504 



SELECTION AND TRAINING 



to be expected, because so many other fac- 

 tors condition success and failure in school 

 that men with desirable personality makeup 

 often fail, and men with poor personality 

 makeup pull through. This is not to say 

 that personality may be ignored. Some per- 

 sonalit}' test might predict accurately a rat- 

 ing on "Annoys fellow crew members" or 

 "Loses nerve in face of danger," and that 

 prediction would be important. 



Even if good multiple criteria were de- 

 veloped, no method is now available for 

 handling them efficiently. Suppose there 

 are 20 aspects to the job of operating a Loran 

 transmitter; suppose there are tests which 

 predict how well each man will do each as- 

 pect of the job. How can such data be 

 used effectively in classification? It is im- 

 portant to develop the mathematics of pre- 

 diction with multiple criteria. One suitable 

 approach is factor analysis of the criteria. 



This analysis may in turn have implica- 

 tions for billet analysis, if a job is found to 

 break down into several psychologically -un- 

 related aspects (cf. 28, pp. 20-24, 69-80). 

 Speculating, it appears possible that we 

 should find a widespread mathematical fac- 

 tor, running through such skills as deter- 

 mining a depth-charge setting, interpreting 

 data from triangulation detectors, planning 

 an interception course, etc. A man with 

 superior mathematical ability might be 

 better at all these functions than the person 

 not so superior. Other shipboard functions 

 might fall into other categories: listening 

 jobs, visual jobs, jobs requiring mechanical 

 insight. It might pay to consider redesign- 

 ing billets so that the best mathematical 

 minds were available for mathematical 

 tasks, so the best ears were trained to do the 

 auditory tasks from monitoring a tor- 

 pedo to listening to the engine valves, 

 and so on. 



The most meaningful type of psychological 

 measuring scale represents one thing and 

 one thing only. A composite estimate which 

 combines a man's knowledge of equipment, 

 skill in detecting targets, speed in making 

 repairs, alertness, and pleasantness, is a very 



poor basis for judging the man's fitness for a 

 particular job. Training might be seriously 

 deficient in one aspect, and yet all men might 

 earn superior ratings at sea because of other 

 assets. A man could be sent to sea lacking 

 some one basic aptitude, yet this weakness 

 might be ignored for years if his record 

 reports only a general estimate of all-round 

 quality. The only effective plan for assess- 

 ing men is to measure separately each man's 

 proficiency in each vital aspect of the job. 

 The proper criterion for evaluating psychia- 

 tric screening is emotional stability on duty, 

 not all-over proficiency. The criterion for 

 evaluating target detection is the man's abil- 

 ity to pick up targets, not some other in- 

 dex such as the accuracy of bearings when 

 he has a target. One of the few attempts 

 to obtain multiple criteria for use in evaluat- 

 ing selection programs is to be credited to 

 Navy aviation psychologists. They ob- 

 tained nominations of good and poor fliers 

 in carrier squadrons, and had the nominator 

 tell why each man was good or poor. This 

 led to a pool of men known to be very good 

 in, say, emotional adequacy under pressure, 

 and a second pool of men poor in this re- 

 spect. A second pair of groups differing 

 in teamwork or in flying skill was selected, 

 and so on (53). The predictors could then 

 be correlated with each aspect of the criter- 

 ion. Final reports on this type of study are 

 not available, but the approach is most 

 promising and should be extended even 

 though wartime criteria are lacking. Re- 

 search can at least set up efficient plans 

 for obtaining and using multiple criteria. 



Qualitative Criteria 



No measurement or rating gives as effec- 

 tive a picture of how a man works out on a 

 job as does a description. This is especially 

 true of officers, who can show a great varietj' 

 of assets: daring, judgment, steadiness, re- 

 sourcefulness, and so on. If we are ever to 

 assess personal attributes, these assessments 

 must be judged against a suitable descrip- 

 tion of performance on duty. One cannot 

 predict a rating without knowing the man, 



