This course offers participants an opportunity to design and program their own experiments. It provides an overview of the most popular approaches, namely lab, field, and survey experiments. It is intended for researchers who have a basic understanding of statistics and quantitative methods but have not designed and/or fielded their own experiment yet. The course will cover theoretical foundations and practicalities of experiments, placing special emphasis on design choices (external and internal validity). The instructor will also illustrate some of the typical modelling approaches using data from his own experiments, as well as publicly available replication data. He will also offer some pointers on common approaches to visualize experimental data.
Participants will have the opportunity to acquire the skills needed to randomize questionnaire blocks or items in Qualtrics. They will also familiarize themselves with access panels and online marketplaces (Lucid, MTurk, Prolific, etc.) where the bulk of survey experiments are fielded, critically comparing aspects such as data quality, accessibility, and cost efficiency.
One day will be dedicated to conjoint experiments, a versatile form of survey experiments with growing popularity, given their cost efficiency and ability to approximate real-world decision making while limiting social desirability bias. We also cover considerations of research ethics and best practices for debriefing experimental subjects when deceptive research designs are used. To avoid underreporting of outcome variables and experimental manipulations, pre-registration of experimental studies have become the norm. The course will teach you how to pre-register your future study and present best practices for pre-analysis plans (PAPs).
