Perceptions of Islam, Migration, and Citizenship in the United States: A List Experiment
Universitat Pompeu Fabra
Home page: https://sites.google.com/site/mathewjcreighton/home
Sample size: 1897
Field period: 4/20/2010-7/8/2010
Social Identity Theory predicts that members of immigrant-receiving societies are favorably biased toward immigrants of their own religious ingroup. Conversely, immigrants from a religious outgroup are viewed neutrally or less favorably. Using a population-level list experiment, we show that this is not entirely the case in the United States. Opposition to citizenship for legal Muslim immigrants is not greater than for legal Christian immigrants as Social Identity Theory would predict, just more openly expressed. The appearance of bias in favor of Christian immigrants reflects a greater reluctance to appear prejudiced, but does not reflect greater underlying tolerance. We show that being a Christian can insulate immigrants from overt anti-immigrant sentiment. In contrast, Muslim immigrants are afforded no such protection. We conclude that Social Identity Theory does not determine bias, but rather how it is articulated.
Theoretical (H) and operational (h) hypotheses
To better understand the extent to which Social Identity Theory offers a theoretical basis for interpreting attitudes toward religious immigrant groups, we operationalize the following hypotheses. The first derives from the straightforward prediction that members of an ingroup will openly favor members of that ingroup.
H1: Membership in an ingroup offers protection from the overt expression of intolerance.
To operationalize this hypothesis, we compare the level of openly expressed (i.e., direct) opposition to the offer of citizenship to the two immigrant groups considered in the experiment – Muslims and Christians. Because legal Christian immigrants share a religion with the host society, we expect that they will be relatively protected from overt opposition.
h1a: Overt opposition to citizenship for legal Muslim immigrants is greater than overt opposition to citizenship for legal Christian immigrants.
The second hypothesis is based on the way in which Social Desirability Bias affects the expression of support or opposition to citizenship for immigrants by members of the host society (i.e., ingroup members). This hypothesis is rooted in the idea that Social Desirability Bias in part determines variation in the expression of opposition to citizenship for certain immigrants, but this overt expression is distinct from true opposition. Expressed opposition may benefit an ingroup (e.g., Christians), but when Social Desirability Bias is taken into account, the expression of this opposition is no longer moderated by social pressure to express greater tolerance to ingroup members.
H2: Accounting for Social Desirability Bias results in a reduction in the level of ingroup bias that directly favors the perceived ingroup.
The operationalization of this hypothesis also reflects the theoretical prediction that Social Desirability Bias will result in less pressure to appear overtly tolerant toward Muslim immigrants who are outgroup members, regardless of the underlying level of opposition. That is, opposition to immigrants (whether Christian or Muslim) is similar, but members of the host society are more likely to express more acceptance of Christian immigrants because they are seen as closer to the ingroup than are Muslim immigrants. Members of a perceived ingroup (i.e., Christian immigrants) benefit from social desirability pressure. In contrast, outgroup members (i.e., Muslim immigrants), do not benefit from social desirability norms. Therefore, the level of opposition to Muslim immigrants is not sensitive to the same social pressure and remains the similar whether or not Social Desirability Bias is taken into account.
h2a: Because Christian immigrants share a religious identity with the host society, the actual level of opposition to citizenship for legal Christian immigrants in the US is higher than expressed opposition.
h2b: Because Muslim immigrants do not share a religious identity with the host society, the actual level of opposition to citizenship for legal Muslim immigrants in the US is closer or identical to expressed opposition.
The third set of hypotheses assesses variation in the attitudes of ingroup members towards immigrants. Specifically, we look at members who identify themselves most strongly with the identity of the ingroup, such as those who attend Christian services more frequently or are evangelical in orientation. These members, according to Social Identity Theory, are subject to greater social pressure to be or appear tolerant toward immigrant members of their religious ingroup. Conversely, these same respondents would be less accepting of immigrant members of an outgroup.
H3: Members of an ingroup to whom ingroup membership is more salient express greater opposition toward outgroup members.
The operationalization of this hypothesis requires the consideration of additional measures of group identity – denomination and religiosity. Respondents who are more active in terms of denomination (i.e., evangelical) or more frequently practice are considered to be subject to greater social desirability pressure to appear more tolerant to ingroup members (i.e., Christian immigrants) and less tolerant toward outgroup members (i.e., Muslim immigrants).
h3a: Evangelical Christians, relative to mainline Protestants, are more opposed to citizenship for legal Muslim immigrants.
h3b: Christians who practice more frequently are more opposed to citizenship for legal Muslim immigrants.
h3c: Evangelical Christians, relative to mainline Protestants, are less opposed to citizenship for legal Christian immigrants.
h3d: Christians who practice more frequently are less opposed to citizenship for legal Christian immigrants.
Direct assessment of anti-immigrant sentiment assumes that a response to a direct question reflects the respondent’s true sentiment. This interpretation has been shown to be vulnerable to Social Desirability Bias which causes expressed attitudes can and do diverge significantly and systematically from those attitudes which are truly held. One method to elicit truthful responses is the list experiment, which allows respondents to permanently and unconditionally conceal their individual responses from researchers. The greatest advantage of this approach is that individual-level responses are not only concealed from the interviewer, but they cannot be known. Respondents are divided between a control group and, in this case, two treatment groups. The control group is asked a single question about the following list of items. The question reads:
Below you will read three things that sometimes people oppose or are against. After you read all three, just tell us HOW MANY of them you OPPOSE. We don’t want to know which ones, just HOW MANY.
(1) the federal government increasing assistance to the poor
(2) professional athletes making millions of dollars per year
(3) large corporations polluting the environment
Two independently sampled treatment groups are asked an identical question, but of a list that includes the original three items above and a fourth item that queries opposition to Muslim and Christian immigrants respectively.
(4) granting citizenship to a legal immigrant who is Muslim
(4) granting citizenship to a legal immigrant who is Christian
In its most basic incarnation, the comparison of the mean of the responses to the control list with the mean of the responses to each of the treatments offers an estimate of the proportion opposed to the additional list item.
The outcome of interest is opposition to citizenship for legal Muslim and legal Christian immigrants. This outcome is measured directly and via a list experiment.
We conclude that the appearance of targeted opposition reflects a greater reluctance to appear prejudiced toward Christian immigrants, than towards Muslim immigrants, but does not reflect an actual difference in opinion. When Social Desirability Bias is taken into account, opposition to Christian and Muslim immigrants is nearly identical. This is not attributable to a lower level of opposition toward Muslim immigrants, but instead reflects greater underlying opposition to ingroup members (i.e., Christian immigrants).
An additional experiment was included to test overall opposition to immigration or support for a closed border. This experiment is similar, but not identical to that described above. The main finding from that experiment is that subsequent to the 2007-2008 financial crisis, opposition to immigration has not increased. Instead, the expression of anti-immigration sentiment has increases suggesting greater public tolerance or, conversely, less social desirability pressure to appear tolerant.
Two papers currently under review.