Correlates of Political Tolerance and Intolerance in the United States

Abstract

Scholars have suggested a number of different causes of political tolerance and intolerance. Using data from the 2008 American National Election Studies (ANES), we test seven of these causes frequently mentioned in the literature (age, gender, education, religion, ideology, so­cial capital and perceived threat) while controlling for class and ethnicity. We find that reli­gion, ethnicity and especially ideology explain the variance in attitudes toward fringe groups quite well, while class and especially social capital appear to be rather poor predictors of an individual's level of tolerance. We also find that of our nine independent variables, only gen­der affected tolerance levels precisely as predicted. For all other independent variables, the relationship with political tolerance is not entirely as previous research has suggested; in some cases, it is even the complete opposite. Thus, our results partially challenge the findings of previous studies and demonstrate the need for further research on political tolerance.

Introduction

Over the last couple of years, the struggle for equal rights for homosexuals has be­come a major issue in the United States (and also in other parts of the world). When it comes to questions such as whether or not gay people should be allowed to serve in the military, adopt children or marry, the public is deeply divided between people who want to grant ho­mosexuals these rights and those who do not. What characterizes this public debate is not only its rigidity but also the fact that individuals and groups intuitively associated with tolerance are among the people who are most intolerant of gay people. This especially applies to nu­merous church officials and religious groups, which relentlessly oppose the establishment of equal rights for homosexuals (no author 2008).

This struggle for equal rights for homosexuals can be viewed in the larger context of political tolerance and intolerance. Over the last couple of decades, numerous fringe groups have been subject to intolerance, which has entailed negative consequences both for these groups and for society as a whole. Thus, this topic has attracted a lot of scholarly attention, which is why by today a fair amount of research on political tolerance exists, especially on its causes. This study goes along the same lines, as it also aims at examining the factors which cause people to be tolerant or intolerant. In order to do so, we first examine the existing litera­ture on the topic. We identify seven factors frequently mentioned by scholars as causes of political tolerance, which build the basis for our seven hypotheses. This list should not be un­derstood as exclusive, as we do not claim to have included every single potential cause of political tolerance in our study; however, we are confident that our analysis entails the major sources of tolerance and intolerance suggested by previous studies. The literature review is followed by a methodological section, in which we define our dependent and independent variables, explain the measures by which we judge them and lay out the statistical analysis through which we assess the relationship between them. What follows is an illustration of our findings, which shows that some of the seven potential causes examined are indeed good pre­dictors of an individual's level of tolerance, while others are less accurate in explaining vari­ance in attitudes toward different fringe groups. A concluding section summarizes the main findings of this study, connects them with the broader theoretical debate and indicates the direction for future research on this topic.

Why examine the causes of political tolerance and intolerance in the United States? What is the merit of such a study? From a normative point of view, political tolerance is a desirable phenomenon, as vital democracies require citizens to respect others' views and tol­erate their efforts to participate in politics, even if they promote unpopular views (Sullivan and Transue 1999). According to scholar John L. Gibson, the loss of respect for dissent and nonconformity in nominally democratic regimes is perhaps one of the greatest threats to polit­ical freedom (1992b). Mass political intolerance establishes a culture of conformity that con­strains individual political liberty and deprives those outside the centrist mainstream of their political opportunities. From this point of view, then, studies on political tolerance seem nec­essary to explore its causes and thus be able to promote policies and behavior which will spread tolerance among the public. Moreover, there is a second, academic reason to study political tolerance. Even though numerous studies on political tolerance exist which greatly enhanced our understanding of this phenomenon, their findings are still far from being uni­form. Also, it is possible that public attitudes toward fringe groups change over time. There­fore, continuous analysis of the topic seems necessary to add more clarity to the academic discussion on political tolerance and keep it up to date.

Literature Review

Scholarly research on political tolerance began more than half a century ago. Early studies like Adorno's The Authoritarian Personality (1950) and especially Stouffer's Com­munism, Conformity, and Civil Liberties (1955) set the tone for future studies by defining the concept, exploring its theoretical background, and, in Stouffer's case, developing models for conducting empirical research. During the past 60 years, countless studies have been under­taken on political tolerance, which is generally defined as the ability to accept people of dif­ferent backgrounds and beliefs and the willingness to extend civil liberties to all members of society, including those who hold opinions and ideas one opposes (Stouffer 1955; Nunn, Crockett and Williams 1978; Sullivan, Piereson and Marcus 1993; Sotelo 2000a).

