Termin: Sonntag, 16.09.2018, 9-17h
Ort: SH 1.106 (Seminarhaus)
Einen Lageplan der Gebäude am Campus Westend finden Sie hier.
This workshop targets methods to deal with response styles in self- and other-reported data. Response styles refer to the systematic tendency to answer questionnaires on some basis other than the target construct. Individual and cultural differences in response styles (e.g., acquiescent, extreme, midpoint and socially desirable responding) can introduce bias and invalidate conclusions. Therefore, it is important to check their presence and gauge their impact, especially in cross-group comparisons. This workshop first provides a systematic review on the nomological network of response styles and their impact on cross-group comparisons. Then, procedures to deal with response styles at the design and data analysis phase of a study are introduced with empirical examples. These procedures consist of (1) design-based procedures including direct assessment of response styles, alternative response formats, measurement of overclaiming, and anchoring vignettes; (2) Post-hoc construction and detection of response styles (with Likert-scale data collected) such as response style indices with the conventional counting procedure, bi-factor models, and latent class analysis. Example data from the Programme for International Student Achievement (PISA) are provided to have a better understanding of these procedures and for hands-on practice. Participants are also encouraged to bring their own data to check the effects of response styles.