Very interesting! Curious — how much of this is selection bias? As in, parents whose kids apply and kids who follow through are generally already more conscientious/engaged, and that explains the difference? (Charters’ results are often faulted for this same phenomenon). Echoing Edward below, I’d love to know what makes this work.
Good question. Because the program is over-subscribed, the authors were able to compare METCO applicants who got into the program versus those who applied but were waitlisted due to space constraints. So all the effects I cited were based on comparing applicants who go in versus those who did not.
Ah, I'm realizing you mentioned that Setren compared applicants who either did or did not get accepted/enroll — Assuming acceptance was random, I suppose that isolates the effect very nicely. (Though still curious what causes that effect!)
We don't for sure which of these elements is the secret sauce, but the paper I cited finds that, "participation in METCO results in a large shift in peer achievement. METCO participants’ classes have 0.45 to 0.61 standard deviations higher lagged test score averages than their counterparts. METCO participation leads students to have on smaller class sizes on average, better paid teachers, more experienced teachers, and better guidance counselor-to-student ratios. The program also shifts students to an environment with higher high school graduation rates where aspiring 4-year college is the norm. This environmental shift has the largest effect for boys and for students whose parents did not go to college."
Very interesting! Curious — how much of this is selection bias? As in, parents whose kids apply and kids who follow through are generally already more conscientious/engaged, and that explains the difference? (Charters’ results are often faulted for this same phenomenon). Echoing Edward below, I’d love to know what makes this work.
Good question. Because the program is over-subscribed, the authors were able to compare METCO applicants who got into the program versus those who applied but were waitlisted due to space constraints. So all the effects I cited were based on comparing applicants who go in versus those who did not.
Fwiw, here's an ungated version if you'd like to dig in more: https://blueprintcdn.com/wp-content/uploads/2024/08/Blueprint-Discussion-Paper-2024.07-Setren.pdf
Ah, I'm realizing you mentioned that Setren compared applicants who either did or did not get accepted/enroll — Assuming acceptance was random, I suppose that isolates the effect very nicely. (Though still curious what causes that effect!)
Question: Why, specifically, is this program successful?
We don't for sure which of these elements is the secret sauce, but the paper I cited finds that, "participation in METCO results in a large shift in peer achievement. METCO participants’ classes have 0.45 to 0.61 standard deviations higher lagged test score averages than their counterparts. METCO participation leads students to have on smaller class sizes on average, better paid teachers, more experienced teachers, and better guidance counselor-to-student ratios. The program also shifts students to an environment with higher high school graduation rates where aspiring 4-year college is the norm. This environmental shift has the largest effect for boys and for students whose parents did not go to college."
Has the number of applicants dropped??