Previous research in the spatial assimilation of racial and ethnic groups has not assessed trends over time due to methodological difficulties and data limitations. I use an innovative method to assess the intercensal changes in neighborhood spatial attainment for African Americans, Hispanics, and non-Hispanic whites in Houston, Texas, between 1970 and 2000. I extend the current literature by showing that an accepted and commonly used method for assessing longitudinal change in spatial attainment is flawed and yields incorrect results. I highlight an alternative approach which makes use of data readily available in Census Summary Files to estimate individual-level spatial attainment regressions. I also show that the choice of neighborhood size affects estimates of spatial attainment effects. Although the influence of spatial scale has been demonstrated in the segregation literature, its consequences for spatial attainment research have not. I investigate and report findings from four geographic scales useful to and commonly used by spatial attainment researchers: the block group, the Census tract, the Zip Code Tabulated Area, and the Public Use Micro Data Area. I compare the benefits and drawbacks of estimating spatial attainment at each level of geography.