Within this tremendous amount of literature, two major scholarly debates can be iden­tified. The first one revolves around the question whether or not the public has become more tolerant over time. Some researchers find a substantial increase in political tolerance since the 1950s (Cutler and Kaufman 1975; Davis 1975; Nunn, Crockett and Williams 1978; McClosky and Brill 1983; Mueller 1988), while others find the increase to be illusory - the public is more tolerant now of groups on the left side of the political spectrum, but has simply found other targets on which to vent its intolerance (Sullivan, Piereson and Marcus 1979; Sullivan, Marcus, Feldman and Piereson 1981; Sullivan, Piereson and Marcus 1993). Related to this debate is the methodological question of how to properly measure political tolerance. The second major scholarly debate revolves around the question which factors cause people to be tolerant or intolerant. Just like the first argument, the second one also remains unsolved; up to today, numerous factors have been named as potential causes of political tolerance, with some studies confirming them while others rejecting them. The following paragraphs lay out the causes most often mentioned by scholars studying political tolerance.

Age

One of the potential causes of tolerance often mentioned in the literature is age. More precisely, there appears to be a negative relationship between an individual's age and level of tolerance: the older a person is, the less tolerant he or she is. The theoretical argument is that “aging is associated with a shift to more conservative political attitudes. Generational differences in political socialization and certain correlates of the aging process are the factors most commonly cited to account for this trend toward greater conservatism. According to this perspective, successive cohorts are socialized to different political attitudes, values and ideologies as the content of the political culture changes. Associ­ated with the subsequent movement of individual members of the cohorts through the life cycle are psychologically based, age-related changes in the direction of greater ri­gidity, cautiousness, and increasing resistance to change. Further pressures against the acceptance of social and political change become manifest with increasing integration into the social system, which leads to a greater stake in the maintenance of status quo. Presumably, therefore, older cohorts are not only the bearers of the more traditional political culture, but their members increasingly adhere to the content of their earlier political socialization as they age” (Cutler and Kaufman 1975: 69-70). Some scholars, however, have questioned this kind of relationship between age and tolerance, especially its strength and causal nature. These critics of the age-argument suggest that it is rather methodological problems and intervening variables like confounding effects of cohort (generational) differences in socialization than the aging process itself which affect tolerance (Ryder 1965; White Riley 1973). However, the number of studies which yield empirical evi­dence that there is a correlation between age and tolerance is overwhelming (Stouffer 1955; Berelson and Steiner 1964; Coale 1964; Cutler and Kaufman 1975; Owen and Dennis 1987; Karpov 1999a; Sotelo 2000a). Therefore, the first hypothesis of our study reads as follows: Hypothesis 1: The younger an individual is, the more tolerant he or she is.

Gender

A second potential cause of political tolerance appears to be gender. Not only do men and women vary in their target groups of intolerance,¹ but they also vary in their overall level of tolerance. Early studies found that men tend to be more tolerant than women (Stouffer 1955; Nielson 1977; Crocket and Williams 1978; Jones 1980; Avery 1988; Gibson 1992a; Sullivan, Piereson and Marcus 1993). Theorists suggested a variety of different reasons for this phenomenon, including different child-bearing practices for boys and girls, women's ma­ternal instincts, the different gender roles assigned to both sexes, and the patriar- chal/repressive structure of society. However, more recent studies yielded converse results, as they found women to be slightly more tolerant than men (Golebiowska 1999; Sotelo 1999). This turn appears to be caused by an increasingly equal structure of society, which offers greater opportunity to women, and thus supports the patriarchy-argument as an explanation of the previously lower tolerance level among women. Thus, we would expect women to be more tolerant than men when using contemporary data.

Hypothesis 2: Women are more tolerant than men.

Education

Education is another factor that is said to affect political tolerance. The theoretical ar­gument is that education imparts certain qualities, including reasoning processes, levels of information and cognitive sophistication, which foster an individual's ability to accept and embrace people of different backgrounds and beliefs (Bobo and Licari 1989). Even though this link between education and tolerance has been challenged by a few scholars,² it is widely accepted that higher educated individuals tend to be more tolerant. The overwhelming majori­ty of the evidence supports the education-tolerance hypothesis (Stouffer 1955; Martin and Westie 1959; McClosky 1964; Davis 1975; Lawrence 1976; Nunn, Crockett and Williams 1978; Bobo and Licari 1989; Peffley and Sigelman 1990; Golebiowska 1995; Karpov 1999a). This leads some scholars to embrace the optimistic notion that the overall level of tolerance will increase over time, as the overall level of education is also increasing (Stouffer 1955). Hypothesis 3: The more educated an individual is, the more tolerant he or she is.

