A Factor Analytic Approach to Human Engineering Analysis and Prediction of System Maintainability

摘要

The study is introduced by reviewing previous attempts to quantify and predict systems maintainability. A theoretical formulation is outlined which treats the measurement and prediction of system maintainability as a 'components-of-variance' model. The sources of variance are identified as: inter and intra man (personnel and social variables); machine (weight, volume, reliability, etc.) and man-machine interaction (human engineering design criteria). A restricted case of the formulation was investigated empirically in an attempt to predict maintainability of a sample of Air Force equipment from a questionnaire evaluation of the human engineering design features. Questionnaire responses, factor-analyzed by the Wherry-Winer Method, yielded eight orthogonal (independent) maintainability design factors. On the basis of the obtained factor loadings, seven of the original eight factors were selected as the most potent predictors. Several regression analyses were performed using different assumptions concerning both heterogeneity of sampled equipment and questionnaire scale stability. Coefficients of determination ranged from .27 to .47 depending on the assumptions made. Generally it was found that predictive efficiency increased and factor validity patterns changed when equipment was classified to make more homogeneous sets (Author)