History Questionnaire. An example of a general variable would be computation 

 of the mean response on hometown size or family size over the entire age range. 

 Another type of general variable would be the number of changes found on a 

 variable over a specified time period--for example, changes in health. More 

 detailed predictors can also be created by looking at responses for a limited 

 age range or by computing the direction and magnitude of differences between 

 one age range and another- -an example of the latter would be a variable formed 

 by taking the signed difference in religious activity between ages 6-12 and 

 12-18. Using program LIHAN, we can form variables to test many theoretical 

 hypotheses about the relative importance of age periods and about the effects 

 of perturbations in experience on later behavior. 



The decision to cover ages through 18 is perhaps not novel, but it is an 

 important feature of the LHQ . This age span was chosen because it is long 

 enough to give a good picture of how a person has developed. The case for the 

 importance of formative years has been well made by psychodynamic writers and 

 others. Also, for many samples which the present investigators intend to study, 

 subjects will be close to their 19th birthdays. Thus, comparable information 

 can be obtained on a complete sample using the 19th year as the stopping point. 

 Finally, experiences tend to be more similar and more structured across groups 

 and societies during childhood and adolescence, making it easier to compare life 

 patterns. Later experiences may be quite specifically related to particular 

 criteria. For example, amount of marine science training was of interest in 

 TEKTITE. Supplemental questions covering such specific later experiences can 

 and should be added to the basic core of information from the LHQ. 



Clearly, the number of conceptual variables that can be formed from the LHQ is 

 almost infinite, tempting the researcher to ever finer cuts of the data. How- 

 ever, at this stage in our research in the use of the LHQ, it has seemed wiser 

 deliberately to resist this temptation and first to concentrate on rather gross 

 conceptual variables. Accordingly, we have, on the basis of a priori assumptions, 

 created three sets of conceptual variables. 



The 39 predictors formed are listed in Table 10, which also shows their corre- 

 lations with the total marine science criterion. The first conceptual set, which 

 we have designated E or Environmental variables, consists of 17 variables which 

 appear to reflect important environmental influence on the developing individual. 

 These include such items as size of hometown, family size, parental employment, 

 and health. The statistic employed is the mean computed for the maximum range 

 of ages available. The second group, 1 or Individual variables, is composed of 

 12 predictors showing individual reactions during youth. This set includes items 

 such as school performance, clashes with authority, and financial independence. 

 Again, the statistic computed is the mean of responses during all relevant years 

 (for example, school performance covers ages 6-18). The assignment of particu- 

 lar variables to either the E or I^ category is arbitrary and is done for conveni- 

 ence in preliminary analysis. Some variables, such as work during childhood, 

 could reflect either environmental pressures, such as poverty, or individual 

 motivation. For the present analysis, this distinction is not crucial. 



The third set of variables, C or Change variables, is composed of 10 items 

 showing changes on theoretically important variables such as family size, aca- 

 demic performance, and health. Each score in this set shows the total number 

 of changes on the selected item for the entire range of ages measured. 



VIII-52 



