Theses Doctoral

Identity in American Politics: A multidimensional approach to study and measurement

Spry, Amber

This dissertation, “Group Identity in American Politics: A Multidimensional Approach to Study and Measurement,” offers novel measurement strategies and theoretic insight toward the study of group politics. My research examines how best to explain group political preferences in a society that is becoming more culturally pluralistic while at the same time experiencing an increase in within-group heterogeneity.
The first chapter frames the dissertation as an exploration group politics with both critical and positive implications. I outline the intellectual lineage of theories in group politics, addressing the tension between research in political behavior which often links behavioral outcomes to fixed categorical identity variables, and research in political psychology which often treats identity as fluid and malleable. I argue that a full understanding of relationship between an individual’s self-identification and sense of shared outcomes with the identity group requires us to understand the different ways individuals may self-categorize and link their identities to political attitudes and behavior.
In chapter one I introduce the type-predictor framework, a theoretical contribution that rests on the notion that it is not ascriptive identity that we should think of as being tied to particular trends in attitudes and behaviors, but the way individuals understand their relationship to group identity when group membership becomes individually salient. I argue that different types of individuals may process group consciousness in a variety of different ways when group membership becomes salient at the individual level, leading them to express different conclusions about the notion of shared fate with in-group members. The type-predictor typology consists of five “types” that describe how individuals self-categorize (non-affiliation, abstract conceptualization, non-conformist identification, multiple identification, and strong identification) that correspond with four “predictors” of how an individual will conceptualize and respond to questions about group consciousness (disassociation, group membership, group identity, and group attachment). My typology clarifies the theoretical discussion by providing a framework that considers the subjective nature of the relationship between demographic characteristics and their political correlates, which can vary by groups.
Drawing on a series of 40 in-depth, semi-structured interviews, the second chapter of this thesis presents support for the type-predictor framework, and demonstrates how individuals link their sense of identity to their political attitudes and behaviors when given the opportunity to explain the process in their own words. One contribution of my work has been to provide an analysis of a linked fate measure based on an open-ended question that allows interview subjects to respond in reference to the group with which they primarily self-identify rather than having subjects answer the linked fate question in reference to an ascribed social category such as race or gender. The linked fate measure is frequently cited to explain the seemingly homogenous political attitudes and behaviors of African Americans and has been used increasingly in the past two decades to argue for a sense of shared outcomes leading to political solidarity among other groups such as Hispanics and Latinos, women, and the LBGT community. The question asks, “Do you think what happens to [people in your group] will have something to do with what happens in your life?” My interviews reveal that people interpret the linked fate question quite differently, with a wide degree of variation the range of responses. These findings are consistent with existing empirical research showing consistent statistical support for linked fate, yet substantial variation between and even within groups. Moreover, the open-ended responses to group consciousness questions in my interviews provide support for the type-predictor framework. I find examples of all five “types” and all four “predictors” in the typology, and the relationships between types and predictors are consistent with the directions I expected. Thus, my analysis of the interview data emphasizes the theoretical underpinning of the type-predictor framework: it is not ascriptive identity that we should think of as being tied to particular trends in attitudes and behaviors, but rather the way individuals understand their relationship to group identity when group membership becomes individually salient. In other words, among the interview sample, a sense of shared outcomes is related to the different degrees of group consciousness individuals may hold at the individual level depending on the group categories with which they do (or do not) self-identify.
The third thesis chapter further explores the multidimensionality of identity by using survey data to examine how group identity matters to individuals across policy areas, with particular attention paid to the politics of immigration and welfare policy. In the 2015 Identity Measurement Survey (IMS) (N=3,010) I introduce a point allocation system for measuring identity that allows subjects to allocate a fixed number of “identity points” to a number of socially relevant identity categories and compares this new approach with conventional survey methods by randomly assigning respondents to one of six methods of identity measurement and assessing the differences in policy-related attitudes across the six randomly assigned groups and across identity categories. Existing empirical work relies almost universally on a set of fixed, categorical measures that fail to reflect the multidimensionality many scholars associate with racial, ethnic, socioeconomic, and other forms of identity. The identity point allocation system allows survey respondents to identify with multiple group identities and to weight the strength of their association across groups. In addition to racial categories, the identity point allocation design includes class, religion and gender as categories to which respondents allocate points, and the random assignment of individuals to different measurement conditions allows us to understand how different approaches to measurement may reveal different outcomes on important identity-related questions. The design also allows us to explore whether the attitudes observed when individuals select a primary identity are different from the attitudes we observe when using conventional measures of demographic correlation.
Data from the IMS reveal that attitudes across policy areas differ according to the primary identity offered by respondents, and differ for some groups from what we might observe using the conventional “checked box” measure of ascriptive group identity. In particular, individuals who primarily identify as white, male, or Protestant consistently stuck out as having distinctive views from the population average, but also as having stronger views than what we would observe under conventional correlation between ascriptive categorization and attitudinal outcomes. Those who ascriptively identified as Protestant, male, or as a white person are most likely to have colder feelings toward immigrants and more conservative views toward providing welfare than people who more strongly associate with other groups. These attitudinal differences were even more pronounced when the analysis considered policy views according to the primary identity offered by respondents rather than through ascriptive categorization alone. These results underscore the opportunities afforded by alternative measurement strategies to reveal additional information about the links between identity and expressed policy attitudes when we allow individuals to tell us which identities matter most to them.
Taken together, the chapters of this dissertation provide perspective on individual thresholds for self-identification, and offer a novel measurement strategy to understand how individuals subjectively relate to group identities. Continued work will shed light on the relationships between individuals, their subjective identities, and the empirical correlates of identity such as inequality, intergroup conflict and violence, coalitional politics, and descriptive representation. The implications of these, to be sure, are not limited to the study of American politics.

Geographic Areas


  • thumnail for Spry_columbia_0054D_14639.pdf Spry_columbia_0054D_14639.pdf application/pdf 1.65 MB Download File

More About This Work

Academic Units
Political Science
Thesis Advisors
Green, Donald P.
Ph.D., Columbia University
Published Here
May 16, 2018