 4.2 Political Tolerance in the Tunisian Context . . . . . . . . . . . . . . . 87 4.3 Data, Methodology, and Expectations . . . . . . . . . . . . . . . . . . 90 4.3.1 Dependent Variable . . . . . . . . . . . . . . . . . . . . . . . . 91 4.3.2 Key Independent Variables and Controls . . . . . . . . . . . . 96 4.3.3 Experimental Design . . . . . . . . . . . . . . . . . . . . . . . 98 4.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 101 4.4.1 The Structure of Tunisian Political Tolerance . . . . . . . . . 102 4.4.2 The Maleability of Tunisian Political Tolerance . . . . . . . . 108 4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Appendices 115 4.A Additional Figures and Tables . . . . . . . . . . . . . . . . . . . . . . 115 4.B Question Wording . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 References 119 vii List of Figures 2.1 Hate Groups in America, 2017 . . . . . . . . . . . . . . . . . . . . . . 17 2.2 Least-Liked Group Choice (Percent of Total), by Assignment . . . . . 28 2.3 Political Tolerance, by Assignment and Chosen Group . . . . . . . . . 34 2.4 Egocentric Threat, by Assignment and Chosen Group . . . . . . . . . 34 2.5 Marginal Effect of Egocentric Threat, by Dogmatism . . . . . . . . . 39 2.6 Proportion of Respondents Changing Initial Tolerance Judgment . . . 41 2.7 Attitude Change among the Initially Neutral – Tolerance Prompt . . 43 2.8 Attitude Change among the Initially Neutral – Intolerance Prompt . 43 2.9 Disliked Group Choice, by Nativity . . . . . . . . . . . . . . . . . . . 47 3.1 Coding Process for the Dependent Variable . . . . . . . . . . . . . . . 71 3.2 Percent Distribution of Dependent Variable, by Year and Country . . 71 3.3 Predicted Probability of a Tolerant Response, by Sharia Support . . . 79 3.4 Predicted Probability of a Tolerant Response, by Religiosity . . . . . 79 3.5 Predicted Probability of a Tolerant Response, by Multilingualism . . 81 4.1 Distribution of Disliked Groups . . . . . . . . . . . . . . . . . . . . . 103 4.2 Mean of Political Intolerance, by Disliked Group . . . . . . . . . . . . 103 4.3 Mean of Tolerance and Threat Levels, by Disliked Group Type . . . . 104 4.4 Determinants of Political Tolerance, Islamist vs. Secular . . . . . . . 108 4.5 Attitude Change, by Initial Attitude and Prompt . . . . . . . . . . . 109 4.6 Attitude Change, by Strength of Initially Tolerant Attitude . . . . . . 110 4.7 Attitude Change, by Strength of Initially Intolerant Attitude . . . . . 110 4.8 Attitude Change, by Islamist Status . . . . . . . . . . . . . . . . . . . 111 viii 4.9 Attitude Change, by Disliked Group Type (Initially Tolerant) . . . . 111 4.10 Attitude Change, by Disliked Group Type (Initially Intolerant) . . . . 111 ix List of Tables 2.1 Determinants of Political Tolerance–Full Sample . . . . . . . . . . . . 36 2.2 Determinants of Political Tolerance–Split Samples . . . . . . . . . . . 38 2.3 Conditional Effect of Education on Political Tolerance . . . . . . . . . 45 2.4 Effect of Immigrant Socialization on Tolerance . . . . . . . . . . . . 46 3.1 Descriptive Statistics (Non-Islamists vs. Islamists) . . . . . . . . . . . 74 3.2 Change in the Probability of a Tolerant Response, Full Sample . . . . 75 3.3 Change in the Probability of a Tolerant Response, by Country and Year 77 3.4 Change in the Probability of a Tolerant Response, by Islamist Status 78 3.5 Change in the Probability of a Tolerant Response, by Multilingualism 80 4.1 Determinants of Political Tolerance–Full Sample . . . . . . . . . . . . 105 x Abstract This three article dissertation explores the determinants of political tolerance by examining understudied groups and settings, emphasizing in each case how context is a key driver of the observed attitudes. First, drawing on an original survey of American Muslims, I theorize and test how the considerations motivating intolerance differ between the general public and minorities in established democracies. My analysis highlights: 1) the enhanced role of perceived threat to personal safety in motivating targeted intolerance among marginalized groups; and 2) the prevailing (if not entirely homogenous) effect of education and acculturation in augmenting overall tolerance levels. These findings thus undermine charges of American Muslims’ inherent resistance to integration, while also suggesting an asymmetric distribution in the burdens of democratic citizenship in America. In Chapter 3, I examine intolerance in a Muslim-majority setting. Although prior work has offered useful insight into the Middle East’s general propensity for intol- erance, I provide a more refined empirical test of political tolerance in the region. Specifically, I construct a novel tolerance scale using Arab Barometer data and use it to test whether Muslim religiosity (variably defined) consistently maps onto intol- erant attitudes, or if there is instead a contingent association based on proximate social and political factors. The analyses demonstrate several findings in support of a more dynamic under- standing of the relationship between Islam and democracy. First, although analysis of the pooled sample indicates that Islamists are, on average, more intolerant than non-Islamists, partitioning the data by year reveals that this division emerged only after the initial transitional periods in these countries. This conditional relationship xi suggests that the observed intolerance is a function of changing political circum- stances rather than a deep-seated aversion to particular democratic norms. Second, it is only among Islamists that increased personal religiosity and sup- port for implementing sharia are correlated with higher intolerance, implying that enhanced in-group identity mediates this link. Last, educational attainment—one of the most robust predictors of tolerance-–exerts a significant influence on Egyptian and Tunisian attitudes towards civil liberties, although not through the usual path- ways. Taken together, these results demonstrate that Muslims in the Middle East appear susceptible to the same factors that generally enhance or impede political tolerance among other populations. In the final empirical chapter, I continue the examination of tolerance in emerg- ing democratic settings through more precise observational data coupled with ex- perimental findings. Drawing on an original, nationally representative survey of Tunisians, I find that the structure of intolerance in this nascent democracy bodes well for its short-term prospects. Specifically, rather than target major political ac- tors, Tunisians focus their intolerance largely on those groups agitating for social change from outside the formal political arena. To be sure, this is neither ideal nor sustainable if the long-term goal is broad-based pluralism. Yet, in terms of the more acute threat of authoritarian backsliding, it would be far less desirable if large swaths of the public wished to curtail the rights of entities participating in the (still quite fragile) political system. The experimental findings further buttress this result. When presented with counter-arguments for their initial position toward a group they strongly dislike, initially intolerant respondents were significantly more likely to be persuaded to tol- erance if their target group was a formal political actor rather than a social agitator. All respondents, however, were equally as likely to move away from an initially tol- erant position and, moreover, did so at a rather high rate (with nearly three in four swayed toward intolerance). Thus, even though the analysis indicates grounds for cautious optimism at this point in time, Tunisian attitudes appear quite malleable and therefore susceptible to stark shifts should conditions on the ground or elite discourse undermine the current equilibrium. xii Chapter 1 Introduction Early on in the summer of 2013, before military strongmen abruptly halted the country’s democratic progress, it was clear to any Egyptian observer that the space for political disagreement was rapidly narrowing. Today in Tunisia, the country’s designation as one of the world’s newest democracies (and the only one in the Arab Middle East) obscures the deep and enduring social cleavage pitting Islamists against their detractors. On the other side of the Atlantic Ocean, the state of political discourse in the world’s oldest democracy is no less fraught, as divisions sharpen along the same lines that nearly tore the American republic apart 150 years ago. Across these varied settings, both intolerance and Muslims feature prominently. What can these contexts and subjects teach us about the drivers of democratic norms in emerging and established democracies? This is the question that motivates this three article dissertation. Before elaborating the specifics of each study, this introductory chapter 1) examines the value of tolerance, 2) discusses the concept’s operationalization, 3) reviews the ways in which scholars have studied political tol- erance, in particular, and 4) surveys the legacy of skepticism over the compatibility 1 of Islam and democracy. 1.1 The Value and Conceptualization of Tolerance Tolerance is a fundamental liberal democratic norm. Yet, whereas other core qualities of a well-functioning democracy (the safeguarding of human rights and a deference to the rule of law, for example) are universally lauded (at least in the abstract), it is not uncommon for politicians and other elites to speak of tolerance in derisive terms. On the one hand, leaders from both sides of the political aisle routinely underscore the intrinsic value of a tolerant society. During the 2004 Demo- cratic National Convention, for example, then-Senator Barack Obama spoke proudly of his parents’ faith in a “tolerant America” (Obama 2004). Praise for this vision of American society is similarly echoed by staunch conservatives. Ronald Reagan, for example, invoked the norm on multiple occasions during his presidency, including at a meeting with Asian and Pacific-American Leaders where he endorsed the “re- sponsibility of parents and schools to be teachers of tolerance” (Reagan 1984). Such reverence is all the more prevalent in the wake of an event that highlights America’s continuing struggle with intolerance, as with the August 2018 White supremacist rally in Charlottesville, Virginia. 1 On the other hand, elites in established and emerging democracies do not shy away from downgrading or even rebuking tolerance as a societal ideal. For instance, during a 2018 commencement address, Canadian Prime Minister Justin Trudeau ad- 1. The chorus of prominent voices calling for a renewed commitment to tolerance (or, relatedly, vigilance against intolerance) in the wake of the violence in Charlottesville included politicians (Kennedy 2017), military leaders (Cohen and Starr 2017), and corporate executives (Barra 2017). 2 vocated for a more robust commitment to pluralism, stating: “There’s not a religion in the world that asks you to ‘tolerate thy neighbor,’ so let’s try for something like acceptance” (Timsit 2018). In a wholly different context, Chawki Tabib, the for- mer head of Tunisia’s National Anti-Corruption Commission, echoed the sentiment when asked about the state of tolerance in his country, asserting that we should stop talking about tolerating one another and instead aim for mutual acceptance. 2 Us- ing far more blunt language, Senator Cory Booker, speaking in front of an audience four months after Donald Trump took office, underscored love as the country’s true moral objective rather than tolerance, which he called a “repugnant ambition for any nation” (Klein 2017). Along these same lines, during a speech on deteriorating race relations toward the end of the 2016 presidential campaign, the CEO of AT&T remarked that “tolerance is for cowards” and that his company’s employees should instead “move into uncomfortable territory and understand each other” (Holland 2016). These divergent takes on the value of tolerance are due, in part, to varying inter- pretations of the concept. In everyday parlance, tolerance often connotes a respect for, or even appreciation of, different groups and ideas. This conception is at the heart of claims that we are a more tolerant society than we once were (“Getting Nicer” 2015; Goldberg 2018). Proponents of this view often point to the steady reduction in overt racism since the Civil Rights era and the more recent (and more 2. Interview with the author. January 4, 2016. Tunis, Tunisia. 3 dramatic) shift in attitudes toward LGBTQ individuals. 3 These swings, however, would fall under the umbrella of what is more precisely called social tolerance. An altogether different set of considerations and outcomes emerges when we consider, instead, political tolerance. Whereas social tolerance is based on affect (primarily) toward racial, ethnic, religious, or sexual minorities–that is, one’s level of prejudice, broadly construed– political tolerance is typically defined as the willingness to afford equitable political rights to one’s ideological opponents. As a case in point, the latter’s systematic study can be traced back to Samuel Stouffer’s (1955) seminal work on Americans’ readiness to limit communists’ rights during the height of McCarthyism. Along these same lines, virulently racist groups, such as the Ku Klux Klan, are frequently the main targets of intolerance in more recent studies. Fundamentally, then, political tolerance denotes the capacity to support (be- grudgingly or otherwise) the civil liberties of groups that one strongly disagrees with– or even detests. It is this understanding of tolerance that implicates first amendment rights to freedom of speech and assembly. Far from the ideals of acceptance and re- spect, this variant of tolerance merely requires that we “put up with” those ideas and groups we find deeply disagreeable. This particular conception of tolerance somewhat validates both those who regard 3. It is important to note that while the trend in direct questions on, for example, anti-gay and racist attitudes has clearly decreased over time (Marsden 2012), there is no way to account for the degree of social desirability bias in these surveys. That is, respondents (particularly when being interviewed in person) may be reticent to provide their true beliefs on sensitive questions, offering instead what they believe to be a socially acceptable response (see, e.g., Coffman, Coffman, and Ericson 2013). Thus, polling on these issues may reflect a change in societal norms rather than individual attitudes. 4 the norm as a high ideal or a low standard. In the first instance, asking citizens to accept that even groups they detest deserve the same civil liberty protections as anyone else may be close to the upper bound of what is reasonable. In this respect, tolerance is clearly aspirational. However, this framework of tolerance quickly begins to yield diminishing returns when instead of exclusively guiding citizens’ interactions with groups that they truly find repellent, it starts informing how they deal with those with whom they merely disagree. When it comes to these general interpersonal relations, tolerance perhaps more closely resembles a bare minimum standard for interpersonal relations. Wherever it may fall on the normative spectrum, however, political tolerance is often empirically elusive. 1.2 Research on Political Tolerance 1.2.1 The Stouffer Agenda While tolerance is certainly not a panacea and is far from the most desirable guideline for interacting with fellow citizens, it is also not a norm that can be taken for granted. The modern study of political tolerance arguably began with Stouffer’s (1955) seminal work in the midst of America’s Red Scare. Examining whether the American public’s attitudes were in line with the country’s widespread political re- pression at the elite level, Stouffer found several troubling parallels. For example, over two-thirds of his respondents supported denying an avowed communist the right to give a public speech. Although the survey’s hypothesized restrictions varied qual- itatively (ranging, as they did, from firing admitted communists from a job to not 5 buying a brand of soap they endorsed), the general finding was unequivocal: Ameri- cans were more than willing to limit the rights and freedoms of groups whose beliefs challenged their own–not just communists, but atheists and socialists, as well. To gauge political tolerance, Stouffer relied on one of the literature’s two primary measurement techniques. His method required asking each respondent a battery of questions about the civil liberties they would be willing to extend to each of several groups from a static set. This approach gained prominence among social scientists, particularly once it was incorporated into the National Opinion Research Center’s flagship General Social Survey (GSS). Over time, however, scholars began to question the validity of Stouffer’s scale, which came to be known as the “fixed-group” method. In particular, critics of this measure argued that it was susceptible to falsely classifying a response as tolerant when, instead, it may actually be a function of indifference. That is, if respondents do not sufficiently dislike a group, it makes little sense to label them as tolerant of that group’s speech and activity. To account for this ambiguity, Sullivan et al. (1979) introduced a “content-controlled” measure of tolerance. With this approach, respondents are first asked to identify their “least-liked” group from a list (or write one in) and are subsequently asked roughly the same battery of questions as in the fixed-group method. For added layers of comparison, scholars will sometimes ask respondents to apply the political tolerance battery to their second or even their third least-liked group (see, e.g., Sullivan, Piereson, and Marcus 1982; Gibson 2013). Marshaling either of these generally accepted measures of tolerance, Stouffer’s work spawned a multitude of studies that sought both to replicate his findings and 6 expand on them. Scholars have, for instance, examined whether elites are systemati- cally more tolerant than the mass public (Jackman 1972; McClosky and Brill-Scheuer 1983) and delineated a range of individual and contextual determinants of intoler- ance (e.g., Marcus et al. 1995; Eisenstein 2006). Subsequent studies also found that the public’s relative degree of tolerance changed little over time (Sullivan, Piereson, and Marcus 1979; Mondak and Sanders 2003), even as the targets of their intolerance may have shifted. Notably, however, Americans’ unremarkable level of forbearance actually rep- resents something of a high water mark globally. In their cross-national study of political tolerance in seventeen countries, Peffley and Rohrschneider (2003) find that citizens in established democracies (such as the United States, Australia, and Swe- den) are significantly more tolerant overall than those living in anocracies or fully autocratic states. Their theory, that it is engagement in the “rough and tumble” of democratic politics that engenders tolerant attitudes, offers an intriguing research avenue for scholars studying either largely immigrant populations in Western Europe and North America or states transitioning away from autocratic rule. This disserta- tion engages both these fronts through the lens of Muslims in America, Tunisia, and Egypt. 1.2.2 Muslims and Democratic Norms There is a long lineage of philosophers, scholars, politicians, and public opin- ion questioning (or outright disavowing) the compatibility of Islam and democracy. This skepticism goes back at least to Montesquieu, who devotes a chapter in the 7 second volume of Spirit of the Laws to explaining, “That A Moderate Government Is Most Agreeable To The Christian Religion, And A Despotic Government To The Mahometan” (Book XXIV, Chapter III). In more recent times, Samuel Huntington famously pitted the “West” against (among others) “Islamic civilization” (Hunt- ington 1993) in large part because “Islam...has not been hospitable to democracy” (Huntington 1984, 253). Other scholars have similarly asserted that “the idea of democracy is quite alien to the mind-set of Islam” (Kedourie 1992, 1) and frequently rested such claims on the relative lack of separation between “church” and state in Islamic political history (Lewis 1996). Even those academics who are not nor- mally associated with this literature nonetheless presuppose that “there does seem to be something about Islam...that makes Muslim societies particularly resistant to modernity” (Fukuyama 2001). 4 The remarkable resilience of authoritarianism in Muslim-majority states, partic- ularly across the Middle East, offers a common point of reference for those charging that tenets of Islam are inherently undemocratic and/or that democratic norms are incompatible with Muslim beliefs. Indeed, up until Tunisia’s recent democratic open- ing, Lebanon in the early 1970s (that is, just prior to its devastating civil war) was the only country in the Arab Middle East to earn Freedom House’s designation of “Free.” Outside this “heartland of Islam” (as some have dubbed the region (Potrafke 2013)), only Senegal currently ranks as “Free” among Muslim-majority states (Freedom in the World 2018). Indeed, even after accounting for a suite of factors that typi- cally depress a country’s level of democracy, study after study has found a strongly 4. Here, Fukuyama uses the term “modernity” to refer to the twin pillars of liberal democracy and capitalism that, for him, characterize the end point of human progress. 8 negative correlation between “Islam” (variably operationalized as a function of the national legal system, geographic historical legacy, or contemporary demographics) and “democracy” (assessed through a handful of common datasets, such as Freedom House and Polity IV) (Fish 2002; Ross 2001; Herb 2005; Rowley and Smith 2009; Potrafke 2012; Fish 2011; but, c.f., Robertson and Stepan 2003; Hanusch 2013). Moreover, this presumed state-level link also informs elite and public perceptions of Muslims living in the West. Evidence of official skepticism over whether Muslims were capable of integrating into American society can be found at least as far back as the early 1940s when, in his decision denying citizenship to a Yemeni immigrant, a federal district court judge commented: Apart from the dark skin of the Arabs [a relevant point given that only Whites were eligible for naturalization at the time], it is well known that they are a part of the Mohammedan world and that a wide gulf separates their culture from that of the predominately Christian peoples of Europe. It cannot be expected that as a class they would readily intermarry with our population and be assimilated into our civilization (In Re Hassan 1942, 844). These concerns tended to emerge only occasionally, however, as few Muslims in the United States were foreign-born prior to the 1965 Immigration and Nationality Act, which abolitioned national origin quotas. In the decades following the enactment of this landmark bill, an influx of immigrant Muslims from the Middle East, South Asia, and Africa, changed the landscape of Islam in America. Buoyed by the cultural depiction of Muslims as a foreign threat (Shaheen 2014; Shaheen 2008) and propelled 9 by the seeming affirmation of that depiction with the terrorist attacks of September 11 th , skepticism over Muslims’ ability to integrate into a liberal democracy gained renewed and widespread salience at the turn of the 21 st century. Although the years immediately following the 2001 attacks saw some evidence of increasingly negative attitudes toward Muslims and Islam, 5 a clear shift (rather than a momentary deviation) in elite rhetoric, public opinion, and hate crimes was not evident until the following decade. To be sure, reported assaults on Muslims in the US increased exponentially in 2001; yet, this tally shrank nearly as dramatically the following year, increased incrementally beginning in 2010, and reached its apex in 2017 (Kishi 2017). The contrast among political elites over this period is similarly dramatic. Famously, President George W. Bush gave a speech only days after 9/11 underscoring that the coming fight was not with Islam as a religion nor with Muslims in general. These sentiments stand in stark contrast with those of Donald Trump, whose presidential campaign featured numerous explicitly negative comments about Muslims and their beliefs, including the blunt proclamation: “I think Islam hates us” (Johnson and Hauslohner 2017). The broader public, too, is also now consistently quite dubious of Muslims. For instance, in a 2017 poll, Americans reported feeling “colder” toward Muslims than any other religious group (including atheists), and fully 44% held that there exists “a natural conflict between the teachings of Islam and democracy” (Pew Research Center 2017). 5. In 2003, the proportion of the public believing that “Islam is more likely than other religions to encourage violence among its followers” increased to 44% from 25% the previous year (Pew Research Center 2003) 10 1.3 Overview of The Dissertation This dissertation builds on the substantive foundations of these renewed dis- courses on tolerance and Muslims’ democratic bona fides. In turn, the following three-article compilation advances the comparative study of political tolerance by developing theory, introducing new methodological strategies, and broadening the literature’s demographic and geographic cache of inferential findings. 