Good principals or good peers?

"School choice policies may improve productivity if parents choose well-run schools, but not if parents primarily choose schools for their peer groups. Theoretically, high income families cluster near preferred schools in housing market equilibrium; these need only be effective schools if effectiveness is highly valued. If it is, equilibrium effectiveness sorting' will be more complete in markets offering more residential choice. Although effectiveness is unobserved to the econometrician, I discuss observable implications of effectiveness sorting. I find no evidence of a choice effect on sorting, indicating a small role for effectiveness in preferences and suggesting caution about choice's productivity implications"--National Bureau of Economic Research web site.