Model 1 predicts math and reading scores based on family income and considers the control variables. It was found that family income has a significant but small effect on scores; for every 10,000 dollar increase in family income, there was a .2 score increase with a standard deviation of 1 percent.

Model 2 compares income segregation among families who enroll their children in public school. This Model doesn't predict reading or math achievement.

Model 3 compares the association between family income and math achievement. Owens found no association between family income and math scores in integrated MSA's but did find a significant association between math scores and family income in highly segregated areas.

Model 4 controls in MSA A. Race B. Racial segregation between districts C. Median household income D.Private school enrollment It was found that the association between family income and test scores becomes greater in highly income segregated areas.