Personality traits and subjective well-being

Abstract

In a panel study, the relationship between Emotional Stability (ES), Extraversion (E) and Subjective Well-Being (SWB) was tested against questionnaire data from 264 Norwegian folk high school students. After a careful reading of recent studies concerned with relationships between personality and subjective well-being, it was hypothesized that the effect from ES on SWB indicators (Life Satisfaction, presence of Positive Affect and absence of Negative Affect) is stronger than the corresponding effect from E. Moreover, it was anticipated that if ES was controlled for, the effect from E on SWB would decrease substantially. In several multiple regression analyses, it was found that, on average, the amount of SWB variance accounted for by ES was 34%, while similar figures for E were 1%. The results are discussed with reference to cultural values and traditions.

Introduction

More than 2000 years ago, Aristotle claimed that “It is contemplation alone that yields happiness” (cited in McGill, 1967, p. 18). Despite the popularity of paraphrasing Aristotle in current happiness research, his postulate about contemplation has had no impact on the literature. Contemplation is almost never mentioned as a correlate of happiness or subjective well-being (SWB). Reviews since the days of Wilson (1967) point to such factors as self-esteem, optimism and, not least, sociability and extraversion as the primary sources, or at least correlates, of SWB. Moreover, by many authors extraversion is held to be the cardinal trait of happiness and well-being (Argyle & Lu, 1990, Argyle & Martin, 1991, Baumeister 1991, Costa & McCrae, 1980, Diener & Larsen, 1993, Hotard et al., 1989, Larsen & Ketalaar, 1989, Lu & Shih, 1997, Lu et al., 1997, Magnus et al., 1993, Myers & Diener, 1995, Pavot et al., 1990, Rusting & larsen, 1997). However, in a recent meta-analysis DeNeve and Cooper (1998) found that when personality traits were grouped according to the Five Factor Model, emotional stability (i.e. the positive pole of neuroticism) was the strongest predictor of both life satisfaction and happiness, although extraversion contributed somewhat in explaining the variance in positive affect. Similarly, studies in which both extraversion and emotional stability are included as independent variables reveal that the effect on satisfaction from emotional stability normally outweighs the effect from extraversion (David et al., 1997, DeNeve & Cooper, 1998, Hotard et al., 1989, Pavot et al., 1997, Ryan & Frederick, 1997). In other words, unless emotional stability is controlled for, there is a need for care when researchers interpret the relationship between well-being and extraversion.

The purpose of this paper is to question the status that extraversion has attained as the cardinal subjective well-being trait, and to ask whether emotional stability may not actually be a more salient personality trait in predicting SWB. SWB is a broad category of phenomena including people's emotional responses, domain satisfaction and global judgment of life satisfaction (Diener, Suh, Lucas & Smith, 1999). Quite often, SWB is presented as a three-dimensional phenomenon, comprising perceived life satisfaction, presence of positive affect and absence of negative affect (Andrews & Withney, 1976, Argyle, 1996, Diener, 1984, Diener, 1994), and all three aspects will be employed in the current article.

Personality is regarded as one of the strongest predictors of subjective well-being (for reviews, see Argyle & Martin, 1991, DeNeve & Cooper, 1998, Diener, 1996, Diener & Lucas, 1999a). Among the personality traits, extraversion is the one that has received the most theoretical and empirical attention. Already a generation ago, Warner Wilson claimed that the happy person is extraverted (Wilson, 1967). Since Wilson's seminal review, a huge body of literature has reported significant associations between extraversion and well-being (Argyle & Lu, 1990, Argyle & Martin, 1991, Baumeister 1991, Diener & Lucas, 1999a, Emmons & Diener, 1985, Fujita, 1993, Larsen & Ketalaar, 1989, Larsen & Ketalaar, 1991, Lu et al., 1997, Pavot et al., 1997, Rusting & larsen, 1997). However, a closer examination of these studies reveals that at least two issues remain ambiguous in the treatment of extraversion as the most dominant predictor of SWB.