6 Each of these standalone studies examines tolerance amid change, highlighting how contextual fac- tors generally drive the dynamics of this democratic norm. Turning first to the state of tolerance in established democracies, I probe one way in which the scope of tolerance is expanding. Specifically, as America’s demographic landscape becomes increasingly diverse, intolerance from (not just toward) minority groups is coming to the fore. Do the considerations motivating political tolerance vary between majority and minority populations? Chapter 2 takes on this largely ne- glected question through an original survey of American Muslims—a socially salient community for which comparatively little quantitative data exist. My analysis highlights: 1) the enhanced role of perceived threat to personal safety in motivating targeted intolerance among marginalized groups; and 2) the prevailing (if not entirely homogenous) effect of education and acculturation in augmenting overall tolerance levels. These findings thus undermine charges of American Muslims’ inherent resistance to integration, while also suggesting an asymmetric distribution in the burdens of democratic citizenship in America. 6. Given that they are meant to stand as individual studies, each chapter features its own corre- sponding appendix. 11 Next, I capitalize on the recent breakdowns in authoritarian rule across the Arab Middle East and North Africa to examine political tolerance in a region wholly ne- glected by the literature. Specifically, I test whether Muslim (in)tolerance is constant or conditional. That is, does Muslim religiosity (variably defined) consistently map onto intolerant attitudes, or is there a contingent association based on proximate so- cial and political factors? To address this question, I draw on data from the second and third waves of the Arab Barometer (the first cross-national, multi-year project surveying public opinion in the Arab Middle East) to model political tolerance in a Muslim-majority setting. In addition to introducing a novel method for gauging tolerance levels absent the typical survey battery, I present several findings in support of a more dynamic understanding of the relationship between Islam and democracy. First, although analysis of the pooled sample indicates that Islamists are, on average, more intol- erant than non-Islamists, partitioning the sample by year reveals that this division emerged only after the initial transitional periods in these countries. This conditional relationship suggests that the observed intolerance is a function of changing political circumstances rather than a deep-seated aversion to particular democratic norms. Second, it is only among Islamists that increased personal religiosity and sup- port for implementing sharia are correlated with higher intolerance, implying that enhanced in-group identity mediates this link. Last, educational attainment—one of the most robust predictors of tolerance-–exerts a significant influence on Egyptian and Tunisian attitudes towards civil liberties, although not through the usual path- ways. Taken together, these results demonstrate that Muslims in the Middle East 12 appear susceptible to the same factors that generally enhance or impede political tolerance among other populations. Finally, in chapter 4, I continue the examination of tolerance in emerging demo- cratic settings through more precise observational data coupled with experimental findings. Drawing on a nationally representative sample of Tunisians, I find that the structure of intolerance in this nascent democracy bodes well for its short-term prospects. Specifically, rather than target major political actors, Tunisians focus their intolerance largely on those groups agitating for social change from outside the formal political arena. To be sure, this is neither ideal nor sustainable if the long-term goal is broad-based pluralism. Yet, in terms of the more acute threat of authoritarian backsliding, it would be far less desirable if large swaths of the public wished to curtail the rights of entities participating in the (still quite fragile) political system. The experimental findings further buttress this result. When presented with counter-arguments for their initial position toward a group they strongly dislike, initially intolerant respondents were significantly more likely to be persuaded to tol- erance if their target group was a formal political actor rather than a social agitator. All respondents, however, were equally as likely to move away from an initially tol- erant position and, moreover, did so at a rather high rate (with nearly three in four swayed toward intolerance). Thus, even though the analysis indicates grounds for cautious optimism at this point in time, Tunisian attitudes appear quite malleable and therefore susceptible to stark shifts should conditions on the ground or elite discourse undermine the current equilibrium. 13 As interpretative lenses go, “change” can often be quite uninformative. The well- known dictum, “change is the only constant,” 7 speaks to this concept’s potentially limited utility. Yet, there is growing evidence that the demographic and attitudinal foundations of Western democracies are reconfiguring. 8 Similarly, political life in the Middle East has undergone unmistakable upheaval in recent years. In these times of flux, tolerance is often in short supply. This dissertation aims to better explain the dynamics of this democratic norm at a time when, arguably, it is most needed. 7. Heraclitus of Ephesus is often credited as the original source this saying. 8. For a review of the ways in which demographic change is affecting political attitudes and behaviors, see Klein 2018. 14 Chapter 2 Are Minorities (In)Tolerant Like Everyone Else?: American Muslims and the Enhanced Role of Personal Threat 2.1 Introduction Political theorists have long viewed intolerance as deleterious to democracy. From John Locke’s call to abandon the imposition of religious conformity to J.S. Mill’s ad- vocacy for a “marketplace of ideas,” the liberal foundations of tolerance in the public sphere are well-established. Yet, despite being a liberal democratic cornerstone, po- litical tolerance remains a particularly difficult norm to inculcate. It is perhaps due to 15 the gulf between its significance and its incidence that, for over sixty years, scholars have sought to better understand the dynamics that drive tolerant attitudes. Despite the breadth of literature on this topic, however, some gaps remain. Thus far, for example, the vast majority of studies stemming from Samuel Stouffer’s (1955) seminal work, Communism, Conformity, and Civil Liberties, wholly neglect the role of minority status (and the particular experiences and worldviews that accompany it) in moderating tolerance judgments. There are important differences between majority and minority citizens across various domains of public opinion (see, e.g., Kinder and Winter 2001; Kinder and Kam 2009; Peffley and Hurwitz 2010), and there is no reason to think that intolerance should constitute an exception to this tendency. Yet, studies on the determinants of tolerance in America almost always limit their inferences to the general population (often relying on General Social Survey (GSS) data), or otherwise draw upon samples representative of majoritarian characteristics (e.g., religious affiliation (Eisenstein 2006)). This lacuna in the literature is all the more conspicuous at a time when racial, ethnic, and religious minority groups, long the targets of intolerance, are themselves being charged with fomenting intolerant attitudes toward others. A recent FBI re- port, for instance, warned of the violent threat posed by “Black Identity Extremists” (Winter and Weinberger 2017), an assessment frequently promoted by conservative commentators and politicians regarding the Black Lives Matter (BLM) movement (Arrowood 2015; Phillips 2017). Moreover, although the Southern Poverty Law Center (SPLC) has clarified that BLM is not a hate group according to their cat- egorization (Cohen 2016), Figure 2.1 shows that “Black Nationalists” nonetheless 16 Figure 2.1: Hate Groups in America, 2017 constitute the largest proportion of active hate groups in the organization’s database (Southern Poverty Law Center 2018). 1 In addition to the charges leveled against voluntary associations, whole communi- ties are also alleged to harbor (or even actively promote) intolerance. This presumed association is perhaps most acute when American Muslims are invoked. Speaking directly to this perception, a 2010 poll found that only one-third of US respondents would characterize Muslims as “tolerant” (Pew Research Center 2011). 2 This senti- 1. To be sure, this distinction is largely due to a disaggregation of groups that preach White supremacy and racial bigotry. The SPLC’s combined tally for “Neo-Nazis,” “White Nationalists,” the “Ku Klux Klan,” and “Racist Skinheads” far surpasses the total for “Black Nationalists. ” More- over, the SPLC tracks only distinct groups and does not gather data on membership size, which likely further masks the true prominence of White Supremacist organizations. 2. This survey asked about Muslims generally, however data from other surveys demonstrate that the public differs little in their attitudes towards Muslims as opposed to Muslim-Americans (Sides and Gross 2013). 17 ment is found not only among the general public, but also elected officials (SAALT 2014; Schleifer 2016). Indeed, anti-Muslim rhetoric in political speech has spiked periodically during recent electoral cycles (Lajevardi 2016), indicating the potency and prevalence of these provocations. Taken together, these trends can have concrete policy implications, as is evident in the meteoric rise of anti-Sharia legislation across the US (Hatewatch 2017). It is, therefore, an opportune time to ask: Are minorities (in)tolerant like ev- eryone else? This question can be thought of as inquiring either about the relative degree of (in)tolerance or, alternatively, its determinants. Given the data I draw upon, my analysis will be limited to the latter. Specifically, I argue that, contrary to prior findings gleaned from studies of the general public, perceived threats to personal safety should significantly predict political intolerance among minorities. Analyzing original survey data from a national sample of Muslims in America, I find strong support for this hypothesis across various model specifications, support which is further bolstered by experimental findings. Thus, in addition to the typ- ically inhibiting effect of threats to one’s way of life, minority citizens must clear an additional psychological hurdle to express forbearance toward their “ideological” opponents. 3 Conversely, I also test whether two of the more robust predictors of tolerant atti- tudes, education and acculturation, have a similar impact among minority groups. I 3. To be sure, when it comes to minorities, the opposition they face is often not simply a matter of divergent ideologies, but is instead often predicated on exclusionary doctrines. That is, rather than a disagreement over, say, economic or social policy, the antipathy between minority groups and their opponents is frequently a function of the latter wishing to impose a hierarchical societal structure (whether on the basis of race, gender, or religion) to the detriment of the former. 18 find that education has a modest, positive effect on tolerance among the full sample, with a more pronounced effect among those respondents born in the United States. Along these same lines, years in residence significantly influence tolerance among foreign-born respondents. Taken together, these results run counter to the narrative that Muslims are not integrating into American society or that they seek to funda- mentally change, rather than adopt, its societal norms (Carroll and Jacobson 2015; Stump 2016). To situate the study’s findings, I first review the relevant political tolerance lit- erature. Subsequently, I outline a set of hypotheses, describe my sample, and detail the model specifications. After discussing the results, I conclude by considering the study’s implications and highlight fruitful avenues for future research. 2.2 Determinants of Political Tolerance The modern study of political tolerance arguably began with Stouffer’s (1955) foundational work in the midst of America’s Red Scare. Seeking to gauge whether Americans’ attitudes were in line with the country’s widespread political repression, Stouffer found that a wide swath of the general public was more than willing to limit the rights and freedoms of groups whose beliefs challenged their own. Among the dozens of studies that subsequently set out to model the determinants of polit- ical tolerance, those elaborating the role of threat perception, minority status, and religiosity inform much of the analysis in this chapter. 19 2.2.1 Threat Perception Perhaps the most consistent finding on political tolerance is the negative and powerful influence of threat perception. Indeed, Stouffer (1955) made this connection in his original study. In turn, numerous studies of varying empirical scopes, whether comprehensive (Sullivan, Piereson, and Marcus 1982; Marcus et al. 1995; Gibson and Gouws 2003) or bounded (Davis and Silver 2004; Sniderman, Hagendoorn, and Prior 2004), confirmed this finding. Notably, however, it is a particular type of threat that consistently influences whether or not one is tolerant of a particular group. Specifically, threats that target one’s way of life (i.e., those that are sociotropic) are a far more reliable and powerful determinant of intolerance than threats perceived to endanger one’s personal safety (i.e., those that are egocentric). Indeed, the latter are rarely found to exert any meaningful impact on tolerance levels. Gibson (2006, 25) explains this seemingly counterintuitive result by highlighting the foundations of political tolerance as a construct: Political tolerance is a social, not an individual attitude.... Intolerance increases not necessarily when people feel their own security is at risk, but rather when they perceive a threat to the larger system or group (or normative community) which they are a part of. Speaking directly to this point, Stouffer observed that Americans were more con- cerned with Communists’ subversive influence on society than they were with the prospect of espionage or infiltration (1955, 177-179). That this sentiment prevailed at a time when the fear of a nuclear conflict with the Soviet Union was ever-present 20 highlights the acutely social underpinnings of political intolerance (at least in this context). Americans were willing to limit the freedoms of communists largely due to the latter’s association with incendiary traits (arguably chief among them “godless- ness”) rather than any perceived personal danger they posed. While this particular processing of threats may be true of the general public, minorities may render their tolerance judgments via a different mental pathway. This alternate route is somewhat speculative at this point, however, given the dearth of studies examining anything other than majoritarian attitudes. 2.2.2 Minority Status Indeed, little attention has been paid to the role minority status may play as a moderator of political tolerance. One notable exception is Davis (1995), who conducted the only systematic analysis of intolerance among African Americans. His findings call into question the perception that, due to cultural and socio-economic disparities, Blacks in America are socially pre-disposed to authoritarian beliefs (Dahl 1956; Lipset 1960). Specifically, Davis concludes that Black intolerance is focused on the Ku Klux Klan, but is not discernibly different from typical levels of intolerance exhibited by non-Blacks toward “everyday racists” (1995, 12). This suggests that, at least among one racial minority in the US, intolerance is not a blanket and involuntary reaction stemming from cultural proclivities, but a conscious decision to secure the group from hatred and violence. Scant scholarship comparably elaborates the attitudes of other minority groups in America. In the main, these works only tangentially address minority mass atti- 21 tudes while focusing on a more particular population. Gibson and Bingham (1984), for example, address American Jews’ tolerance in light of an infamous legal dispute involving neo-Nazis marching in Skokie, Illinois, but these authors are mostly con- cerned with the role of elite organizations in framing civil rights conflicts. Likewise, Gibson (1987) analyzes a case involving a Ku Klux Klan rally in Houston, Texas to gauge homosexuals’ political tolerance, but his sample is drawn from a popula- tion of activists whose attitudes likely differ systematically from their non-activist counterparts. More comprehensively, Djupe and Calfano (2012) conducted the only academic study of tolerance–and one of the very few in any domain of public opinion–focused on American Muslims. This rarity makes it all the more notable (and normatively troubling) that these authors’ findings relate a largely negative tale. After prompt- ing their sample to consider the views of someone who (in the abstract) is against Christianity and Islam, respectively, Djupe and Calfano report a marked dearth of tolerance. Only about 30% of their American Muslim respondents chose the tolerant option in any of the three anti-Christian scenarios, while even fewer were tolerant of the same actions when directed against Islam: 20% would permit an anti-Muslim speech in their community, 9% would allow a person against Islam to teach at a uni- versity, and 6% would accept a book critical of Muslims in their local public library. Moreover, the authors find that both mosque attendance and scriptural literalism predict greater intolerance. These latter results highlight a final set of relevant de- terminants linking religious belief and practice to the expression of (in)tolerance. 22 2.2.3 Religiosity From the earliest studies, scholars have consistently correlated religious convic- tion and intolerance. Stouffer’s (1955) original results initially suggested that reg- ular churchgoers were generally less tolerant than those who attended services in- frequently. Subsequent studies have added greater nuance by examining additional dimensions of religious life, yet the overall association of “more religious” to “more intolerant” has remained remarkably stable (Nunn, Crockett, and Williams 1978; Sullivan, Piereson, and Marcus 1982; McClosky and Brill-Scheuer 1983). Despite these steady results, not all scholars are convinced of an invariably nega- tive correlation between religiosity and tolerance. Busch (1998) contends that prior findings, given their reliance on a fixed-group tolerance battery and blunt gauges of religiosity, are effectively measurement artifacts rather than genuine representations of underlying attitudes. In light of this critique, Eisenstein (2006) opts for content- controlled (or “least-liked”) political tolerance measures (Sullivan, Piereson, and Marcus 1982) 4 and couples her analysis with better specified religion variables. After also accounting for oft-neglected psychological variables in her sequential equation model, she concludes that neither “doctrinal orthodoxy” nor “religious commitment” has a direct effect on political tolerance. 5 More recently, Burge (2013) disentangled the effect of the “three B’s”–behavior, belief, and belonging–and found that bibli- 4. See section 2.4.1 for the distinction between a “fixed group” versus a “least liked” measurement approach. 5. Notably, however, Eisenstein (2006) does find an indirect effect for these two variables. Specif- ically, doctrinal orthodoxy increased threat perception, which in turn decreased tolerance, whereas religious commitment negatively correlated with a secure personality, the latter being a positive predictor of tolerance. 23 cal literalism decreases political tolerance, while more frequent church attendance actually exerts a countervailing positive effect. Thus, while the preponderance of ev- idence still suggests that religiosity (variably defined) negatively influences tolerance (Gibson 2010), this link does not appear to be as unyielding nor as comprehensive as once believed. 2.3 Hypotheses In light of the literature reviewed above, and the current political and social at- mosphere in the US, two central hypotheses guide this study. The first contends that American Muslim intolerance is, in part, a function of perceived egocentric threat from a target group. Here, I draw on Davis’ (1995, 17) conclusion that “Blacks, in order to defend and protect themselves from anxiety and fear that comes from racial hatred, consciously and selectively deny procedural rights to such groups that threaten their existence directly.” His findings thus implicitly underscore the height- ened role of threats to personal safety in African-American tolerance judgments. It is likely that this same dynamic operates among American Muslims given the level of hate crimes against this community, which has dramatically increased since the start of the 2016 presidential campaign (Williams 2017), and the concurrent rising prominence of hate groups in America (Southern Poverty Law Center 2018). H 1 : Political tolerance among minority groups significantly decreases as egocentric threat perception increases. Turning to the other side of the tolerance ledger, a second set of hypotheses test 24 whether the factors that typically enhance tolerance among the general public func- tion similarly within minority populations. Specifically, I examine whether higher education levels and acculturation to liberal democratic norms positively correlate with increased tolerance. Although immigrant status has not received much consideration in the tolerance literature, it is an increasingly relevant factor in analyses of political attitudes and behavior. At a macro level, prolonged experience with the “rough and tumble” of democratic governance enhances a society’s aggregate expression of democratic norms (Peffley and Rohrschneider 2003). At a micro level, however, the same process and effects cannot be taken for granted. Samuel Huntington, for example, famously lamented the findings of a study on Mexican-origin citizens, noting that “those more incorporated into mainstream society, the native-born Mexican-Americans, are less supportive of core American values than are the foreign-born” (2004, 242). Since the vast majority of Muslims in America are either first- or second-generation im- migrants, 6 there has been ample speculation (and skepticism) in recent decades over whether this community can truly integrate into American society. Therefore, I an- alyze the observed effect on tolerance levels associated with nativity and, among foreign-born respondents, years in residence. H 2a : Political tolerance significantly increases as education level increases. H 2b : Political tolerance is significantly higher among US-born respon- 6. The latest polling numbers find that about 75% of the American Muslims are either immigrants or the children of immigrants (Pew Research Center 2017). To be sure, however, African Americans constituted the majority of Muslims in the US for all the years prior to the Immigration and Nationality Act of 1965, which paved the way for successive waves of Muslim migrants from across the globe. 25 dents. H 2c : Political tolerance significantly increases among foreign-born respon- dents the longer their terms of residence in the US. 2.4 Data and Methodology To test the preceding hypotheses, I draw upon data from an original online survey of American Muslims. Recruiting a sample of respondents from a low-incidence population poses myriad challenges (Berry, Chouhoud, and Junn 2016). Appendix 2.A details the strategy used to address one glaring threat to internal validity, namely the possibility of respondents falsely identifying as Muslim in order to participate in the survey. With regards to external validity, efforts were made at both the collection and analysis stages to ensure that the survey sample did not stray too far from known American Muslim demographics. In the first instance, quotas for age and education were set to guard against dramatic skews from baseline population estimates. No- tably, these target proportions could not be too constraining given the already shal- low pool of potential respondents. As a result, the final unweighted sample of 682 respondents still tilts younger, more educated, and more female, as is common with online polling (see, e.g., Heckathorn 1997; Roster et al. 2004; Fricker et al. 2005). 7 To partially correct for this bias, I calibrated the data to the Pew (2017) proportions for American Muslims’ age, education, and sex. The resulting weights ranged from 7. See Table 2.A.1 in Appendix 2.A for a comparison of the unweighted sample’s key marginals with those found in the most recent Pew (Pew Research Center 2017) poll of American Muslims. 26 .23 to 4.95. 