UC Berkley Probelm

There was a study done in 1973 on UC Berkley over their school admissions process and if there was gender discrimination happening. In this study they admitted 43% of their male applicants, and 35% of their female ones which would show a clear discrimination. In fear of a lawsuit they hired a statistician named Peter Bickel to assess the data. Below is a graph that better shows this data:

The Truth of Admissions

The truth of the story is that the data isn’t what it actually seems. In reality, after Bickel took a closer look the females were actually favored in a majority of the departments. In fact females had a lower acceptance because they applied to the departments that had a lower acceptance rate overall. While the males did the opposite. This is the explanation behind the admissions problems.

The Simpson’s Paradox

In conclusion, much of this problem is related to Simpson’s paradox. This paradox is stated to be,“A trend or result that is present when data is put into groups that reverses or disappears when the data is combined”(Grigg). The UC Berkley admissions example, further explores this idea. It has a lurking variable, which is a variable that changes the data. Before looking at this variable it seemed that the university favored males over females. However after adding the “departments” variable, it showed that the females applied to lower accepted departments. Resulting in the less female acceptance overall. This is a strong example of Simpson’s paradox and the lurking variable.

Work Cited

Hoffman, B. (2020, November 27). Simpson’s paradox and interpreting data. Towards Data Science. https://towardsdatascience.com/simpsons-paradox-and-interpreting-data-6a0443516765