Evangelical or Born-Again Christian: Unpacking a Double-Barreled Question

Download data and study materials from OSF

Principal investigator:

Michele Margolis

University of Pennsylvania

Email: mmargo@sas.upenn.edu

Homepage: https://www.michelemargolis.com/


Sample size: 2902

Field period: 07/26/2019-10/18/2019

Abstract
Public opinion research often identifies evangelical Christians—one of the largest and most politically consequential social groups in the United States—based on a double-barreled question asking respondents whether they are an evangelical or born-again Christian. This paper uses a survey experiment to demonstrate the implications of this measurement strategy among both White and Black Christians. Among White (Black) Americans, more than one third (just under two thirds) of those whom researchers classify as evangelical using the standard double-barreled question actually eschew the evangelical label. Additionally, these born-again non-evangelical Christians hold less conservative political outlooks compared to the self-identified evangelicals with whom they are grouped, and, in fact, more closely resemble those who reject both the evangelical and born-again labels. In other words, “born-again” and “evangelical” are not synonymous to respondents and different sorts of people accept and reject the different labels. Moreover, asking about the two identities in one question changes response patterns. The double-barreled identification question produces an “evangelical or born-again” group that is both demographically and politically distinct compared to a composite “evangelical” or “born-again” group based on two questions asking about each identity separately. The double-barreled question therefore affects both evangelicals’ size and makeup. Finally, important experimental differences appear across race, suggesting that racial differences in histories, experiences, and relationships to politics affect how people interpret and respond to double-barreled questions. Consequently, scholars should not assume that double-barreled questions will produce comparable shifts in response patterns among different groups of survey takers.
Hypotheses
In what ways does the self-identification question of evangelicals on surveys affect who researchers determine to be an evangelical?
Experimental Manipulations
Randomize whether respondents answered a double-barreled question about identifying as an "evangelical or born-again Christian" or whether they answered each question separately.
Outcomes
Policy attitudes, vote choice, religious beliefs
Summary of Results
Abstract above includes a summary of findings.