Designing Multi-Factorial Survey Experiments : Effects of Presentation Style (Text or Table), Answering Scales, and Vignette Order

Abstract

Multi-factorial survey experiments have become a well-established tool in social sciences as they combine experimental designs with advantages of heterogeneous respondent samples. This paper investigates three under-researched design features: how to present vignettes (running text vs. table), how to measure responses (rating vs. open scale), and how to sort vignettes (random vs. extreme-cases-first, to prevent censored responses). Experiments were conducted in a 2 x 2 x 2 between-subject design with 408 university students rating decks à 20 vignettes. Analyses of 7,895 ratings showed no differences of whether vignettes were presented as running texts or tables. Open scales revealed more measurement problems, e.g., missing values, than rating scales. Finally, vignettes presented randomly sorted produced similar results compared to sorting extreme vignette cases first. Recommendations based on the findings are to use random orders of vignettes and rating scales. Table vignettes provide an alternative to text vignettes but should be further evaluated with heterogeneous samples.