Obama Pays Women 18% Less Than Men (Is a Fake Statistic)


The Washington Free Beacon, a website that appears to be run by a high school chapter of the Young Republicans, recently lobbed a juvenile charge at President Obama: He is sexist because he pays women 18% less than men. They determined this by taking the total salaries of women and dividing that number by the total number of women employed and comparing that number to the total salaries of men divided by the number of men. 

If you think this is an accurate method of comparing salaries, then you either failed 6th grade math, don't understand statistics, or are feigning ignorance to peddle a weak story.

We have a crisis of civic participation in America, and I blame the level of stupidity in our political discourse. Intelligent people knowingly pass off nonsense like this as truth and less sophisticated people lap it up, infuriating or deflating the civic efficacy of people with moderate levels of intelligence. It pains me to debunk this idiocy. As I write this piece, I feel like a gigantic rock is pressing down on me, slowly squeezing out my will to participate in increasingly moronic political discussions initiated by people who either know better or should be forced to wear beanie caps with propellers on them at all times.

It's an open question whether the editors of the Free Beacon are drooling imbeciles or moderately clever ideologues, but my continued faith in humanity forces me to accept the latter. They can't possibly be dumb enough to believe this rubbish, can they? Should I even bother explaining why this study is dumb? If you don't already know, then I urge to you delete your accounts on this site and then either read a book on statistics or have your brain scanned for a debilitating malfunction.

When performing a statistical analysis, you need to isolate variables that might bias your analysis. You also need to compare apples to apples. The laziest possible analysis of compensation discrimination would compare the compensation for similar jobs. This would be a horrendously flawed method of comparison, but the Daily Beacon failed to account for a basic variable like "type of job." Employment discrimination studies need to account for job descriptions, qualifications, tenure, and other factors. There are some very good studies that show that women make much less than men when you hold these factors steady.

Using some crude estimates and the Free Beacon's methods, I sought to determine whether the Washington Redskins are a sexist employer. I found that - on average - the Redskins pay female employees ~98% less than male employees. Does this mean that the Redskins are a sexist organization? No.

Do you get it? If you don't, I'll be happy (not really) to explain this further.