2.4.1 Dependent Variable Contrary to the fixed-group approach, which gauges one’s willingness to limit the civil liberties of a static (if presumably salient) set of associations, I employ a content- controlled measure of political tolerance (Sullivan, Piereson, and Marcus 1982). The survey first asks respondents to indicate their least-liked group (as well as another strongly disliked group from the remaining options) and subsequently populates the tolerance battery with their choices. For an added layer of comparison, the survey randomly assigned respondents to consider extending tolerance toward either their first or second most disliked group. The least-liked options comprised an equal number of groups derived from both ends of the ideological spectrum. Figure 2.2 displays respondents’ group choice by assignment. White Supremacists were far and away the top choice among those assigned to consider their least-liked group. The variance for the “Second Choice” assignment was much narrower, although White Supremacists still maintained the top spot. Political tolerance in the sample is gauged through four questions. Respondents were asked the extent to which they would support: 1) banning their chosen group from running for office; 2) putting their chosen group under government surveillance; 3) allowing their chosen group to hold a rally in their city [REVERSE CODED]; and 4) banning a member from their chosen group from speaking at a local college. Possible response options ranged from “Strongly Disagree” to “Strongly Agree” along 27 Figure 2.2: Least-Liked Group Choice (Percent of Total), by Assignment 7 points. The answers to these four questions formed the additive tolerance scale (a=.72). 2.4.2 Key Independent Variables and Controls The core explanatory variable for the first hypothesis is also tied to the respon- dents’ assigned group. Egocentric Threat is tapped through two questions: 1) [AS- SIGNED GROUP] pose a threat to my own and/or my family’s personal safety; and 2) If [ASSIGNED GROUP] had their way, they would reduce my own and/or my family’s personal freedom. Sociotropic Threat similarly scales two indicators: 1) [ASSIGNED GROUP] are a threat to our society’s way of life; and 2) [ASSIGNED GROUP] are dangerous to the lives of normal people in this country. 28 Three variables are used to test the second set of hypotheses. Both College Grad- uate and Born in the US are dichotomous variables, indicating whether respondents earned a college degree or are native born, respectively. For the sample’s foreign- born, Years in the US measures the length of their residence (grouped into four categories with the following ranges in years: 0-5, 6-15, 16-30, 30+). Beside these key variables of interest, the full model of political tolerance includes three sets of controls. First, two multi-item indicators capture the psychological dispositions that tend to depress tolerance levels. Dogmatism indicates the extent to which respondents hold a closed-minded view of the world, tending to categorize social stimuli in an unnuanced manner. Along these same lines, Order vs. Freedom taps the inclination to limit liberty when granting it may undermine social order. Prior studies have found that both these cognitive states negatively correlate with political tolerance (Rokeach 1973; Sullivan, Piereson, and Marcus 1982; Marcus et al. 1995; Eisenstein 2006). Second, the questionnaire features a religiosity battery gauging Mosque Atten- dance, Importance of Religion (in respondents’ lives), and Frequency of Prayer, draw- ing on the standard dimensions of behavior, belief, and belonging (with adjustments made to fit the phrasing within an Islamic frame). A scale of Quran Literalism sim- ilarly accords with the usual factors tested in tolerance studies, but goes beyond the typical dichotomy. Respondents are asked to choose which of the following comes closest to their belief: 1. The Quran is the actual word of God and should be taken literally, word for word 29 2. The Quran is the actual word of God, but has some content that is merely symbolic 3. The Quran is an ancient book of history and moral guidance authored by men A final set of demographic controls–namely, Age, sex (Female), and Race/Ethnicity–complete the full model specification. The question wording for all independent variables not already elaborated can be found in Appendix 2.B. 2.4.3 Experimental Analysis To further probe the dynamics of political tolerance among American Muslims, the survey includes a set of experimental manipulations. Once respondents have completed the tolerance battery, they are presented with two follow-up questions conditional on their willingness to allow a political rally for their disliked group. Specifically, those who offered an initially intolerant response were provided with arguments in favor of a more tolerant attitude and vice versa. By gauging the pliability of respondents’ tolerance judgments, we can better understand whether and what kind of engagement may shift American Muslims away from their original position when presented with these scenarios in their everyday lives. Acknowledging that the attitudes recorded in any one survey are not necessarily immutable is an important component of political science research in general, but is especially pertinent when examining the dynamics of tolerance. One of the more prominent models of opinion formation posits that survey responses emerge from a sampling of (at times, opposing) considerations that happen to be momentarily 30 salient (Zaller and Feldman 1992). This salience, in turn, is not simply a result of happenstance, but is often driven by issue framing and the consequent activation of particular values (Sniderman and Theriault 2004; Chong and Druckman 2007). Put differently, as Sniderman et al. (1996, 55) assert: “Where people start off politically matters, but what counts is where they wind up after the pushing and shoving of political argument.” This dictum is all the more apt when the subject is political tolerance. After all, if an individual’s initial inclination is toward intolerance, as some scholars suggest (Marcus et al. 1995), then tolerant attitudes will often emerge only after citizens give the matter what Stouffer famously dubbed a “sober second thought” (1955, 223). By that same token, however, an initially tolerant position may abate in the face of heightened threat salience. The survey featured three different sets of prompts to assess American Muslims’ persuasibilty. Following their completion of the tolerance battery, respondents were presented with counter-arguments based on their initial attitude toward a disliked group holding a public rally in their neighborhood. Within each of these conditions, the argument primed considerations of social order and rights reciprocity, respec- tively. The wording was as follows: 1) Those who were initially tolerant: Earlier, you responded that [ASSIGNED GROUP] should be allowed to hold a public rally in your town or city. Such rallies, however, could attract disorder. Suppose that community leaders warn that there would be a heightened risk of violence between 31 the organizers’ supporters and opponents, would you still feel that the rally should be allowed? Some argue that [ASSIGNED GROUP] should not be allowed to hold public rallies because they WOULD NOT be willing to extend those same liberties to those that oppose them. Considering this argument, do you still feel that the rally should be allowed? 2) Those who were initially intolerant: Earlier, you responded that [ASSIGNED GROUP] should NOT be al- lowed to hold a public rally in your town or city. What if local community leaders assured the public that there would be extra security to minimize the risk of violence between the organizers’ supporters and opponents? Given this assurance, would you still feel that the rally should NOT be allowed? Some argue that [ASSIGNED GROUP] should be allowed to hold public rallies because limiting the rights of those whom we disagree with may motivate others to limit OUR rights. Considering this argument, would you still feel that the rally should NOT be allowed? 3) Those who were initially neutral: 32 Earlier, you responded that you were not sure whether [ASSIGNED GROUP] should be allowed to hold a public rally in your town or city. What if local community leaders assured the public that there would be extra security to minimize the risk of violence between the organizers’ supporters and opponents? Given this assurance, would you feel that the rally should or should not be allowed? Some argue that [ASSIGNED GROUP] should not be allowed to hold public rallies because they WOULD NOT be willing to extend those same liberties to those that oppose them. Considering this argument, would you feel that the rally should or should not be allowed? 2.5 Results and Discussion For ease of interpretation, all variables (including political tolerance) in the anal- ysis are re-scaled from 0-1. In effect, this means that the coefficients in the OLS regression correspond to the mean point difference on the tolerance scale between the lowest and highest values on any given variable, holding all other variables con- stant. Beginning with a descriptive breakdown of the data, Figure 2.3 presents the mean level of political tolerance in the sample by assignment and chosen group. As we would expect, tolerance is significantly lower among those who were asked 33 Figure 2.3: Political Tolerance, by As- signment and Chosen Group Figure 2.4: Egocentric Threat, by As- signment and Chosen Group about their least-liked group compared to those asked about another group they deeply dislike. Similarly evident is the distinction between subjects of intolerance; White Supremacists are significantly less likely to elicit a tolerant response when compared to all other groups. Figure 2.4 points to a likely explanation for this trend. Comparing levels of threat perception across groups and assignments, a similar pattern emerges with respondents fearing for their personal safety and freedoms far more when White Supremacists are invoked. This finding, too, is fairly intuitive given that the first year of Donald Trump’s presidency not only featured a resurgence of White Supremacist public presence, but also a marked rise in violence credited to these groups (Greenblatt 2018). Turning to the OLS regression analysis, 8 the relationship between personal threat and intolerance among minority groups is brought into starker relief. Table 2.1 demonstrates a statistically and substantively significant negative effect for Egocen- tric Threat across multiple model specifications. To be sure, the magnitude of this 8. Although Race/Ethnicity and Choice Assignment are included in each of the following regres- sion models, these variables are excluded from the tables for ease of presentation. 34 impact is somewhat blunted once Sociotropic Threat is taken into account. Nonethe- less, a pattern of perceived personal threat weakening tolerance is clearly evident. While the role of education and nativity will be covered in detail in the next section, it is worth pausing here to consider the other determinants of American Muslim political tolerance. Firmly in line with prior research, those who are more dogmatic in their outlook and, relatedly, have a preference for social order over indi- vidual freedom are both significantly less tolerant than their counterparts. In terms of demographics, women are less tolerant than men, a fairly consistent finding in tolerance studies (Sullivan et al. 1985; Gibson 2013), while Age demonstrates no significant effect. Although Age is generally not a meaningful predictor of tolerance, future research on American Muslims should be mindful of a possible cohort effect taking root. Those whose attitudes were shaped during the dozen or so years follow- ing the attacks of 9/11 may be systematically different from those who came of age during the Trump presidency and its aftermath. Given the abiding suspicion of Muslims’ commitment to democratic norms (see Section 1.2.2), it is notable that religiosity has little influence on their expression of political tolerance. Of the four measures included in the full model, only Quran Literalism is a significant predictor. Those who believe that Muslim scripture should be interpreted literally tend to be more intolerant. Notably, Mosque Attendance exhibits an equally strong, if not quite statistically significant, countervailing positive effect on tolerance. This finding is in line with recent research highlighting the integrative role of mosques in America (Jamal 2005; Dana, Barreto, and Oskooii 2011; Schoettmer 2015). 35 Table 2.1: Determinants of Political Tolerance–Full Sample Full Model (w/o Soc Threat) Full Model (w/o Religiosity) Full Model (w/o PD) Full Model Threat Egocentric Threat −0.51 ∗∗∗ −0.13 ∗ −0.12 ∗ −0.14 ∗ (0.03) (0.06) (0.06) (0.06) Sociotropic Threat −0.46 ∗∗∗ −0.52 ∗∗∗ −0.46 ∗∗∗ (0.06) (0.06) (0.06) Democratic Learning College Graduate 0.03 0.03 0.03 0.02 (0.02) (0.02) (0.02) (0.02) Born in the US 0.01 0.01 0.02 0.01 (0.02) (0.02) (0.02) (0.02) Psych Dispositions Dogmatism −0.11 ∗∗ −0.09 ∗ −0.08 ∗ (0.04) (0.03) (0.04) Order vs. Freedom −0.24 ∗∗∗ −0.21 ∗∗∗ −0.21 ∗∗∗ (0.04) (0.04) (0.04) Religiosity Quran Literalism −0.04 −0.08 ∗∗ −0.05 ∗ (0.03) (0.02) (0.02) Mosque Attendance 0.06 ∗ 0.04 0.05 (0.03) (0.03) (0.03) Prayer Frequency −0.02 −0.02 −0.02 (0.03) (0.03) (0.02) Importance of Religion 0.01 −0.00 0.00 (0.04) (0.04) (0.04) Demographics Female −0.04 ∗ −0.05 ∗∗∗ −0.04 ∗ −0.04 ∗∗ (0.02) (0.01) (0.02) (0.02) Age 0.02 0.01 0.02 0.01 (0.02) (0.02) (0.03) (0.02) Constant 0.86 ∗∗∗ 0.92 ∗∗∗ 0.84 ∗∗∗ 0.95 ∗∗∗ (0.05) (0.04) (0.04) (0.04) N 680 681 680 680 Standard errors in parentheses ∗ p< 0.05, ∗∗ p< 0.01, ∗∗∗ p< 0.001 36 Returning to the influence of interpersonal threat, the split-sample analysis in Table 2.2 takes a deeper look at the role of societal marginalization in moderating tolerance. Effectively modeling the interaction of a binary identifier with each predic- tor in the full model, this analytical approach is driven by the theoretical expectation that intersecting minority identities will alter the impact of additional covariates in the tolerance model beyond just Egocentric Threat. Some scholars of race and ethnic- ity have explicitly employed this methodology based on the same logic (e.g., Jamal 2005; Chong and Kim 2006), but many others implicitly rely on this technique when drawing inferences from samples composed exclusively of a particular sub-population to determine whether and how they differ from findings derived from the general pub- lic. In their comprehensive assessment of attitudes towards immigration, Masuoka and Junn (2013) vigorously advocate for what they call a “comparative relational” analytical approach when theoretically appropriate. By estimating models for each racial or ethnic group separately, they argue, a researcher can better represent how hierarchical identities structure multiple explanatory variables compared to the tra- ditional means of accounting for these identities within a single model and assuming an “equality of individual agency” (32). I concur with these scholars, and apply the comparative relational approach to examine the influence of several theoretically relevant identities. This analysis under- scores that although all the respondents in this survey are Muslim, not all Muslims experience the discrimination aimed at this community to an equal degree. Indeed, as Table 2.2 demonstrates, when comparing those who are more marginalized to those who are more societally privileged, the effect of personal threat on tolerance 37 Table 2.2: Determinants of Political Tolerance–Split Samples US Born Foreign Born Whites Non-Whites Males Females Threat Egocentric Threat −0.10 −0.20 ∗ −0.09 −0.14 ∗ −0.08 −0.18 ∗ (0.07) (0.09) (0.13) (0.06) (0.09) (0.07) Sociotropic Threat −0.51 ∗∗∗ −0.41 ∗∗∗ −0.41 ∗∗ −0.50 ∗∗∗ −0.59 ∗∗∗ −0.36 ∗∗∗ (0.08) (0.09) (0.13) (0.06) (0.08) (0.07) Demographics Age −0.00 0.03 0.01 0.01 0.01 0.02 (0.04) (0.03) (0.06) (0.03) (0.03) (0.04) Female −0.04 ∗ −0.04 −0.04 −0.04 ∗ (0.02) (0.02) (0.03) (0.02) Democratic Learning College Graduate 0.05 ∗ −0.01 −0.01 0.04 ∗ 0.02 0.02 (0.02) (0.02) (0.03) (0.02) (0.02) (0.02) Born in the US −0.03 0.00 −0.00 0.01 (0.04) (0.02) (0.03) (0.02) Psych Dispositions Dogmatism −0.04 −0.16 ∗∗ 0.01 −0.10 ∗ −0.05 −0.11 ∗ (0.05) (0.06) (0.08) (0.04) (0.05) (0.05) Order vs. Freedom −0.25 ∗∗∗ −0.12 ∗ −0.23 ∗∗ −0.19 ∗∗∗ −0.17 ∗∗ −0.25 ∗∗∗ (0.04) (0.06) (0.08) (0.04) (0.05) (0.05) Religiosity Quran Literalism −0.06 −0.02 −0.05 −0.05 −0.06 −0.04 (0.03) (0.04) (0.05) (0.03) (0.03) (0.03) Mosque Attendance 0.05 0.03 0.07 0.05 0.05 0.04 (0.04) (0.04) (0.05) (0.03) (0.04) (0.04) Prayer Frequency −0.02 −0.02 0.02 −0.03 −0.03 0.00 (0.03) (0.03) (0.05) (0.03) (0.04) (0.03) Importance of Religion −0.02 0.05 −0.01 0.00 0.05 −0.06 (0.05) (0.05) (0.08) (0.04) (0.05) (0.05) Constant 0.95 ∗∗∗ 0.93 ∗∗∗ 0.86 ∗∗∗ 0.95 ∗∗∗ 0.95 ∗∗∗ 0.91 ∗∗∗ (0.06) (0.06) (0.09) (0.04) (0.06) (0.05) N 402 278 184 496 275 405 Standard errors in parentheses ∗ p< 0.05, ∗∗ p< 0.01, ∗∗∗ p< 0.001 38 levels markedly shifts. Whether comparing US-born to foreign-born (columns 1 and 2), Whites to non-Whites (columns 3 and 4), or men to women (columns 5 and 6), the pattern is identical: the threat of personal harm decreases tolerance only among those who are more vulnerable. The same tendency does not hold for Sociotropic Threat, however, signifying that the processing of group-level risk among Muslims is fairly uniform, even as the effect of interpersonal threat varies. Another consist pattern that emerges in Table 2.2 is that Dogmatism is negatively correlated with tolerance only among the more marginalized in each pairing. This suggests that within these more vulnerable sub-groups, individuals may feel that they do not have the luxury of considering threats in a nuanced manner, and so render their tolerance judgments in more zero-sum terms. To further examine this finding, Figure 2.5 models a three-way interaction between Egocentric Threat, Dogmatism, and the relevant binary identifiers. The pattern in each interaction plot aligns with that in the regression table. The effect of Egocentric Threat on tolerance is enhanced as Dogmatism increases, but only among the more vulnerable in each dyad. Figure 2.5: Marginal Effect of Egocentric Threat, by Dogmatism 39 Indeed, for minorities who feel targeted by particular groups, the psychological response could be “defensive intolerance. ” This reaction can be thought of as the tendency to re-conceptualize the (ideally) infinite game of rights reciprocity as, in- stead, a finite game predicated on preservation. That is, rather than fidelity to an abstract principle of respect for civil liberties, which would yield some indirect long- term marginal benefit, minorities may prefer to guard against what they perceive as a more acute and existential short-term risk. Put differently, for targeted minori- ties, expressing tolerance in certain situations would require that they prioritize a commitment to democratic norms over their own immediate security. It is worth emphasizing again, however, that not all interactions with disliked groups undergo this distinct processing. Three related factors structure American Muslim intolerance: 1) a tendency to choose White Supremacists as a least-liked group, 2) a higher likelihood of expressing intolerance toward this group compared to other strongly disliked choices, and 3) an abiding perception that the KKK and their ilk are personally threatening (see Figures 2.2, 2.3, and 2.4). On this latter point, fully 42% of those asked about White Supremacists reported the highest level of Egocentric Threat on the scale, which is more than three times the proportion among those asked about all other groups. Taken together, these points underscore that intolerance among minorities is not an involuntary impulse, but a considered response. Presented with groups that are merely ideologically threatening (atheists, for example) there is no a priori reason to expect Muslims to be significantly more intolerant that a sample of the general public. The heterogenous pliability of American Muslims’ political tolerance further 40 demonstrates this selective processing at work. Figure 2.6 displays the results of the persuasibility experiment. The overall finding that respondents were more likely to change their original attitude if it was initially tolerant accords with prior re- search, although this sample’s initial judgments appear to be stickier than those in other studies (see, e.g., Marcus et al. 1995; Gibson 1996). Another apparent pattern is the nearly identical average effect of the social order prompt relative to the reci- procity prompt regardless of initial attitude or group choice. The reason for this may be that the response options were not sufficiently gradated to provide for adequate variation or it could be a function of the prompts appearing one after another. In any case, it seems prudent to withhold final judgment on whether particular arguments are more or less effective in persuading minority respondents toward (in)tolerance. Figure 2.6: Proportion of Respondents Changing Initial Tolerance Judgment Another key conclusion from the experimental data is the stubbornness of in- 41 tolerance toward White Supremacists. Whereas those who were asked about other highly disliked groups were just as likely to change their attitude regardless of their initial judgment, respondents asked about White Supremacists were significantly less likely to shift from an intolerant position, compared both to their initially tolerant counterparts and to all others with an initially intolerant response. Thus, not only does perceived personal threat drive intolerance toward White Supremacists, but the resulting attitude is also quite stubborn. This pattern holds even among those who were initially neutral in their judg- ments. Although the sample size precludes the observation of significant differences (the number of respondents in this condition totaled 151), the distributions in Figures 2.7 and 2.8 demonstrate that it is at least marginally easier to persuade ambivalent respondents toward intolerance when considering the rights of White Supremacists while at the same time being harder to convince them to tolerate these groups. Taken together, then, the experimental and observational data appear to support Hypoth- esis 1, even if the evidence oscillates between compelling and merely suggestive. Turning to the second hypothesis examining the factors that typically augment tolerance levels, the results are not as clear cut. As shown in Table 2.1, both College Graduate and Born in the US are marginally positive across all the models, although only the former crosses into statistical significance. However, when we examine the effect of earning a college degree conditioned on nativity, a different pattern emerges. Table 2.3 shows that among foreign-born respondents, graduating from college has no effect on political tolerance, whereas earning a post-secondary degree significantly increases the tolerance of US-born Muslims. This finding (confirmed through both 42 Figure 2.7: Attitude Change among the Initially Neutral – Tolerance Prompt Figure 2.8: Attitude Change among the Initially Neutral – Intolerance Prompt 43 split sample analysis and an interaction model) indicates that it is not simply the level of educational attainment, but the broader context in which that education is attained that matters for democratic norm adoption. With regard to nativity as a standalone predictor, although being born in the United States does not, in itself, push one toward more tolerant attitudes, this finding masks a trend that is potentially just as normatively positive. Specifically, Table 2.4 demonstrates that, among foreign-born Muslims, Years in the US has a strongly correlates with political tolerance. On the one hand, this result further challenges the notion that Muslims are culturally pre-disposed to intolerance and unable to adopt America’s democratic norms. Those Muslim immigrants more acculturated to American society appear to exhibit a higher degree of tolerance. On the other hand, there could be a more normatively ambiguous interpreta- tion of this result. Specifically, tolerance among foreign-born Muslims in America may be a function of the target groups that are most salient to this subpopulation. Indeed, Figure 2.9 demonstrates that although respondents born outside the US choose White Supremacists at a rate comparable to their native born counterparts, the former are significantly less likely to choose groups on America’s ideological fringe (namely, Christian Fundamentalists and The Tea Party) as among their least-liked. Conversely, a higher percentage of foreign-born Muslims chose Atheists (16% vs 11% among US-born respondents) and Extremist Muslims (17% vs 8% among US-born respondents). Given these systematic differences, a less charitable explanation of the headline finding in Table 2.4 would be that immigrant Muslims are not more tolerant, per 44 Table 2.3: Conditional Effect of Education on Political Tolerance Foreign Born US Born Full Sample College Graduate −0.01 0.05 ∗ −0.01 (0.02) (0.02) (0.02) Born in the US −0.01 (0.02) College Grad X US Born 0.05 † (0.03) Threat Egocentric Threat −0.20 ∗ −0.10 −0.14 ∗ (0.09) (0.07) (0.06) Sociotropic Threat −0.41 ∗∗∗ −0.51 ∗∗∗ −0.47 ∗∗∗ (0.09) (0.08) (0.06) Psych Dispositions Dogmatism −0.16 ∗∗ −0.04 −0.08 ∗ (0.06) (0.05) (0.04) Order vs. Freedom −0.12 ∗ −0.25 ∗∗∗ −0.20 ∗∗∗ (0.06) (0.04) (0.04) Religiosity Quran Literalism −0.02 −0.06 † −0.05 ∗ (0.04) (0.03) (0.02) Mosque Attendance 0.03 0.05 0.05 † (0.04) (0.04) (0.03) Prayer Frequency −0.02 −0.02 −0.02 (0.03) (0.03) (0.02) Importance of Religion 0.05 −0.02 0.00 (0.05) (0.05) (0.04) Demographics Female −0.04 † −0.04 ∗ −0.04 ∗∗ (0.02) (0.02) (0.02) Age 0.03 −0.00 0.01 (0.03) (0.04) (0.02) Constant 0.93 ∗∗∗ 0.95 ∗∗∗ 0.96 ∗∗∗ (0.06) (0.06) (0.04) N 278 402 680 NOTE: Race was included in the analysis but excluded from the table for ease of presentation. † p< 0.10, ∗ p< 0.05, ∗∗ p< 0.01, ∗∗∗ p< 0.001 45 Table 2.4: Effect of Immigrant Socialization on Tolerance Democratic Learning Years in the US 0.09 ∗∗ (0.03) College Graduate −0.00 (0.02) Threat Egocentric Threat −0.20 ∗ (0.09) Sociotropic Threat −0.40 ∗∗∗ (0.09) Psych Dispositions Dogmatism −0.15 ∗∗ (0.06) Order vs. Freedom −0.10 † (0.06) Religiosity Quran Literalism −0.01 (0.04) Mosque Attendance 0.02 (0.04) Prayer Frequency −0.01 (0.03) Importance of Religion 0.04 (0.05) Demographics Female −0.04 (0.02) Age −0.02 (0.03) Constant 0.88 ∗∗∗ (0.06) N 273 Standard errors in parentheses † p< 0.10, ∗ p< 0.05, ∗∗ p< 0.01, ∗∗∗ p< 0.001 46 Figure 2.9: Disliked Group Choice, by Nativity se, the longer they stay in America, but that they are not sufficiently updating their set of disliked groups. This alternative assertion somewhat mirrors the argument that Americans have not necessarily become more tolerant toward, say, communists over the years, but rather that this particular group has simply become less salient and thus antipathy toward them has dissipated (i.e., the objection pre-condition no longer holds). In the case of Muslim immigrants, the salience of certain groups, seemingly as a function of their background from a (presumably) Muslim-majority context, appears steady. US-born Muslims, on the other hand, appear better attuned to more temporally (and, arguably, geographically) proximate group threats. Just as political knowledge makes one more susceptible to motivated reasoning, targeted minorities who have experienced social and political discrimination from birth seem better able to process societal threat cues, which in turn increases their propensity for defensive intolerance. 