Looks like there’s a lot of hand-wringing on the Internet over new reports that the average SAT scores have fallen by a couple points on all subsections this year. Some of the articles point out that this could easily be due to changing demographics of who actually takes the test. Some states have far more students taking the ACT rather than the SAT. There are also well known effects where SAT scores correlate with family income, so falling average scores could even be a good thing in the end if it really means that more lower income students are preparing to apply to college. It could also just mean that some recent tests were somewhat more difficult than usual,leading to artificially low scores. The test is always being tweaked in some way, with the return to the two section (math and verbal) format being a notable upcoming change, so it doesn’t seem too concerning if the average score moves around a bit from year to year. I would guess that this isn’t really much of a story in the end. It would be much more interesting if we had (1) a national exam taken by all students to get a more complete cross section of the population, (2) much more detailed demographic information (not just race/ethnicity and income but also type of high school coursework, high school demographics, future plans, etc) and (3) more statistical information than just the average score.