Search

The research described here moves personnel psychology beyond the blind empirical search for correlates and predictors of job performance and toward an understanding of the psychological processes and mechanisms that determine the wide variation in job performance that has long been observed and measured for all jobs studied. Based on studies of workers' output in jobs of many kinds and levels, this variation has recently been formally quantified and shown to be very large. On a typical lower level job, a worker at the 85th percentile in performance produces about 20% more than average worker. On the typical medium-complexity job, this figure is 32%. For professional and managerial jobs, it is about 48%. Thus, improvements in selection, placement, and training that findings from the sort of research described here may make possible could lead to large and economically significant improvements in output and economic competitiveness.

The practice of predicting performance by means of psychological tests began in the modern age with World War I. If an employer were to use only intelligence tests and selected the highest scoring applicant for each job, training results would be predicted well regardless of the job, and overall performance from the employees selected would be maximized. Many people criticize the current test batteries, which they feel are too dependent on academically learned content, the knowledge of which can be affected by the extent of the examinee's experience as well as ability. Despite differential individual investment, the current test batteries work; their validities are evidence of their efficacy. Virtually all individuals are exposed over a number of years to basic mathematics and reading and vocabulary. Measurement of this common investment has formed the basis of intelligence tests. Contemporary cognitive psychology offers tests that appear to be content free.