Religion

Multiple studies have shown that religious commitment and denominational affiliation play a significant role in influencing whether individuals want to extend civil liberties to fringe groups. When it comes to denominational affiliation, studies generally find that Jews tend to be the most tolerant religious group, followed by Catholics, whereas Protestants score rather low on tolerance scales. However, every religious group scores lower on tolerance scales than people without any religious affiliation; hence, atheists/agnostics appear to be the most tolerant (Beatty and Walter 1984; Katnik 2002; Kim and Zhong 2010). When it comes to religious commitment, people with a higher degree of religious activity seem to be less tolerant. For example, numerous studies have shown that there is a negative relationship be­tween church attendance and tolerance (Martin and Westie 1959; Sullivan, Piereson and Mar­cus 1993; Sotelo 2000a; Sotelo 2000b; Froese, Bader and Smith 2008; Rowatt, LaBouff, Johnson, Froese and Tsang 2009). Thus, despite a few studies that challenge the link between religion and political tolerance (Karpov 1999b; Eisenstein 2006), we hypothesize as follows: Hypothesis 4: The less religiously committed an individual is, the more tolerant he or she is.

Ideology

Another potential cause of political tolerance often mentioned in the literature is ideol­ogy. Ideology is most frequently measured along a single dimension, with liberalism and con­servatism at its left and right endpoints, respectively. Liberals and conservatives can be dis­tinguished by their attitudes toward social change and equality among people: whereas con­servatives are rather resistant to change and prefer a more hierarchical organization of society, liberals are more likely to seek change and to embrace egalitarian ideas. Early research (Adorno, Frenkel-Brunswik, Levinson and Sanford 1950) suggests that these differences be­tween liberals and conservatives are rooted in personality differences, and later research has borne out these suppositions (White-Ajmani and Bursik 2011). By definition, then, liberals tend to be more tolerant than conservatives; a notion that has been confirmed by an array of studies (Sullivan, Piereson and Marcus 1993; Pratto, Sidanius, Stallworth and Malle 1994; Christiansen and Lavine 1997; Altemeyer 1998; Tygart 2000; Crowson 2004; McFarland and Mathews 2005; Jost, Nosek and Gosling 2008; White-Ajmani and Bursik 2011).

Hypothesis 5: Liberals are more tolerant than conservatives.

Social capital

Another potential cause of political tolerance appears to be social capital. In recent years, the social capital-concept has been most prominently promoted by Robert Putnam (1993; 1995a; 1995b; 2000), who defines social capital as “features of social organizations such as networks, norms, and social trust that facilitate coordination and cooperation for mu­tual benefit” (1995a: 65-66). Hence, an individual's social capital is tantamount to the number of voluntary organizations of which he or she is a member. There appears to be a positive relationship between social capital and political tolerance: the more organizations an individ­ual is a member of, the more tolerant he or she is. The reason for this is that “the give-and-take and face-to-face internal group interaction, coupled with the organ­izational imperative for cooperative endeavors, encourages the development of norms such as reciprocity and trust. Compromise is often called for and willingness to respect the views and affiliations of others becomes a social necessity within the group. [...] orientations developed within such ‘little democracies’ affect the larger political con­text, forming the basis of diffuse support for higher-level social and political institu­tions” (Cigler and Joslyn 2002).

Several studies have confirmed this link between social capital and political tolerance (Cigler 1991; Baumgartner and Leech 1998; Finkel, Sigelman and Humphries 1999; Sullivan and Transue 1999; Cigler and Joslyn 2002). Thus, we hypothesize as follows:

Hypothesis 6: The more social capital an individual has, the more tolerant he or she is.

Perceived threat

Previous studies have shown that an individual's level of tolerance is directly related to perceptions of threat (Stouffer 1955; Martin and Westie 1959; Sullivan, Piereson and Mar­cus 1979; Sullivan, Marcus, Feldman and Piereson 1981; McClosky and Brill 1983; Gibson and Bingham 1985; Green and Waxman 1987; Marcus, Sullivan, Theiss-Morse and Wood 1995; Sullivan and Transue 1999; Hutchinson and Gibler 2007; Canetti-Nism and Hirsch- Hoefler 2009; Van der Noll, Poppe and Verkuyten 2010). One dimension of perceived threat is that intolerance arises from perceptions that dissident/fringe groups threaten important val- ues or constitute a danger to the constitutional order. That explains why, for example, intoler­ance of communists was significantly higher in Cold War-America than it is in the United States of today (Peffley and Sigelman 1990; Sullivan, Piereson and Marcus 1993). A second dimension of perceived threat is that in times of distress, the overall tolerance among the pop­ulation decreases. When, for example, the economy goes bad or the well-being of the nation is threatened by outside actors such as other states or terrorists, people's uncertainty is more likely to transform into intolerance (Shamir and Sullivan 1983; Wang and Chang 2006). Hence, there is ample reason to hypothesize that individuals who feel overly threatened by current events or certain groups are less tolerant.

Hypothesis 7: The less threatened an individual feels, the more tolerant he or she is.