47 2.6 Conclusion What then, are we to make of American Muslim intolerance? The preceding findings indicate that intolerance in this community is not a socially embedded trait, invariably operative when a disliked group is invoked. Instead, this disposition is more akin to a latent attitude that can get activated under the right conditions– which is to say that Muslims function psychologically in the same way as everyone else. Where they (and, I argue, other targeted minority groups) do differ is in the particular considerations that determine their tolerance judgments. Broadly speaking, Muslims are particularly driven to intolerance when faced with the prospect of interacting with groups that pose a distinct threat to their personal safety. This dynamic is fairly intuitive, given the increased prevalence of anti-Muslim rhetoric and behavior since the start of the 2016 presidential campaign (Levin 2017; Johnson and Hauslohner 2017; Crunden 2017), yet it is also fairly atypical. That is, studies of tolerance among the general public find that threats to one’s “way of life” drive up intolerance, but consideration of personal safety factors little. Why, after all, would a member of the majority White, majority Christian public fear for their personal safety from a member of the Ku Klux Klan or a Neo-Nazi (two perennial “least-liked” groups in academic studies of tolerance)? Yet, one fruitful extension of this research would be to examine whether egocen- tric threat may now similarly influence political tolerance among (at least certain segments of) the general public. One reason for this shift in the determinants of tolerance would be a commensurate realignment of the types of groups that are particularly salient given the current political landscape. Whereas Stouffer (1955) 48 rightly sought to gauge Americans’ tolerance in the 1950s by measuring their atti- tudes toward communists and atheists, such groups are fairly anachronistic political threats today, and so it is no wonder that the general public has exhibited increas- ing “tolerance” 9 toward them (Twenge, Carter, and Campbell 2015). Today, what one regards as extreme and worthy of intolerance may largely be motivated by a respondent’s attitudes toward racial, ethnic, religious, and sexual minorities. Indeed, social tolerance may be partially structuring political tolerance, though its effects may vary. On the one hand, we would expect the more socially tolerant to be much more politically intolerant of bigoted groups. Trends in GSS data show that respondents who are more liberal (who also tend to be more socially tolerant) have generally shown increased political tolerance over the years, except toward racists (Yglesias 2018). This pattern also aligns with Harell’s (2010) finding that youth with more diverse networks tend to generally be more politically tolerant, except when the subject of their tolerance judgment is a group with an explicitly exclusionary ideology. On the other hand, those who hold more prejudiced attitudes towards minorities would likely hold politically intolerant views of groups that are readily associated with minority communities (e.g., Black Lives Matter or “fundamentalist” 10 Muslims). This ordering would run parallel to research emerging in the wake of the 2016 presidential election outlining the impact of racial resentment in conditioning 9. Although researchers continue to use the GSS to measure tolerance, it is difficult to ascertain whether or not they are instead measuring some function of indifference (see discussion in Section 1.2.1). 10. Admittedly, the adjectives used to signal a particular ideological reading of Islam are quite fraught. The GSS offers respondents a lengthy description (“Muslim clergymen who preach hatred of the United States”) while Gibson (2008) opts for the shorthand “Radical Muslims. ” Without a doubt, more research is needed to determine what considerations are primed with each formulation. 49 political attitudes (Lopez 2017). Egocentric threat could prove more predictive among those whose social toler- ance highly correlates with their political tolerance, regardless of the correlation’s directionality. At one end of the spectrum, the socially tolerant could perceive their safety to be threatened by White Supremacists, even if they themselves are not a minority, and thus feel more inclined to limit the rights of these groups. After all, the one fatality in the 2017 Charlottesville rally, Heather Heyer, was a White woman and the vast majority (74%) of murders by domestic extremists from 2007-2016 were at the hands of right-wing groups. The socially intolerant would also likely feel personally threatened by groups they highly dislike. The shootings perpetrated by Black nationalists in recent years (Johnson 2017) coupled with the protest violence that has become a hallmark of ANTIFA and like-minded anti-racist groups (Beinart 2017) doubtless informs this potential sense of threat. The equivalence of perceived threats to personal safety from White Supremacists, on the one hand, and Black na- tionalists, on the other, is certainly debatable, yet the operative point is that, once again, a better contextualizing of our analyses could reveal deeper insights into the dynamics of tolerance. The belief that Muslims are particularly intolerant relies in part on a faulty expectation that this targeted minority group assesses the costs associated with being tolerant in the same way as any other member of the broader society. Thus, when American Muslims fall short of this standard–one that is divorced from their lived experience–skeptical observers can explain this deficiency with reference to some embedded disposition in the community. This tendency to discount extenuating 50 circumstances in lieu of a presumably more fundamental attribute speaks to how uncharitable assessments of minority behavior can be, especially when compared to the more generous appraisals socially privileged groups often receive. 11 Ultimately, the findings presented here further underscore the need to jettison such blanket, unsubstantiated claims about Muslims’ (and numerous other minority communities’) “nature” and instead focus on the contexts that shape their attitudes and actions. 11. The tendency to more likely ascribe religious intent to violent acts committed by Muslims as opposed to white non-Muslims is one way in which this prejudiced processing manifests (Mercier, Shariff, and Norris 2017). 51 Appendices 2.A Sampling Methodology To populate the sample, Qualtrics solicited the participation of respondents who had indicated that they were Muslim on previous surveys. The invitation itself was generic and made no mention of the specific topic. The second stage of this double opt-in process (i.e., the fielding of this survey) took place from August 25, 2017- October 16, 2017. Potential respondents were once again asked their religion with those indicating that they were Muslim invited to continue. A battery of religious knowledge questions provided an additional check on the sample’s fidelity. While there is no readily available benchmark to compare with these responses, the results in Figure 2.A.1 do not appear to raise any acute concerns. That is, a relatively high percentage of the sample was able to correctly answer questions on fairly basic items, while a lower proportion (though still clear majority) responded correctly to the slightly more intermediate-level item asking about the Prophet Muhammad’s destination when he made the migration out of Mecca. 52 Figure 2.A.1: Proportion Correctly Answering Religious Knowledge Questions Table 2.A.1: Comparison of Key Unweighted Marginals Marginal Current Study Pew (2017) % Female 59 49 % College graduate 49 32 % Less than age 30 48 35 % Born in the US 59 42 % Attending mosque weekly 31 42 % Never wearing hijab 48 42 53 2.B Question Wording Variable (type) Composite Question(s) Answer Options Order vs Freedom (index) It is better to live in an orderly society than to allow people so much freedom that they can become disruptive. Strongly Disagree – Strongly Agree (5pt) Society shouldn’t have to put up with those who have political ideas that are extremely different from the majority. Strongly Disagree – Strongly Agree (5pt) Public hate speech should be banned; it should not be protected under free speech laws. Strongly Disagree – Strongly Agree (5pt) No matter what a group’s political beliefs are, they are entitled to the same legal rights and protection, as anyone else. Strongly Disagree – Strongly Agree (5pt) Dogmatism (index) There are two kinds of people in this world: those who are for the truth and those who are against it. Strongly Disagree – Strongly Agree (5pt) A group that tolerates too many differences of opinion among its own members cannot exist for long. Strongly Disagree – Strongly Agree (5pt) To compromise with our political opponents is dangerous because it usually leads to the betrayal of our own side. Strongly Disagree – Strongly Agree (5pt) Mosque Attendance (single indicator) Aside from weddings or funerals, how often do you go to a mosque? Almost never; A few times a year; Once or twice a month; once a week; more than once a week Prayer Frequency (single indicator) In general, how often do you pray salah or namaz (formal prayer)? Almost never; Only during Eid; Once or twice a month; Once or twice a week; Daily Importance of Religion (single indicator) How important is religion in your life? Not at all important; Not too important; Somewhat important; Very important 54 Chapter 3 (When) Are Muslims (In)Tolerant?: The Conditional Expression of Democratic Norms in Tunisia and Egypt 3.1 Introduction The enduring debate over whether the religion of Islam is fundamentally at odds with democracy hinges, in part, on a static conception of this association. Those doubting that Muslims and/or Muslim-majority nations can ever be fully demo- cratic believe that there is something intrinsic to Islam that bars its adherents from sufficiently embracing democracy’s core values. Proponents of this position hold that 55 the “effect” 1 of Islam (variably defined and often either bluntly measured or simply assumed) on democratic norms is negative and constant, both across populations and over time (Kedourie 1992; Huntington 1993). Other scholars of history and re- ligion counter such skepticism by arguing that there exists no basis in Islam for an irresolvable antagonism with democracy (Lawrence 2000; Esposito 2003). Therefore, according to this line of thought, any observed negative relationship at either the micro or macro level is better understood as a function of path dependencies brought on by foreign intervention, or regional pathologies (particularly in the Arab Middle East). Neither of these views, however, leave much theoretical space for adherence to Islam exerting a conditional effect on democratic attitudes. That is, as a result of interactions with more proximate social and political factors, religious variables among Muslims may either positively or negatively correlate with the expression of democratic norms, or demonstrate no discernible impact at all. This dynamic perspective on the role of Islam in motivating political attitudes and behavior is all the more relevant given the recent spread of political uncertainty and flux across the Middle East and North Africa (MENA). To empirically examine this alternative approach to the Islam-democracy debate, I use data from the second and third waves of the Arab Barometer to map the determinants of political tolerance, a foundational democratic value, in a Muslim- majority setting. I focus on Tunisia and Egypt, two countries at the heart of the 1. To be sure, “Islam” cannot “do” anything, and therefore cannot have an effect. The use of this admittedly crude shorthand is not meant to insinuate otherwise, but is merely employed for the sake of both brevity and recognition that the discourse is (however erroneously) around the force of Islam and not the socially conditioned actions of Muslims. 56 “Arab Spring,” that were still battling through the ebbs and flows of democratic transition during the time that the surveys were fielded in 2011 and 2013. In ad- dition to introducing a novel method for gauging tolerance levels absent the typical survey battery, my analysis yields several findings in support of a more conditional evaluation of Muslims’ propensity to adopt democratic attitudes and behaviors. First, although the pooled sample indicates that Islamists are, on average, more intolerant than non-Islamists (see below for how I code these two group designations), partitioning the sample by year shows that this division emerged only after the ini- tial transitional periods in these countries. This contingent relationship indicates that ideological support for political Islam does not invariably impede the expression of tolerance toward one’s political opponents. Second, it is only among Islamists that increased religiosity and support for a broader implementation of sharia are correlated with higher intolerance, while the inverse is true for non-Islamists. This asymmetry suggests that even though stronger religious adherence does not, in itself, negatively impact tolerance, it may indirectly diminish one’s propensity to tolerate opposing groups by activating a stronger political identity among Islamists. Last, the factors that generally augment political tolerance, namely education and accul- turation to civil libertarian norms, appear to operate similarly in a Muslim-majority context, although not through the typical pathways. Taken together, these results demonstrate that Muslims in the Middle East appear susceptible to the same factors that tend to enhance or impede political tolerance among other populations. To situate this study’s insights, I review the limited comparative literature on political tolerance. In particular, I highlight the sparse findings relative to either 57 the Middle East or democratizing states. Thereafter, an outline of the relevant political context highlights key attributes of the two countries and the span of time covered in this analysis. The subsequent section elaborates the study’s hypotheses and model specifications, which are thereafter tested and followed by a discussion of the results. I conclude by considering the limitations of this project and highlighting the implications we can nonetheless draw from the analysis. 3.2 Political Tolerance: An Overview 3.2.1 Comparative Studies of Political Tolerance For decades, social scientists have sought to better understand the determinants of political tolerance (see Section 1.2.1 for a review). One area, however, where the study of this core democratic norm has lagged is in research outside North Amer- ica and Europe. That is to say, the samples buttressing the tolerance literature are largely drawn from Western, Educated, Industrialized, Rich, and Democratic (“WEIRD”) (Henrich, Heine, and Norenzayan 2010) societies. To be sure, some of the discipline’s seminal works on tolerance are explicitly comparative in nature (e.g., Sullivan et al. 1985; Duch and Gibson 1992) and non-American country stud- ies are becoming less rare (e.g., Kim and Zhong 2010; Lee 2013; Peffley, Hutchison, and Shamir 2015). Yet, there are two distinct disadvantages to studying political tolerance outside the US, particularly in non-WEIRD settings. First, there is the abiding (if perhaps narrowing) gulf between both the quantity and quality of public opinion data produced in developing countries versus OECD 58 member states. In the U.S., for example, there is a long lineage of regular, nationally- representative political and social surveys, such as the American National Election Study (ANES) and the General Social Survey (GSS). The gathering of comparable data began only relatively recently in Canada and select countries in Western Europe, and is still quite rare outside these regions. To be sure, however, the spread of massive, cross-national surveys (such as the World Values Survey (WVS), European Social Survey (ESS), and various regional barometers) has brought a number of public opinion blind spots into view. In addition to the general obstacles associated with comparative public opinion research, there is a second, more idiosyncratic, difficulty in surveying political toler- ance outside the United States. Specifically, neither of the two primary methods of measuring political tolerance (outlined above) draws on variables commonly found in either national or cross-national surveys. Since they require a specific battery of items that, for all intents and purposes, is particular to a single domain of public opinion, direct measures of tolerance are generally excluded from the vast majority of questionnaires. There are, however, some qualified exceptions. The GSS, for instance, has fea- tured a fixed-group measure of tolerance in its (occasionally annual, but mostly biennial) survey of American social attitudes since its inception in 1972. In terms of comparative data, the WVS questionnaire is the only nationally representative source to include a content-controlled tolerance battery; however, these added ques- tions appeared only in the third wave of the project, fielded between 1995 and 1998. The few remaining data sources that feature even a modified tolerance battery cer- 59 tainly provide rare insight into this largely overlooked branch of comparative public opinion, yet feature significant tradeoffs in either coverage 2 or measurement. 3 In line with these data limitations, the scholarship on tolerance and democrati- zation is similarly bounded. Two of the three cross-national studies in this genre of research utilize the third wave WVS data (Peffley and Rohrschneider 2003; Gu´ erin, Petry, and Crˆ ete 2004), while the remaining study leverages the 1988 European Om- nibus Survey questions on attitudes toward fascists as a proxy measure (Duch and Gibson 1992). National studies on tolerance judgments following democratic transi- tions are even rarer. This scarcity is not surprising given that the polling infrastruc- ture in such countries is often underdeveloped. Yet, even when such organizational capacity exists in however nascent a form, inefficient physical and technological in- frastructure can make an already resource-intensive endeavor that much costlier. Given these challenges, Gibson and Duch’s work on Russia (1992) and Gibson and Gouws’ work on South Africa (2003) constitute the only two sets of research in this domain. 3.2.2 Political Tolerance in the MENA Region With regard to the other substantive emphasis of this study, there has been little work done on Arab or Muslim political tolerance. To be sure, scholars have 2. For example, the first wave of the Comparative Youth Project, a joint endeavor of McGill and Leuven Universities, asked respondents a close approximation of the typical fixed-group battery of questions, but limited the sample to 10th and 11th grade students in Canada and Belgium. 3. The 1988 European Omnibus Survey, for example, gauges only respondents’ level of tolerance toward fascists while the AmericasBarometer eschews any reference to specific groups and asks respondents to consider “government dissidents” in the abstract when answering the tolerance battery. 60 not altogether ignored the Middle East, as there is a relatively robust literature on Israeli tolerance (Shamir 1991; Sullivan et al. 1985; Peffley, Hutchison, and Shamir 2015). In terms of the rest of the region, however, the available works are mostly descriptive and strictly speaking, examine social tolerance (Jamal and Tessler 2008; Tessler, Jamal, and Robbins 2012). Indeed, a reliance on measures of prejudice as a substitute for the typical political tolerance battery is found throughout the scant literature on Muslim political tolerance (e.g., Moaddel 2006; Sarkissian 2012; Spierings 2014) despite the repeated finding that these two dimensions of tolerance do not necessarily run parallel (Gibson 2006). There are a few studies of Muslim attitudes that do contain dedicated tolerance batteries, yet their results run the gamut on key measures of religiosity. At one end of the spectrum, Djupe and Calfano (2012) find that among American Muslims, Quran literalism and increased mosque attendance both correlate with intolerance, while, at the other end, Verkuyten et al. (2014) report that religious identification increases tolerance among their sample of Muslims in Germany and the Netherlands. Countering both these findings, Fletcher and Sergeyev’s (2002) analysis shows no discernible link between “Islamic religious beliefs” and tolerance levels among Kyrgyz Muslims. 3.3 Political Context Given the dearth of research on political tolerance in either societies transitioning from autocratic rule or the Arab Middle East, and in light of the disparate and limited 61 findings on Muslim populations in any context, it is an opportune time to study the mechanics of this norm in the MENA region. Studies of political tolerance tend to focus on settings that are at least nominally democratic since the diversity of thought and association needed to activate the relevance of tolerance (both in terms of mass media and civil society) is often superficial or altogether absent in authoritarian environments. Accordingly, much of the MENA area remains empirically outside this study’s scope. Logistical constraints similarly preclude examining most of the countries featured in the Arab Barometer, given that the composite items I use to measure political tolerance (outlined in the next section) were asked in less than a handful of sites. Due to these general constraints, the two countries that lend themselves most readily to examination and comparison are Egypt and Tunisia. This assertion, while not overlooking several notable disparities between these two settings, 4 is borne out by a number of similarities. First, both countries began their transitions away from authoritarian rule at roughly the same time (early 2011). Second, the nature of the mass uprisings in both these settings was not as protracted nor as violent as those that took place in other countries across the region–whether momentarily successful (Libya, Yemen), failed (Bahrain), or ongoing (Syria). Relatedly, a third parallel between Egypt and Tunisia is their relatively homogenous populations, which staved off the destabilizing force of sectarian strife seen most acutely in Syria and most consistently in Iraq. Last, both the Egyptian and Tunisian regimes prior to 2011 4. Among the most notable ways in which Egypt and Tunisia differ include: land area (Tunisia is less than one-fifth the size of Egypt), population size (Egypt’s is approximately nine times larger), and military history (while Tunisia has largely remained absent from the region’s conflicts, Egypt has historically played a central role). 