Understanding the Sensitivity of Party Identification Questions in Polarized African Contexts

Abstract The global rise of political polarization has generated new insights into how polarization affects political behavior. But political polarization may also shape the dynamics of researching such behavior in the first place. This paper argues that in polarized contexts, asking about party identification (PID) during academic surveys can be highly sensitive. We demonstrate this in sub-Saharan Africa, using three forms of evidence. First, a survey of enumerators in Côte d’Ivoire shows that over 50% feel uncomfortable asking about PID, linking their discomfort to the polarized political climate. Enumerators also believe that such questions reduce honest responses and influence the research encounter itself. Second, data from the Afrobarometer reveals that respondents in polarized countries often self-censor on PID questions more than on traditionally sensitive topics like clientelism, violence, or identity. Finally, interviews and focus groups with Ivorian citizens support these findings, highlighting the sensitivity around PID. These results emphasize that in politically polarized environments, partisanship questions may be uniquely sensitive and impact data quality and the research experience in addition to holding ramifications for political behavior. We conclude with a discussion of implications for survey design and analysis as well as a general call for political scientists to consider what is sensitive, in what context, and for whom when designing surveys.