These salaries on the image on the right seem extremely high. Half of Colgate Social Science majors are in the top 5% of income earners? Half of Tufts humanities majors? These are great colleges, but it doesn't pass the smell test.

And, looking at the actual data collected, the image on the right doesn't hold up. The study must have collected data from hundreds of colleges, yet there were only 7,300 observations. Sounds like a lot, but assuming that there were 200 universities (probably a low estimate) and 8 categories of majors, you're only looking at 4 to 5 observations per college per category. And these observations were probably unevenly distributed. So it's inevitable that a few colleges will have extreme "median" reported incomes, as the table on the right shows the highest incomes observed in each field.

So while the percentile distinctions cited in the article between selective and less-selective schools are likely accurate, the numbers on the image of the right-hand side - the really striking part of the article - have almost no statistical validity.

The WSJ is leading people to a shocking conclusion even though the real finding - that there's a 12%-18% difference in earnings between selective and less-selective colleges - is much less extreme. (It's also likely making a lot of people feel poor... one of the strange selling points of the WSJ that in this case simultaneously gets eyeballs on PayScale.com)