Key points are not available for this paper at this time.
is an identity between the contextual effect specified in eq. 1 and that identified in the analysis of covariance method of assessing contextual effects (Hauser, 1971; see also Alwin, 1976b, 1976c). Notice that if the between-school regression coefficient and the within-school coefficient are the same, i.e. byKj = b, there can be no school context effect. This issue can be resolved quite easily by examining the statistical significance of the difference (see Schuessler, 1971). In the present analysis we estimate several models of the form:
Sorenson et al. (Sat,) studied this question.