Page 27, Table 2.3 SAS program for for null model

proc mixed data=imm10; model math = / solution ; run;

Results for null model

Covariance Parameter Estimates Cov Parm Estimate Residual 124.00 Fit Statistics -2 Res Log Likelihood 1989.0 AIC (smaller is better) 1991.0 AICC (smaller is better) 1991.0 BIC (smaller is better) 1994.6 Solution for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 51.3000 0.6906 259 74.28 <.0001

SAS Program for "With Homework" model

proc mixed data=imm10; model math = homework / solution ; run;

Results

Covariance Parameter Estimates Cov Parm Estimate Residual 93.7318 Fit Statistics -2 Res Log Likelihood 1915.6 AIC (smaller is better) 1917.6 AICC (smaller is better) 1917.6 BIC (smaller is better) 1921.2 Solution for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 44.0739 0.9886 258 44.58 <.0001 homework 3.5719 0.3882 258 9.20 <.0001

Table 2.4 (skipped)

Table 2.5 equations (for model "with homework"). The "meanhomework" is the mean of homework by school, computed within MLwiN via the "data manipulation" and "multilevel operations" menu.

skipped for now.

Page 30, Table 2.6

Results (these results have a different intercept and different standard errors)

Table 2.7, page 31, skipped