Wartime sexual violence (WSV) is often stigmatised, and in survey questionnaires, respondents may under-report their victimisation if the question is asked directly. To mitigate this risk, we developed a list experiment allowing victims to report sexual violence experienced during the Kosovo war in a less intrusive, less direct way. While the responses collected through a direct question suggest a prevalence of WSV roughly comparable with previous estimates, our item list revealed an insignificant and negative reporting of WSV. Lower reports of WSV in a non-intru- sive question compared to a direct question suggest that the item-list failed. Following a critical discussion of our research design, we explore the underlying patterns in the data and reveal an issue with the data collection itself: when interviewers helped respondents with entering their answers, respondents provided systematically different answers to the count questions. We con- clude that variations in the implementation of the interview protocol interfered with our design. Overall, the behaviour of interviewers and interviewees appear to translate complex societal re- lationship with victimisation. It suggests a possible degree of “fleeing”, a term used to describe a situation where respondents actively avoid being associated with a stigmatised behaviour, but also, and more clearly, an unexpected greater reporting of non-stigmatised victimisation when answers are filled in by an interviewer rather than provided privately. We conclude with suggestions for future work.
The Item Count Technique (ICT) is an experimental method aimed at eliciting truthful responses to sensitive questions. In its basic form, the ICT randomly assigns respondents to one of two conditions: a control group, which receives a list of neutral statements, and a treatment group, which receives the same list with an additional sensitive item. Respondents then indicate the total number of statements that apply to them without specifying which ones. The difference in the average counts between the two groups is used to estimate the prevalence of the sensitive behavior. This study examines discrepancies in reporting the purchase of sexual services by comparing direct questioning with the ICT using an online survey of 1603 panelists in Spain. The findings reveal that the ICT was ineffective in estimating the prevalence of purchasing sexual services, underscoring the need to thoroughly test proposed designs and to further develop strategies for obtaining accurate self-reported data on sensitive topics.
The impact of discrete emotions such as anger, anxiety or happiness is an important and emerging field of research in political science. For priming happiness, several different techniques (emo- tional recall, visual stimuli, imaginary) have so far shown consistent effects in western countries. In this paper, we examine whether these methods for experimentally inducing happiness can also be used to induce happiness in an extremely unfavorable context: Ukraine during the Rus- so-Ukrainian War. Our results from four studies conducted in three panel waves in 2022/23 (N = 1980), 2023 (N = 1542) and 2024 (N = 1639) show that a) none of these techniques significantly increased happiness and that b) some of these techniques actually can have the opposite effect on Ukrainians and reduce happiness in respondents. We discuss possible reasons for this finding and conclude with recommendations for future research.
In the first wave of the UK’s Generations and Gender Survey, conducted as an online-only, probability-based survey, an experiment was carried out to examine the impact of providing Quick Response (QR) codes with URLs versus just URL links in the invitation and reminder letters. The aim of the experiment was to investigate differences in participants’ likelihood of engaging with the survey, as well as in the quality of the data collected, between the two experimental conditions. The results suggest that respondents who received letters with a QR code were more likely to participate and to use smartphones, rather than PCs/desktops, to complete the survey compared to those who received letters with only URL links. The finding is important because QR codes could be leveraged to encourage respondents to use smartphones for surveys that require additional tasks to be completed via mobile devices. No significant differences in data quality were observed between the two conditions, indicating that QR codes show promise for future survey data collection. However, the experiment did not determine whether the observed effects were specifically associated with participants using the QR code to access the questionnaire. Future studies offering QR codes for online survey access should ensure that the data on how respondents access the survey (via QR code or URL) is collected. This will allow for a more in-depth analysis of the factors associated with the different methods of survey access.
Survey experiment studies show that corruption can reduce a voter’s willingness to vote for a candidate but that positive partisanship for that candidate can lessen the detrimental impact of corruption on vote likelihood. How corruption affects voting among nonpartisans and negative partisans, though, remains understudied. Using a preregistered survey experiment fielded in Brazil between the first and second rounds of the 2022 presidential elections, we examine how corruption information affects vote intention for president across different partisan types. However, we find no different effects of corruption on vote choice across the three partisan types. Through exploratory analyses of our survey data, we show that our experiment most likely produced null findings because Brazil’s 2022 election was highly polarized environment where corruption perceptions and partisanship were strongly intertwined, and our choice to use real corruption information about actual candidates instead of hypothetical candidates drastically weakened our treatment in such an environment.
