Change and Stability during the Third Age: Longitudinal Investigations of Self-Rated Health and Religiousness with the Terman Sample1
McCullough, Michael E., Polak, Emily L., Annual Review of Gerontology & Geriatrics
One can conceptualize the Third Age variously as a distinct age range (nominally, ages 65-79 years), a discrete developmental stage (i.e., the temporal nexus of a variety of interesting developmental challenges and opportunities), or the years following the occurrence of a distinct developmental event (i.e., reduced or terminated engagement in the world of paid work). When one asks questions about how a distinct age, a distinct stage, or a distinct developmental event influences peoples characteristics-their personality, their relationships, their health, their income, or their well-being-one inevitably confronts questions about how best to describe and measure change and stability in those characteristics.
One approach to studying change and stability in psychosocial constructs in light of the Third Age is to make use of long-term longitudinal studies that allow us to examine development within individuals as they approach, enter, and leave the Third Age. As one of the longest running longitudinal studies in history, the Terman Life Cycle Study of Children with High Ability (Terman & Oden, 1947) is a remarkable resource for studying psychosocial and social development before, during, and after the Third Age (i.e., the years from age 65-79) and can shed light on some of the unique psychosocial processes that might take place during this stage of life. Begun in 1921-1922, the Terman study comprises data from 1,528 gifted boys and girls (all the students had IQs of 135 or more) from the state of California. The average birth year for children in the original sample was 1910.
By early adulthood, the Terman participants were extremely welleducated. By 1940, approximately 99% had high school diplomas; 89% had at least some college experience, 70% had at least a bachelors degree; 45% had at least a masters degree, and 8% had one or more doctoral degrees. By 1986, about 90% of participants had been married at some point in their lives, and nearly two-thirds of those who married managed to stay married well into old age or until their spouses died. Socioeconomic status was generally high and individuals were remarkably long-lived. Nearly all of them were White and from middle-class backgrounds.
Since the sample was assembled in 1922, the members of the Terman cohort have been re-contacted for over a dozen follow-up surveys (as recently as 1999). Attrition has been remarkably low. In addition, Howard Friedman at the University of California, Riverside, and his colleagues have worked assiduously to obtain death certificates for the aging Terman participants so that the timing and causes of their deaths could be ascertained (Friedman, Tucker, Schwartz, Tomlinson-Keasey, Martin, Wingard, & Criqui, 1995). Taken together, the work of Louis Terman and his successors, along with work with the data by scholars including Friedman, as well as Leslie Martin, Carol Tomlinson-Keasey, Glen Elder, Carole Holahan, George and Caroline Vaillant, and many others, has made the Terman study a truly extraordinary resource for studying the Third Age.
The work we have done with the Terman Study that might be of interest in developing a scientific understanding of the Third Age has focused on two sets of issues related to stability and change in the Terman participants' psychosocial development. First, we have been trying to understand the longitudinal development of self-rated health over the life course. Many researchers have noted that self-rated health appears to decline with age (Idler, 1993; Roberts, 1999) but the field has not given enough attention to 1) the shape and timing of those changes or 2) individual differences in how self-rated health changes as people age. In addition, we have been trying to understand how the Terman participants' religious lives unfolded. Few researchers have tried to examine the Terman participants' religious lives and how their religious development might have influenced, or been influenced by, other aspects of their psychological, relational, and physical functioning (but see Clark, Friedman, & Martin, 1999; Holahan, Sears, & Cronbach, 1995 for some analyses involving religious variables in the Terman data set). …