Diagnostic Tests and Robustness Checks

As noted in the previous lesson, whether the estimation and inference results are valid depends on whether the Gauss-Markov conditions are satisfied or not. Therefore, it’s important that we double-check whether the results in Braithwaite are based on violated assumptions and whether they’re sensitive to robustness checks.

Following the example in the previous chapter, we first use the augment_columns() function in the broom package to add various diagnostic statistics as new variables to the original data mid and then create a new diagnostic dataset mid.v2.