Differences in pay between women and men is a troubling and well-documented phenomena that occurs around the globe. In the United States, it is estimated that women make roughly $0.80 for every dollar that men make.
Such patterns are particularly troubling when they are observed in public sector jobs, as governments should not only enforce but also demonstrate fair and balanced employment practices.
Here, we investigated wage discrepancy within the City of Seattle.
The City of Seattle largely pays men and women equal wages for the same job (median wage difference = $0).
However, the city employs vastly more men than women. The imbalance in employment is particularly noticeable in a few large departments, which contain a majority of the city's workforce.
Of important note is that the gender imbalance is particularly striking when comparing jobs with high salaries. As seen in the plot below, the proportion of women in a given income bracket decreases as hourly wages increase.
On average, men are paid ~$3.00/hour more than women (P-value < 0.001). Assuming a 40 hour work week and 50 week work year, men are paid, on average, ~$6000 more per year than women.
All analyses were performed using Python v3.X and R 3.4.3
Data downloaded from City of Seattle repository.
Gender was identified using the Python package gender_guesser, which uses first names to categorize genders into the following categories: andy (androgynous), male, female, mostly_male, mostly_female, and unknown. We combined mostly_male and male into a single male category and mostly_female and female into a single female category.
We explored the data for patterns of pay vs sex within jobs, departments, and compensation levels. The statistical significance of differences in pay between men and women was determined using three methods: T-test, Wilcoxon rank sum test, and linear regression. All tests were highly significant and indicated that men receive more compensation than women (even after corrections for multiple comparisons).