ESRA 2027 sessions by theme

Back to Overview of Sessions

Session Details

Selective nonresponse and panel attrition remain among the most consequential threats to the validity of survey estimates. This session brings together methodological and empirical contributions that advance our ability to (1) predict which sample members are at risk of nonresponse and (2) implement innovative strategies to increase their response propensities. Neither task is sufficient on its own: prediction without intervention leaves the risk unaddressed, and intervention without prediction may raise response rates without improving sample composition. Understanding why sample members do not participate sharpens both what we can predict and where we can meaningfully intervene.
We invite papers addressing, among other topics:
- nonresponse prediction that goes beyond respondent characteristics, drawing on data generated within the survey process, such as paradata, interviewer observations, or open-ended text responses
- nonresponse prediction drawing on externally linked data, such as geospatial, administrative, or digital trace data
- novel statistical and machine-learning approaches to nonresponse prediction, including model transferability across surveys, waves, and populations
- the development of new response enhancement strategies that address sample members' reasons for nonparticipation
- the comparison of response enhancement strategies, such as differential incentives, contact strategies, or questionnaire design
- the implementation of such strategies, for example in adaptive or responsive survey designs, and their consequences for sample composition and response quality
The session aims to foster discussion of how advances in predicting nonresponse can be translated into strategies that improve participation and sample composition in practice. We welcome submissions from researchers working on nonresponse and survey design across the social and behavioral sciences.