It should be pointed out that the relationships which have been discussed above 

 are only those where correlations with the criterion were significant at better 

 than the 5% level. Many other variables show strong trends (often curvilinear) 

 and discriminate between subgroups of the aquanaut population. 



Comparison of Top and Bottom Aquanauts 



Another example of the discriminatory power of the variables employed is a 

 comparison of scores on predictive variables of the 10 scientist-aquanauts 

 composing the top 25% of the sample on the criterion with the scores of the 10 

 with the lowest marine science output. An analysis of variance contrasting the 

 top 25% of the sample with the bottom 25% was computed for each predictive vari- 

 able. The obtained results reinforce the argument that demographic predictors 

 can effectively differentiate between those who will perform adequately in such 

 an environment and those who will do extremely well. The two groups differed 

 significantly (with a probability less than .05) on 25% of the items while 12% 

 of the items in the variable pool discriminated between the two groups with a 

 probability less than .01. 



Many of the variables which significantly differentiated the two groups have 

 been discussed in references to overall correlations. However, a number of 

 variables which have not yet been mentioned are significant predictors of 

 extremes of performance. These include health, family mobility (as shown by 

 number of changes in hometown size), changes in family size, economic orienta- 

 tion (measured by the Allport-Vernon-Lindzey Study of Values), and birth order. 



Health during childhood and youth differentiates among aquanauts but in a counter- 

 intuitive way. The top performers had significantly poorer health than the low- 

 est 25% (F = 20.44, £<.001. Again, a more detailed analysis of life history 

 clarifies this result. Although the groups show a highly significant overall 

 difference, this is caused by large differences during early school years (ages 

 6-12), as the groups do not differ significantly in earliest childhood or in 

 adolescence. A preliminary hypothesis is that physical restriction because of 

 ill health in early school years may stimulate interest in intellectual and sci- 

 entific pursuits (perhaps a comment on the intellectual excitement engendered by 

 typical elementary education?). Later good health may be associated with 

 heightened interest in outdoor activities, perhaps as a psychological compensa- 

 tion for early restriction. We have tenatively christened this phenomenon the 

 "Teddy Roosevelt Effect" and will explore its implications more fully in subse- 

 quent research. 



Families of the high achieving 10 moved significantly more than those of the 

 lower group (F = 4.11, 2.<.05), but the top group experienced fewer changes in 

 the composition of the nuclear family (F = 4.62, £ C05) . Thus both disruptions 

 due to geographical change and disruptions or change in family composition appear 

 to be associated with performance, but in opposite directions. The low perform- 

 ing group of aquanauts scored significantly higher on the economic scale of the 

 A-V-L (F = 7.89, 2. ^01)- One interpretation of this finding is that those with 

 high economic values were more motivated by the fame and profit potential 

 involved in being an aquanaut than by scientific interest and sought more the 

 designation "aquanaut" than the opportunity to conduct research in the undersea 

 environment. 



VIII-56 



