ERIC Number: ED232594
Record Type: Non-Journal
Publication Date: 1983-May
Use of Automatic Interaction Detector in Monitoring Faculty Salaries. AIR 1983 Annual Forum Paper.
Cohen, Margaret E.
A university's use of the Automatic Interaction Detector (AID) to monitor faculty salary data is described. The first step consists of examining a tree diagram and summary table produced by AID. The tree is used to identify the characteristics of faculty at different salary levels. The table is used to determine the explanatory power of the selected predictors and to determine the total amount of variation explained by the analysis. The second step consists of identifying the faculty whose salaries differ significantly from the means of their final groups and then reviewing their individual situations. The final step involves the use of a pairing method in situations where apparent inequity has been flagged. The main advantage of AID is its ability to detect interactions among variables. The application of the AID program was faculty data for 1979-1980 through 1981-1982 indicated that field was the most explanatory predictor of salary. Since most of the female faculty were in the fields having the lower salaries, the all-university average salary for females was lower than that of the male faculty. However, within fields there were still differences between salary averages. Race, ethnic origin, or sex were not selected as significant predictors of salary. (SW)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A