Problems in needs assessment data: discrepancy analysis.

Needs assessment (NA) is generally based on the discrepancy between two conditions-the desired and present states. To date, there has not been an extensive research regarding a number of subtle problems in discrepancy analysis. One such example is missing data for one or both the two states. This leads to highly varied item n's for calculating discrepancy scores. Concerns like this arose in a NA study of minority students in science, technology, engineering, and mathematics programs in universities. A number of problems observed in this context are discussed as well as possible solutions for them. The results should be valuable to needs assessors and evaluators responsible for assessing needs.