I came across this article “Politics and Personality: Most of What You Read is Malarkey” in the New Yorker.
Overall, I think this is an interesting piece. It summarizes the source document quite well. Given that there are a lot of problems with the reproducibility of some social science experimentation (especially in the field of social psychology, see the reproducibility project, this summary from science, and this Atlantic piece about Brian Nosek’s version of the project).
As Ms. Konnikova notes about the bulk these types of studies they use correlation, and as we all should know “correlation does not equal causation.” However, and this is where I push back on the ultimate thesis of this piece, these types of analyses in psychological science are somewhat outmoded. To hold modern psychological science to the methods of the past in light of OLS regression, multiple regression, structural equation modeling, and other methods which can establish both correlation and causation better than those used when Zimbardo did the “Stamford Prison Experiment” is somewhat disingenuous.
