November 20, 2017

Statistical Tests Find Support for Pay Disparity

This is one article in a series explaining changes at Gordon College, called Define: Gordon

By: Liam Adams ‘17 and Colin Ponzani ‘17

Staff Writer and Contributor

A document was provided to the Tartan that contained approximate salaries of every full-time and part-time employee at Gordon. After a series of tests, the Tartan has found statistical evidence suggesting women may historically have been paid less than men in comparable positions..

The document that was provided to the Tartan was a “Final Check Register” for the pay period of October 30, 2015. The document was accidentally emailed to many members of the community at about that time. The “Final Check Register” contains direct deposit payments of every faculty and staff member for that single pay period.  

Tartan reporters confirmed the accuracy of the document by asking employees who worked at the college in October 2015 to verify information in the document.

The results derived from these tests are not conclusive because there are several variables that could not be controlled.  One of the uncontrolled variables is withholdings that were  taken before the payment was made. The salaries provided in the document are only direct deposits, they are not the whole payment given to the faculty and staff members.

The withholdings that could not be controlled for include income tax, pre-tax retirement, pre-tax health insurance, and programs such as Medicare. It also must be noted that the document provided to the Tartan is a snapshot in time. The results do not represent a perfect picture of Gordon’s payment of faculty and staff.

A Tartan reporter grouped employees based on categorical, independent variables of position (faculty, staff, administration), full time or part time, sex, and race in order to measure potential differences in employee salary across these categories. The statisticians removed what appeared to be reimbursements for travel expenses as they were unrelated to salary.

One of the statistical tests student reporters and statisticians ran was a factorial analysis of variance. An analysis of variance is employed in order to expose differences between group means and averages. More specifically, this test can measure the influence and interaction of multiple independent variables on a dependent variable.

There were expected differences in employee salary in terms of employee position or full-time versus part-time status, but the most notable statistic was the difference in mean salary of male and female employees. The female mean salary for this pay period was $1,785.85 and the male mean salary was $2,006.88.

This pay discrepancy based on sex is made even more robust by the factorial analysis of variance, which folds the other independent variables of position, race, sex, and full-time or part-time status into the statistical test. Thus, the magnitude of difference in male and female salary means remains substantial even after the variable of sex is fed through the the other categorical variables of position, race and full-time or part-time status.

It’s important to note that the majority of these salaries were inherited, not generated, by the current administration. If pay discrimination exists, it is not a recent issue.  

Though there is evidence to support the claim that there is pay discrimination on the basis of sex at Gordon, there is no definite, conclusive proof. At this point, there is no clear conclusion about the reality of equal pay (or the lack of equal pay) at Gordon College. Advocates for pay equity say the only way to achieve clarity is to bring in a third-party review to perform a thorough evaluation.

Administration’s Response to “Define: Gordon”

Be the first to comment

Leave a Reply

Your email address will not be published.


*