What We're Reading: New Evidence on College Prep and Financial Aid

The National Bureau of Economic Research has released a couple of new papers worth reading.

(1) A new study from Kirabo Jackson at Northwestern considers a Texas college-prep program called APIP: Advanced Placement Incentive Program. It provides cash incentives for student and teachers to help students perform well on AP tests, and it provides teacher training, curricula oversight, and test-prep sessions.  While this clearly fits the model of "teaching to the test" it's about way more than paying students for grades.  The cash incentives for the teachers are real money-- depending on the school, as much as $500 for each student in their course scoring a 3 or better on an AP exam, and bonuses up to $1000 at year's end, and for the "lead" teachers, annual bonus payments up to $15,000 per year. Payments to students are also large-- not only is the fee for the exam defrayed, but they can also get up to $500 for good scores.

Now, this is clearly not an entirely public program-- private donors are paying 70% of the costs, which can range up to $200,000 per school per year, which an average cost of $225 per student. Moreover, currently opportunities aren't extended to all schools in Dallas, where the program operates--these are schools that seek out the chance and sign up. In fact, Jackson carefully notes, "Many districts are interested in the program but there are no donors. So there is always a shortage of donors.”

APIP was not randomly assigned to students, but Jackson uses a rigorous quasi-experimental diff-in-diff approach to try and sort out program effects from selection effects.

Jackson identifies program impacts on college attendance and completion, with attendance effects more prominent for white students, and completion effects more pronounced for black and Hispanic students. While he draws attention to these heterogeneous impacts in the paper, he does not discuss the potential mechanisms underlying the heterogeneity-- I hypothesize one reason is that while succeeding in AP coursework may not be enough to move more black and Hispanic students into college, at the institutions when they commonly attend, having the AP credit may well accelerate time to degree for those who manage to go.  On the other hand, a lack of AP courses may be among the few barriers faced by the white students, but since they are more likely to attend selective institutions that use AP in admissions but don't apply those credits against college coursework it may not accelerate time to degree.

In any case, this is a carefully executed study well worth-reading.  The author concludes,"Because there has been little credible evidence on the efficacy of college-prep programs despite large public  and private expenditure on such programs, the results of this study are encouraging about the potential efficacy of college-preparatory programs at improving the educational outcomes of disadvantaged students who are consigned to inner-city schools."  I would only add, the results are only encouraging if you believe that there will ever be enough private donors and/or sufficient resources devoted to helping students at disadvantaged schools succeed. Perhaps living under the Walker-regime in Wisconsin simply makes me a little too pessimistic.


2. The second fascinating study of the morning tests a hypothesis I've floated in several places recently-- perhaps giving promotes giving, especially when it comes to financial aid.  Jonathan Meer and Harvey Rosen examine this at a single (anonymous) research university (which I'm betting is Princeton).  Overall, they find the answer is no-- people don't tend to return the generosity they are given.  There's virtually no effect for students who received scholarships or campus jobs, and in fact students who took out/ are saddled with loans are less likely to give to their institution post-graduation.  In other words, duh -- this strategize that state institutions are being forced to rely on, putting the burden of costs of attendance on their students, is going to drive down the size of their endowments over time.  It contributes to a vicious cycle.

What the paper doesn't do is examine the effects of widely-respected, beloved programs like the federal Pell Grant. In my research I detect substantial gratitude among students for that program, and I strongly suspect that given the opportunity to realize the economic benefits of college attendance--e.g. when people reach their mid-40's-- they are more likely to give back.  A hypothesis still open for testing.

In addition, it's worth considering the possibility that effects of aid on giving are heterogeneous-- much like the effects of aid on college persistence, and the effects of college on wages.  This anonymous university isn't very diverse-- it's 74% white and 31% of students went to a private high school. The authors test for some race interactions and find a few but the more interesting question is whether students more likely to depend on need-based financial aid are more likely to give-- and this school has very few truly poor students attending it.

Overall, a nice start for a literature with plenty of room to grow.

For a free copy of each paper, click the link in the first paragraph of each description and type in a .edu email address.