Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

Problem Set 1

RCT Simulation

Due Tuesday, November 15th at midnight

Using Python, R or Stata (not Excel), design a simulation study that illustrates the increased precision (smaller standard errors) afforded by including additional covariates in a fully randomized experiment. Then design a second simulation study that shows the bias that obtains when treatment is conditionally randomized but mitigated once one controls for that bias.

Build your context around a A/B test of two different website designs, where the outcome of interest is spending on products sold by the advertising firm (e.g. Warby Parker glasses). See the articles above for some examples of what this context might look like.

The finished product should be a short (2-4 page) writeup describing how you set up the simulation, what assumptions you made about the data generating process and empirical  context, and the key results (sets of parameter estimates) along with the code you used to produce the data and the estimates.