Earl R. Misanchuk

Analysis of multi-component educational and training needs

ConclusionThe adaptation of a proportionate reduction in error approach to the analysis of needs assessment data is relatively straightforward. In addition to permitting a greater sensitivity to differences in frequency distributions than other analytic approaches while providing a single statistic, ∇N, for the comparison of bivariate distributions, there is a potential for extension of the statistic to the multivariate case. A scheme of error weights specified by the researcher permits differential weighting of need components, although a standard one proposed. The researcher must also specify the expected distribution of responses, and again, a standard distribution is proposed. Although experience with application of the procedure is still limited, its apparent advantages make it the method of choice for the analysis of two-component needs assessment data.