Wednesday, August 26, 2009

On Those SAT Scores

The new summaries of SAT scores by gender, ethnicity and income have come out. They are not that different from the past, even if the average scores for many groups have slipped a little. The tests are being changed over time and the population taking the tests is changing, too. For instance, recent immigrants are not going to score as well on something which is very culture-dependent than those who were born and grew up in the country. The larger the percentage of recent immigrants among the test-takers, the lower the average scores will be.

The average score for girls/women is lower than for boys/men for the American SATs. The reasons for this discrepancy have been debated, especially given that girls do better on other criteria which predict college success, but one of those reasons certainly is the fact that a larger percentage of female school-leavers takes the test when compared to male school-leavers. If those most likely to take some test first are the ones who expect to do well on it, then the average score will drop as more and more people from a certain group starts participating, always assuming that other reasons for greater participation (such as increasing income levels of families who only now can afford college for their children) are held constant.

That's why I found this statement a little puzzling:

•Average scores dropped 5 points for females and 2 points for males. While females represent more than half (53.5%) of test takers, their total average score (1496) is 27 points below that of males (1523).

There's a lot more to be written about the gender gap in the U.S. SAT scores. For example, the tests have been adjusted in the past in ways which raised the average male score and lowered the average female score and the experiences of other countries differ from the American one. But understanding the effect of more females taking the test is important.