Personality

As valuable as they are, mean-level and rank-order methods often conceal the extent of individual differences in personality trait stability. A trait may show constancy in its mean over time and display strong rank-order consistency, but these statistics are based on groups and may mask change that occurs at the level of the person. Even if a correlation is strong between two time points, this does not mean that everyone retains the same rank order position over the time period. Some people may be changing even as others remain the same. This is essentially a question of individual differences in stability and change. Recently, researchers in this area have turned their attention to stability and change as manifested at the person level, and many are supporting the notion of personality stability and change as a phenomenon of individual differences. These investigators have argued that even if the majority of people show stability over time, not everyone does, and it is thus best to think of personality trait stability in an individual differences framework.

The idea of individual differences in stability and change is not a new one, and can be traced back to various conceptual works in life-span developmental theory, especially the writings of Paul B. Baltes and John R. Nesselroade. In particular, they and other life-span developmentalists have championed the concept of individual differences in intra-individual change. That concept combines the ideas of variability among people (individual differences) and of change that occurs within persons (intra-individual change). The latter requires the estimation of within-person change or stability. Historically, this has been difficult to study, hampered by a lack of adequate statistical techniques. However, beginning in the 1980s, such methods were developed, and by 2000 there were several approaches available to the researcher who wished to assess and analyze within-person change longitudinally.

These techniques are variously known as growth-curve modeling or random effects modeling, among other names. At their most effective, these models require that people be measured at a minimum of three time points, which allows the calculation of growth curves for each individual. The revolutionary aspect of these approaches is that the notion of individual growth curves allows conceptualization of a given trait within a given person as a personality trajectory. Previous research on personality stability and change (whether focusing on rank-order or mean level stability) typically considered only two time points (which allows calculation of only linear trajectories), yielding an impoverished representation of the possible complexity of individual differences in personality change. By the late 1990s, these shortcomings were apparent, and scholars in this area were turning their attention to the various models that allow estimation of trajectories.

The concept of the individual trajectory possesses theoretical relevance because it is the foundational metaphor of the developmental perspective. Before the advent of methods that allowed the estimation of individual trajectories, the basic building block of developmental research, including aging research, was the difference score. Difference scores are obtained by subtracting the score at one time from the score at another. However, difference scores are inherently limited as the basis for developmental analysis, and with techniques now available for modeling multiple occasions of measurement, the trajectory is replacing the difference score.

In addition, the notion of the trajectory allows scholars to link the idiographic and nomothetic traditions in the social sciences. The idiographic approach focuses on information about individual people to better understand unique persons, while the nomothetic position prefers data derived from samples of people in order to discover general laws. The tension between the two is partially resolved by the concept of the trajectory. Trajectories, such as growth curves, can be estimated for large samples but also can be calculated for individuals, allowing for both idiographic and nomothetic analyses of data. Some have argued that the ideal personality theories are those which are broadly applicable to all persons, yet can still be used to explain the behavior of any single person, in essence integrating the idiographic and nomothetic approaches. The trajectory model may prove to be such an integrating force in the area of personality and personality development in the future.

By the late 1990s, investigators had begun applying trajectories techniques to longitudinal personality data. Trajectories can be studied in the aggregate (e.g., the overall trajectory for a sample), but are most interesting when applied to the study of multiple individuals, to permit an analysis of variation among people in trajectories. The key focus, as noted by Baltes and Nesselroade, is the question of interindividual differences in intra-individual change. Persons can differ on trait trajectories in the average level (how much of the trait a person has), the rate of change, and the direction of change. In a small number of studies, investigators have shown that there are significant differences in trajectories among individuals with respect to a number of personality dimensions, including traits. Some people have trait trajectories that rise; some, that fall; and some, that stay stable. Investigators have even found evidence of significant variability in trajectories among older adults. Thus, even as the rank-order stability of personality traits becomes higher among older adults, studies have shown that the ability to change remains. Even if the majority of people remain the same, or at least retain the same relative position within the distribution, it appears that many people do not. Trajectory modeling thus has shed valuable light on the issue of personality trait stability and change. Some people remain stable; others change. Investigators in this area have begun to search for the reasons why some people change and others remain stable.