The Sad, Sad SAT Factor
By Fawn Johnson

The College Board reported last week that 43 percent of college-bound students are academically ready for college. This means that less than half of those who took the test this year are likely to maintain a B- average or higher during their freshman year of college. The figure shouldn't be a surprise to anyone involved in higher education. In community colleges, it isn't unusual for three-quarters of the entering students to need some sort of catch-up course. Still, it's a problem for a country that seems to be in agreement that an increase in college graduates would help grow the economy and shrink the poverty rate.

Let's look at these numbers a little bit more closely. Math scores have remained stable over the last four years. That in itself is good news, since falling behind in high school math is the surest way to eliminate the most lucrative of college majors--the science, technology, and engineering fields that both President Obama and Republican nominee Mitt Romney are encouraging. Moreover, educators are well aware that reading and writing is harder to teach and harder to test than math.

Writing scores have declined by four and five points respectively. That's not good, but it could be worse. And the population of test takers is also expanding, largely in disadvantaged populations. The SAT test takers grew from 1.56 million in 2008 to 1.66 million this year, making 2012 the largest class of test takers in history. The number of test takers who qualify for a fee waiver has increased by 61 percent over four years. Almost half of the test takers this year were minorities (45 percent), up from 38 percent in 2008. The proportion of test takers who came from non-English speaking or bilingual homes increased by 10 points over 10 years.

How significant is the 43 percent figure in judging the quality of the future workforce? Does the expanded population of test takers explain the decline in reading and writing scores? How could the SAT test be improved? Are there other measures that can predict a student's success in college? What can be done to improve tests on reading and writing? What can be done to improve reading and writing instruction?

Jeanne Allen’s Confusing Numbers
By Kevin Welner

The below post from Jeanne Allen includes some confusing contentions. In particular, I don’t understand how she is using her numbers to contend that the expanded population of test takers does NOT help explain the decline in reading and writing scores.

Page 13 of main report from the College Board presents an easy-to-read section entitled, “The Increasing Diversity of SAT Takers.” The main information presented – in can’t-miss, bold type – is as follows:

754,922 minority students in the class of 2012 took the SAT, up from 600,830 in 2008.

457,971 class of 2012 SAT takers spoke a first language that was not exclusively English, up from 347,512 in 2008.

Unlike the National Assessment of Educational Progress (NAEP), which randomly samples students throughout the U.S., the SAT is taken by a self-selected group. As the nation becomes more successful in encouraging under-represented subgroups to consider college, we can expect that those taking the SAT will include more and more lower-scoring students from all groups (male, female, White, Black, etc). That is, when only small numbers of, e.g., Latino students take the SAT, those students will be clustered toward the top of the curve. As that pool grows larger, the average score for the subgroup will likely decrease. Indeed, this appears to be happening. Of course, other factors may be in play as well, but unfounded speculation about those possible factors isn’t particularly helpful.

Viewed charitably, perhaps Ms. Allen could be understood to be arguing that the overall flat numbers cannot be completely accounted for by shifts in group participation. That is, Simpson’s Paradox cannot completely explain these results. This is true. But see here and here for nice discussions of how understanding long-term SAT trends does require an understanding of Simpson’s Paradox.

In any case, where Ms. Allen and I can agree is that achievement gaps are unacceptable and should be addressed. Children learn when they have opportunities to learn. That is, achievement gaps arise from opportunity gaps. We as a nation have decades of strong research about the sources of these inequitable opportunities to learn. And we’re starting to see signs that lawmakers are indeed paying attention to some of these inequities – with a renewed focus on things high-quality early-childhood education and so-called ‘children’s zones.’ Let’s hope this trend continues.