62 shared the institutional legacy of de facto one-party, nominally republican rule (as opposed to de jure monarchical rule elsewhere in the region). There is, of course, one glaring difference between the two states: the outcome of their transitions to date. Naturally, any path toward democratization will be fraught with political and economic hazards. Yet, given the similarities mentioned above, it would not have been obvious ex ante that the two countries would diverge so dramatically following their initial push toward democracy. The Freedom House rankings underscore this point, with Egypt and Tunisia both being designated as “not free” on the eve of their uprisings, subsequently making incremental progress to “partly free” in the 2013 report, but then drastically parting ways, with Egypt regressing to “not free” and Tunisia earning the first “free” designation for an Arab country since Lebanon in the 1970s. To explain this variation, some scholars have highlighted the differing historical role of the military in each (Bellin 2014; Brown 2013), while others have underscored Tunisia’s more developed civil society (Marks 2015; Masoud 2015). The divergent ways in which these two countries’ Islamist parties practiced politics, with Tunisia’s Ennahda lauded for its more inclusive and moderate approach relative to that of Egypt’s Muslim Brotherhood has also featured prominently in political post mortems (Bellin 2014). Regardless of the factor one chooses to emphasize, however, it is clear that in both Egypt and Tunisia the masses (and, consequently, mass attitudes) played a pivotal role. For instance, while it may have been members of the secular elite in Egypt that on July 3, 2013 went “knocking on the barracks door” (Stepan 1988), the calculus that facilitated the coup would not have so plainly been in favor of military 63 intervention had it not been for millions of citizens taking to the streets four days earlier to call for President Muhammad Morsi’s ouster. Similarly, although it was the national dialogue quartet that helped pull Tunisia back from a political precipice in 2013 and 2014, ultimately earning this group a Nobel Peace Prize, their efforts at elite reconciliation would have been for naught had the citizenry not felt confident in the system and the choices it presented them (as evident by the remarkable 69% voter turnout in the 2014 parliamentary elections). The relevant point is that a citizenry’s political tolerance is a necessary (if far from sufficient) component to keep a transition on track. In this sense, a politically tolerant public is perhaps less of a facilitator for progressing along a democratic path and more of a bulwark against authoritarian backsliding. Ultimately, we would want to understand what moves the needle in societies transitioning away from autocratic rule. In the case of transitioning Muslim-majority states, the first step is to establish whether that needle can move at all. 3.4 Hypotheses and Model Specification 3.4.1 Hypotheses I test three hypotheses corresponding to the contention that an inverse, and in- variable, relationship exists between Islam and the expression of democratic norms. These hypotheses examine this invariability, in turn, with regards to: 1) the role of Islam as a political identifier; 2) the effect of increased adherence to the tenets of Islam or the desire for society to be better aligned with them; and 3) the pre- 64 sumed resistance to democratic learning among Muslim populations generally. Each hypothesis features a “static” and “dynamic” variant, corresponding to the broad belief sets that regard Islam’s effect on democratic norms as either 1) negative and constant, or 2) conditional. Whereas this latter set of hypotheses center the role of proximate political context and experience, the former set discounts or altogether dismisses such explanations. The first hypothesis compares the tolerance levels of Islamists to non-Islamists (see Figure 3.1 and accompanying discussion below for the factors that determined this desgination). Those affiliated with Islamist groups or ideologies, particularly under the Muslim Brotherhood umbrella, have been long thought to hold illiberal beliefs. In part, this association is due to Islamists’ poor track record when holding power in Afghanistan, Iran, and Sudan. Additionally, aside from these concrete test cases, the Brotherhood’s own particular history and rhetoric more than amply fueled suspicion of the group’s democratic bona fides across the Middle East (Wickham 2013; Masoud 2013). Accordingly, even as Islamists took the political reigns in Egypt and Tunisia early on in the these countries’ transitions, there was lingering concern over whether the political process would take the form of “one man, one vote, one time” (a phrase that gained currency in the early 1990s as the Islamic Salvation Front (FIS) won clear pluralities in Algeria’s municipal elections). Given these abiding suspicions, the static expectation is that Islamists will be more intolerant, regardless of time period or setting, with the dynamic corollary postulating that this effect is conditioned on the political environment, either tem- porally or geographically. Thus, Islamist intolerance relative to non-Islamists may 65 have been significantly higher in 2013 than it was in 2011. The shift would parallel a reconfiguration of political power in both Tunisia and Egypt. Specifically, there was less ideological enmity during the early transitional period, given the lingering opti- mism from having unseated an entrenched autocrat and a (theoretically) broad array of possible political futures. With the election of, and subsequent calls to unseat, Islamist governments, however, tensions were quite high early in 2013, 5 and perhaps altered a suite of political attitudes as a result, including tolerance. The different settings may also condition the degree to which Islamists express tolerance, given that not all off-shoots of Islamism are equivalent in terms of the political outcomes they seek or the means by which they endeavor to reach them (Mozaffari 2007), and Tunisia’s variant is often characterized as particularly moderate on both counts (Cavatorta and Merone 2013). H 1-static : Political intolerance is greater among Islamists. H 1-dynamic : Political intolerance is greater among Islamists, conditioned on a) survey year, or b) country. H 1-null : Political intolerance is unrelated to Islamist status. The second hypothesis tests the influence of religiosity and support for imple- menting sharia. These variables, both of which are composite measures, tap into 5. The power grab that galvanized opponents of then-Egyptian President Mohamed Morsi, and set in motion events that led to the July 2013 military coup, took place in late 2012 (Kirkpatrick and Sheikh 2012). In Tunisia, the first of two high-profile political assassinations took place on February 6, 2013, furthering a political crisis that ended in the country’s Islamist-led government voluntarily stepping down by the end of the year (see Section 4.2 for a more detailed review of this period’s key events). 66 one’s regard for Islam in both one’s personal life and in society. On the one hand, from the earliest studies and for decades thereafter, scholars have consistently found a link between religious conviction and intolerance (Stouffer 1955; Nunn, Crockett, and Williams 1978; Sullivan, Piereson, and Marcus 1982; McClosky and Brill-Scheuer 1983; Gibson 2010). On the other hand, the influence of Islam on tolerance is thought to be especially pernicious both in the eyes of mass publics across the West (Pew Research Center 2011) and among numerous prominent commentators (e.g., Pipes 2003; Spencer 2005). In light of these general findings and particular indictments, the static expectation is that both religiosity and support for sharia will consistently correlate with greater intolerance. However, this correlation may also be a function of either the political environment or, alternatively, political identification. H 2-static : Political intolerance increases as religiosity and support for sharia increases. H 2-dynamic : Political intolerance increases as religiosity and support for sharia increases, conditioned on a) survey year, b) country, or c) Islamist status. H 2-null : Political intolerance is unrelated to religiosity and support for sharia. The last hypothesis examines whether the Muslim-majority populations of these countries are particularly resistant to the forces that promote tolerance. One of the strongest determinants of tolerance is education, which speaks to the importance 67 of being socialized to this norm. To this end, education may be the most effective means through which citizens build the cache of considerations that allow for a “sober second thought” to win out against an initial, almost involuntary (Marcus et al. 1995), urge toward intolerance. If the mere adherence to Islam impedes this democratic learning among Muslims, then we would expect education not to have an effect on political tolerance. Yet, it could be that Muslims in these and other Muslim-majority countries are not receiving the “right” education. That is, political tolerance is unlikely to be promoted in an authoritarian context, and so the mechanism (i.e., socialization) by which education can moderate tolerance judgments may not be operable in (osten- sibly) nascent democracies. In this case, we would similarly observe no relationship between respondents’ level of education and the propensity to grant their political enemies the same liberties they themselves covet. To examine this possible detour (or disruption) in the education to tolerance pipeline, I account for whether or not a respondent speaks a second language in addition to Arabic. Although far from a comprehensive proxy, multilingualism should indicate whether a respondent has access to information from other societies that promote tolerance and/or the type of education that values these ideas within their own society. H 3-static : Political tolerance does not increase as a function of education level or knowledge of a second language. H 3-dynamic : Political tolerance increases as a function of a) education level, or b) multilingualism. 68 3.4.2 The Data and the Dependent Variable I use the Arab Barometer’s second and third wave to model the determinants of political tolerance in Tunisia and Egypt. The second wave was fielded in 2011, shortly after the ouster of each respective country’s longtime dictator, while the third wave was fielded in early 2013, preceding the military coup in Egypt. These timings are fortunate as there are very few surveys that capture both countries at multiple points in the years following the initial uprisings of late-2010/early-2011 and before the region’s authoritarian backsliding became clear. Among all the potentially viable datasets, however, none include questions that directly address political tolerance. Recalling that political tolerance denotes “a willingness to permit the expression of ideas or interests one opposes” (Sullivan, Piereson, and Marcus 1982), the two ways that this construct is typically measured is either by: 1) asking respondents about their willingness to extend certain civil liberties to a set number of pre-selected groups; or 2) asking the respondent to select their least-liked group and, thereafter, administering the tolerance battery. As neither of these methods is featured in any surveys conducted within the geographic and temporal parameters of interest, I utilize a novel technique to measure tolerance that nonetheless maintains construct validity. The tolerance scale I construct is the product of a two-step process. First, a question measuring respondents’ level of trust in the Muslim Brotherhood (and, ad- ditionally for Egypt, the religiously ultra-conservative Salafis) is used to establish the objection pre-condition by dividing the sample into those who at least lean Islamist (trusting the Brotherhood/Salafis to any degree) and ardent non-Islamists (respond- 69 ing that they “absolutely do not trust” the Brotherhood/Salafis). For the sake of brevity, I will simply refer to the first group as “Islamists” and the latter group as “non-Islamists” for the remainder of the article. Second, a question gauging whether respondents would be willing to exclude those on the opposing end of the Islamist/non-Islamist spectrum effectively measures their level of tolerance. For Islamists, tolerance is gauged by the extent to which they believe that “a parliamentary system in which only Islamist parties compete in par- liamentary elections” is appropriate for their country while, conversely, non-Islamist tolerance is scaled according to the degree to which they believe “a parliamentary system in which only non-religious parties compete in parliamentary elections” is appropriate. As with the first step, I dichotomize this measure, coding as tolerant only those who responded that it would be “absolutely inappropriate” to ban certain groups from participating in the political process. Thus, rather than measuring po- litical tolerance as a function of mere affect toward particular groups, this scale more accurately aligns with the construct’s central component, namely the willingness to limit one’s enemy’s civil liberties. Figure 3.1 visually presents the logical flow for this coding procedure. Figure 3.2 shows the distribution of political tolerance for both Egypt and Tunisia during the two years captured by the surveys. One feature of this graph that im- mediately stands out is the seemingly high degree of tolerance exhibited in these countries, with a clear majority of respondents in both countries reporting that a system of government that effectively excluded their political opponents from run- ning for elected office was “Absolutely Inappropriate. ” On its face, this seems like a 70 Figure 3.1: Coding Process for the Dependent Variable Figure 3.2: Percent Distribution of Dependent Variable, by Year and Country 71 remarkable level of tolerance for societies with little acculturation to this norm and where the stakes of political loss are magnified. There are two points to consider, however, when processing these data. First, across all contexts, prompts to consider the outright exclusion of disliked groups from running for public office elicit only a moderate level of intolerance–not as low as removing a book written by a member of the disliked group from the local library, though not nearly as high as prohibiting a member of the disliked group from teach- ing at a local public school (see, e.g., (Peffley and Rohrschneider 2003; Gibson 2008; Mather and Tranby 2014; Djupe, Lewis, and Jelen 2016). Second, given that, for decades, the expression of sensitive opinions was not only frowned upon, but poten- tially dangerous in the settings for these surveys, these data potentially contain some unknown measure of social desirability bias. Thus, the level of intolerance expressed in this sample is likely a conservative estimate, perhaps representing something close to this attitude’s lower bound in these societies. 3.4.3 Additional Controls Beyond the predictor variables already mentioned, the models include two sets of relevant controls. The first set features measures of key democratic attitudes and behaviors. One such variable gauges a respondent’s belief that their country is fit for democratic governance. Measures of general support for democracy are often included in tolerance models since this belief likely drives relevant norms at a lower level of abstraction, including support for freedom of speech and assembly (Sulli- van et al. 1985; Duch and Gibson 1992; Peffley and Rohrschneider 2003). Another variable measures civil libertarian propensities more directly, assessing whether re- 72 spondents feel that the limiting of human rights is justified in their country in order to maintain security. Such sentiments favoring social order over liberty tend to nega- tively correlate with tolerance (Gibson 2013). A dummy measure indicating whether respondents engaged in public protests any time over the previous three years is also included in this set of variables. Political participation is generally considered a positive predictor of tolerance (a relationship at the crux of the elite-mass attitu- dinal divide in this domain (Gibson and Duch 1991)), while engaging in protest, in particular, has been found to have a robust, if substantively marginal effect (Gu´ erin, Petry, and Crˆ ete 2004). Last, a set of demographic and fixed effects variables–namely, age, sex, urban status, income, year and country–round out the specifications. For ease of interpre- tation, each of the variables (excluding Country and Year) were rescaled from 0-1. Table 3.1 provides key descriptive statistics for each of these variables as they cor- respond to non-Islamists and Islamists in the sample. Full question wording for the core explanatory variables not already elaborated upon can be found in Appendix 3.B. 3.5 Results and Discussion Given that the dependent variable is binary, I use logit regression models to estimate the determinants of political tolerance. To further facilitate interpretation, the coefficients in each model represent the predicted change in the probability of giving a tolerant response when the value of a variable is increased by one unit which, 73 Non-Islamist Islamist mean sd mean sd Political Tolerance 0.70 0.46 0.58 0.49 Islam Implement Sharia 0.57 0.22 0.58 0.20 Religiosity 0.73 0.23 0.76 0.21 Democratic Learning Education 0.47 0.32 0.42 0.32 Knows Second Language 0.38 0.49 0.39 0.49 Democratic Norms Democracy Fit for Country 0.83 0.25 0.85 0.21 Limit HR for Security 0.21 0.33 0.23 0.32 Protested w/in Three Years 0.13 0.34 0.13 0.34 Demographics and Fixed Effects Age 0.47 0.33 0.50 0.33 Female 0.49 0.50 0.48 0.50 Urban 0.55 0.50 0.54 0.50 Income 0.31 0.29 0.29 0.26 Country (Egypt=1; Tunisia=2) 1.42 0.49 1.53 0.50 Observations 2097 2275 Table 3.1: Descriptive Statistics (Non-Islamists vs. Islamists) 74 given the rescaling, reflects a move from each variable’s lowest to its highest value. The bulk of the evidence in the following analyses points toward a more dynamic interpretation of the Islam-democracy relationship. Turning to Table 3.2, the pooled sample (combining both waves of the Arab Barometer for both Tunisia and Egypt) appears to confirm the common wisdom that Islamists are indeed more intolerant than their non-Islamist counterparts. Both the limited and full regression models demonstrate a statistically and substantively significant correlation between Islamist support and intolerance. Table 3.2: Change in the Probability of a Tolerant Response, Full Sample Political Tolerance Political Tolerance Islam Islamist −0.11 ∗∗∗ (0.02) −0.12 ∗∗∗ (0.02) Implement Sharia 0.02 (0.07) Religiosity 0.02 (0.05) Democratic Learning Education 0.04 (0.04) Knows Second Language 0.10 ∗∗∗ (0.03) Democratic Norms Democracy Fit for Country 0.14 ∗∗ (0.05) Limit HR for Security −0.06 (0.03) Protested w/in Three Years 0.01 (0.03) Demographics and Fixed Effects Age 0.03 (0.04) Female 0.05 ∗ (0.02) Urban 0.01 (0.02) Income −0.01 (0.05) Year (2013) 0.06 ∗∗∗ (0.02) 0.07 ∗∗ (0.03) Tunisia 0.04 ∗ (0.02) −0.03 (0.03) N 3887 2882 Standard errors in parentheses ∗ p< 0.05, ∗∗ p< 0.01, ∗∗∗ p< 0.001 75 Disaggregating the data by country and year, however, reveals a more nuanced association. Table 3.3 demonstrates that the negative relationship between an Is- lamist designation and political tolerance emerges only in the third wave of the Arab Barometer. This finding is notable for two reasons. First, it indicates that there is no invariable mapping of Islamism onto political intolerance; there is clearly a more conditional relationship. Second, this conditionality is arguably a function of the political ebbs and flows in society, given the events in Egypt and Tunisia’s early transitional periods–including, in particular, Islamist parties coming to power in each country, opponents contesting the legitimacy of their rule, and (still only a looming possibility at the time that the third wave of the Arab Barometer was fielded) not so veiled threats to remove them from office through extra-democratic means. The shift that Table 3.3 demonstrates thus underscores that changing societal circum- stances and threat assessments (that is, two standard drivers of political attitudes and behavior) are likely motivating Islamist opinions and actions rather than any intrinsic aversion to a particular set of norms. Splitting the sample between Islamists and non-Islamists exposes yet another conditional finding on the drivers of intolerance. Specifically, Table 3.1 reports that the two factors expected to uniformly depress tolerance in H 2-dynamic , namely religios- ity and support for sharia, actually exert a mildly positive effect on tolerance levels (though, notably, without crossing a conventional level of statistical significance). Table 3.4, however, demonstrates a stark difference in the way these variables oper- ate across the Islamist/non-Islamist divide. Figures 3.3 and 3.4 plot the predicted probabilities of a tolerant response con- 76 Table 3.3: Change in the Probability of a Tolerant Response, by Country and Year Egypt 2011 Tunisia 2011 Egypt 2013 Tunisia 2013 Islam Islamist −0.12 (0.19) −0.25 (0.19) −1.05 ∗∗∗ (0.25) −0.59 ∗∗ (0.19) Implement Sharia 1.22 ∗ (0.60) −3.26 ∗∗∗ (0.81) 2.07 ∗∗ (0.64) −0.77 (0.44) Religiosity 1.21 (0.75) −0.33 (0.37) 0.08 (0.75) 0.60 (0.46) Democratic Learning Education −0.24 (0.30) 0.59 (0.40) 0.80 (0.44) −0.20 (0.45) Knows Second Language 0.09 (0.25) 0.32 (0.26) 0.37 (0.36) 0.84 ∗∗∗ (0.24) Democratic Norms Democracy Fit for Country 0.51 (0.54) 0.03 (0.51) 0.29 (0.53) 1.24 ∗∗∗ (0.38) Limit HR for Security −0.47 (0.32) −0.39 (0.31) −0.06 (0.38) −0.65 ∗ (0.27) Protested w/in Three Years 0.63 ∗ (0.31) 0.13 (0.22) −0.81 ∗ (0.34) 0.34 (0.30) Demographics Age 0.43 (0.31) 0.10 (0.33) 0.24 (0.41) −0.02 (0.35) Female 0.67 ∗∗ (0.22) −0.17 (0.19) 0.19 (0.27) 0.52 ∗ (0.20) Urban 0.09 (0.19) −0.21 (0.21) 0.69 ∗∗ (0.25) −0.35 (0.20) Income −0.15 (0.44) 0.18 (0.51) 0.17 (0.45) −0.24 (0.52) Constant −1.93 ∗ (0.86) 1.85 ∗∗ (0.63) −1.50 (0.96) −0.17 (0.53) N 616 609 828 807 Standard errors in parentheses ∗ p< 0.05, ∗∗ p< 0.01, ∗∗∗ p< 0.001 77 Table 3.4: Change in the Probability of a Tolerant Response, by Islamist Status Non-Islamist Islamist Islam Implement Sharia 0.45 ∗∗∗ (0.08) −0.47 ∗∗∗ (0.09) Religiosity 0.06 (0.07) 0.00 (0.08) Democratic Learning Education 0.07 (0.06) 0.04 (0.06) Knows Second Language 0.05 (0.04) 0.15 ∗∗∗ (0.04) Democratic Norms Democracy Fit for Country 0.04 (0.06) 0.27 ∗∗∗ (0.08) Limit HR for Security −0.05 (0.05) −0.08 (0.05) Protested w/in Three Years −0.01 (0.04) 0.07 (0.04) Demographics and Fixed Effects Age 0.01 (0.05) 0.04 (0.05) Female 0.05 (0.03) 0.06 (0.03) Urban 0.03 (0.03) −0.01 (0.03) Income −0.00 (0.07) −0.04 (0.07) Year (2013) 0.02 (0.03) 0.15 ∗∗∗ (0.03) Tunisia 0.10 ∗ (0.04) −0.13 ∗∗∗ (0.04) N 1425 1457 Standard errors in parentheses ∗ p< 0.05, ∗∗ p< 0.01, ∗∗∗ p< 0.001 78 Figure 3.3: Predicted Probability of a Tolerant Response, by Sharia Support Figure 3.4: Predicted Probability of a Tolerant Response, by Religiosity ditioned on support for sharia and personal religiosity, respectively. In both cases (though more starkly in Figure 3.3), the probabilities for the Islamist/non-Islamist sub-samples noticeably diverge, with Islamists moving in a more intolerant direction in both cases (see Appendix 3.A for corresponding regression tables). This finding may be a result of these variables’ eliciting increased group identity among Islamists, triggering a bunker mentality among those that score higher on these scales. In any case, however, the fact that non-Islamists move in a decidedly opposite direction is enough to reject the static variation of the second hypothesis. Last, Table 3.5 provides support for the dynamic variant of the third hypothesis. The first model demonstrates that education is a strong positive driver of tolerance, yet this effect is wholly subsumed in the second model within the variable measure of second language proficiency. This suggests that socialization to democratic norms is not a moot point when Muslim-majority states are concerned, but that the “official” routes through state education systems may not be conducive to such processes. Additionally, as Figure 3.5 demonstrates, the effect of multilingualism is especially 79 strong among Islamists, altogether closing any statistically discernible tolerance gap between them and non-Islamists. This finding is in line with accounts chronicling the moderating influence of time abroad in exile for Islamists from the region (Stepan 2012; Marks 2015). Table 3.5: Change in the Probability of a Tolerant Response, by Multilingualism Political Tolerance (No Second Language) Political Tolerance (w/ Second Language) Islam Islamist −0.12 ∗∗∗ (0.02) −0.12 ∗∗∗ (0.02) Implement Sharia 0.02 (0.06) 0.02 (0.07) Religiosity 0.02 (0.05) 0.02 (0.05) Democratic Learning Education 0.09 ∗ (0.04) 0.04 (0.04) Knows Second Language 0.10 ∗∗∗ (0.03) Democratic Norms Democracy Fit for Country 0.14 ∗∗ (0.05) 0.14 ∗∗ (0.05) Limit HR for Security −0.06 (0.03) −0.06 (0.03) Protested w/in Three Years 0.02 (0.03) 0.01 (0.03) Demographics and Fixed Effects Age 0.02 (0.04) 0.03 (0.04) Female 0.05 ∗ (0.02) 0.05 ∗ (0.02) Urban 0.02 (0.02) 0.01 (0.02) Income 0.00 (0.05) −0.01 (0.05) Year (2013) 0.07 ∗ (0.03) 0.07 ∗∗ (0.03) Tunisia 0.03 (0.03) −0.03 (0.03) N 2882 2882 Standard errors in parentheses ∗ p< 0.05, ∗∗ p< 0.01, ∗∗∗ p< 0.001 80 Figure 3.5: Predicted Probability of a Tolerant Response, by Multilingualism 3.6 Conclusion This study contributes to the debate over the compatibility of Islam and democ- racy through an atypical mode of inquiry. Rather than asking whether Islam (vari- ably measured) depresses the expression of democratic norms, I examine when reli- gious variables may either have a positive or negative association with democratic norms. Overall, the analyses support a more dynamic perspective on the relationship between Islam and democracy, with political tolerance in the sample being condi- tioned by shifting political circumstances, expressed political identity, and exposure to socializing forces–which are, incidentally, the factors that significantly determine tolerance among any other given population. Some caution is, of course, warranted. The generalizability of these findings is partially limited by the bluntness of the survey instrument. Although I argue that my tolerance scale offers improved construct validity, the single indicator metric lags 81 in measurement validity behind the typical 3-4 question indices used in dedicated tolerance studies. It is left to future research to corroborate with still more robust measures the patterns I report here. Nonetheless, this analysis provides at least two meaningful contributions to the study of political tolerance. First, it addresses both the MENA regional gap and the emerging democracy contextual gap in the literature, providing researchers with a wider body of evidence to draw upon for future comparative work. Second, the method I employed to construct the tolerance scale can be fruitfully put to use in analyzing other data. These datasets can be “salvaged” in a sense, to offer a window into the state of tolerance in settings and time periods when no other sources are available. With tolerance seemingly in ever shorter supply around the globe, there is a great and growing need to understand what drives this fundamental norm of democratic–and, more to the point, peaceful–coexistence. 