4 ms·

Also, it doesn't "prove" anything . . . no statistician would ever make that kind of assertion. Correlation is not causation. The conclusion has an error margin

• VHRanger 9y ago
  Depends on the statistical method used. Some quasi-experimental methods like instrument variables or difference-in-differences can plausibly tease causal effect sizes from messy economic data That said, this paper seems to be mostly simple descriptive statistics