Conferences

Panel surveys generally combine the measurement of stable traits of the respondent with characteristics of the respondent that change over time.

The fact that panel surveys measure change over time induces the risk of specific forms of measurement error. Wave non-response, attrition, panel conditioning and the occurrence of the so called ‘seam effect’ can lead to measurement error that is specific to the longitudinal character of panel surveys.

Longitudinal measurement errors are frequently studied, and some methods to limit these measurement errors have been successfully implemented in the design of panel surveys.
This session covers questionnaire designs that try to prevent measurement errors in panel surveys. Papers that focus on, but are not necessarily limited to any of the following topics related to questionnaire design in panel surveys are welcome:

- Examples of the reconstruction of histories of respondents. Event History Calendars have shown to be successful in the reconstruction of past events in a respondent’s life course.

- The occurrence of spurious change that occurs when respondents are repeatedly asked the same question over time. Methods like Dependent Interviewing have shown to be effective in reducing these spurious changes in various circumstances, but questions remain about the implementation of Dependent Interviewing.

- The problem of wave non-response. How to deal with missing information of an individual respondent? What are the effects of filling informational gaps in later waves or using proxy information?

- The study of attrition in panel surveys and methods to prevent attrition in a panel survey, such as decreasing the amount of effort a respondent experiences being a member of a panel study.