List Experiments in Survey Research on Corruption Engagement: Evidence from Eurasia

Surveys with nationally representative citizen samples are increasingly being used not just to ask about perceptions of corruption but about citizens’ own involvement in corrupt or potentially corrupt interactions with bureaucrats. While asking citizens about their own experiences provides a powerful tool, bribery is universally illegal and, to varying degrees, stigmatized, so questions about personal corruption experience are open to social-desirability and other biases, including more frequent non-response. These surveys therefore commonly include list experiments, which estimate the magnitude of the biases from direct questions about corruption engagement (Glynn 2013; Rosenfeld, Imai and Shapiro 2016). Yet no studies have examined list experimental results to see how social desirability bias and non-response vary cross-culturally. We will address this gap using data from five nationally representative surveys conducted in China (2018), Georgia (2015), Russia (2015 & 2018), and Ukraine (2015). Each included direct questions about citizens’ involvement in street-level bureaucratic corruption as well as list experiments measuring bribery engagement. We will explore crossnational differences in how the direct responses differ from list experimental results, using techniques for multivariate analysis of list experiment results (Blair and Imai 2012) to assess how far each country’s covariates of corruption engagement differ.