- About 16 to 21 per cent of the total variance in English, mathematics, and science is between school variance, with the remaining 84 to 79 per cent being variance between individuals.
- The mean SES of the students in the school, their mean prior achievement scores and the like are related to performance over and above the relationship found at the individual level. Schools with larger proportions of students with high initial achievement, larger proportions of students with high initial achievement, larger proportions of students with high initial achievement, larger proportions of students with high socioeconomic status, fewer race minority students, stable rolls and the like are at an advantage, and students will perform better in them than they will in schools with the opposite mix of students.
- Other non-mix variables, such as stability of the school roll, are also related to achievement. If it is assumed that these independent variables are causal the conclusion that must follow is that it does make a difference which schools students attend.
- Check table 7.4 in page 124 for more detailed results.