USING GROWTH CURVE ANALYSES TO ASSESS PERSONALITY CHANGE AND STABILITY IN ADULTHOOD

A. George Adler and Steven J. Scher
University-College of the Cariboo


In this chapter, we have attempted to explain how the techniques of growth curve analysis can be applied to longitudinal personality data. We began by clarifying the definitions of two types of stability, absolute stability and relative stability. The former refers to the amount of stability in the absolute level of a trait over time: the latter refers to the degree of stability in the rank order of the trait over time. After reviewing several approaches to measuring these two types of stability, we argue that several problems in the measurement of change could best be solved by adopting a different perspective on change than has usually been adopted. Rather than focusing on discrete jumps (or drops) in a measured attribute, we argue that it may be preferable to conceptualize change as a continuous process for each individual. That is, we can think of personality change as forming a continuous function in time for each individual, and questions about the degree of stability in the population are examined by aggregating the various parameters that describe these individual functions. Through simulated data, we have shown that several of the methods of analysis discussed were able to accurately capture the level of absolute stability in the data. However, of the methods used, growth curve analysis was the only one that came close to accurately indicating the level of relative stability in the data. These findings suggest that further comparative study of these methods is clearly needed. In the meantime, however, these findings lead us to suggest that we should be cautious in offering an answer to the question in the title of this book. Although a fairly large research literature supports the notion that personality is stable in adulthood (see Costa & McCrae, this volume), this literature is based largely on data analyzed with the techniques that, if our simulations are to be believed, tend to overestimate the degree of relative stability present. What conclusion, then, can we reach? We would suggest that caution be exercised when choosing a method of analysis for future research on personality change. When designing studies, it is desirable to obtain measurements on more than two occasions in order to allow the use of growth curve analysis. At the very least, data from studies of personality change should be analyzed with several of these techniques. If different techniques suggest different conclusions about the degree of personality stability, then the question must remain open. However, when we can truly feel confident about and answer to the question, Can personality change?


E-mail Steve Scher to request a reprint (cfsjs@eiu.edu)

Return To Steve Scher's Vita