calculate_ai: Calculate Adverse Impact Metrics
In samuelkaminsky/SKTools: SK's Tools

Calculates adverse impact metrics

calculate_ai(
  df,
  groupings,
  stage1,
  stage2,
  n_min = 0,
  prop_min = 0,
  only_max = FALSE,
  correct = TRUE
)

`df`	Dataframe with passrate information. Must include columns for each grouping of interest and columns with logicals indicating success at each stage
`groupings`	Character vector with names of each column containing grouping variables (e.g, race, gender, etc.)
`stage1`	Unquoted column name of column indicating whether or not a row should be counted in the calculations (i.e., the denominator). Column must be logical.
`stage2`	Unquoted column name of column indicating whether or not a row should be counted as a 'Pass' in the calculations (i.e., the numerator). Column must be logical.
`n_min`	Minimum n count for a group to be included in the calculation
`prop_min`	Minimum percentage for a group to be included in the calculation
`only_max`	Only calculate adverse impact using the group with the highest selection ratio as the denominator
`correct`	a logical indicating whether Yates' continuity correction should be applied where possible.

List with two data frames. The first includes all selection ratios, the second includes all adverse impact calculation

`Selection.Ratio`	Dataframe with selection ratios
`Adverse.Impact`	Dataframe with adverse impact metrics, including adverse impact ratio, Cohen's H, Z score test of two proportions, Pearson's chi-squared test of proportions, and Fisher's exact