Discriminant Function Job Analysis in Three Army Technical MOS

Abstract

Abstract : This paper presents a technique to facilitate the determination of task training priority in Military Occupational Specialties (MOS). The job data analyzed consisted of scale ratings on tasks performed in three aviation maintenance MOS. Job incumbents rated MOS tasks they performed on a Relative Time Spent Performing scale. Supervisors rated all tasks in their MOS on four scales: Task Learning Difficulty, Consequences of Inadequate Performance, Immediacy of Task Performance, and Type of Training. An examination of the criterion Type of Training scale revealed that the data produced were not normally distributed and that the Type of Training scale categories were at the nominal level of measurement. These Type of Training scale characteristics made standard multiple regression analysis less desirable than discriminant function analysis, which is more compatible with these particular characteristics. For the discriminant analysis, the seven response categories for the Type of Training scale were collapsed into two new categories: tasks to be trained at local units to be trained in a formal school setting. The results indicated that the discriminant functions could classify tasks in the appropriate training category by incorporating the mean ratings per task on the four predictor scales. The discriminant function categorization agreed about 80% of the time with supervisor classifications based on the raw frequency of training choices. When there was disagreement in task training categorization, the supervisors' priority rating was typically anomalous.