First, the effect size of prediction from extraversion to SWB is normally quite small compared with the corresponding effect size from emotional stability. For instance, in a landmark article published in 1980, Costa and McCrae proposed a model relating positive and negative affect to the personality traits of extraversion and neuroticism respectively (Costa & McCrae, 1980). Within the framework of 1 year (four measures, 3 months apart) and likewise after 10 years, neuroticism correlated highly with negative affect, whereas extraversion correlated more moderately with positive affect. In both studies, the correlation coefficients were about 0.40 between neuroticism and negative affect and about 0.20 between extraversion and positive affect. Despite the fact that extraversion accounts for only about 4% of positive affect in a 10-year time span, and despite the fact that neuroticism predicts both positive and negative affect, the Costa and McCrae study is no doubt the single most widely cited support for the hypothesis that extraversion predicts positive affect.

Suh, Diener and Fujita (1996), however, presented data showing that neuroticism, like extraversion, predicts both negative and positive affect. Contrary to common belief, correlations between extraversion and negative affect in this study were −0.40 and −0.39 (for two separate measures collected 2 years apart) and between neuroticism and positive affect −0.40 and −0.21. The correlation between extraversion and positive affect was 0.41 (0.33) and between neuroticism and negative affect it was 0.69 (0.47) (Time-2 figures in parentheses). Furthermore, using the Affectometer (Kammann & Flett, 1983), Kammann and his colleagues did not find any relationship between extraversion and either positive or negative affect (Kammann, Farry & Herbison, 1984). They did, however, detect a correlation between neuroticism and both positive and negative affect. In a study among Norwegian university students, no relationship was found between extraversion and either SWB or positive affect (Wallentin, 1996). No measure of neuroticism was available in this investigation. In an Australian study, the relationship between extraversion and positive affect was 0.15, while the correlation between neuroticism and negative affect was 0.32 (Headey & Wearing, 1989). Again, neuroticism explains more than 5 times as much variance in the affect measures as extraversion does. Using multiple regression, Brebner, Donaldson, Kirby and Ward (1995) found that for several measures of well-being and happiness, neuroticism (as measured from the Eysenck Personality Questionnaire) explained about three times as much variance than did extraversion. With regard to optimism, Brebner and his coworkers discovered that extraversion did not predict any of the Life Orientation Test (LOT, see Scheier & Carver, 1985) variance at all, while neuroticism accounted for about 22% of it. Finally, interaction studies conducted by McFatter (1994) makes the author conclude that “…the common claim that extraversion is related to measures of PA but not NA, whereas neuroticism is related to measures of NA but not PA, is probably misleading” (p. 577).

A second anomaly within the field relates to the problem of confounding variance. Because extraversion and emotional stability are normally found to be correlated, and particularly so within the Five Factor Model (FFM) approach (e.g. David et al., 1997, Diener et al., 1995, Francis et al., 1991, Lu et al., 1997, Magnus et al., 1993, Stelmack, 1981, Suh et al., 1996) it is challenging to sort out the unique contribution from extraversion in subjective well-being. Indeed, in studies controlling for emotional stability the effect of extraversion often attenuates or disappears altogether. Thus in two studies reported by Hotard et al. (1989), although initial analysis showed a significant regression coefficient from extraversion to SWB, that effect disappeared after emotional stability was introduced into the equation. Similarly, in a recent study by Pavot and colleagues it was shown that the relation between self-congruence (i.e. the discrepancy between the real-self and the ideal-self) and the four main traits other than emotional stability dropped to non- significance when emotional stability was controlled. Emotional stability, on the other hand, showed an extraordinarily large correlation with the congruence measures (Pavot et al., 1997). In the same study, emotional stability remained a significant predictor of SWB even after partialling out congruence (but not the other way around).

In a study conducted by David and collaborators (David et al., 1997), the zero-order correlations reported between positive affect and extraversion and neuroticism were 0.26 and −0.28 respectively. However, the standardized regression weights from the two traits were reduced to 0.03 (for extraversion) and −0.20 (for neuroticism) when both were treated as independent variables in a multiple regression analysis. Positive mood was the dependent variable, and some other variables were also included as independent variables.

Finally, in a study by Ryan and Frederick (1997), the effect from extraversion to the authors' concept of subjective vitality disappeared when emotional stability was entered into the equation. Nevertheless, the authors retained extraversion for further analysis, and not emotional stability, although the data seem to show that emotional stability is much more important.