ESRA 2027 sessions by theme
Back to Overview of Sessions
Methods to mitigate item nonresponse in probability-based, cross-sectional self-completion surveys |
| Coordinator 1 | Mr Nathan Reece (European Social Survey HQ (City St George's, University of London)) |
| Coordinator 2 | Dr Marika De Bruijne (Centerdata) |
| Coordinator 3 | Professor Peter Lugtig (Utrecht University - Methodology and Statistics) |
| Coordinator 4 | Ms Victoria Salinero-Bevins (European Social Survey HQ (City St George's, University of London)) |
Session Details
As many high-quality general social surveys transition from face-to-face interviewing towards self-administered data collection modes, item nonresponse is expected to increase significantly for certain types of survey questions (Jabkowski, Cichocki, & Piekut, 2026). Higher item nonresponse is usually observed at questions that elicit higher cognitive burden (e.g., household income, employment history), are intrusive (e.g., demographic information of the respondent or household members), or require greater effort compared to simple closed questions (e.g., open response questions). Item nonresponse is also expected to be higher for cross-sectional surveys of the general population compared to panel surveys, as people who do not typically answer surveys may exhibit less desirable responding behaviour.
To mitigate item nonresponse, various methods are employed in self-completion surveys. E.g., 'soft checks' in web surveys prompt respondents that tried to skip a question to answer but ultimately let them proceed without providing one. Follow-up questions in case of item nonresponse (e.g. on income) may ask simpler or more protective versions to encourage a response. 'Hard checks' prevent respondents from proceeding at all. Additional instructions at complex questions (e.g., explaining how to calculate years of education completed) may help respondents formulate a response. However, each intervention also risks that item nonresponse or break-off rates increase. Adding too much non-essential text may obscure the actual question or give the perception that it is overly complex to answer; web surveys overloaded with checks preventing the respondent from proceeding may cause frustration and increase the propensity to skip later questions or drop out of the survey.
This session invites papers that demonstrate the effectiveness (or lack thereof) of methods aimed at reducing item nonresponse in probability-based cross-sectional self-completion surveys. We especially welcome papers that include experimental evidence.
