Members of the Centre for Social Informatics published an original scientific article in Journal of Survey Statistics and Methodology

Vasja Vehovar and Gregor Čehovin have published an original scientific article titled Using Web Survey Paradata to Analyze Response Quality, Respondent Characteristics, and Survey Estimates: A State-of-the-Art Review and Typology in Journal of Survey Statistics and Methodology, one of the most reputable journal for survey methodology. The review showed that the empirical use of web survey paradata remains limited in scope and is concentrated on response quality.

Paradata are digital traces of the response process, such as timestamps, device information, and records of questionnaire navigation. The authors conducted a systematic review of empirical publications for the period from 1998 to 2023, searching Google Scholar, Scopus, and Web of Science. They screened 3,436 abstracts and analyzed the 173 publications that met the eligibility criteria. Each publication was classified by domain and by the paradata indicators used. Response quality was addressed in 92 percent of the publications, respondent characteristics in 25 percent, and survey estimates in 6 percent. No publication covered all three domains. Response times were the dominant indicator (83 percent), followed by device characteristics (38 percent), whereas navigation, answer changes, and multitasking each appeared in less than a tenth of the publications. The annual number of publications remained stable. Advanced indicators that can be recorded only on the respondent's device were rare, because web survey software mainly supports basic server-side paradata. The authors propose an operational framework for integrating paradata into survey design.

The study was funded by research projects J5-3100 (Evaluating probability and nonprobability online panels) and J5-50159 (Assessing and Modelling Unit Nonresponse Bias in Probability-Based Web Surveys), and the research programme P5-0399 (Internet research).