Last week, a story in the NY Times got quite a lot of coverage. It reported that women aged 21-30, working full-time, in big cities like NY, LA, and Dallas were dramatically out-earning their male counterparts (e.g by 17% in NY). This despite lagging behind by 11% nationwide. Feminist bloggers expressed cautious enthusiasm, noting that many of the most serious financial problems for women are associated with motherhood, and that young urban women may not yet have had children. Still, it seemed a trend worth keeping an eye on. However, a more recent article in the Wall Street Journal (mentioned at Broadsheet) points out that the Times story relied solely on medians, and that a closer look told a messier story. For example, the chart in the Times article breaks the data on medians down into 4 racial subgroups; “Asian men and women made the same amount; Hispanic and black women outearned their male counterparts by just 8% and 7%, respectively; and white men made 13% more than white women.” How could this be? Well, medians are very strange and subject to Simpson’s Paradox, which I taught in Critical Thinking a few times (there’s a nice explanation of it in the WSJ article). As a result, they can give a very misleading picture of what’s going on. So, the news may not be quite what it seems. Many thanks to the Jender-Parents for starting me off on this one.