Abstract A21: Effect of forced-response on completion rate in objective Internet-based health survey

Introduction: Forced-response in Internet-based surveys is one solution to item nonresponse in population science research, but forced-response can inflate dropout rate and introduce bias. As a result, forced-response has been described as a trade-off between data quantity and data quality. Studies have examined the effect of forced-response in subjective, self-report data. The impact of forced-response on objective assessment is not clear. This study explores the effect of forced-response on completion rate in an objective, Internet-based survey. Procedure: The 29-item questionnaire, which was hosted on Qualtrics, used multiple-choice items to assess health education students’ health literacy knowledge. The instrument was programmed in forced-response mode so respondents had to provide an answer to a question before they could advance to the next question. Links to the instrument were distributed through participants’ institutional emails. All participants who completed the questionnaire received a $10 gift card. Metadata were extracted and analyzed with SPSS Version 25. Results: Of approximately 844 students who were invited to participate in the survey, 521 clicked on the link and were taken to the screening page. Almost 50% of students (n=234) were screened out because they did not match eligibility criteria. The final eligible sample was composed of 287 participants, 46 of whom did not complete the questionnaire. Of the 46 eligible participants who did not complete the study, 14 (30%) dropped out after completing the screeners and another 25 (55%) dropped out in the first half of the questionnaire. Items associated with the highest rates of dropout were Item 3 (n=10), which asked about the relationship between health literacy and years of schooling; Item 6 (n=4), which asked about consequences of low health literacy; and item 13 (n=3), which assessed respondents’ knowledge of the use of Test of Functional Health literacy. Completion rate for the survey was 83.9%. When the 14 participants who dropped out before encountering the forced-response format were excluded from the analysis, the completion rate increased to 88.3%. Conclusion: Results of our study suggest forced-response may be useful in generating quantity data without compromising quality. Completion rate in this study is comparable to completion rates (80% - 97%) in other studies that did not employ forced-response. A limitation of this work is that we did not examine reactance to determine if answers were random or a true reflection of respondents’ knowledge. It will be useful for future population science research to utilize mixed methods to explore how forcing responses affects response accuracy when survey questions are objective and delivered via electronic platforms. Citation Format: Lenna Dawkins-Moultin, Kimlin Ashing, Lisako McKyer. Effect of forced-response on completion rate in objective Internet-based health survey [abstract]. In: Proceedings of the AACR Special Conference on Modernizing Population Sciences in the Digital Age; 2019 Feb 19-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(9 Suppl):Abstract nr A21.