Predoc's Stata-tistic Guide to Experimental Economics (Updated on 11/13/2024)
This intro-level Stata tutorial is designed for predocs and graduate students in experimental economics. Using an example dataset, I’ll cover key steps in data management, cleaning, and analysis. Topics include data organization, power analysis, pre-analysis plans, randomization, data cleaning, codebooks, descriptive analysis, balance and attrition checks, intent-to-treat models, heterogeneous treatment effects, and robustness checks. I also provide tips on communicating results effectively with supervisors and creating an organized data workflow. This guide reflects my 1.5 years of experience as a predoc at the Behavioral Insights and Parenting Lab, University of Chicago. I hope it prepares future predocs for the challenges of data analysis in experimental economics. Feedback and corrections are welcome!
Intuition Behind Power Analysis (Written on 11/13/2024)
The 5% significance level and 80% power are standard benchmarks in hypothesis testing, but their meanings are often taken for granted. While we distinguish between one-tailed and two-tailed tests for significance, we don't have similar distinctions for power. Why is this? This post explores why power analysis matters and how these familiar thresholds connect to fundamental statistical principles like the Law of Large Numbers and the Central Limit Theorem. Using visualizations, I illustrate the relationship between significance testing and power analysis, making these concepts more intuitive.