82 Appendices 3.A Additional Tables and Figures Table 3.A.1: Change in the Probability of a Tolerant Response – Interactions Islamist X Religiosity Islamist X Sharia Support Islam Islamist −0.00 (0.07) 0.28 ∗∗∗ (0.06) Sharia Support 0.02 (0.07) 0.37 ∗∗∗ (0.08) Religiosity 0.10 (0.07) 0.03 (0.05) Islamist X Religiosity −0.16 † (0.09) Islamist X Sharia Support −0.68 ∗∗∗ (0.09) Democratic Learning Education 0.04 (0.04) 0.06 (0.04) Knows Second Language 0.10 ∗∗∗ (0.03) 0.10 ∗∗∗ (0.03) Democratic Norms Democracy Fit for Country 0.14 ∗∗ (0.05) 0.13 ∗∗ (0.05) Limit HR for Security −0.06 (0.03) −0.05 (0.03) Protested w/in Three Years 0.01 (0.03) 0.02 (0.03) Demographics and Fixed Effects Age 0.03 (0.04) 0.03 (0.04) Female 0.05 ∗ (0.02) 0.05 ∗ (0.02) Urban 0.01 (0.02) 0.00 (0.02) Income −0.01 (0.05) −0.02 (0.05) Year (2013) 0.07 ∗∗ (0.03) 0.07 ∗∗ (0.02) Country (Tunisia) −0.02 (0.03) −0.01 (0.03) N 2882 2882 Standard errors in parentheses † p< 0.10, ∗ p< 0.05, ∗∗ p< 0.01, ∗∗∗ p< 0.001 83 3.B Question Wording Variable (type) Composite question(s) Answer options Implement Sharia (additive index) The government and parliament should enact laws in accordance with Sharia law Strongly disagree – Strongly agree (4pt) The government and parliament should enact penal laws in accordance with Sharia Strongly disagree – Strongly agree (4pt) The government and parliament should enact inheritance laws in accordance with Sharia Strongly disagree – Strongly agree (4pt) Applying Sharia more strictly Strongly don’t support – Strongly support (4pt) Religiosity (additive index) Pray daily Rarely – Always (4pt) Attend Friday prayer/Sunday services Rarely – Always (4pt) Listen to or read the Quran/the Bible Rarely – Always (4pt) Generally, would you describe yourself as... Not religious; Somewhat religious; Religious Democracy Fit for the Country (single indicator) Opinion with regard to the country’s governance: A democratic political system. Very bad – Very good (4pt) Limit Human Rights for Security (single indicator) To what extent do you think the lack of respect for human rights is justified in order to maintain security in your country? Not at all – Great extent (4pt) Protested w/in the Past Three Years (single indicator) During the past three years, did you ... Participate in a protest, march or sit-in. I have never participated; Once; More than once (converted to a binary variable by combining the two positive choices) 84 Chapter 4 Tolerance and the Prospects for Tunisian Democracy 4.1 Introduction Can Tunisia’s nascent democracy survive in the face of mounting challenges? From a stalled economy (Daragahi 2018), to terrorism in and around the country (“Terrorists Returning” 2017; Macdonald and Waggoner 2018), to widening corrup- tion (Yerkes 2017), there is no shortage of domestic and regional perils. Moreover, these individual dangers could also collectively work to undermine the country’s frag- ile political equilibrium. As crises mount, antagonisms between various segments in society become more salient, making it harder for Tunisia’s vaunted consensus (Marks 2017b) to survive. As the case of Egypt’s failed transition makes clear, lingering in- stability and insecurity tend to prime a willingness to constrain political space and 85 create scapegoats. If Tunisians should similarly support limiting (or eliminating) certain speech or activity, such repression could pave the way for unrest, violence, or even authoritarian backsliding. To assess Tunisians’ attitudes toward their ideological opponents’ rights, I draw on an original, nationally representative survey featuring the first dedicated tolerance battery fielded in the Arab Middle East. Analysis of the observational and experi- mental data reveals key insights into the structure and malleability of tolerance in Tunisia. First, intolerance in the sample is focused primarily on groups that threaten “traditional values” (e.g., homosexual activists and atheists) rather than those who are active (or potentially active) in formal politics (e.g., Islamists and old regime elites). Consistent with these descriptive data, regression analysis demonstrates that social threat is the chief driver of intolerance whereas respondents who believe that a disliked group could plausibly wield political power are more likely to provide a tolerant response. Second, in terms of public religiosity, support for a non-religious party is not significantly associated with tolerance while (separately) support for sharia, likely because it taps into conservative attitudes toward society, negatively affects tolerance levels. Last, experimental attempts to persuade respondents to move away from their initial attitude were much more successful in leading subjects from tolerance to in- tolerance than the other way around. Yet, those who chose a group active in formal politics, rather than one informally agitating for social change, were more likely to be moved from their initially intolerant attitude. On balance, I argue that these 86 findings bode well for Tunisia’s democratic experiment in the short term, but may presage long-term difficulties. 4.2 Political Tolerance in the Tunisian Context Scholars researching political tolerance, which is generally conceived of as the willingness to support the political rights of disliked groups (Sullivan, Piereson, and Marcus 1979), have elaborated a suite of factors associated with the expression of this core democratic norm. From the earliest tolerance studies, for example, the positive effect of a secure and open-minded personality has been evident (Stouffer 1955; Nunn, Crockett, and Williams 1978; Marcus et al. 1995; Mondak and Sanders 2003). Along similar lines, religiosity (variably measured) has demonstrated a largely negative relationship to tolerance (Sullivan, Piereson, and Marcus 1982; McClosky and Brill-Scheuer 1983; Gibson 2010; but c.f. Eisenstein 2006). While the psychological components of tolerance have received consistent and ex- tensive treatment in the literature, relevant contextual factors have attracted compar- atively less attention. For example, many early works, measured tolerance through a “fixed group” approach that surveyed respondents’ willingness to limit the rights of a set list of disliked entities. The imposition of a static choice set foreclosed the ability to model one of the most robust contextual determinants of tolerance: threat percep- tion. Sullivan et al.’s (1982) introduction of the “least-liked” measurement technique, which asks respondents to first select a group they strongly dislike to populate the tolerance battery, largely remedied this shortcoming. Still, even those studies that 87 employed this improved method neglected meaningful broader contexts–such as, for example, minority status (see Chapter 2). Additionally, for more fundamental struc- tural reasons, studies outside established Western democracies, particularly in states that had only recently transitioned away from authoritarian rule, remained few in number. 1 As the case of Tunisia makes clear, this latter context offers an especially rich set of opportunities to expand our understanding of political tolerance and its place in democratic theory. From the “Arab Spring’s” heady early days through the crushing disappointments of the years that followed, Tunisia has largely maintained its standing as a beacon of democratic hope in the MENA region. To be sure, however, the country has had its share of fits, starts and setbacks. Although it never really came close to the total state collapse witnessed in neighboring Libya or farther afield in Syria and Yemen, Tunisia’s democratic transition faced a distinct threat of authoritarian reversal in the latter half of 2013 as opposition mounted against the country’s Islamist-led gov- ernment (which was headed by the Muslim Brotherhood-inspired Ennahda Party). Domestically, this turmoil drew on the outrage over the assassinations of Chokri Belaid in February 2013 and Mohamed Brahmi four months later, both of whom were prominent secular opposition figures. In response, the authorities banned a local salafist group, Ansar al-Sharia, which was widely believed to have orchestrated the killings. This action did little to calm tensions, however, as mass protests and general strikes spread across the country amid increasing calls for the government’s immediate resignation. All this took place at a time when, regionally, Egypt’s demo- 1. See Chapter 3, Sections 3.2.1 and 3.2.2 for a review of these constraints and the relevant literature. 88 cratic experiment had recently fallen apart following the July 3, 2013 military coup. Given that the lead up to this critical juncture was similarly characterized by public animosity toward an Islamist-led government, many feared that Tunisian democracy was in its final throes. Despite the dangers, Tunisia was able to resolve its political crisis. A coalition of four prominent civil society organizations, led by the Tunisian General Labor Union (UGTT), intervened to push for a negotiated settlement between the country’s key political actors. The efforts of this “National Dialogue Quartet” resulted in Ennahda’s agreeing to cede power to an interim technocratic government in late 2013, the passing of a widely well-regarded constitution in early 2014, the holding of parliamentary elections in December of that year, and, finally, a presidential election in January 2015. The resulting government that was formed in early 2015 allied Nidaa Tounes, a broad based secularist party that held the most seats in parliament, with the Islamist Ennahda party, which held the second-most seats, highlighting what proponents laud as Tunisia’s commitment to political consensus. Yet, although this elite accord staved off the threat of authoritarian reversal, it also left much of the mass level discontent unresolved. In the survey fielded for this study, a clear majority (61%) believe that their rights and freedoms are better safe- guarded now than they were prior to 2011, however the opposite is true when it comes to the country’s economy and national security, where 60% and 80%, respectively, believe that the situation was better under ousted President Ben Ali’s regime. Many Tunisians today regard the pact between Nidaa Tounes and Ennahda as predicated, not on consensus, but on collusion. The passage of controversial policies forgiving the 89 transgressions of old regime elites (Marks 2017a) and the enactment of a municipal electoral law seen as favoring larger parties and handicapping newcomers 2 (Dhouib 2017) have made citizens all the more skeptical that the current government can bring about meaningful reform (Fassihian and Wilson 2018). Early in 2018, major protests once again rocked the country (Smith-Spark 2018) and talk of coups and even civil war crept back into the political discourse (Qaheri 2018). Given this rising instabil- ity, and against the background of the (albeit brief) experience that Tunisians have had with democratic governance, it is an opportune time to systematically examine political tolerance in this last bastion of the “Arab Spring. ” 4.3 Data, Methodology, and Expectations To assess the prospects for tolerance in Tunisia, I draw on original data from a nationally representative survey fielded in December 2017. The commissioned firm, One to One for Research and Polling, is currently the national partner for both the Afro- and Arab-Barometer projects. Employing a stratified clustered sampling methodology based on Tunisian census data, the firm’s enumerators conducted face- to-face interviews (using the Tunisian dialect of Arabic) in each of the country’s 24 governorates. This effort resulted in a final sample size of 900. The data were subsequently weighted to the governorate’s relative population size, and national age, gender, education level, and urban/rural proportions, according to the National Institute of Statistics (Tunisia). 2. Incidentally, these concerns proved unfounded as independent candidate lists received more votes during the May 2018 municipal election than either Nidaa Tounes or Ennahda (Amara 2018). 90 4.3.1 Dependent Variable One of the primary means of gauging political tolerance is through the fixed-group method, pioneered by Samuel Stouffer (1955) and, most notably, still used in the biennial General Social Survey (GSS). As its name suggests, this approach presents respondents with a static set of groups and asks the extent to which they would be willing to limit each respective entity’s rights in various ways (e.g., by banning their members from running for political office). The validity of this measure can be undermined in two ways, however. First, the menu of groups may not be salient enough to register strong feelings of dislike. A perennial example of this scenario is the seemingly increasing tolerance that Americans have reported towards communists since the initial Stouffer study, which many scholars alternatively attribute to a reduction in this group’s relevance as a target of intolerant attitudes (Gibson 2007; Sullivan, Piereson, and Marcus 1979). Second, given an understanding of tolerance as forbearance, the fixed-group method’s validity also rests on the assumption that respondents are not sympa- thetic, or even indifferent toward the groups they are being asked about. This so- called “objection pre-condition” may be fairly assumed in certain cases (e.g., when asking African-Americans about White Supremacist groups). Yet this supposition is far from a given when, as in the GSS questionnaire, the general public is asked about limiting the rights of atheists even though as many as one quarter of US adults consider themselves religious non-believers (Gervais and Najle 2018). These potential threats to validity are especially acute in the contemporary Tunisian context. For one thing, while some surveys taken in the aftermath of Zine 91 El Abidine Ben Ali’s fall (or, more accurately, flight (Chrisafis 2011) from power document Tunisians’ affect towards certain institutions (e.g., parliament) and or- ganizations (e.g., Islamist parties), there is no comprehensive catalog of attitudes towards all the notable social and political groups in the country. Thus, putting to- gether a salient fixed-group list for Tunisian respondents would rely far more heavily on interpolation (based on the limited data available) than inference (based on a synthesis of public opinion toward all potentially relevant groups). Additionally, any static list of presumably disliked groups would include ideologies or associations that a large portion of Tunisians support. Examples include the leading Islamist party in Tunisia, Ennahda, and a number of perennially (especially for this region) contentious groups–atheists, feminists, socialists–that surely have their detractors but doubtless also have a considerable number of proponents, as well. Thus, it would be all but certain that a (notably, unknown) segment of the sample would not sufficiently object to the groups they are asked about. Given the fixed-group measure’s shortcomings (both in general and particularly in the context of Tunisia), I gauge political tolerance using the respondent-driven “content-controlled” method (Sullivan, Piereson, and Marcus 1982). First, those taking the survey were asked to indicate the group they most dislike in Tunisian society from the following (randomly ordered) list: UGTT members Old regime elites Atheists Homosexual rights activists Socialists Feminists Ennahda members Hizb ul Tahrir 92 Taking into account the local context, the list was purposefully populated to in- clude four options from both ends of the conservative–liberal ideological spectrum, re- spectively. A conscious effort was also made to have a mix of formal political entities (whether they be parties or influential coalitions), on the one hand, and those repre- senting groups agitating for social change from outside the formal political arena, on the other. Indeed, in addition to guarding against threats to measurement validity, allowing respondents to first choose the groups they highly dislike provides valuable insights in and of itself. Specifically, mapping the potential targets of Tunisian intol- erance can reveal whether the citizenry is more wary of the groups associated with the current political consensus, or those social non-conformists pushing to change society largely through informal means. Which way this balance tips, as I argue below, could indicate the near and long term challenges this nascent democracy will face. It is worth briefly reviewing the rationale behind each group’s inclusion in the choice set: • UGTT may initially seem an odd choice to include in a set of potentially least-liked groups given that they initiated the National Dialogue Quartet and helped stabilize Tunisia’s faltering democracy. Yet, early in the 2013 political crisis, they drew quite a bit of ire for calling a general strike and leading the charge for the Islamist government’s resignation when tensions were nearing their peak. Today, they still wield significant political power. • Old regime elites would naturally be a target of animosity for many who lived during Ben Ali’s tenure and participated in the protests that helped topple 93 him. Their continued salience owes to Nidaa Tounes, one of the country’s two major political blocs, being deeply rooted in the old regime (while, to be sure, also boasting support from other segments in society). The debate over how to properly balance between holding these elites accountable and incorporating them into the political system continues to inform Tunisian politics. • Atheists in the Middle East are generally forced to keep their beliefs hidden and lead a conformist public life. While Tunisia is not known for religious conservatism, atheism can still draw sanctions from the citizenry and even the authorities (Khlifi 2018). • Homosexual rights activists in Tunisia have been more vocal in the years since the uprising (“Gay-Rights Activists” 2018), however that may simply make them a more salient target for intolerance. In a country where 94% of the populace believe that society should not accept homosexuality (The Global Divide on Homosexuality 2013), those who fight for homosexual rights are likely to attract animosity and repression. • Socialists have had a storied history in the Arab world, but one that features no shortage of detractors. A number of socialist political parties are active in Tunisia, with most of the notable ones allied with other leftist groups under the Popular Front banner. • Feminists in Tunisia may have fewer restrictions on their rights than their counterparts across the region, but that does not mean that they do not face any backlash for their effort. This de facto sanctioning is particularly evident 94 whenever their agenda bumps up against firmly rooted religious doctrine, as is the case with recent efforts to amend the country’s inheritance law to grant men and women equal shares (Associated Press 2018). • Ennahda members, as with all Islamists in the region and elsewhere, are fre- quently viewed with suspicion. Skeptics not only doubt Islamists’ democratic commitment, but often lump them together with extremist elements that may indeed deserve sanction. • Hizb ul Tahrir avowedly seeks to establish a caliphate, and frequently advocates for this goal through incendiary rhetoric. Yet, its activities have remained non- violent (at least in Tunisia). Nonetheless, several efforts have been made to ban the group outright since 2011. Once respondents indicate their least-liked group, they are then asked to choose their next most disliked group from the remaining options. Of these two choices, one is randomly assigned to populate each individual’s respective tolerance battery. The survey asked whether an assigned group: 1. Should be banned from running for any public office. 2. Should be under government surveillance. 3. Should be allowed to hold public rallies in my city/town. [REVERSE CODED] 4. Should be outlawed. 95 The composition of this additive index (a=.56) is generally in line with previous research, although it is common for there to be some discrepancy between any two given studies. One pertinent reason for this divergence is that context can blunt the relevance of a particular item in the typical tolerance battery. For example, although asking respondents about the removal of a book from their local public library is a relatively standard query in the US given the institution’s civic importance (Fallows 2016), libraries do no have the same significance in Tunisia. 4.3.2 Key Independent Variables and Controls I model the determinants of political tolerance (capitalized and italicized here- after) using four sets of explanatory variables and one set of controls. This analysis draws on the constructs that prior work has empirically and/or theoretically linked to the expression of tolerant attitudes, but also includes variables that are relevant to studying this topic in a Muslim-majority context. The first set of variables assesses the impact of threat perception (in reference to the respondent’s assigned group), which is one of the strongest and most consis- tent correlates of intolerance (e.g., Sullivan, Piereson, and Marcus 1982; Marcus et al. 1995; Gibson and Gouws 2003). Primarily, this attitude manifests itself through the belief that a target group could undermine societal norms or the social order. This Sociotropic Threat is tapped through two questions: 1) [ASSIGNED GROUP] are a threat to our society’s way of life; and 2) [ASSIGNED GROUP] are dangerous to the lives of normal people in this country. A more common conception of threat would likely call to mind the danger to 96 one’s personal status and safety. Surprisingly, however, this mode of threat has rarely been found to meaningfully affect tolerance levels (Gibson 2010). Yet, in the context of democratization the political stakes can feel heightened, increasing the odds of violence and even the sense of being existentially threatened (Hamid 2014). This Egocentric Threat is similarly scaled via two indicators: 1) [ASSIGNED GROUP] pose a threat to my own and/or my family’s personal safety; and 2) If [ASSIGNED GROUP] had their way, they would reduce my own and/or my family’s personal freedom. A third, less common measure of threat rests on the potential for a disliked group to gain political power. Presumably, the more one believes that a threaten- ing group can actualize their vision of society, the likelier one is to be intolerant towards them. This seemingly straightforward proposition has found surprisingly little empirical backing, however (Marcus et al. 1995; Gibson and Gouws 2003). We would nonetheless expect Group Power to negatively correlate with tolerance in the Tunisian context given the still stark Islamist/non-Islamist societal divide. The next set of variables evaluates two factors that should augment tolerance through a process of democratic learning. Education is generally the strongest pos- itive predictor of tolerant attitudes given that the more educated are both better socialized in the relevant norms and can more readily draw on a cache of civil lib- ertarian considerations to counteract instinctual intolerance. It could be, however, that Tunisians have not received the type of education that would tend to engender tolerance, given that at the time this survey was in the field, they were only seven years removed from authoritarian rule. To account for this possibility, I also include 97 a dichotomous measure indicating whether or not a respondent is Multi-lingual. Al- though an uncommon component in models of political tolerance, in this context, language skills may serve as a proxy for access to the environments (academic or otherwise) that may facilitate socialization to tolerance norms. Complementary sets of religiosity variables are included to address both the par- ticular and general expectations that these factors will influence tolerance. First, a dichotomous variable indicating whether respondents Prefer a Secular Party and an additive index capturing various dimensions of Sharia Support together address the impact of public religiosity. Second, the private manifestations of adherence to Islam are tapped through a variable gauging Quran Literalism and Prayer Frequency, two stock determinants of intolerance from the earliest studies in the literature (Stouf- fer 1955; Nunn, Crockett, and Williams 1978; Sullivan, Piereson, and Marcus 1982; McClosky and Brill-Scheuer 1983). The final set of variables includes a standard psychological control (Dogmatism) along with a suite of demographic variables (Female, Age, Income, and Rural), and the respondent’s Choice Assignment. 