Cluster Analysis of Occupational Data with Focus on Task Rather than People

The job analyst is concerned with identifying and systematically recording the behaviors performed by job incumbents. From the collection and analysis of these data by the job analyst, inferences can be drawn and useful recommendations made regarding such matters as personnel selection policies, training programs, planning manning tables and force studies. One of the important needs of the job analyst for accomplishing these ends is a method of grouping tasks into meaningfully useful clusters. One such method is cluster analysis, i.e., a technique by which entities are formed into relatively homogeneous groups, based on similarity measures. The usual procedure in such analysis is to measure a number of attributes of the entities and by pairwise comparisons of the entities andor subclusters of the entities form clusters based on the similarity of their respective attributes. When applied in occupational analysis these techniques can cluster individuals entities on the basis of the tasks attributes they perform. The results of this process are clusters of people who perform similar jobs. Observe cluster analysis is a modification of the usual clustering procedures so that clusters of tasks are constructed on the basis of individuals who perform them. The task measurements used are the same as in a traditional method of clustering. However, in observe clustering a task is clustered with another task depending upon how many, or few, individuals perform both tasks.