Organizational Diagnosis: Issues and Methods.
Abstract
This report presents different approaches to data based organizational diagnosis (in particular, distance function, discriminant function, decision tree, and a Bayesian approach) and their effectiveness when applied to approximately 6,000 work groups. In particular, several different criteria are used to evaluate the classification techniques: proportion of correct classification, ability to reproduce the typology, a zero-one closeness measure, severity of misclassification, the information required to make classifications, the amount of data required to develop the procedure, and the ease of implementing the procedure. Discussion of these issues, and, as a consequence, suggestions for further investigation are included. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1979
- Accession Number
- ADA065289
Entities
People
- Alan S. Davenport
- David G. Bowers
Organizations
- University of Michigan