Results
Multilevel models were estimated separately for the two out-come measures of interest – the PPVT and the EOWPVT. Standards cores of both measures were used in all analyses. For each outcome variable, a random intercept model with no predictors (i.e., a random effects ANOVA) was first estimated to determine how much of the variance in the outcome was associated with the teachers. ForPPVT, 00= 6.09 and 2= 70.36, resulting in ICC = .08. For EOWPVT,00= 12.10 and 2= 79.66, resulting in ICC = .13. For both outcome measures, the variance of the random intercepts (00) was not significantly different from zero, but given that the sample size was small, resulting in low power for these tests, and the fact that the ICC values were substantial, multilevel models were still used for the main analyses.