Control variables

In addition to the seven hypotheses outlined above, we include two control variables in our analysis which are quite common in the social sciences: class and ethnicity. Controlling for these two factors makes a great deal of sense in the context of this study, as a few scholars have suggested a connection between political tolerance on the one side and class and ethnici­ty on the other. According to these authors, ethnic minorities and economically disadvantaged groups tend to be less tolerant of other fringe groups as a reaction to their own experiences of discrimination (Mansel and Spaiser 2012).

Methodology

In order to test the hypotheses outlined above, we use data from the 2008 American National Election Studies (ANES). ANES is the leading academically-run national survey of voters in the United States, conducted before and after every presidential election. The con­sistency of the study, asking the same questions repeatedly over time, makes it very useful for academic research. Another advantage of the study is its magnitude and the variety of ques­tions being asked,³ with the result that it is frequently cited in works in the social sciences. The question, then, becomes: which of the ANES survey items are suitable measures of our dependent and independent variables?

A dependent variable is a phenomenon thought to be influenced, affected, or caused by some other phenomenon; in our case, it is an individual's level of political tolerance. As out­lined above, political tolerance is the ability to accept people of different backgrounds and beliefs and the willingness to extend civil liberties to all members of society, including groups who hold opinions and ideas one opposes. In order to create a rather comprehensive account of an individual's level of political tolerance, we included groups from different societal spec­tra in our study. These groups are: homosexuals, atheists, environmentalists, feminists, illegal immigrants, people on welfare, big business, the military, blacks, Muslims, Christians and Christian fundamentalists. We assess an individual's tolerance (or intolerance) of these groups by looking at the ANES feeling thermometers, which allow respondents to rank their attitudes toward specific groups on a scale from 0 to 100.

An independent variable is a phenomenon thought to influence, affect, or cause some other phenomenon. As the literature review has shown, there are seven factors possibly influ­encing an individual's level of tolerance: age, gender, education, religion, ideology, social capital and threat. Most of these independent variables are straightforward and rather easy to measure when applying the 2008 ANES data set. In the 2008 ANES, every respondent is asked for his or her age and gender. In order to test the education-tolerance hypothesis, we use the item “summary of respondent's educational attainment”, which is a combination of three education-related items: the highest grade of school or year of college a respondent has com- pleted, whether or not a respondent has a high school diploma, and the highest degree a re­spondent has earned. Consistent with other studies, we measure an individual's religious commitment by looking at how often he or she attends religious services. For the 2008 ANES item, low values represent a frequent attendance of religious services, whereas high values indicate a weak religious commitment. To assess an individual's ideology, we employ a 7- point scale on which respondents can rank themselves from extremely liberal (1) to extremely conservative (7). Social capital is also rather straightforward, as the 2008 ANES includes a question which asks for the number of organizations in which a respondent is a member. The only independent variable that is not as easy to measure is perceived threat, as the 2008 ANES does not include questions that specifically ask how threatened a respondent feels by certain groups or events. Thus, we have to resort to a question which asks about the importance of combating international terrorism as a U.S. policy goal. Through this question, we intend to assess an individual's perception of terrorism as a threat. Compared to previous studies, this is an imperfect measure; however, we still consider it to sufficiently grasp the concept of per­ceived threat. As for our control variables, we take an individual's household income as an indicator for class, and we measure ethnicity by employing a dummy variable which distin­guishes between whites (1) and non-whites (2).

To examine the relationship between the dependent and independent variables, we use ordinary least squares regression. By doing so, we are able to assess which of the seven poten­tial causes outlined above does indeed have an effect on tolerance and if, how strong these effects are. Therefore, we are able to evaluate which of our independent variables is the best predictor of an individual's level of tolerance. Moreover, ordinary least squares regression allows us to assess the goodness of fit, which means how well our overall model can explain the observed variations in attitudes toward different fringe groups. Thus, ordinary least squares regression is the appropriate statistical analysis for the purpose of our study.

[...]

¹ Men's and Women's choices of intolerance targets are largely pluralistically distributed and parallel, with only a few exceptions. Women exhibit a greater preference for the Ku Klux Klan and abortion groups, whereas men evince a preference for militarists and homosexuals (Golebiowska 1999).

² Some scholars did not only criticize studies which suggested a link between education and tolerance for meth­odological problems (Jackman 1973), but they also came up with explanations as to why the education-tolerance hypothesis is wrong. One of these arguments is that higher education might generate an increased support of the general idea of tolerance, but it does not affect individual attitudes toward specific groups or policies (Jackman 1978). Another argument is that the greater support for democratic values observed among the highly educated is merely a superficial advocacy of individual rights that provides a principled basis for rejecting group-based claims on society (Jackman and Muha 1984).

³ For the 2008 ANES, the sample size is 2,322, with almost 2,000 items being included.