Two Implications of Survey Research Mode during War: Evidence from Russia's Full-Scale Invasion of Ukraine

Conducting social science research during an active war raises distinct challenges. To investigate the relationship between mode and both coverage and social desirability bias during wartime, I conduct a multi-mode study (web, telephone) with two independently drawn samples of the Ukrainian population during Russia's full-scale invasion of Ukraine in the summer of 2022. I employ identical demographic and behavioral questions in both surveys to examine coverage bias and pre-register a framing experiment to investigate social desirability bias in reported in reported volunteering activities after the onset of the full-scale invasion. Observationally, I find that because web studies cannot reach the oldest and most rural Ukrainians, they likely contain much more significant coverage bias with respect to war-related demographic variables compared to the telephone sample. Experimentally, consistent with other studies that report greater social desirability bias in interviewer-mediated models, I find evidence of inflation in reporting volunteering activities only in the telephone survey. Our results demonstrate that, given modern survey techniques, wartime attitudes and behavior in Ukraine can be reliably measured, but there is a trade-off in survey modes between coverage bias and social desirability bias.