190 FUTURE SELECTION OF AQUANAUTS 



that some men may report what they feel is the desired reaction or the appropriate reaction 

 rather than their true reaction. 



Quality of Sleep— An Adjustment Criterion — Also on the daily-activities checklist each 

 man was asked to indicate how well he had slept the night before. These reactions were given 

 on a five-point scale. Scores were computed within teams, since there were apparently some 

 differences in the atmosphere in the capsule between teams which may have affected quality of 

 sleep. It should be noted that this is another self-report variable. 



Times Up During the Night— An Adjustment Criterion —A fourth indicator of overall ad- 

 justment, closely correlated with quality of sleep, was the number of times a man got up dur- 

 ing the night. This variable was corrected for between-team differences for the same reason 

 as the quality-of-sleep variable. It is, of course, a self-report variable. 



Leader Rating — The team leaders were asked to rate the performance of each man in their 

 teams at the end of the Sealab experiment. These ratings were given on two 4-point scales. 

 On one scale, the leaders were asked to rate the men on overall performance as divers. On 

 the second scale, they were asked to rate the willingness of the man to perform his share or 

 more of common work. The results of these two scales were summated to produce one leader- 

 rating score. For purposes of analysis, the two leaders themselves were given the top ratings 

 available on this variable. No corrections were made for between-team differences on leader 

 ratings. One man who was rated by both leaders was given a score which was the average of 

 the two leader-rating scores. 



Teammate Choices —At the end of his 15 days in Sealab, each man was asked to name the 

 five men he would most like to have with him on a future hypothetical Sealab submersion. In- 

 structions were that men could be chosen from among the 28 Sealab aquanauts. For this vari- 

 able, each man was given a score determined in the following way; if he was chosen first by a 

 teammate, he was given 5 points; if chosen second, 4 points; if chosen third, 3 points; if chosen 

 fourth, 2 points; and if chosen fifth, 1 point. If he was not chosen, of course he was given no 

 points. The weighted choices were then added for each man, and this score was used in deter- 

 mining the favorability of teammates toward each man. It should be noted that within-team 

 choices increased from a pre- to postmeasure for all three teams. However, the choices 

 within teams were by no means unanimous, and there were still a large number of choices 

 across teams. 



PREDICTOR VARIABLES 



As indicated above, only a limited number of predictor variables have been used for this 

 preliminary analysis. Only two types of predictor variables will be reported here. They are 

 demographic characteristics and basic values as measured by the AUport-Vernon-Lindzey 

 Study of Values 



Demographic Characteristics 



Six variables of this type were used. They are age, diving experience, education, birth 

 order, family mobility, and size of home town. At first glance this may seem to be an odd mix 

 of variables; however, they were not chosen capriciously, and an explanation of their choice 

 may be in order. 



First of all, demographic characteristics have one appealing advantage over such variables 

 as interests, attitudes, personality, and the like for the purpose of developing criteria. The ad- 

 vantage is that they are objective and therefore highly reliable. A problem with measures of 

 interests and attitudes is that they are frequently of low reliability. This is particularly true 

 if a man knows that such measures will be used as selection criteria. In such circumstances 

 he may say to himself, "How will it affect my chances of being selected if I answer this ques- 

 tion 'yss' vs 'no' or 'strongly agree' vs 'moderately agree'," and answer accordingly. Even 

 without such conscious or unconscious biasing influences there is the problem that some men 

 are just more candid about themselves than others, or that they know their own thoughts and 



