Some Advances in the Design of Survey Experiments

This article calls attention to some designs in survey experiments that give new leverage in hypothesis testing and validation. The premise of this review is the modesty of survey experiments—modesty of treatment, modesty of scale, modesty of measurement. The focus of this review, accordingly, is the compensating virtues of modesty. With respect to hypothesis testing, I spotlight ( a) cross-category comparisons, ( b) null-by-design experiments, ( c) explication, ( d) conjoint designs, and ( e) sequential factorials. With respect to validation regimes, I discuss ( a) parallel studies, ( b) paired designs, and ( c) splicing. Throughout, the emphasis is on moving from experiment in the singular to experiments in the plural, learning as you go.