Ideological Constraint in the American Public (with Thomas J. Wood)Are Americans ideologically constrained? Historically, the conventional wisdom in political science would suggest not per Converse (1964) and subsequent updates (Kinder and Kalmoe, 2017). Recent scholarship challenges these widely read works, contending that the modern American public hold stable ideological beliefs that comport with a traditional liberal/conservative scale (Fowler et al., 2022; Warren and Yan, 2022). We leverage consistently asked policy item and group evaluation items in the ANES and GSS from the 1990's up to today to investigate the evolution of observed ideological constraint in the public. Our contribution to this literature clarifies the ideological coherence among individuals of varying educational attainment in the American public and speaks to the need to continue to reassess and reaffirm classic findings in American political scholarship.
|
Pre-Election Polling and the Effects of Interviewer PartisanshipWhen conducting live telephone interviewing for election polls, interviewers are trained to conduct standardized, neutral interviews with respondents so as to increase the reliability of the survey. However, interviewers are not blank slates, and may conduct interviews in such a way as to tip off a respondent to their demographics or beliefs, thereby influencing how a respondent engages with the survey (Davis and Silver, 2003; West and Blom, 2017). Interviewers and respondents with equivalent partisanship see higher levels of corresponding vote choice than those interviewer/respondent pairs with opposing partisanship. However, support can be attributed to Republican interviewers garnering higher rates of corresponding partisan respondents during the survey introduction, thus explaining the increased rates of corresponding vote choice. This analysis contributes to the growing body of literature regarding various interviewer effects, as well as speculates on how to address and minimize such effects in an effort to continue to refine the art of public opinion polling.
|
Police Perceptions and Race-of -Interviewer EffectsAfter controlling for several mediating factors, black interviewers remain more likely to collect responses in which respondents agree that police brutality is something they personally worry about. This finding affirms previous literature in the field; that race effects can be present when asking racially charged questions. Additionally, a respondent's age and education can also affect responses, which is why weighting data post-collection is essential. Interestingly, whether the interviewer was male or female did not affect respondent answers.
Nonrandom measurement error can occur when not factoring interviewer race into presentation of survey data. Generalizing to the public becomes much more difficult when bias is present. Researchers should consider accounting for race effects if conducting surveys with racially charged questions, either on the front end or by weighting on the back end. |