Some years ago I read this NBER working paper. (Note – a couple years later a slightly modified version appeared in the American Economic Journal but I will quote from the earlier, non-paywalled version since it is available to everyone.)
Here’s the issue, in a nutshell:
In this paper, we use administrative data from the Houston Independent School District and the Louisiana Department of Education to examine whether the influx of Katrina and Rita students adversely affected the academic performance, attendance and discipline of their new peers.
Later in the paper:
…the arrival of low achieving peers hurts all native students, but this effect is more negative for low achieving natives in elementary and high achieving natives in secondary schools. By contrast, the arrival of high achieving evacuees benefits everyone, though the biggest benefit is for the low achieving natives.
If you missed that, later on the same page they write:
…we find that high achieving evacuees increase native performance and low achieving evacuees reduce native performance.
But it isn’t just performance…
By contrast, the results for discipline and attendance do show that it is enough to have 1 or 2 misbehaving evacuee children to worsen the attendance and behavior of native kids in elementary schools. In middle- and high-schools, only having many undisciplined kids in a classroom worsens native behavior.
And it isn’t just because more kids = less resources:
These results show no statistically significant effect of the fraction of evacuees on class-size in elementary schools. In middle and high-schools there is little evidence that the influx of evacuees significantly increased class-size, except for class-sizes in social studies which shows a marginally significant effect…. The results once again show no statistically significant effect of the influx of evacuees on either operating or instructional expenditures per student. This is likely because the Federal and State Governments seemed to have reimbursed schools and districts almost fully. Also, interviews with principals in Houston, suggested that schools received substantial aid from a number of foundations around the country.
Jumping to the conclusion, just to repeat the findings in case someone is tempted to misread them:
Non-linear models show that high achieving natives are significantly positively affected by high achieving evacuees and significantly negatively impacted by low achieving evacuees. Low achieving natives also generally benefit from high achieving evacuees and are hurt by low achieving evacuees in terms of their own test scores…
Of course, any parent who isn’t blind knows that a big determinant of the quality of his/her kids’ education is the quality of his/her kids’ peers. Still, its a well constructed and well executed paper. I also happen to think this situation makes a fine allegory for immigration.