Experimental Vignette Designs for Factorial Surveys

The factorial survey (vignette study) as a tool for complex investigations has been rediscovered in the recent past. The planning of a vignette design is often a challenge to the researcher because, in general, the vignette population is too large to be presented to each respondent. Hence, in most cases the number of vignettes presented to each respondent must be drastically reduced. This can be done either by selecting a vignette subpopulation or by grouping the whole vignette population into sets, also called decks. There are several strategies for the selection & grouping of vignettes, which are associated with different implications for the analysis & interpretation of the data. We discuss the advantages of fractional & confounded factorial designs compared to not experimentally planned designs like random selection or random grouping strategies. We demonstrate the use & analysis of a specific experimental design, the randomized block confounded factorial design, with a small pilot study on normative attitudes towards giving foreigners Austrian citizenship. Tables, Figures, References. Adapted from the source document.