the criterion of total marine science work performed by scientist-aquanauts . 

 The total marine science score is the sum of time spent in direct marine 

 science and marine science support expressed as a percentage of total mission 

 time. This criterion seems most appropriate for a first cut of the data since 

 the expressed goal of the project was the accomplishment of marine scientific 

 work by scientists. 



Selection of Aquanauts 



Because of our interest in obtaining a wide range of individual differences on 

 criterion measures, no psychological selection of aquanauts was undertaken 

 other than superficial screening for gross psychopathological disorders. No 

 use was made of available predictive measures. Scientist-aquanauts were 

 selected on the basis of the merits of their proposed research while engineers 

 were chosen to provide a broad range of experience and professional skills. 

 Employing these selection processes resulted in enormous variability on all 

 behavioral measures, as we have illustrated earlier. 



Rationale for Use of Life History 



One of the most widely accepted truisms in psychology is that "The best pre- 

 dictor of future behavior is past behavior." Research evidence supports this; 

 for example, the best predictor of college grades is high school grades; 

 previous income predicts success in selling life insurance (Tanofsky et al., 

 1969) ; completion of high school predicts completion of service school and Navy 

 enlistment (Flag and Goffman, 1966); it is assumed that the best jet pilots 

 will be the best astronauts (Voas, 1961). These simple examples provide basic 

 support for the use of life history information in prediction. 



Our own previous research has also convinced us of the value of such informa- 

 tion. Life history items were very successful in predicting performance in 

 SEALAB II, especially in contrast with personality and interest inventory data 

 (see Radloff and Helmreich, 1968). At a more general level, theoreticians have 

 argued the potential power of life history information (see Guthrie, 1944, for 

 an especially compelling argument) . Finally, a recent conference of experienced 

 users of biographical information asserted that it is "The best single predictor 

 of future behavior where the predicted behavior is of a total or complex nature" 

 (Henry, 1966). Thus, there appears to be compelling arguments for the use of 

 life history information. The general proposition is straightforward, well 

 accepted, and documented. However, it proved to be deceptively simple in its 

 application to the research goals of project TEKTITE. 



In TEKTITE, we were attempting to predict complex criteria of performance and 

 adjustment. Our goal was to understand and explain differences among TEKTITE 

 participants in their ability to work well, get along with fellow team members, 

 complex real-life behavior; it followed that the best predictive information 

 would be a total record of prior experiences. We looked for and failed to find 

 extant measuring instruments which would yield such information. 



Forming Conceptual Predictors 



We reported in our introductory discussion that the investigator can create any 

 number of conceptual variables from the matrix of yearly responses to the Life 



VIII -51 



