Personality

Together, the three main approaches to personality (trait, cognitive-behavioral, and psychodynamic) have given rise to wide-ranging definitions of personality, spurring confusion regarding the stability of personality over time. Some aspects of personality, such as traits, have been conceptualized as relatively stable over time. Other aspects have been described as malleable and changeable over the life span, including many cognitive-behavioral concepts, such as self-efficacy. To give structure to the issue of personality stability and change, especially in adulthood, Dan P. McAdams has proposed a three-tier model that separates the elements of personality that should remain stable over time from those which are likely to change. This model gained popularity during the 1990s, and by 2001 was a widely cited framework for understanding personality stability and change.

The first tier of the model is comprised of traits, which are defined as relatively enduring patterns of behavior and feelings. In the 1980s and 1990s, a trait model emerged that many consider definitive: the five-factor, or Big Five, model. Via the application of factor analysis (a statistical technique) to large pools of trait questionnaire items or descriptive adjectives, many psychologists concluded that the universe of personality traits can be subsumed under five broad dimensions. These five—extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience—generally make up the cornerstone of the first tier of McAdams's model. Variables at this level often have a physiological or genetic basis, and are hypothesized to change little, if at all, over long periods of time in adulthood.

The second tier is comprised of characteristic adaptations. This is a varied category that is made up of many different kinds of psychological constructs which in general are not as stable or biologically based as traits. Like traits, they are dispositions. However, they are more likely to change over comparable periods of time. Characteristic adaptations include such well-known constructs as coping strategies, goal orientations, defense mechanisms, self-efficacy, control beliefs, mastery, and optimism.

The third tier, life narrative, contains broad elements of personality, such as one's life story, autobiographical memory, and sense of identity. This is the level that most people think of when they contemplate their personalities. Few people see themselves in terms of traits or characteristic adaptations, but most think of their identities or their personal histories when asked to reflect on their personalities. The life narrative tier is hypothesized to change the most over time. As people live their lives, they constantly revise their autobiographies. One's sense of identity, which is often continuous over time, can be revised and modified to better fit one's life and experiences. Thus, at this level, personality should show more change over time than the other two levels (although such change has proven more difficult to assess).

This three-tier model is useful because it provides a framework for organizing the many types of personality variables that have been proposed, and because it allows one to think more clearly about which personality variables should be more stable and which should display more change. Traits should change the least, life narrative the most, and characteristic adaptations should fall in between.

However, there has been no research on stability and change at the life narrative level. Studies of characteristic adaptations have mostly examined stabilities over short periods (a year or less), mainly to establish test-retest reliability. Few studies have considered characteristic adaptations over the long term or have tried to estimate the extent to which they change or are stable.

Traits, on the other hand, have been studied extensively over the long term, with some studies following the same people for 30 years or more. These longitudinal investigations generally focus either on mean-level stability or rank-order stability. The former refers to constancy in the absolute level of a trait over a period of time, as assessed by the arithmetic mean. The latter indexes constancy over time in the relative ordering of people on a trait, and is usually assessed via the correlation coefficient (sometimes adjusted for the reliability of the measure). It gauges the extent to which people retain the same rank within the sample distribution. For example, everyone may decrease over time on a given trait, but if they all decline by roughly the same amount, then the rank ordering of people within the distribution will be preserved, hence giving rise to high rank-order stability. The higher the correlation coefficient between two points in time, the greater the level of preservation of the rank order over that period.

Mean-level stability. The literature focusing on mean-level stability generally has reported change in the Big Five traits over the course of adulthood. The means of extroversion, neuroticism, and openness usually decrease with age. The results with respect to agreeableness and conscientiousness are mixed. It is not clear whether agreeableness changes or not, because some investigations show increases; some, decreases; and others, no change. With regard to conscientiousness, most studies have documented increases, but some studies have shown no change over time.

Rank-order stability. There is a fair amount of consensus with respect to rank-order change. High correlations between measurements of personality on the same people over time means that the relative positions, or ranks, within the sample distribution remain the same. That is, if person A is higher than person B on extroversion at age 30, it is likely that A will still be higher than B at age 60 or 70. Most studies that have followed adults over long periods of time (five to thirty years) have found impressive rank-order constancy. However, these results depend on the length of the longitudinal follow-up period, as a meta-analysis of the literature on the rank-order stability of personality traits concluded (Robertson and Del Vecchio). The correlation coefficients tend to be in the range of .60 to .80 over short periods (less than seven years), but fall to .40 to .60 over longer periods of time (ten years or more). Nonetheless, even over periods of thirty years or more, investigators typically report impressive levels of rank-order constancy. Using extroversion as an example, this implies that a person who is extroverted at age thirty is likely to remain extroverted at ages forty, fifty, sixty, and beyond.

However, the aforementioned meta-analysis concluded that rank-order stability is higher within samples of older adults when compared with samples of midlife or young adults. This would suggest that personality traits become more stable as people grow older. More specifically, relative positions within the distribution appear to become more fixed as older adulthood approaches. Personality traits, therefore, may be more likely to change during youth or middle age than in older adulthood. This does not mean that personality cannot or does not change during older adulthood; it simply means that it is less likely to do so.