Home -> Law Blog Directory -> Legal Commentary Blogs -> Discriminations

Legal Commentary

: Discriminations

Here We Go Again: New Debate Over SAT ?Bias?

A new study in the Harvard Educational Review purports to confirm a highly criticized older study that purported to find racial ?bias? in a handful of questions on the SAT. (On that earlier study, see my ?Biased ?Bias?? from October 2003, which has a couple of links to thorough, and thoroughly critical, analyses of it.)

The alarmist conclusions of the new study are amplified through the educational megaphones of Jay Matthews? influential Washington Post blog (?New evidence that SAT hurts blacks?) and Scott Jaschik?s article on Inside Higher Ed (?New Evidence of Racial Bias on SAT?), which has a number of interesting comments.

Those two cheerleading posts were severely, and effectively, criticized by the invaluable blogger, JunkScienceMom, on Andrew Breitbart?s Big Journalism (?Junk Reporting ? the Hidden Truth Behind Claims Of Racial Bias In the SATs?) and on her own blog (?Blatant Bias?), where she notes all the evidence for test bias comes from ?two questions from four tests.?

Those two questions with a significant DIF [differential item functioning, the technical term for test items said to be biased] were also on one of the other SATs used in the study, but they were presented in a different order. Those exact same questions on a different test showed no significant DIF. Neither Mathews nor Jaschik found these facts worthy of inclusion in their articles which makes me wonder: Were they were deliberately ignoring them because calling the SAT racist will attract a lot more readers than a headline reading, ?SAT is pretty much fair to everybody,? or did they just not read the study or understand what they were reading?

The SAT does seem to reveal a good deal of bias, at least in the studies and commentary about it.

Full post as published by Discriminations on June 24, 2010 (boomark / email).