… a statistical test of selections from a pool
square calculates the statistical significance of disparities in an employment selection process when applicant flow data is available.
square computes probability values using the Chi-Square test and, for 2x2 tables, Fisher's Exact Test. For example, in a race/hiring case, square determines if there is a significant disparity in the rates at which blacks and whites are hired.
When this probability is sufficiently low (i.e., .05 or less), the result is said to be statistically significant. Disparities that are statistically significant are generally recognized as evidence of discrimination.
Overview of Process
square creates tables of rows and columns to analyze a selection process. By default, a 2x2 table is created, but you may create a table with up to 5 rows and 6 columns. You provide a title for the table, select one of the Display Options, and enter descriptive labels (e.g., Black, White, Hired, Not Hired) for the rows and columns.
Finally, you enter data into the individual cells and click the Calculate button to have square compute the results, which are reported in the Statistics box. The results may be printed or saved.
In the example below …
A charge has been filed against the XYZ Co. alleging race discrimination in hiring. We know who applied and who was hired: about 16 percent of black applicants were hired compared with 38 percent of white applicants.
square shows this disparity to be statistically significant, with a probability of 3 in 10,000 for the two-tail Fisher's Exact Test. As a rule, probabilities less than 0.05, or 5 in 100, are considered statistically significant.
— click image to expand