Using the standard ANOVA, we estimated
between-group variance ([CV.sub.G]) = 42.4%.
Between-group variance was associated with two moderator variables: method of data collection (observation less than interview) and rater (reported by others less than self-reported).
First, we conducted null model analyses with no predictors to identify any significant
between-group variance in the variables.
Based on the distribution characteristics of the clinical data of each group, t -tests were performed to compare data showing a normal distribution and homogeneity of
between-group variance, while analysis of variance (ANOVA) was used to compare data from multiple groups.
As shown in Model 1, for PEG-AA, the ICC (ratio of the
between-group variance to total variance) was .03, indicating that only 3% of the variance in access aspirations was attributed to between-group (i.e., classroom) differences and that 97% of the variance in the access aspirations measure was attributable to differences in student-level characteristics.
Instead, we performed a series of one-way ANOVA tests to estimate ICC and calculate within- and
between-group variance in cultural values for countries and other clustering dimensions to see what percentage of differences between subjects reside within and between groups, or how similar subjects in groups are to one another, and how distinct groups are from one another.
The first predictor variable selected for inclusion by DISCRIM is that variable that explains most of the total
between-group variance (i.e., group separation), while any remaining variables are selected sequentially based on the amount of residual variance they explain.
Between-group variance in the number of teacher-reported hours of English language development instruction with ELL students was not statistically significant.
Accurate specification of boundaries in research is necessary to meet the basic statistical and measurement principal of maximizing
between-group variance and minimizing within-group variance.
The modelling revealed that three discriminant functions are significant and when combined they explain 94.3 per cent of the
between-group variance across the Voting Groups.