Previous studies found that the small size of schools or school districts mitigated the negative influence of poverty on academic achievement in California, Alaska, and West Virginia. The Matthew Project extends this research in four additional states selected to provide varied settings: Ohio, Georgia, Texas, and Montana. Data from each state were used in regression equations that predict overall school or district achievement from measures of size, socioeconomic status (SES), and the product of size and SES. These equations illuminate possible "excellence effects" of size by showing which communities (based on SES-level) may benefit or lose from increases in school or district size. Equity effects of size on achievement were also tested by computing the correlation between SES and achievement in groups of larger and smaller schools and districts. Strong evidence of an interaction effect of school size was found in Ohio, Georgia, and Texas, such that academic achievement benefited from smaller schools in more impoverished communities and from larger schools in more affluent communities. A weaker interactive effect was found in Montana, which maintains many small schools. Across all four states, a strong equity effect was found at all grade levels, whereby small size reduced the negative influence of poverty on school and district performance. Strong evidence of an interaction effect of district size was found only in Ohio. The Matthew Project studies indicate that a one-best, everywhere "optimal," school size is a figment. Four policy questions related to school and district size are discussed. (Contains 15 references.) (SV)