Gregor Čehovin successfully completed doctoral studies

Gregor Čehovin, a member of Centre for Social Informatics, successfully completed doctoral programme in the Humanities and Social Sciences and the young researcher training financed by Slovenian Research Agency. He successfully defended the doctoral thesis addressed "The contribution of meta-analysis to knowledge on survey research operations on nonresponse".

The essential contribution of this dissertation is the systematic and integrated investigation of nonresponse in web vs. other survey modes, as well as according to different levels of personalization of survey invitations. In the first part, a systematic review is performed to analyze the current state of meta-analyses in the field of survey methodology and to identify gaps in current research. In the second part, two meta-analytic studies are performed. The first meta-analysis examines how the data collection mode (i.e., web vs. other survey modes) influences the difference in item nonresponse rate. The second meta-analysis explores how the level of personalization of survey invitations (i.e., more personalized vs. less personalized) affects the difference in unit nonresponse rate.

The systematic review reveals that the thematic areas of existing meta-analyses in the survey methodology field apply to only two of the seven total survey error (TSE) dimensions, namely measurement error and nonresponse error. The results thus suggest that there are several key research gaps in current meta-analytic research in survey methodology. The results of the first meta-analysis reveal that the difference in web vs. other survey modes disappears once units respond to the survey and item nonresponse is measured. The second meta-analysis shows that, on average, the unit nonresponse rate is 2.8 percentage points lower in personalized survey invitations compared to a less personalized approach. In this context, the two meta-analytic studies contribute to knowledge about how to maximize the quality of survey data with regard to nonresponse by making proper decisions on the use of survey mode and level of personalization when conducting survey projects.