Impact of Field Period Length in the Estimates of Sexual Victimization in a Web Based Survey of College Females
Marcus Berzofsky
RTI International
Chris Krebs
RTI International
Christine Lindquist
RTI International
When administering a survey on a sensitive topic one needs to be concerned about the potential for bias due to the length of the survey's field period. If persons who have a greater interest in the survey topic are more likely to respond quickly then a short field period may lead to upwardly biased results. However, a long field period may negatively impact estimates when there is a fixed starting point for the reference year. The Campus Climate Survey Validation Study (CCSVS) Pilot Test was a web-based survey administered at nine colleges measuring the prevalence and incidence of sexual victimization among undergraduate students during the 2014-15 academic year. The survey was in the field for approximately 60 days even though most schools achieved their targeted sample size within 28 days. We present a comparison of response rates, design effects, and the estimates for sexual victimization outcomes based on three different field periods. We discuss how the use of incentives and other strategies in the CCSVS may have contributed to mitigating potential bias in terms of how long it took victims and non-victims to decide to participate in the survey.