close× Call Us
Title Identity as Dependent Variable: A Survey Experiment
Post date 11/06/2018
C1 Background and Explanation of Rationale I hypothesize that the high salience of partisanship and ideology as social identities in the U.S. can lead some liberal Democrats and conservative Republicans to shift their demographic identities to better align with the demographic prototypes of their political groups. The hypothesis comes from social identity theory, which holds that when an identity is salient, group members engage in a process called self-categorization in which they modify their attitudes and behaviors to become more similar to that of the prototypical group member. Having found evidence for this hypothesis with observational panel survey data drawn from the General Social Survey (in a working paper entitled “Identity as Dependent Variable,” currently under review), I now seek to extend this research to ascertain whether such change also occurs in the context of an experiment in which partisan prototypes are primed.
C2 What are the hypotheses to be tested? to think about the different demographic groups who are parts of the Democratic and Republican Parties, shift how they describe themselves on these demographic dimensions to become more similar to their party’s demographic prototype. The quantity of interest is the difference in shifts in treated subjects’ claimed identities along the dimensions of race, ethnicity, religion and sexuality compared to subjects in the control group.
C3 How will these hypotheses be tested? *

I will test my hypothesis about politicized identity shifting on the 2018 U.S. Cooperative Congressional Election Study (CCES), in which subjects are interviewed twice—once before the general election and once afterward. In the pre-election wave, all subjects are asked a series of demographic questions about their race, national origin, religion, and sexual orientation. In the post-election wave, treated subjects perform a task designed to prime partisan demographic prototypes. In the experiment, treated subjects are first given a list of racial and ethnic groups (Blacks, Hispanics, Asians and Whites) and asked “Check whether you think the group’s members tend to be Democratic or Republican.” Treated subjects then perform the same task with religious groups (Protestants, Catholics, Jews, Born-Again Christians, and atheists) and finally with sexuality groups (Gays, lesbians & bisexuals and heterosexuals). Subjects in the control group perform a task of the same format, except they are provided lists of activities of the same length and asked to describe them as either “enjoyable” or “boring.” All subjects then immediately answer the same battery of demographic questions administered in the pre-election wave.

The hypothesis will be confirmed if treated subjects’ claimed identities along the dimensions of race, ethnicity, religion and sexuality shift in directions closer to the prototype of their partisan and ideological group compared any shifts in claimed identities among subjects in the control group. Specifically, the net probability of liberal Democrats shifting into identification as black, Hispanic, Asian (or of African, Asian, or Hispanic national origin), Jewish, nonreligious, and lesbian, gay or bisexual (LGB) is expected to be higher than that for conservative Republicans. Conversely, conservative Republicans are expected to be more likely to shift into calling themselves white, Protestant, Catholic, Born-Again Christian, or heterosexual than liberal Democrats.

C4 Country United States
C5 Scale (# of Units) Expected: 1,000 (subject to any attrition of panelists in post-election wave of CCES).
C6 Was a power analysis conducted prior to data collection? No
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval? Yes
C8 IRB Number NYU IRB-FY2016-1291
C9 Date of IRB Approval August 13, 2018
C10 Will the intervention be implemented by the researcher or a third party? The Cooperative Congressional Election Study (CCES)
C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research? No
C12 If relevant, is there an advance agreement with the implementation group that all results can be published? not provided by authors
C13 JEL Classification(s) J150