510 



SELECTION AND TRAINING 



of a particular type should be as similar as 

 possible: position and appearance of con- 

 trols and indicators, speed of train and stiff- 

 ness of controls, marking of reticles, and so 

 on. 



Error Analysis 



For effective training it is necessary to 

 determine the errors men make on the job 

 and to remove them. Observing men as 

 they perform their duties shows not only 

 how competent they are, but reveals the 

 specific errors they make. Similar studies 

 of men at the end of training reveal their 

 weaknesses. For example, it has been found 

 that gunners tend to overestimate opening 

 ranges (27); when such a systematic weak- 

 ness is identified, specific training to correct 

 the error can be introduced. Studies of 

 operational difficulty (e.g., studies of rela- 

 tive difficulty of code characters (41)) show 

 how much time should be spent on each re- 

 sponse in the training course. Sometimes 

 attention is drawn to aptitude factors to be 

 considered in future selection. Occasion- 

 ally, such studies lead to recommendations 

 for modifying gear so that errors will be 

 reduced. 



Error analysis requires effective methods 

 of evaluating performance. In all training 

 schools, one of the major contributions of 

 psychologists was their assistance in de- 

 veloping ways of measuring learning, by 

 written tests, performance tests, and product 

 ratings (44), In making these tests, they 

 drew on every known principle of test design 

 and developed a few new ones. The im- 

 portance of these tests as a means of im- 

 proving training cannot be overestimated. 

 They are a sufficient reason for encouraging 

 research on test design. 



It is difficult to point out types of research 

 needed in this area, since testing methods are 

 somewhat specific to particular situations. 

 Performance tests are especially important, 

 however, since they have received less atten- 

 tion than their validity warrants. How a 



man actually does a job can be evaluated by 

 an observer, by automatic scoring devices 

 such as a phototube indicator of time on 

 target, by motion pictures of his perfor- 

 mance, and by other methods. It would be 

 helpful to have a comparison of methods, 

 such as the study of checksight scoring 

 methods (6), as a guide to future perfor- 

 mance measures. 



SUMMARY 



Selection research has increased the ef- 

 fectiveness of military forces by placing 

 superior men in responsible posts. Selection 

 and placement of men can be improved by 

 mathematical studies of mass processing, of 

 the multiple cutoff method, and of factor 

 analysis. Billets should be studied in terms 

 of their human requirements, and the signifi- 

 cant human aptitudes should be classified. 

 Test design should be deliberate, based on 

 principles of measurement, rather than a 

 trial-and-error process. Basic studies of test 

 design are much needed. Improved techni- 

 cal methods of classifying men must be sup- 

 plemented by careful use of the methods. 



A frontier barely explored is personality 

 assessment. Tests of personality can be im- 

 proved. Studies should consider how nor- 

 mal men function, rather than identifying 

 potential breakdown cases only. 



Research has been handicapped by in- 

 adequate criteria of proficiency. Procedures 

 for judging the effectiveness of men on the 

 job can be developed. 



Training has frequently been bettered by 

 criticism in the light of psychological prin- 

 ciples. Specific studies of local training pro- 

 grams have led to alterations of course con- 

 tent, length, or methods. Further studies 

 of specific courses are advisable, but re- 

 search setting up broad principles is essen- 

 tial. Time allotments, explanation pro- 

 cedures, practice conditions, and the design 

 of training aids are promising subjects for 

 research. 



