These graphs made manifestly clear one fact which we had previously suspected: 

 a "'day" running from midnight to midnight would not do as our next unit of 

 analysis (after the first-half /last-half mission split). It is worth pointing 

 out here that the analyses reported above were indeed based on mission splits 

 --the observations were divided at the midpoint and percentages computed for 

 the various activity variables for all observations in the first-half and then 

 in the second-half of the missions. No assumptions involving days were neces- 

 sary. But a day--defined as encompassing a person's major period of wakeful- 

 ness--is a logical unit for time series analysis. Careful examination of the 

 graphs described above indicates that a "day" running from 3 a.m. to 3 a.m. 

 will, in almost all cases, satisfy the preceding definition. It has accordingly 

 been adopted as the TEKTITE Standard Day for future analyses. 



At this point, an embarrassing wealth of questions suggest themselves; hope- 

 fully, some answers will be forthcoming from future work. In the previous 

 section, for example, perturbations caused by a crew change were discussed. 

 How long these perturbations existed in various types of activity is certainly 

 a researchable question. 



The TEKTITE data indicates, as does most previous work with group processes, the 

 strong way in which groups seem to form and enforce norms for various behaviors. 

 It would be useful to know not only how early in time norms (especially work 

 norms) can be determined but what accounts for later deviations from these 

 norms. 



It is generally conceded that the greater the group cohesiveness the more each 

 member will strive to carry out the group's goals. In the case of TEKTITE II, 

 this should be evidenced by more total marine science work. It seems likely, 

 but not yet tested, that productive groups will evidence patterns of cohesive- 

 ness early, and will increase in cohesiveness over time, while less productive 

 groups will be less cohesive early and will decrease in cohesiveness. 



Clearly, questions are more numerous than answers at this point. Two conclu- 

 sions, however, do seem warranted. The first is simply that a 2-week mission 

 is not a sufficiently good analogue for long-term missions but that a 3-week 

 mission probably is. Given what we now know, it seems reasonable that findings 

 from 3-week missions will generalize to groups confined for periods of longer 

 duration. The second, rather obviously, is the fact that a study of patterns 

 of variables over time can immeasurably deepen our understanding of group proc- 

 esses and can increase the certainty of predictive statements involving the 

 objective variables discussed throughout this report. Not only should this 

 make for more precise theoretical descriptions of the phenomena, but (remember- 

 ing Kurt Lewin's dictum that "there is nothing so practical as a good theory") 

 should also allow for more precise interventions in on-going groups to achieve 

 desired outcomes. 



VIII-49 



