Convergence: An Accelerated Longitudinal Approach

Despite the many well-known advantages of the longitudinal approach, which have been described by Anderson (2) and others, Jones and Bayley (4) have pointed out that very few research centers are making a systematic use of this approach in studies of human development. The difficulty of financing and managing such projects engenders much hesitancy in both the investigators considering the task and the prospective supporting foundation or agency. The investigators need to consider carefully whether their own life plans are compatible with the experimental plan since cooperation of the group studied is difficult to maintain with frequent shifts in personnel. The large expenditure involved leads to prolonged negotiations and extensive planning periods. In addition, the magnitude of the expenditure may result in a tendency of the investigators to attempt to gather entirely too much data. This compulsion to collect measurements on all possible variables could be a function of the feeling that later results may turn up promising leads for which necessary substantiating data were not gathered in earlier periods. It may also emanate from a feeling that for such expenditures of money and effort only complete comprehensiveness is acceptable. Such a view would be consistent with a belief in the importance of basic data accumulation at a given phase in the development of a field. In any event, theory construction and precise testing of hypotheses are less frequent than data amassing operations. In addition to the above problems, long-term longitudinal studies present many methodological difficulties which restrict their utility. The necessary sampling selection involved in restricting a study to cooperative groups limits generalization of findings as Anastasi and Foley point out (I, p. 268). It must be considered also that sampling errors made in the original selection are perpetuated in the continuing study, except to the extent that loss of cases due to death, migration, and other factors, changes the fundamental characteristics of the original sample still further. Another problem is that the repeated contacts and measurements necessary in some projects may lead to concern that the observation has itself introduced important changes