There is a debate going on in the US about whether affirmative action leads to a mismatch effect. The idea is that students who gain entry to better colleges through affirmative action are often made worse off because they struggle to graduate, and struggle to pass professional exams. The claim is made that many of those students would have done better in life if they had attended a college that was a better match for their academic abilities.

A common response to this criticism of affirmative action is that, of course affirmative action students know they will have to work harder than other students to succeed. If they still choose to take up these admissions offers then that shows they think the benefits to be gained are worth the extra struggle, and the extra risk of failure.

A new study based on data from Duke University has just added to this debate. The authors are extremely tentative in all of their conclusions, but there is one obvious feature of the data that will draw a lot of attention. All four of the racial groups examined in the study received lower grades in their first year than they expected at the start of the year, but for the two groups who benefited from affirmative action the gap between expectations and reality was much larger.

In short, affirmative action students didn't just do worse than average, they did much worse than they themselves expected. Which suggests that they did not realise how much harder things were going to be. The authors of the study recommend that admissions letters should include information on grade averages for the racial group that the student belongs to. At least then students will have more realistic expectations when they decide whether to accept an offer.

I think that's a good idea because it will help some students to make better choices. Unfortunately I doubt if it will solve the broader problem. Many US Colleges have de facto racial quotas, and I doubt if Duke is an exception. If current admissions standards are not admitting enough students from a racial group to meet the quota then the standards are lowered. So if the extra information included in admissions letters does turn off some students, the standards will be lowered to make up for it.

Of course, if prospective students were all fully rational, of even equally rational, then lowering admissions standards would not work. Prospective students who just scraped in under the lower standards would look at the information about their prospects in the admissions letters and decline the offers of admission. The problem is that people are not all equally rational. Given the same information some people make better choices than others.

What I think will happen is that colleges who follow the advice of this new study will increasingly wind up trawling for students who are academically weaker and less able to make good life choices.

Via Instapundit and TaxProf.