Comparing Different Methods of Compensating for Survey Breakoffs

• DOI: 10.31235/osf.io/35np8_v1
• Publication Date: 2025-04-12T20:37:08Z
• Authors: Zeming Chen; Alexandru Cernat; Natalie Shlomo

Web surveys are increasingly popular but tend to have more breakoffs compared interviewer-administered surveys. Survey occur when respondents quit the survey partway through. It reduces sample size for analysis, lowers statistical power and may lead biased estimates. Few prior studies investigated how compensate after data collection. This study develops a simulation where breakoff rate cause of manipulated then compensated using five different methods. We find that nonresponse weighting multiple imputation better than other methods investigated. Meanwhile, leads worst compensation results. Overall, we recommend as it has good consistent performance in keeps maximum amount information data, therefore helping with power.