Dr. Parke focused on three key aims: (1) raising issues of the past; (2) elucidating how the same challenges researchers faced in the early days of longitudinal research are the same challenges researchers face today; and (3) exploring what we should do, or what we need now, in terms of longitudinal studies.

The climate of the 1920-1930 period gave rise to longitudinal studies (approximately 12 studies). These studies aimed to (1) describe what was happening across time, and (2) trace individual differences, which can get lost or distorted with cross-sectional research. Characteristics of early longitudinal studies were their long-term duration (now conceptualized as “life span course”); their focus on multiple domains of development; and their interdisciplinary character. Studies that were started between the 1920’s and 1940’s lasted 42 years on average, and 78.6% of them are still ongoing (Friedman et al., 1994). Only 40.6% of longitudinal studies started between 1940 and 1980 are still ongoing, however, and they have lasted 9.3 years on average. In addition, early studies were more likely to follow children through later ages and to cover more domains than studies beginning in later years. One possible reason for this discrepancy is the decrease in available institutional support.

Sam Fels, a philanthropist in Philadelphia, donated money to begin the Fels Longitudinal Study in 1929. This was a multi-site study, and some data collection took place in Berkeley, CA. In order to unconfound age and cohort effects, the Fels study adopted the strategy of enrolling ten families per year, and gradually building the sample size. The staff was interdisciplinary (e.g., including pediatricians, dentists, and physical anthropologists). Measures administered included the Thematic Apperception Test (TAT), Rorschach, MMPI, Kuder Preference Record, and Stanford Binet, again reflecting that measures are “period bound.” Participants were observed in the home and at school (including nursery school and day camp). In addition, parents were interviewed, and parent-child interactions were observed. Prominent findings that emerged from the Fels study included: (1) categorizations of parenting style as autocratic, democratic, or laissez-faire (Alfred Baldwin); (2) stability of personality traits (Kagan & Moss, 1962); and (3) studies of stability and instability of IQ over time (McCall).

There were several advantages of the Fels study: (1) it was long-term; (2) it measured cohort and time of testing effects; (3) it examined multiple domains and cross-domain links; (4) it included multi-disciplinary teams; and (5) it yielded a descriptive normative profile. Features missing in the Fels study included: (1) poor sampling strategy and unknown representativeness; (2) small samples; (3) limited attention to diversity; (4) limited attention to processes driving changes; (5) limited theoretical guides and a lack of specific hypotheses; (6) a global approach was adopted, rather than a focus on specific issues; (7) measurement methods were selected without specification of the problem or question; (8) it adopted a “womb-to-tomb” approach, rather than focusing on a specific period or age; and (9) there was limited appreciation of the context or cohort effects (e.g., economic factors).

Based on early longitudinal studies such as the Fels study, we gained several historical lessons for the future: (1) how to retain participants (e.g., through continuity of staff, sending holiday cards); (2) the cohort sensitized us to the timing of measurement effects; (3) measuring multiple domains of development; (4) the value of multi-disciplinary perspectives; (5) limits of an atheoretical approach to development; (6) reminds us of the need for process-oriented approaches. The Fels study raises the question, Will there be—and should there be—another long-term project like this? One answer Dr. Parke suggested is: no. Instead, an alternative would be to conduct short-term, transition-sensitive longitudinal projects that are theory-based and process-oriented. Nested studies, and studies that identify problems and target these problem-areas, are also recommended.

Ross Parke's presentation "Historical perspective on longitudinal studies: Berkeley and Beyond" can be viewed in PDF format, using Adobe® Acrobat® Reader®.