This paper reviews the use of a segregation ratio in analyzing changes in the pattern of socioeconomic segregation between schools in England and Wales, addressing how the modifiable areal unit problem affects results. Researchers are developing methods for assessing socioeconomic stratification among school admissions and for comparing those across time and place. The best measure shows that school admissions have become more even since 1988. While there are variations in levels of change between areas of England and Wales, findings are consistent at all levels of aggregation, from school to national. The remaining problems are whether the precise definition of the area of analysis affects the results, whether researchers should use economic and administrative borders, and whether it would make any difference if researchers used natural markets or approximate areas of competition for each school. This paper examines the effects of the modifiable areal unit problem on the segregation index and segregation ratio, concluding that using different levels of analysis in calculating the segregation ratio actually focuses on different segregation levels. Therefore, differences between them are equally valid and informative. However, in the case study examples, year-to-year changes are similar no matter what levels of analysis are used. (Contains 15 references, 7 figures, and 4 tables.) (SM)