“Pre-Analysis Plan: Does the causal inference we make depend on the research design we have? An experimental analysis into the determinants of causal reasoning”
Dr. Felix Bethke (CCCP)
Monday, 29 May 2017
12:00 - 13:30
The next talk in our Lecture Series on Microfoundations of Politics THIS Monday, May 29 at noon. Felix Bethke will present a joint project/pre-analysis plan with Ingo Rohlfing.
Please feel free to bring your own lunch!
This paper describes the pre-analysis plan for a survey experiment which seeks to analyze variation in causal reasoning among social scientists. Causal reasoning refers to how one identifies causality, i.e. the ways in which individuals use their beliefs about cause and effect to make inferences, decisions, and predictions about events and phenomena. Researchers doing empirical research in the social sciences either subscribes to a set-relational or correlational perspective on causation. The survey experiment we describe in this paper aims to uncover how researchers decide between these two perspectives. Using a factorial experiment, we confront participants with a vignette of a research design that varies on three dimensions: (1) the number of cases (small-n vs. large-n), (2) the measurement of the cause (binary vs. continuous) and, (3) measurement of the effect (binary vs. continuous). After reading about the research design, participants are asked whether they think a set-relational or correlational perspective is appropriate. Thereby, we seek to evaluate whether researchers are flexible in their causal reasoning, i.e. they adjust their reasoning regarding features of the research design, or whether causal reasoning is fixed.