S.I. 8 (PART) STUDY III

The third study was also run online on a representative sample of adult Norwegians. We conducted a conjoint experiment with a rating assignment embedded in the 2018 fall wave of the Norwegian Citizen Panel.

The conjoint design allows us to explore scope conditions of findings. The advantage of conjoint designs, a method increasingly used in experimental political science research (Leeper, Hobolt, and Tilley 2020), is the possibility to include a large number of experimental treatments and estimate their effects simultaneously (Hainmueller, Hopkins, and Yamamoto 2014). The stimulus material for each respondent consists of randomly combined levels of the given features, resulting in sufficient observations for each feature while only a small share of the possible combinations is actually shown to participants. This enables us to include context variables into our design and to test our relationships of interest across a range of other factors that could potentially interfere with the treatment. As a result we gain insights into generalizability and robustness of our findings. Our conjoint design is presented as a text vignette with rating outcome measures like in studies 1 and 2. For similar applications in conjoint designs, see for example Huff and Kertzer (2018).

References

Hainmueller, Jens, Daniel J Hopkins, and Teppei Yamamoto. 2014. “Causal Inference in Conjoint Analysis: Understanding Multidimensional Choices via Stated Preference Experiments.” Political Analysis 22 (1): 1–30.

Huff, Connor, and Joshua D Kertzer. 2018. “How the Public Defines Terrorism.” American Journal of Political Science 62 (1): 55–71.

Leeper, Thomas J, Sara B Hobolt, and James Tilley. 2020. “Measuring Subgroup Preferences in Conjoint Experiments.” Political Analysis 28 (2): 207–21.