Betsey Stevenson and Justin Wolfers have an interesting column today in Bloomberg View:
The real crisis in American higher education is that our best colleges never see a large chunk of our smartest students. In an important recent study, the economists Caroline Hoxby and Christopher Avery found that very few high achievers from low-income families ever apply to top colleges, and that the missing applications from these kids largely explain why they’re underrepresented at our leading universities.
Sarena Goodman, a job market candidate from Columbia, provides additional evidence that many disadvantaged students who would earn sufficiently high scores to go to great schools aren’t taking the test and don’t end up in college.
ABSTRACT: In the last decade, five U.S. states have adopted mandates forcing high school juniors to take the ACT college entrance exam. Using microdata on ACT test-takers, I demonstrate that, in the two earliest-adopting states (Colorado and Illinois), between one-third and one-half of students were induced into testing by these policy changes, and that 40-45 percent of them, many from disadvantaged backgrounds, earned scores high enough to qualify for competitive-admission schools. Moreover, selective college enrollment rose by about 20 percent following implementation of the mandates in these states, with no effect on overall college enrollment. I argue that this combination of results is incompatible with unbiased decision-making about test participation and instead must reflect a large number of high-ability students who dramatically underestimate their candidacy for selective colleges. The results thus demonstrate that lack of information about one’s competitiveness is an important determinant of college outcomes, and that policies aimed at reducing this information shortage are likely to be effective at increasing human capital investment for a substantial number of students.
owenzidar 
I wonder if Sarena were to compare the ACT scores of those students forced to take the test who score well with their state high stakes testing data how it would compute. I also wonder if she delved further into her claim that “many from disadvantaged backgrounds” scored highly enough to be considered for selective admission colleges. Based on research that shows a correlation between high poverty and low test scores, it would not seem possible. I also wonder if she actually investigated the reason for the 20% jump in college enrollment that she believes is a direct result of the ACT scores. Again, simply scoring well does not pay the tuition. Do Colorado and Illinois have state programs that fund college tuition for students who meet criteria including ACT scores? Then again, “selective” enrollment colleges are often private and do not qualify for those programs. I look forward to further research by Sarena or those who are interested in her study.
Pingback: College Costs & Enrollment for Low-Income Students | owenzidar
“Based on research that shows a correlation between high poverty and low test scores, it would not seem possible.”
This is a surprising statement. There will be a bell curve around the median for both high income and low income groups, with the latter median being lower. That doesn’t mean there are not substantial overlaps in the curves. This simply suggests that, left to their own devices, many of those in the low income group who actually can score in the same range as many in the high income test assume they will not qualify for competitive colleges and do not take a test. Requiring the test reveals who those students are, and opens up possibilities they might never be aware of. You can call it “forcing” them to take the test, which is kind of like “forcing” a young driver to stay within the speed limit.
There is no reason to believe high income kids as a group have higher native intelligence than low income kids. This report makes perfect sense.