f.adverse_impact: Compute adverse impact statistics

In awmeade/psymetlab: Provides useful functions for APA formatting and writing output to Excel

Description

Accepts grouping vector d.group, vector of outcome (e.g., pass/fail) as d.pass, and value for majority group (e.g., "male"), assumes only two groups

Usage

f.adverse_impact(
  d.group,
  d.pass,
  majority.group,
  minority.group,
  min_percent_minority = 0.05
)

Arguments

d.group	is a vector of grouping variable values
d.pass	is a vector of outcome values
majority.group	is the label of value assigned to the majority group
min_percent_minority	is the minimum percent of the sample that must be in the minority group in order to report results

Value

Returns a dataframe with columns...

total.n = overall sample size

majority.n = sample size of the majority group

minority.n = sample size of the minority group

percent.minority = percentage of sample in minority group

SR.total = selection ratio of the entire sample

SR.majority = selection ratio of the majority group

SR.minority = selection ratio of the minority group

impact.ratio = SR.minority / SR.majority

fishers.p = p value of the Fisher Exact Test

SD2.test = results of the 2 Standard Deviation Test

lower_95CI, upper_95CI = lower and upper confidence intervals around Impact Ratio

shortfall = the number of additional minority group members needed to pass to have an impact ratio = 1

Author(s)

Adam Meade [e-mail]

Examples

require('psych')
pass <- ifelse(sat.act$ACT > mean(sat.act$ACT) , 1, 0)
table(sat.act$gender)
f.adverse_impact(sat.act$gender, pass, majority.group = 2, minority.group = 1)

awmeade/psymetlab documentation built on Nov. 21, 2020, 11:39 p.m.