Longitudinal research is the main road to fundamental and valid knowledge of living organisms' development. Because the longitudinal approach implies collecting data about the same subjects across time, the demands on the researchers to be accurate and scrupulous are high for several reasons. First, a mistake made at one assessment occasion can seldom be satisfactorily remedied at a later stage. Second, the complexity of information makes it necessary to treat and analyze data with special care.

The importance of issues of methodology and research strategy in longitudinal research is evident from the fact that two of the eight workshops organized by the European Science Foundation's Network on Longitudinal Studies on Individual Development have exclusively dealt with methodology. The first of these was concerned with data quality and resulted in one of the volumes that have come out of the workshop series (see page iv).

This volume emanates from the second methodology workshop, which was held at Soria Moria conference center in Oslo, Norway, April 2-5, 1989. It concentrates on issues that have to do with the intricate task of applying methodological tools which are appropriate to the problem at hand. Two overriding and related concepts are central to longitudinal research: stability and change. In this book these concepts subsume multiple aspects of the human condition: for example, depression, health development, cognitive development, behavior genetics, and adjustment problems. An array of statistical methods is introduced and discussed in relation to the content areas and the problems studied.

INTRODUCTION

In this chapter we investigate in turn the following issues.

1. We test systematically hypotheses concerning the conceptual differentiation of reliability of intelligence measures and stability of intelligence constructs as a feature of the developmental process itself.

2. We test hypotheses about longitudinal changes in structure of the intelligence constructs over time indicated by sets of observables. This is of general interest in intelligence longitudinal research, particulary concerning the differentiation/reintegration hypothesis of intelligence.

We study both problems with intelligence data from the Bonn Longitudinal Study of Aging (BOLSA).

The Bonn Longitudinal Study of Aging (BOLSA) was started in 1965 with a sample of 222 women and men born in 1890-5/1900-5. The younger cohort (1900-5) consisted of 114 persons and the older one (1890-5) of 108. Of the 114 younger subjects, 55 were women and 59 men and in the older cohort there were 49 women and 59 men. Each measurement period took five days; during this period three semistructured interviews were administered, each of which focused on different sections of our subjects' lives. Furthermore, subjects underwent a series of intelligence tests (WAIS/HAWIE, Raven test), psychomotoric, and personality tests, and were examined by a specialist for internal medicine.

The Bonn Longitudinal Study of Aging

The Bonn Longitudinal Study of Aging (BOLSA) was started in 1965 with a sample of 222 women and men born between 1890 and 1895 and between 1900 and 1905. The majority of the subjects (97%) lived in their own households in different parts of West Germany. Average length of education was 11.2 years, slightly above the average education of the German population, lower-middle-class status prevailing. In agreement with the planners of the Berkeley, Baltimore (Shock, 1984), and Duke longitudinal studies (Busse & Maddox, 1986), our design did not focus on testing specific hypotheses. Rather, we tried to assess the psychological, social, and physical situation of our subjects at each measurement point as comprehensively as possible. Each measurement session lasted 5 days; during this period, three semistructured interviews were conducted, each of which focused on different aspects of our subjects' lives. At the first interview, the current situation in all of its social, psychological, and physical aspects was emphasized, whereas the second and third interviews were directed toward the past and the future. Furthermore, subjects underwent a series of cognitive, psychomotor, and personality tests and were examined by a specialist for internal medicine. Data were collected between 1965 and 1984 in eight steps: first measurement in 1965/66, second in 1966/67, third in 1967/68, fourth in 1969/70, fifth in 1972/73, sixth in 1976/77, seventh in 1980/81, and eighth in 1983/84.