Recent research demonstrates that the test score gap between relatively advantaged and relatively disadvantaged students is much higher in some school districts than it is in other districts. But measured school quality often varies dramatically within a school district, and therefore it is important to know whether individual schools differ in the relative success of advantaged and disadvantaged students. We make use of detailed, linked birth and school records in Florida to investigate the degree to which this is true. We find that schools vary dramatically in the relative success of advantaged and disadvantaged students, and that different schools within the same school district differ substantially in terms of their advantaged-disadvantaged success gaps. In some schools, both advantaged and disadvantaged students fare especially well; while in other schools, both fare especially poorly; while in still others, one group does relatively well and the other group does relatively poorly. We investigate whether these differences across schools can be explained by differences in relative kindergarten readiness of advantaged and disadvantaged students, and we find that pre-school preparation is unlikely to explain the cross-school differences that we find. Moreover, we find that overall school advantage levels are unrelated to differences between the success levels of advantaged and disadvantaged students. Our findings indicate that policymakers should pay much closer attention to the practices of individual schools rather than concentrating exclusively on policies and interventions typically enacted at the district level. Moreover, the cross-school differences in the relative success of advantaged and disadvantaged students argue for enacting school accountability policies that shine the light on the success of specific populations, rather than concentrating solely on overall schoolwide performance levels or gains.