4.3.3 Experimental Design To examine the elasticity of tolerance in Tunisia, I presented respondents with two arguments designed to persuade them to move away from their initial judgment. After completing the tolerance battery, respondents were coded as either initially tolerant or initially intolerant based on their response to the public rally question outlined above. I chose this question since 1) it has clear salience in contemporary 98 Tunisia, given the recurring demonstrations throughout the country in the years since Ben Ali’s ouster, 3 and 2) it is quite amenable to realistic and reasonable counter arguments in either direction (as opposed to, say, arguing for feminists to be outlawed outright). Regardless of respondents; initial tolerance judgment on the rally question, they were prompted with both a “social order” and a “rights and freedoms” counter ar- gument that read as follows: Initially Tolerant: Earlier, you responded that [ASSIGNED GROUP] should be allowed to hold a public rally in your town or city. 1. Such rallies, however, could attract disorder. Suppose that community leaders warned that there would be a heightened risk of violence between the orga- nizers’ supporters and opponents, would you still feel that the rally should be allowed? 2. Some argue that [ASSIGNED GROUP] should not be allowed to hold public rallies because they are a threat to democracy in Tunisia. Considering this argument, would you still feel that the rally should be allowed? Initially Intolerant: Earlier, you responded that [ASSIGNED GROUP] should NOT be allowed to hold a public rally in your town or city. 1. What if local community leaders assured the public that there would not be any violence during the demonstration? Would you support allowing the rally? 3. In 2017 alone, for example, more than 200 protests took place in nearly all governorates (Carboni 2017). 99 2. Some argue that [ASSIGNED GROUP] should be allowed to hold public rallies because limiting the rights of those we disagree with would make it easier for the government to limit the rights of those we support-or even limit OUR rights. Considering this argument, would you support allowing the rally? Compared to other domains of public opinion, experimental methods are not as common in political tolerance studies. Yet, there are still some expectations that can be drawn from extant research. Notably, the normative valence of these expected outcomes is largely negative. That is, prior findings consistently point to the resilience of intolerant attitudes and the relative malleability of initially tolerant appraisals. Given the tensions and uncertainty that remain in Tunisian society, it is likely that this empirical pattern persists. This is not to say, however, that the results will necessarily be invariable across treatment conditions and sample subsets. Although there is no compelling a priori reason to expect the “social order” prompt to prove more effective in convincing Tunisians to tolerate their political enemies than the “rights and freedoms” prompt (or vice versa), each outcome has its own meaningful implications. On the one hand, if Tunisians are more persuaded by a promise to stave off violence, then it would indicate that short-term risk is the paramount consideration in respondents’ tolerance judgments. If, on the hand, appeals to democratic reciprocity prove more persuasive, then it could mean that Tunisians have begun making more systemic, long-term calculations for their actions. There is also reason to believe that heterogeneous treatment effects would be present within certain subsets of the sample. For instance, we may observe a divide 100 between Islamists and secularists. Despite conceding power at the national level, En- nahda remains a formidable force in Tunisian politics, especially at the sub-national level, which is one reason that municipal elections were delayed for three years until June 2018. Fears of a reversion to Islamist dominance may drive secularists to ini- tially more intolerant views than those respondents who prefer a religious party and, subsequently, more firmly entrench those views. Another divide may emerge between those who strongly dislike formal political entities rather than groups agitating for social change primarily from outside the formal political framework. If the former attract more intolerance, then we would expect those respondents who were asked about these groups to more stubbornly hold on to their initial judgments. The same scenario would apply should social agitators be found to be the primary targets of intolerance. 4.4 Results and Discussion To facilitate interpretation and presentation, all variables (including political tol- erance) are rescaled from 0-1. In effect, this means that the coefficients correspond to the mean point difference on the tolerance scale between the highest and lowest values on any given variable, holding all other variables constant. I use OLS re- gression to estimate the determinants of political tolerance and, given the complex survey data, present robust standard errors. 4 4. I also estimated the models using negative binomial regression, given the large proportion of wholly intolerant responses (i.e., “0” scores on the tolerance index). This analysis yielded identical results to the OLS regression in terms of statistical significance and directionality. 101 4.4.1 The Structure of Tunisian Political Tolerance Figure 4.1 shows the distribution of target groups across the sample. 5 Over one- fifth of respondents chose either atheists or homosexual rights activists, respectively, as groups they highly dislike. Ennahda members come in third followed by old-regime elites and feminists in approximately equal proportions. Surprisingly, a relatively low percentage of Tunisians chose Hizb ul-Tahrir as a least-liked group despite multiple attempts to ban the organization outright in recent years (AFP 2016; AFP 2017). UGTT members fall slightly behind Hizb ul-Tahrir, with socialists garnering the least animosity among Tunisians. 6 As Figure 4.2 shows, atheists and homosexual rights activists are not only the most disliked, but also the least tolerated. The remaining groups mostly elicit statis- tically similar levels of intolerance, with socialists and UGTT members (particularly the latter) marginally more tolerated than the others. Turning to the types of groups that Tunisians are unwilling to tolerate, a clear pattern emerges: disliked groups active in formal politics enjoy significantly less in- tolerance than those agitating for social reform from outside the formal political arena. Figure 4.3 compares respondents’ mean levels of tolerance and threat percep- 5. Note, this distribution represents the groups that respondents were assigned from among their first or second choices, not the total proportion of either choice. 6. Figure 4.A.1 in Appendix 4.A shows little change between the distributions of group selection conditioned on choice assignment. Two notable exceptions are homosexual rights activists and Ennahda members, both of whom have a higher likelihood of being chosen as a respondent’s most disliked group. 102 Figure 4.1: Distribution of Disliked Groups Figure 4.2: Mean of Political Intolerance, by Disliked Group tion (both egocentric and sociotropic) toward these two broad sets of target groups. 7 7. The “Social Agitator” group comprises those respondents who chose feminists, atheists, and homosexual rights activists, with all other remaining groups categorized under “Political Actor. ” Figure 4.A.2 in Appendix 4.A shows that the significant differences are robust to an alternative coding scheme with only the two most disliked groups under each label populating the category. 103 On each of these metrics, formal political entities score significantly lower than their informally associated counterparts. Figure 4.3: Mean of Tolerance and Threat Levels, by Disliked Group Type The OLS regression analysis further reinforces this finding. Table 4.1 models political tolerance in Tunisia as a function of key contextual, psychological, and demographic variables. Turning to the first set of predictors, Sociotropic Threat and Egocentric Threat each significantly decrease tolerance, as expected. Of these two factors, however, the former is far more influential. To be sure, Egocentric Threat exerts a strongly negative influence on tolerance in its own right, an effect that is all the more noteworthy given that it is at odds with prior research that finds threats to personal safety tend not to meaningfully predict intolerance (Marcus et al. 1995; Gibson 2006). Still, across all the models, moving from the lowest to the highest value on the Sociotropic Threat scale effectively halves a respondent’s estimated level of tolerance, all else equal. It is the third variable in this set, however, that most enhances our understanding 104 Table 4.1: Determinants of Political Tolerance–Full Sample (1) (2) (3) (4) (5) Threat Sociotropic Threat −0.24 ∗∗∗ (0.03) −0.25 ∗∗∗ (0.03) −0.25 ∗∗∗ (0.03) −0.26 ∗∗∗ (0.03) −0.26 ∗∗∗ (0.03) Egocentric Threat −0.14 ∗∗∗ (0.03) −0.14 ∗∗∗ (0.03) −0.13 ∗∗∗ (0.03) −0.12 ∗∗∗ (0.03) −0.11 ∗∗ (0.03) Group Power 0.08 ∗∗∗ (0.02) 0.08 ∗∗∗ (0.02) 0.08 ∗∗∗ (0.02) 0.07 ∗∗∗ (0.02) 0.08 ∗∗∗ (0.02) Democratic Learning Education 0.14 ∗∗∗ (0.04) 0.17 ∗∗∗ (0.04) Multi-lingual −0.05 ∗ (0.02) −0.05 ∗ (0.02) Public Religiosity Prefer Secular Party 0.01 (0.02) 0.02 (0.02) Sharia Support −0.08 ∗ (0.04) −0.09 ∗ (0.04) Private Religiosity Quran Literalism −0.04 (0.02) −0.02 (0.02) Prayer Frequency −0.01 (0.03) 0.00 (0.03) Controls Dogmatism −0.08 ∗∗ (0.03) −0.07 ∗∗ (0.03) −0.06 ∗ (0.03) −0.09 ∗∗ (0.03) −0.07 ∗ (0.03) Female −0.01 (0.02) −0.01 (0.02) −0.01 (0.02) −0.01 (0.02) −0.02 (0.02) Age −0.03 (0.03) 0.00 (0.03) −0.02 (0.03) −0.02 (0.03) 0.01 (0.03) Income 0.07 (0.04) 0.02 (0.04) 0.06 (0.04) 0.04 (0.04) −0.02 (0.04) Rural 0.03 (0.02) 0.03 (0.02) 0.03 (0.02) 0.03 (0.02) 0.03 (0.02) Choice Assignment 0.00 (0.02) 0.00 (0.02) 0.00 (0.02) 0.00 (0.02) −0.00 (0.02) Constant 0.51 ∗∗∗ (0.04) 0.49 ∗∗∗ (0.04) 0.55 ∗∗∗ (0.05) 0.55 ∗∗∗ (0.05) 0.53 ∗∗∗ (0.05) N 710 710 623 677 600 Standard errors in parentheses ∗ p< 0.05, ∗∗ p< 0.01, ∗∗∗ p< 0.001 105 of contemporary Tunisian tolerance. Specifically, respondents were significantly less willing to limit the freedoms of a disliked group if they felt that its members had a chance to gain political power. This relationship runs counter to what we would expect in theory, but points to an on-the-ground reality that (at least tepidly and relatively) grants the benefit of the doubt to those entities that have shown a will- ingness to work within the political system. Gibson and Gouws (2003, 69) report a similar result in their study of South African intolerance, but suggest that the association signals a reluctance to exercise intolerance for fear of the repercussions. Although it is difficult to dismiss this interpretation altogether, it is not clear why respondents would volunteer that they highly dislike a group but would be fearful of even hypothetically limiting their rights. A more likely basis for this finding is fear of systemic, rather than personal repercussions. That is, respondents may be calculating that the risks of undermining the country’s fragile institutions outweigh the benefit (whether psychological or societal) of neutralizing one’s political enemies. To be sure, blunting the desire to be intolerant by considering the implications of such actions can be cognitively demanding. Yet, there is some evidence that Tunisians can meet this challenge. Specifically, Table 4.1 shows that Education has an impressively robust positive correlation with tolerance. While a shift in the nature of formal education may in part explain this strong relationship, experiencing the “rough and tumble” of democratic politics (Peffley and Rohrschneider 2003) likely accounts for much of the alignment between educational attainment and tolerant attitudes. That is, the higher citizens’ education levels, the likelier they are to process the ebbs and flow of a transition from democratic rule and connect probable 106 causes to likely effects. Curiously, Multilingualism is negatively associated with tolerance. A difference in worldview likely drives this effect, given that urbanity, income, and education level (three of the more likely confounding variables) are all held constant, it appears that the effect of multilingualism is indeed driven by a difference in worldview. Last, the influence of religiosity (or lack thereof) similarly points to a shift in the structure of tolerance in Tunisia. Neither of the private religiosity variables registers as significant in the full model (Model (5) in Table 4.1). Yet, whereas Islamists were conditionally more intolerant in the initial years following the Arab Uprisings, 8 the tendency to Prefer a Secular Party now appears unrelated to tolerance one way or another. Indeed, as Figure 4.4 shows, there is little divergence in the significance, directionality or magnitude of the determinants of tolerance between Islamists and secularists. Moreover, in terms of group choice, tolerance levels and threat percep- tion, partisans of religious parties are statistically on a par with their more secular counterparts (see Figures 4.A.3 and 4.A.4 in Appendix 4.A). One of the more no- table findings in this plot is the invariable impact of Sharia Support. This equivalency, particularly in light of the findings thus far regarding group type, suggests that sup- porting sharia may serve as a proxy for a general conservative/traditional outlook on society. 8. See Chapter 3, Tables 3.2 and 3.3 and accompanying discussion. 107 Figure 4.4: Determinants of Political Tolerance, Islamist vs. Secular 4.4.2 The Maleability of Tunisian Political Tolerance The preceding analysis offered the first comprehensive rendering of political tol- erance in Tunisia, yet it was still only a cross-section. Particularly in a region so characterized by flux in recent years, an initial attitude may prove ephemeral. How easily can Tunisians be pushed toward in/tolerance? Which arguments are the most effective? And what contextual or dispositional factors moderate this effect? Turning to the experimental findings, Figure 4.5 demonstrates that, in line with prior research, Tunisians are much more likely to move from tolerance to intolerance than the other way around. In terms of the respective prompts’ effectiveness, there was virtually no difference between the Social Order and Rights and Freedoms argu- ments among those who were initially tolerant. Those that were initially intolerant 108 were slightly more persuaded to change their judgment by a Social Order prime, though not to a statistically significant degree. Figure 4.5: Attitude Change, by Initial Attitude and Prompt This aggregate plot masks some variance based on the strength of respondents’ initial attitude. Unfortunately, as Figure 4.6 shows, this variability is largely absent among the initially tolerant, who shift toward intolerance at relatively similar rates regardless of whether their original position was weakly or strongly held. The same pattern does not emerge among the initially intolerant, who are much more willing to shift toward a tolerant position if they were only weakly intolerant to start. Along the same lines as the descriptive and inferential findings, the Is- lamist/secularist divide does not appear to moderate the experimental effects very much. Figure 4.8 shows that, regardless of initial attitude or prompt, the propor- tion of respondents willing to shift their tolerance judgment is independent of their non-/religious party preference. The one partial exception is that Islamists do not appear to be as easily swayed by threats to their Rights and Freedoms, either relative 109 Figure 4.6: Attitude Change, by Strength of Initially Tolerant Attitude Figure 4.7: Attitude Change, by Strength of Initially Intolerant Attitude to the Social Order prompt or to secularists, overall. This tendency could be a result of power differential, with Islamist party supporters feeling relatively secure that a group could not repress their views under the current political balance of power. It could also be that Islamist partisans are simply unconvinced that the type of group they were initially tolerant toward (in the main, the kind of groups unlikely to gain a foothold in relatively conservative society) could ever realistically threaten their rights. As the following plot shows, this latter interpretation is supported by the moderating effects attendant to group type. Whether respondents chose an Informal Social Agitators or Formal Political Ac- tors as their target group moderated the elasticity of their tolerance judgment to a noticeable degree. Figures 4.9 and 4.10 plot the treatment effect of each persuability prompt controlling for the strength of respondents’ initial attitude. In line with the findings thus far, those with initially tolerant responses move away from this original position to a large and nearly uniform degree. The one exception to this trend is the relative ineffectiveness of the Rights and Freedoms prompt among respondents who 110 Figure 4.8: Attitude Change, by Islamist Status Figure 4.9: Attitude Change, by Dis- liked Group Type (Initially Tolerant) Figure 4.10: Attitude Change, by Dis- liked Group Type (Initially Intolerant) chose a Informal Social Agitator as their target group. Again, this may be due to a belief that either such groups simply would not be able to attain the power necessary to repress or, more charitably, that such groups would not choose to repress, even if given the power to do so. A much starker divide is evident among initially intolerant respondents. Within this subset, those who supported banning politically active groups from holding a rally in their town were significantly more likely to be persuaded toward tolerance 111 than those respondents wishing to ban atheists, homosexual rights activists, or fem- inists. Thus, not only are Tunisians more likely to single out the latter set of groups as their most disliked, but they are also more intolerant towards them and less likely to be persuaded away from that intolerance. 4.5 Conclusion This study presented the first comprehensive analysis of political tolerance in the post-Arab Uprisings context and the Arab Middle East in general. I find those groups that challenge traditional societal norms are more likely to engender the ex- treme dislike of Tunisian citizens compared to those groups that pursue their agenda from within the formal political system. Moreover, Social Agitators attract higher levels of intolerance and are perceived to be more threatening (both in sociotropic and egocentric terms) than their formal political counterparts. In line with this di- vide, a target group’s potential for holding political power negatively correlates with intolerance. Tunisians, it seems, are less eager to limit the civil liberties of those working within the political system, perhaps out of fear that it would undermine the an already fragile political structure. What implications do these findings hold for Tunisian democracy? The answer to this question may depend on the time horizon. Certainly, it is troubling to find that those advocating for social change in terms of gender, sexuality, or religiosity face such high and resilient levels of intolerance. This particular constellation of targeted groups is especially worrying given Inglehart and Welzel’s (2005) belief that it is nec- 112 essary for countries to move away from more traditional values in order to advance and maintain democracy. Yet, I argue that the prospects for Tunisian democracy, at least in the short term, would be much bleaker if the citizenry’s intolerance were focused on the parties and organizations vying for influence in the formal political structure. This latter scenario could indicate the sort of zero-sum political environ- ment that would be ripe for authoritarian reversal. That Tunisians do not regard their ideological enemies in the formal political sphere as warranting exclusionary action should be seen as a bulwark against authoritarian backsliding and, therefore, grounds for tempered optimism. Longer term, of course, the patterns documented in this study could prove more problematic. For one thing, the focused intolerance of those who challenge “tra- ditional values” may pave the way for avowedly illiberal policies. It was indeed a focused intolerance of communists that buttressed the various McCarthyite attacks on this group’s civil liberties in the 1950s. Aside from the threat that this prospect poses for particular subsets of the population, this latent willingness to go along with the repression of social non-conformists could provide a ready-made scapegoat for politicians seeking to rally conservative populist sentiment. Social Agitators of various stripes in Tunisia could therefore server the same political function as “im- migrants” in North America and Western Europe have (and still do) in the 2000s. Along these same lines, another challenge is the malleability of Tunisian tolerance. Tolerant respondents were far easier to move away from their initial attitude than their intolerant counterparts. Thus, even though the structure of Tunisian political tolerance at the moment bodes well for the country, this optimism is tempered by 113 the possibility of reconfiguration should conducive circumstances develop. If, for example, the economic and security situation in Tunisia continues to fall betray citizens’ expectations, the political environment could prove ripe for a charismatic authoritarian seeking to consolidate power by constraining political space in return for promises of order and prosperity. Indeed, the increasing prominence of right-wing populists across the globe suggests that we may have overestimated the resilience of established democracies in the face of intolerant elites willing to activate social threat in service to their political ends. In light of these enduring challenges, Tunisia’s sustained vigilance against the forces that may undermine its democratic gains– including intolerance–seems not only prudent, but paramount in our current age. 114 Appendices 4.A Additional Figures and Tables Figure 4.A.1: Group Distribution, by Choice Assignment 115 Figure 4.A.2: Mean of Tolerance and Threat Levels, by Group Type (Alternative Coding) Figure 4.A.3: Group Distribution, by Islamist/Secularist Status 116 Figure 4.A.4: Mean of Tolerance and Threat Levels, by Islamist/Secularist Status 117 4.B Question Wording Variable (type) Composite Question(s) Answer Options Mosque Attendance (single indicator) Aside from weddings or funerals, how often do you go to a mosque? Almost never; A few times a year; Once or twice a month; once a week; more than once a week Prayer Frequency (single indicator) How often do you preform the five daily prayers (salat)? Almost never; Occasionally; Most days; Almost always Dogmatism (index) There are two kinds of people in this world:,those who are for the truth and those who are against it Strongly Disagree – Strongly Agree (5pt) A group that tolerates too many differences of opinion among its own members cannot exist for long Strongly Disagree – Strongly Agree (5pt) To compromise with our political opponents is dangerous because it usually leads to the betrayal of our own side Strongly Disagree – Strongly Agree (5pt) Sharia Support (index) The government and parliament should enact criminal laws in accordance with Islamic law Strongly Disagree – Strongly Agree (5pt) Banks charging interest contradicts the teachings of Islam Strongly Disagree – Strongly Agree (5pt) The government and parliament should enact personal status laws (marriage, divorce) in accordance with Islamic law Strongly Disagree – Strongly Agree (5pt) When religious teachings guide public life, it does more harm than good for society Strongly Disagree – Strongly Agree (5pt) 118 References AFP. 2016. Tunisia Calls for Ban on Islamist Party Hizb Ut-Tahrir. https://www.alaraby.co.uk/english/news/2016/9/8/tunisia-calls-for-ban- on-islamist-party-hizb-ut-tahrir, September. . 2017. Tunisia Radical Islamist Party Banned for One Month. https://www.news24.com/Africa/News/tunisia-radical-islamist-party-banned- for-one-month-20170607, June. Amara, Tarek. 2018. “Independent Candidates Get Most Votes in Tunisia’s Municipal Election.” Reuters (May). Arrowood, Emily. 2015. Here Are The Conservative Pun- dits Branding Black Lives Matter A ”Hate Group”. https://www.mediamatters.org/research/2015/09/02/here-are-the- conservative-pundits-branding-blac/205337, September. Associated Press. 2018. “Tunisian Women March for Equal Inheritance Rights.” Washington Post (March). 119 Barra, Mary. 2017. GM Is about Unity and Inclusion, and so Am I. We Must Rein- force Tolerance and Diversity.Pic.Twitter.Com/JgUQZNKzsz. Tweet, August. Beinart, Peter. 2017. The Rise of the Violent Left. https://www.theatlantic.com/magazine/archive/2017/09/the-rise-of-the- violent-left/534192/, August. Bellin, Eva. 2014. Explaining Democratic Divergence: Why Tunisia Has Succeeded and Egypt Has Failed. http://pomeps.org/2014/12/10/explaining-democratic- divergence/. Berry, Justin, Youssef Chouhoud, and Jane Junn. 2016. “Reaching Beyond Low- Hanging Fruit: Surveying Low Incidence Populations. ” In Oxford Handbook of Polling and Polling Methods. Oxford; New York: Oxford University Press. Brown, Nathan J. 2013. “Egypt’s Failed Transition. ”

