Use Regression For Pay Analysis
These individual variables are important because the technique for pay equity analysis is regression. Regression works by systematically plotting the variables of interest against the outcome variable. For pay equity analyses, the outcome variable is always pay. The variable you do not want to explain pay differences is gender. If gender explains pay differences, then you have discriminatory pay practices.
It is helpful to express how you expect variables to predict pay as a hypothesis.
Geography Could Be A Variable in the Analysis
I hypothesize that if you compare the pay of two employees with the same role – one who works in a higher cost of living location and one who works in a lower cost of living location – all other things being equal, the employee that works in a higher cost of living location will have higher pay. This is sometimes called a cost-of-living differential.
The way pay is assessed today, those differentials are usually offered for some cities within a country. In the EPIC methodology we propose that pay equity analyses should be done for employees with the same role across all countries. In this case, there could be cost-of-living differentials within a country, and country differentials to explain pay differences around the world. This topic has gotten complicated in a COVID world. Since employees were working from home, some set of them moved to a different geography. Organizations are now deciding on adjustments to their pay practices given that some employees moved to more or less expensive geographies.
As a reminder, organizations may or may not choose to take different variables into account. It is not Gartner’s intent to state what organizations should value in their pay practices, just that those pay decisions should be transparent. Your organization may not give cost of living adjustments, in which case you would not need to account for this in your pay equity analyses.
There are other variables that can explain pay differences. Stay tuned for the next variable in a future blog.