13.7 Regression tables

  • each column represents a different regression
  • parentheses: usually standard errors (sometimes t-statistics)
  • significance stars –> p-values (author isn’t a fan)
  • below: descriptions of analysis/measures of model quality
  • adjusted \(R^2\) consideres number of variables
  • \(F\)-statistic: null: all the coefficients in the model are all zero
  • RMSE: estimate of the standard deviation of the error term

What can we do with all of these model-quality measures? Take a quick look, but in general don’t be too concerned about these.

If you don’t care about most of the causes of your dependent variable and are pretty sure you’ve included the variables in your model necessary to identify your treatment, then \(R^2\) is of little importance.

13.7.1 Interpretation

A one-unit change in …

13.7.2 Controls

The idea is that by including a control for Year, we are removing the part explained by Year, and can proceed as though the remaining estimates are comparing two inspections that effectively have the same year.