Volume 24, Number 4, October 2013

, Journal of Democracy 24 (4): 45–58. Burge, Ryan. 2013. “Using Matching to Investigate the Relationship between Religion and Tolerance.” Politics and Religion 6 (02): 264–281. Busch, Beverly G. 1998. “Faith, Truth, and Tolerance: Religion and Political Toler- ance in the United States.” Ph.D., The University of Nebraska - Lincoln. Carboni, Andrea. 2017. Tunisia – May 2017 Update. https://www.crisis.acleddata.com/tunisia-may-2017-update/, June. 120 Carroll, Lauren, and Lewis Jacobson. 2015. Trump Cites Shaky Survey in Call to Ban Muslims from Entering US. http://www.politifact.com/truth-o- meter/statements/2015/dec/09/donald-trump/trump-cites-shaky-survey-call- ban-muslims-entering/, December. Cavatorta, Francesco, and Fabio Merone. 2013. “Moderation through Exclusion? The Journey of the Tunisian Ennahda from Fundamentalist to Conservative Party.” Democratization 20, no. 5 (August): 857–875. Chong, Dennis, and James N. Druckman. 2007. “Framing Theory. ” Annual Review of Political Science 10, no. 1 (May): 103–126. Chong, Dennis, and Dukhong Kim. 2006. “The Experiences and Effects of Economic Status among Racial and Ethnic Minorities. ” The American Political Science Review 100 (3): 335–351. Chrisafis, Angelique. 2011. “Zine Al-Abidine Ben Ali Forced to Flee Tunisia as Protesters Claim Victory.” The Guardian (January). Coffman, Katherine B., Lucas C. Coffman, and Keith M. Marzilli Ericson. 2013. The Size of the LGBT Population and the Magnitude of Anti-Gay Sentiment Are Substantially Underestimated. Working Paper 19508. National Bureau of Economic Research, October. Cohen, Richard. 2016. Black Lives Matter Is Not a Hate Group. https://www.splcenter.org/news/2016/07/19/black-lives-matter-not-hate- group, July. 121 Cohen, Zachary, and Barbara Starr. 2017. US Military Leaders Condemn Racism after Charlottesville Violence. https://www.cnn.com/2017/08/16/politics/joint- chiefs-charlottesville-racism/index.html, August. Crunden, EA. 2017. Communities Hit by Rising Hate Crimes Say Trump’s Rhetoric Is Having a Devastating Impact, November. Dahl, Robert A. 1956. A Preface to Democratic Theory. Chicago: University of Chicago Press. Dana, Karam, Matt A. Barreto, and Kassra A. R. Oskooii. 2011. “Mosques as Amer- ican Institutions: Mosque Attendance, Religiosity and Integration into the Po- litical System among American Muslims. ” Religions 2, no. 4 (September): 504– 524. Daragahi, Borzou. 2018. “Belt-Tightening Demands Put Tunisia’s Democracy at Risk.” The New York Times (May). Davis, Darren. 1995. “Exploring Black Political Intolerance. ” Political Behavior 17, no. 1 (March): 1–22. Davis, Darren, and Brian D. Silver. 2004. “Civil Liberties vs. Security: Public Opin- ion in the Context of the Terrorist Attacks on America. ” American Journal of Political Science 48, no. 1 (January): 28–46. Dhouib, Hayfa. 2017. Municipal Elections: The Law and the Consolidation of a Two- Tiered System (Al Intikhabat Albladiyya: Alqanun Wa Takris Al Aistiqtaab Al- thaan‘a). http://nawaat.org/portail/2017/04/24/, April. 122 Djupe, Paul A., Andrew R. Lewis, and Ted G. Jelen. 2016. “Rights, Reflection, and Reciprocity: Implications of the Same-Sex Marriage Debate for Tolerance and the Political Process.” Politics and Religion 9, no. 03 (September): 630–648. Djupe, Paul, and Brian R. Calfano. 2012. “American Muslim Investment in Civil Society Political Discussion, Disagreement, and Tolerance. ” Political Research Quarterly 65, no. 3 (September): 516–528. Duch, Raymond M., and James L. Gibson. 1992. “”Putting Up With” Fascists in Western Europe: A Comparative, Cross-Level Analysis of Political Tolerance.” The Western Political Quarterly 45, no. 1 (March): 237–273. Eisenstein, Marie A. 2006. “Rethinking the Relationship between Religion and Po- litical Tolerance in the US.” Political Behavior 28, no. 4 (December): 327–348. Esposito, John L. 2003. Unholy War: Terror in the Name of Islam. Oxford University Press. Fallows, Deborah. 2016. “The Library Card. ” The Atlantic (March). Fassihian, Dokhi, and Theo Wilson. 2018. Tunisian Democracy Stands at the Precipice. https://freedomhouse.org/blog/tunisian-democracy-stands-precipice, February. Fish, M. Steven. 2002. “Islam and Authoritarianism. ” World Politics 55 (1): 4–37. . 2011. Are Muslims Distinctive?: A Look at the Evidence. Oxford, New York: Oxford University Press, February. 123 Fletcher, Joseph F., and Boris Sergeyev. 2002. “Islam and Intolerance in Central Asia: The Case of Kyrgyzstan. ” Europe-Asia Studies 54 (2): 251–275. Freedom in the World: 2018. 2018. Technical report. Washington, D.C: Freedom House. Fricker, Scott, Mirta Galesic, Roger Tourangeau, and Ting Yan. 2005. “An Experi- mental Comparison of Web and Telephone Surveys. ” The Public Opinion Quar- terly 69 (3): 370–392. Fukuyama, Francis. 2001. We Remain at the End of History. http://www.independent.co.uk/voices/commentators/francis-fukuyama-we- remain-at-the-end-of-history-5363424.html, October. Gervais, Will M., and Maxine B. Najle. 2018. “How Many Atheists Are There?” Social Psychological and Personality Science 9, no. 1 (January): 3–10. Gibson, James L. 1987. “Homosexuals and the Ku Klux Klan: A Contextual Analysis of Political Tolerance.” The Western Political Quarterly 40, no. 3 (September): 427–448. . 1996. “The Paradoxes of Political Tolerance in Processes of Demo- Cratiza- tion.” Politikon: South African Journal of Political Studie 23, no. 2 (December): 5–21. . 2006. “Enigmas of Intolerance: Fifty Years after Stouffer’s Communism, Con- formity, and Civil Liberties. ” Perspectives on Politics 4, no. 1 (March): 21–34. 124 Gibson, James L. 2007. “Political Intolerance in the Context of Democratic Theory.” In Oxford Handbook of Political Behavior, edited by Russell J. Dalton and Hans- Dieter Klingemann, 323–341. Portion of title: Political behavior Bibliography note: Includes bibliographical references and indexes Electronic version: Table of contents only http://www.loc.gov/catdir/toc/ecip079/2007003074.html Control Number: ocm80360489. Oxford: Oxford University Press. . 2008. “Intolerance and Political Repression in the United States: A Half Century after McCarthyism. ” American Journal of Political Science 52, no. 1 (January): 96–108. . 2010. “The Political Consequences of Religiosity: Does Religion Always Cause Political Intolerance?” In Religion and Democracy in the United States: Danger or Opportunity?, edited by Alan Wolfe and Ira Katznelson, 147–175. New York: Russell Sage Foundation. . 2013. “Measuring Political Tolerance and General Support for Pro–Civil Liberties Policies Notes, Evidence, and Cautions. ” Public Opinion Quarterly 77, no. S1 (January): 45–68. Gibson, James L., and Richard D. Bingham. 1984. “Skokie, Nazis, and the Elitist Theory of Democracy.” Political Research Quarterly 37, no. 1 (January): 32–47. Gibson, James L., and Raymond M. Duch. 1991. “Elitist Theory and Political Tol- erance in Western Europe.” Political Behavior 13 (3): 191–212. 125 Gibson, James L., and Amanda Gouws. 2003. Overcoming Intolerance in South Africa: Experiments in Democratic Persuasion. (Cambridge studies in political psychology and public opinion). New York, NY: Cambridge University Press. Goldberg, Jonah. 2018. America Is Not As Intolerant As We Make It Out to Be. https://www.nationalreview.com/2018/04/america-not-most-racist-sexist- nation-progress-made/, April. Greenblatt, Jonathan. 2018. “The Resurgent Threat of White-Supremacist Violence.” The Atlantic (January). Gu´ erin, Daniel, Fran¸ cois Petry, and Jean Crˆ ete. 2004. “Tolerance, Protest and Demo- cratic Transition: Survey Evidence from 13 Post-Communist Countries. ” Euro- pean Journal of Political Research 43, no. 3 (May): 371–395. Hamid, Shadi. 2014. Temptations of Power: Islamists and Illiberal Democracy in a New Middle East. 1 edition. Oxford ; New York: Oxford University Press, April. Hanusch, Marek. 2013. “Islam and Democracy: A Response. ” Public Choice 154, nos. 3-4 (March): 315–321. Harell, Allison. 2010. “Political Tolerance, Racist Speech, and the Influence of Social Networks.” Social Science Quarterly 91, no. 3 (September): 724–740. Hatewatch. 2017. Anti-Sharia Law Bills in the United States. https://www.splcenter.org/hatewatch/2017/08/08/anti-sharia-law-bills-united- states, August. 126 Heckathorn, Douglas D. 1997. “Respondent-Driven Sampling: A New Approach to the Study of Hidden Populations. ” Social Problems 44 (2): 174–199. Henrich, Joseph, Steven J. Heine, and Ara Norenzayan. 2010. “The Weirdest People in the World?” Behavioral and Brain Sciences 33, nos. 2-3 (June): 61–83. Herb, Michael. 2005. “No Representation without Taxation? Rents, Development, and Democracy.” Comparative Politics 37 (3): 297–316. Holland, Patrick. 2016. AT&T CEO Passionately Defends Black Lives Matter. https://www.cnet.com/news/at-t-ceo-randall-stephenson-defends-black-lives- matter/, September. Huntington, Samuel P. 1984. “Will More Countries Become Democratic?” Political Science Quarterly 131 (2): 237–266. . 1993. “The Clash of Civilizations?” Foreign Affairs 72, no. 3 (July): 22–49. . 2004. Who Are We?: The Challenges to America’s National Identity. Simon and Schuster. In Re Ahmed Hassan. 1942, December. Inglehart, Ronald, and Christian Welzel. 2005. Modernization, Cultural Change, and Democracy: The Human Development Sequence. Cambridge, UK ; New York: Cambridge University Press, August. Jackman, Robert W. 1972. “Political Elites, Mass Publics, and Support for Demo- cratic Principles.” The Journal of Politics 34, no. 3 (August): 753–773. 127 Jamal, Amaney. 2005. “The Political Participation and Engagement of Muslim Amer- icans Mosque Involvement and Group Consciousness. ” American Politics Re- search 33, no. 4 (January): 521–544. Jamal, Amaney, and Mark Tessler. 2008. “Attitudes in the Arab World. ” Journal of Democracy 19, no. 1 (January): 97–110. Johnson, Daryl. 2017. Return of the Violent Black Nationalist. https://www.splcenter.org/fighting-hate/intelligence-report/2017/return- violent-black-nationalist, August. Johnson, Jenna, and Abigail Hauslohner. 2017. “‘I Think Islam Hates Us’: A Timeline of Trump’s Comments about Islam and Muslims. ” Washington Post (May). Kedourie, Elie. 1992. Politics in the Middle East. Oxford University Press. Kennedy, Joe. 2017. We Won’t Allow @POTUS’s False Equivalency of #Char- lottesville Divide Us & Will Always Proudly Stand on the Side of Tolerance and Progress.Pic.Twitter.Com/OIO099gMNU. Tweet, August. Khlifi, Roua. 2018. Taboo of Atheism in Tunisia, January. Kim, Junghyoun, and Yang Zhong. 2010. “Religion and Political Tolerance in South Korea.” East Asia: An International Quarterly 27, no. 2 (June): 187–203. Kinder, Donald R., and Cindy D. Kam. 2009. Us Against Them: Ethnocentric Foun- dations of American Opinion. Chicago: University of Chicago Press. 128 Kinder, Donald R., and Nicholas Winter. 2001. “Exploring the Racial Divide: Blacks, Whites, and Opinion on National Policy. ” American Journal of Political Science 45, no. 2 (April): 439–456. Kirkpatrick, David D., and Mayy El Sheikh. 2012. “President Morsi in Egypt Seizes New Powers.” The New York Times (November). Kishi, Katayoun. 2017. Assaults against Muslims in U.S. Surpass 2001 Level, Novem- ber. Klein, Ezra. 2017. Cory Booker on the Spiritual Dimension of Politics, May. . 2018. White Threat in a Browning America. https://www.vox.com/policy- and-politics/2018/7/30/17505406/trump-obama-race-politics-immigration, July. Lajevardi, Nazita. 2016. The Media Matters: Muslim American Portrayals and the Effects on Mass Attitudes. Conference Paper. Philadelphia, PA, September. Lawrence, Bruce. 2000. Shattering the Myth. Edition Unstated edition. Princeton, NJ: Princeton University Press, April. Lee, Francis L. F. 2013. ““Tolerated One Way but Not the Other”: Levels and De- terminants of Social and Political Tolerance in Hong Kong. ” Social Indicators Research 118, no. 2 (September): 711–727. Levin, Brian. 2017. How the Rise in Hate Crimes against Muslims Shows What Politicians Say Matters — Opinion. http://www.newsweek.com/islamophobia- america-rise-hate-crimes-against-muslims-proves-what-politicians-640184, July. 129 Lewis, Bernard. 1996. “A Historical Overview. ” Journal of Democracy 7, no. 2 (April): 52–63. Lipset, Seymour Martin. 1960. Political Man: The Social Bases of Politics. Control Number: ocm00709914. Garden City: Doubleday. Lopez, German. 2017. The Past Year of Research Has Made It Very Clear: Trump Won Because of Racial Resentment. https://www.vox.com/identities/2017/12/15/16781222/trump-racism- economic-anxiety-study, December. Macdonald, Geoffrey, and Luke Waggoner. 2018. “Dashed Hopes and Extremism in Tunisia.” Journal of Democracy 29, no. 1 (January): 126–140. Marcus, George E., John L. Sullivan, Elizabeth Theiss-Morse, and Sandra L. Wood. 1995. With Malice toward Some: How People Make Civil Liberties Judgments. New York: Cambridge University Press. Marks, Monica. 2015. “How Egypt’s Coup Really Affected Tunisia’s Islamists. ” The Washington Post (March). . 2017a. ’Letting Go of Every Principle’: Tunisia’s Democratic Gains un- der Threat. http://www.middleeasteye.net/columns/letting-go-every-principle- tunisia-s-democratic-gains-under-threat-673085990, July. . 2017b. Tunisia’s Unwritten Story. Technical report. Washington, D.C: The Century Foundation, March. 130 Marsden, Peter V. 2012. Social Trends in American Life: Findings from the General Social Survey Since 1972. Princeton University Press. Masoud, Tarek. 2013. “The Muslim Brotherhood in Egypt. ” In The Oxford Handbook of Islam and Politics, edited by John L. Esposito and Emad El-Din Shahin, 475– 502. New York, NY: Oxford University Press. . 2015. “Not Ready for Democracy: Pluralism, Modernization, and the Arab Spring.” San Francisco, CA. Masuoka, Natalie, and Jane Junn. 2013. The Politics of Belonging: Race, Public Opinion, and Immigration. Chicago: University Of Chicago Press. Mather, Darin, and Eric Tranby. 2014. “New Dimensions of Tolerance: A Case for a Broader, Categorical Approach. ” Sociological Science 1:512–531. McClosky, Herbert, and Alida Brill-Scheuer. 1983. Dimensions of Tolerance: What Americans Believe about Civil Liberties. New York: Russell Sage Foundation. Mercier, Brett Gregory, Azim Shariff, and Adam Norris. 2017. “Muslim Mass Shoot- ers Are Perceived as Less Mentally Ill and More Motivated by Religion.” PsyArXiv (October). Moaddel, Mansoor. 2006. “The Saudi Public Speaks: Religion, Gender, and Politics.” International journal of Middle East studies 38 (1): 79–108. Mondak, Jeffery J., and Mitchell S. Sanders. 2003. “Tolerance and Intolerance, 1976–1998.” American Journal of Political Science 47, no. 3 (July): 492–502. 131 Mozaffari, Mehdi. 2007. “What Is Islamism? History and Definition of a Concept.” Politics, Religion & Ideology 8, no. 1 (March): 17–33. Nunn, Clyde Z, Harry J Crockett, and Allen J. Williams. 1978. Tolerance for Non- conformity. San Francisco: Jossey-Bass Publishers. Obama, Barack. 2004. Keynote Address at the 2004 Democratic National Convention. http://www.presidency.ucsb.edu/ws/?pid=76988, July. Peffley, Mark, and Jon Hurwitz. 2010. Justice in America: The Separate Realities of Blacks and Whites. (Cambridge studies in public opinion and political psychol- ogy); Cambridge studies in public opinion and political psychology. Cambridge ; New York: Cambridge University Press. Peffley, Mark, Marc L. Hutchison, and Michal Shamir. 2015. “The Impact of Persis- tent Terrorism on Political Tolerance: Israel, 1980 to 2011. ” American Political Science Review 109, no. 04 (November): 817–832. Peffley, Mark, and Robert Rohrschneider. 2003. “Democratization and Political Tol- erance in Seventeen Countries: A Multi-Level Model of Democratic Learning.” Political Research Quarterly 56, no. 3 (January): 243–257. Pew Research Center. 2011. Muslim-Western Tensions Persist. Technical report. Washington, D.C: Pew Research Center, July. . 2017. U.S. Muslims Concerned About Their Place in Society, but Continue to Believe in the American Dream. Technical report. Washington, D.C: Pew Research Center. 132 The Global Divide on Homosexuality. 2013. http://www.pewglobal.org/2013/06/04/the- global-divide-on-homosexuality/, June. Poll: Growing Number of Americans Say Islam Encourages Violence Among Follow- ers. 2003, July. How the U.S. General Public Views Muslims and Islam. 2017, July. Phillips, Frank. 2017. “GOP Senate Candidates Divided over Proposal to Label Black Lives Matter a Hate Group - The Boston Globe. ” BostonGlobe.com (August). Pipes, Daniel. 2003. Militant Islam Reaches America. New York: W. W. Norton & Company. Potrafke, Niklas. 2012. “Islam and Democracy. ” Public Choice 151, nos. 1-2 (April): 185–192. . 2013. “Democracy and Countries with Muslim Majorities: A Reply and Up- date.” Public Choice 154, nos. 3-4 (March): 323–332. Qaheri, Mahgoob Ahmed. 2018. Will Tunisia Be Seeing a Military Coup or a Civil War? (Hal Tunis Fi Intidhaar Inkilaab ’askari Aw Harb Ahly?) https://www.raialyoum.com/index.php/%D9%87%D9%84- %D8%AA%D9%88%D9%86%D8%B3-%D9%81%D9%8A- %D8%A7%D9%86%D8%AA%D8%B8%D8%A7%D8%B1- %D8%A7%D9%86%D9%82%D9%84%D8%A7%D8%A8- %D8%B9%D8%B3%D9%83%D8%B1%D9%8A-%D8%A7%D9%88- %D8%AD%D8%B1%D8%A8-%D8%A3%D9%87/, April. 133 Reagan, Ronald. 1984. Remarks at a Meeting With Asian and Pacific-American Lead- ers. http://www.presidency.ucsb.edu/ws/index.php?pid=39556, February. Robertson, Graeme B., and Alfred C. Stepan. 2003. “An ”Arab” More Than a ”Mus- lim” Democracy Gap.”

Volume 14, Number 3, July 2003

, Journal of Democracy 14 (3): 30–44. Rokeach, Milton. 1973. The Nature of Human Values. New York: Free Press. Ross, Michael L. 2001. “Does Oil Hinder Democracy?” World Politics 53 (April): 325–361. Roster, Catherine A., Robert D. Rogers, Gerald Albaum, and Darin Klein. 2004. “A Comparison of Response Characteristics from Web and Telephone Surveys.” International Journal of Market Research 46, no. 3 (May): 359–373. Rowley, Charles K., and Nathanael Smith. 2009. “Islam’s Democracy Paradox: Mus- lims Claim to like Democracy, so Why Do They Have so Little?” Public Choice 139, nos. 3-4 (June): 273–299. SAALT. 2014. Under Suspicion, Under Attack. Technical report. Takoma Park, MD: South Asian Americans Leading Together. Sarkissian, Ani. 2012. “Religion and Civic Engagement in Muslim Countries. ” Journal for the Scientific Study of Religion 51, no. 4 (December): 607–622. Schleifer, Theodore. 2016. Donald Trump: ’I Think Islam Hates Us’. http://www.cnn.com/2016/03/09/politics/donald-trump-islam-hates- us/index.html, March. 134 Schoettmer, Patrick. 2015. “Mobilization and the Masjid: Muslim Political Engage- ment in Post-9/11 America. ” Politics, Groups, and Identities 3, no. 2 (April): 255–273. Shaheen, Jack. 2008. Guilty: Hollywood’s Verdict on Arabs After 9/11. Northampton, Mass: Olive Branch Pr, January. . 2014. Reel Bad Arabs: How Hollywood Vilifies a People. 3 edition. Northamp- ton, Mass: Olive Branch Pr, December. Shamir, Michal. 1991. “Political Intolerance among Masses and Elites in Israel: A Reevaluation of the Elitist Theory of Democracy. ” The Journal of Politics 53, no. 4 (November): 1018–1043. Sides, John, and Kimberly Gross. 2013. “Stereotypes of Muslims and Support for the War on Terror.” The Journal of Politics 75, no. 03 (July): 583–598. Smith-Spark, Laura. 2018. What’s behind the Protests in Tunisia? https://www.cnn.com/2018/01/10/africa/tunisia-protests-intl/index.html, January. Sniderman, Paul M., Professor Joseph F. Fletcher, Peter Russell, and Professor Philip E. Tetlock. 1996. The Clash of Rights: Liberty, Equality, and Legitimacy in Plu- ralist Democracy. 1St Edition edition. New Haven, CT: Yale University Press, November. 135 Sniderman, Paul M., Louk Hagendoorn, and Markus Prior. 2004. “Predisposing Fac- tors and Situational Triggers: Exclusionary Reactions to Immigrant Minorities.” American Political Science Review 98 (01): 35–49. Sniderman, Paul M., and Sean Theriault. 2004. “The Structure of Political Argu- ment and the Logic of Issue Framing. ” In Studies in Public Opinion: Attitudes, Nonattitudes, Measurement Error, and Change, 133–165. Princeton: Princeton University Press. Southern Poverty Law Center. 2018. The Year in Hate: Trump Buoyed White Supremacists in 2017, Sparking Backlash among Black Nationalist Groups. https://www.splcenter.org/news/2018/02/21/year-hate-trump-buoyed-white- supremacists-2017-sparking-backlash-among-black-nationalist, February. Spencer, Robert, ed. 2005. The Myth of Islamic Tolerance: How Islamic Law Treats Non-Muslims. Amherst, NY: Prometheus Books, January. Spierings, Niels. 2014. “The Influence of Islamic Orientations on Democratic Sup- port and Tolerance in Five Arab Countries. ” Politics and Religion 7, no. 04 (December): 706–733. Stepan, Alfred. 1988. Rethinking Military Politics. Princeton, N.J: Princeton Univer- sity Press. . 2012. “Tunisia’s Transition and the Twin Tolerations. ” Journal of Democracy 23 (2): 89–103. 136 Stouffer, Samuel Andrew. 1955. Communism, Conformity, and Civil Liberties: A Cross-Section of the Nation Speaks Its Mind. New York: Doubleday & Company. Stump, Scott. 2016. Trump: Brussels Is ‘Catastrophic,’ Waterboarding Paris Sus- pect ‘Would Be Fine’. https://www.today.com/news/donald-trump-responds- brussels-attacks-it-s-very-dangerous-city-t81716, March. Sullivan, John, James Piereson, and George E. Marcus. 1979. “An Alternative Con- ceptualization of Political Tolerance: Illusory Increases 1950s–1970s. ” American Political Science Review 73, no. 03 (September): 781–794. . 1982. Political Tolerance and American Democracy. Chicago: University of Chicago Press. Sullivan, John, Michal Shamir, Patrick Walsh, and Nigel S. Roberts. 1985. Political Tolerance in Context: Support for Unpopular Minorities in Israel, New Zealand, and the United States. Boulder: Westview Press. Tessler, Mark, Amaney Jamal, and Michael Robbins. 2012. “New Findings on Arabs and Democracy.”

Volume 23, Number 4, October 2012

, Journal of Democracy 23 (4): 89–103. “Getting Nicer.” 2015. The Economist (March). “Terrorists Returning Home to Tunisia. ” 2017. The Economist (January). “Gay-Rights Activists Take to the Air in Tunisia. ” 2018. The Economist (March). 137 Timsit, Annabelle. 2018. Justin Trudeau: “There’s Not a Religion in the World That Asks You to ‘Tolerate Thy Neighbor”’, May. Twenge, Jean M., Nathan T. Carter, and W. Keith Campbell. 2015. “Time Period, Generational, and Age Differences in Tolerance for Controversial Beliefs and Lifestyles in the United States, 1972–2012. ” Social Forces 94, no. 1 (September): 379–399. Verkuyten, Maykel, Mieke Maliepaard, Borja Martinovic, and Yassine Khoudja. 2014. “Political Tolerance among Muslim Minorities in Western Europe: The Role of Denomination and Religious and Host National Identification. ” Politics and Religion 7, no. 02 (June): 265–286. Wickham, Carrie Rosefsky. 2013. The Muslim Brotherhood: Evolution of an Islamist Movement. Princeton: Princeton University Press. Williams, Janice. 2017. Hate Crimes against Muslims Are on the Rise in the U.S. http://www.newsweek.com/hate-crime-america-muslims-trump-638000, July. Winter, Jana, and Sharon Weinberger. 2017. The FBI’s New U.S. Terrorist Threat: ‘Black Identity Extremists’. https://foreignpolicy.com/2017/10/06/the-fbi-has- identified-a-new-domestic-terrorist-threat-and-its-black-identity-extremists/, October. Yerkes, Sarah. 2017. Corruption, Not Terrorism, Is Tunisia’s Biggest Threat, August. 138 Yglesias, Matthew. 2018. Everything We Think about the Political Correctness Debate Is Wrong. https://www.vox.com/policy-and- politics/2018/3/12/17100496/political-correctness-data, March. Zaller, John, and Stanley Feldman. 1992. “A Simple Theory of the Survey Response: Answering Questions versus Revealing Preferences. ” American Journal of Polit- ical Science 36 (3): 579–616. 139

Abstract (if available)

Abstract This three-article dissertation explores the determinants of political tolerance by examining understudied groups and settings, emphasizing in each case how context is a key driver of the observed attitudes. First, drawing on an original survey of American Muslims, I theorize and test how the considerations motivating intolerance differ between the general public and minorities in established democracies. My analysis highlights: 1) the enhanced role of perceived threat to personal safety in motivating targeted intolerance among marginalized groups

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Economic policy-making in a changing environment: the politics of success in Peru vis-à-vis China

PDF

Information, public opinion, and international relations

PDF

Identity salience and political violence: centering social identity in the study of intrastate conflict

PDF

Volatile nationalism: nationalism and its influence on maritime disputes

PDF

The role of local actors in international development: how incentive types affect foreign aid effectiveness

PDF

A theory of status-quo terrorism: democracies in conflict and their proclivity to outsource repression

PDF

Authoritarian religion: explaining when and why authoritarian governments instrumentalize religion

PDF

Politics behind appointments: policy, patronage, and party systems in the Andes

PDF

Developing countries in the digital era: state and business interactions for industrial upgrading

PDF

The strategy of firms towards political elites and consumers in polarized western democracies

PDF

Political participation, public opinion, and the law

PDF

Public opinion and international affairs: a multi-method approach to foreign policy attitudes

PDF

Legitimizing self-determination: advancing the sovereignty of separatist movements

PDF

Forced march to modernity: State-imposed cultural change and regime stability in 20th-century east Asia

PDF

Habits and policy: the socio-cognitive foundations of foreign policy stability

PDF

Essays on political economy and corruption

PDF

Descriptive representation in a multi-racial America

PDF

Distress tolerance and mindfulness disposition: associations with substance use during adolescence and emerging adulthood

PDF

Political identity and political participation: China's post-80's generation

PDF

Political decision-making in an uncertain world

Asset Metadata

Creator Chouhoud, Youssef (author)

Core Title The changing state of tolerance and tolerance in changing states

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Political Science and International Relations

Publication Date 04/29/2021

Defense Date 07/15/2018

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag American Muslims,Arab Spring,Egypt,OAI-PMH Harvest,political tolerance,Tunisia

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Brand, Laurie (committee chair), Munck, Gerardo (committee chair), Chong, Dennis (committee member), Jackson, Sherman (committee member)

Creator Email chouhoud@usc.edu,ychouhoud@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-158734

Unique identifier UC11660007

Identifier etd-ChouhoudYo-7342.pdf (filename),usctheses-c89-158734 (legacy record id)

Legacy Identifier etd-ChouhoudYo-7342.pdf

Dmrecord 158734

Document Type Dissertation

Format application/pdf (imt)

Rights Chouhoud, Youssef

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA