THE SELECTION AND TRAINING OF LEADERS 



479 



in each of the group dimensions were asso- 

 ciated with variations in the reported beha- 

 vior of effective leaders. Further investiga- 

 tion of such relationships is obviously 

 necessary, using methods of direct observa- 

 tion and taking into account additional 

 situational differences, not only between but 

 within groups. It is, however, too much to 

 expect extensive development along these 

 lines until more suitable concepts and meth- 

 ods have been developed for stud3ang leader- 

 ship in any particular situation. 



The Selection and Training 

 or Leaders 



The problems of selecting and training 

 leaders are, in principle at least, no different 

 from those arising with various classes of 

 technical personnel. The necessity for job 

 analysis, development of reliable perform- 

 ance criteria, and the development and vah- 

 dation of selection and training techniques, 

 is present in both cases. These general pro- 

 cedures are discussed elsewhere. Here it is 

 sufficient to consider briefly some questions 

 of their specific application to problems of 

 leadership. 



Job Analysis 



Any particular leadership job is defined by 

 the series of typical situations which it in- 

 volves, that is by the formal characteristics 

 of the group (including official specifications 

 of the leader's duties), typical informal char- 

 acteristics, and external situations typically 

 encountered, as these are relevant to the 

 leader himself. Now, the determination of 

 the relevant aspects of situations is actually 

 the objective of much of the needed research 

 previously discussed, so that adequate anal- 

 yses of leaders' jobs are at present difficult. 

 And their significant features cannot be sur- 

 mised with the same ease as in the case of a 

 technician's job, which is defined to such a 

 great extent by relatively inflexible physical 

 equipment. Even now, however, probably 

 considerably more can be done in the way 

 of rough, tentative analysis and classification 



than has been done to date. Suggestions 

 concerning the analysis of general leadership 

 functions which supplement the preceding 

 considerations of situational factors, have 

 been made by Barnard (7) and Coffin (22). 



Leadership Criteria 

 Measures of Group Effectiveness 



A leader's effectiveness is measured by the 

 contribution which he makes to group effec- 

 tiveness. Thus, if other factors contributing 

 to group effectiveness are equal for groups of 

 a given type, the various measures of group 

 effectiveness previously discussed may be- 

 come criteria of leader proficiency. These 

 other factors are rarely equal, but group 

 measures may be valuable nevertheless, if 

 treated with appropriate reservations. And, 

 imperfect as they may be, they provide the 

 ultimate criterion against which other criteria 

 must at least in theory be validated. 



In general it is desirable that criteria be 

 analytic, in the sense of providing separate 

 evaluations of performance on different as- 

 pects of a job as these are identified through 

 job analysis. If meaningful morale dimen- 

 sions can be used as a framework for analysis 

 of leadership functions, a possibility consid- 

 ered earlier, then appropriate measures of 

 group effectiveness might provide analytic 

 criteria. 



Judgments by Associates 



Judgments of a leader by his associates 

 have been used frequently in leadership 

 studies, doubtless for obvious reasons. 

 Questions arise concerning what associates to 

 use and what kinds of judgments to ask of 

 them. With respect to the first question it 

 is a common practice to secure only the 

 judgments of superiors. These are neces- 

 sary, but not always sufficient. A leader 

 must also work with subordinates, who often 

 have quite different expectations, different 

 bases for judgment, and, indeed, different 

 opinions of him. When the Research 

 Branch (46) asked, "What makes a good 



