districts the expenditures per pupil ranged from $7,635 in akron central to $11,416 to ours in lackawanna city, about a 60% livings. the outcomes in the state-administered fourth grade reading, science and math scores was zero. at the school-by-school level, where the data was available school by school, the pupil-teacher ratios in term of class size ranged from 10.4 pupils in buffalo's public school, 72, to 20.8 in west seneca elementary, in other words, 1 100% differential. what was the correlation between class size and outcomes? zero. by contrast, the percentage of low-income children school by school accounted for 75% of the variation of fourth grade math scores, 77% of the variations in fourth grade science scores, and 87% of the variation in fourth grade english scores. as a matter of fact, if you at that time had just told me that percentage of low-income children in this school, i would have predicted the state english scores plus or minus six point is less than 1% on your scale with 95% accuracy. these are not revolutionary findings by me. we've known the relationship of socioeconomi