Limitations of List Experiments in Post-Conflict Zones

Abstract

The list experiment is a survey methodology designed elicit truthful answers to sensitive behavior or attitudes that respondents may under report if asked directly. We attempted to investigate why some individuals took risks to actively support insurgents by providing resources and services to them in the eastern Ukrainian town of Slovyansk, which was controlled by the pro-Russian separatists from April 12, 2014 to July 5, 2014 as part of the Donbas War. Civilian support of an organized rebellion is an illegal behavior, therefore, the true extent of support would be difficult to estimate via direct questioning; we thus relied on a list experiment. However, despite much anecdotal evidence of widespread civilian assistance to the rebels during their control of Slovyansk, the list experiment failed to demonstrate any support for the rebels in a random sample of 610 people. We discuss the unwillingness of respondents to reveal the truth about the key item despite the protections of anonymity provided by the list experiment and the suboptimal choice of non-key items that have exacerbated the probability of our null findings. This research note discusses the limitations of the list experiment as a survey method to elicit candid answers in post-conflict environments.