ABSTRACT
The subjects for the validation exercise are staff from 18 group child care agencies in North and South Carolinas. While the Trieschman Carolinas project instrument contains 102 items, some of these items refer to matters such as occupational group, agency size or location of services on a single site and as such seek demographic information about agencies and respondents. The low Cronbach’s alpha and the non significant Lillefors result for scale 7 indicated that this scale should be omitted from attempts to validate the family centred group care instrument. The standardised factor loadings are equivalent to correlation’s between observed variables and constructs and as such they indicate the construct validity of the observed variables. The magnitude of the correlation between service availability and parental involvement and sen ice availability and staff attitude is low, which may indicate that these variables are only marginally related but statistically significant.
