Author: Richard C. Sorenson

Publisher:

Published: 1967

Total Pages: 64

ISBN-13:

The COMPUTERIZED MANPOWER SYSTEMS Task seeks to integrate the growing body of psychological, mathematical, and computer technology in the solution of problems in military manpower. One of its objectives has been to develop computer-aided research approaches and simulation tools useful in (1) defining general principles of manpower systems and (2) formulation and evaluation of manpower policies and procedures. Technical Research Report 1152 summarizes the research methodology developed or adapted by the U.S. Army Personnel Research Office to deal with manpower information problems. Particular concern is with relationships between the amount and type of information on which assignment decisions are made and the performance of the Army's enlisted personnel. Typical findings are: Total performance of Army enlisted personnel could be enhanced by using as performance estimates in optimal assignment procedures the full 2- or 3-digit scores on qualification measures. The effect of the one-digit score is more serious when the range of ability in the allocation group is narrow and when the general level of ability is lowered. More precise prediction of job performance could be attained by introducing performance estimates more closely resembling least squares regression estimates based on the full Army Classification Battery. Total performance could be increased by refinement of the Army Personnel System to afford additional job categories. Optimization techniques may be beneficially applied in the manpower system even though estimates of performance are ordinal (including various ways of ranking) and the assumption of an interval scale is not strictly met.