Measuring client satisfaction with public education III: group effects in client satisfaction

A contracted research project to evaluate client satisfaction with public education (King and Bond, 2003; Bond and King, 2003) required that the satisfaction of certain groups of clients received particular attention. A number of target groups were specifically identified by the state education department as those requiring separate satisfaction analyses. ConQuest software (Wu, Adams and Wilson, 1997) provided convenient techniques for estimating group mean effects for identified sub-groups of the parent and student samples. This allowed for the analysis of directly comparable satisfaction estimates for groups such as these, as well as for school size, school type, school location and parents according to the year level of the child's school class. Rasch analyses of group mean effects revealed generally, that for students, differences in overall satisfaction levels between identified sub-samples and the whole student sample even if statistically significant, were, at most, substantively marginal. However, for identified groups of parents and caregivers, the results were not so equivocal: one group of indigenous Australians, and parents of children with disabilities showed marked positive group mean effects, while another group of indigenous Australians reported lower satisfaction levels than did the whole parent sample. Moreover, analyses based on groupings of schools according to their relative complexity, appeared to reiterate the finding mentioned en passant in a previous paper (King and Bond, 2003), that client satisfaction decreased as school size increased.