Is dispositional emotional intelligence synonymous with personality?

Abstract

To examine the utility of dispositional models of emotional intelligence (EI), two studies (Ns = 263, 116) examined whether dispositional EI predicted psychological health independent of personality. Participants completed measures of three EI dispositions (perception, understanding, regulation), Big-Five traits, psychological well-being and emotional distress. In Study 2 participants completed the health scales a second time three months later. Results revealed that dispositional EI is related to health outcomes cross-sectionally (Study 1) and predicts changes in health over time (Study 2), after controlling for the Big-Five. These findings suggest that dispositional EI is not synonymous with personality and predicts meaningful life outcomes above and beyond the Big-Five personality traits. These findings have implications for the theoretical conceptualization and measurement of emotional intelligence.

Notes

1. However, it should be noted that ability tests have also been sharply criticized on a variety of methodological and conceptual grounds (Izard, 2001; Roberts et al., 2001; Schaie, 2001).

2. It should be acknowledged that not all theorists agree with this assertion. Numerous prominent theorists (e.g., Block, 2001; McAdams, 1992; Pervin, 1994) have sharply criticized the five-factor model arguing that it is but a rough approximation of personality, and fails to tap the processes underlying behavior or other important constituent components (e.g., motives) that make up human personality. Despite these concerns, most extant work exploring the nature of EI has included the Big-Five (e.g., Brackett & Mayer, 2003; Davies et al., 1998; Roberts et al., 2001), and it is this work that has led perceived EI to be criticized as synonymous with “personality.” Therefore, despite potential disagreements about what constitutes personality, we chose to evaluate whether EI predicts meaningful outcomes independent of the major aspects of personality to which it has been previously linked (i.e., the Big-Five traits).

3. In Study 2 we used only one measure of each EI disposition. Thus there was a possibility that the results might not be comparable to Study 1, which used a composite of multiple measures. To examine this possibility the regression analyses for Study 1 were conducted separately using each individual subscale (e.g., Clarity, Labeling) rather than the averages of the subscales. The results of these analyses were very similar to those presented in the text that used the composites. For instance, as the composite Perception predicted personal growth, so did the constituent subscales Monitoring (β = .17, p < .01) and Attention (β = .30, p < .01). As the composite Understanding predicted positive relations, so did Labeling (β = .26, p < .01) and Clarity (β = .20, p < .01), and as the composite Regulation predicted depression, so did NMR (β = −.15, p < .05) and Repair (β = −.20, p < .01). In addition, the findings emerging with the individual subscales Labeling and Monitoring and with NMR in Study 1 generally parallel findings from Study 2 using these scales. Thus, overall, the results that emerged with the individual subscales of the TMMS and MAS were the same as those emerging with the composite scales.

4. There is a possibility that significant relationships between EI and health (after controlling for the Big-Five) emerged because the emotional intelligence scales used here tap content more similar to the health scales than does the Big-Five measure. If this was the case, we would expect that the magnitude of the correlations between health and emotional intelligence would be larger than the correlations between health and the Big-Five personality variables. To evaluate this possibility we first did a visual inspection of the correlations among dispositional EI, health and the Big-Five. The correlations presented in Table 2 reveal that in Study 1, 9 of the 11 health measures (82%) have a stronger correlation with at least one of the Big-Five traits than with any of the EI variables. In Study 2, 10 of the 11 (91%) Time 1 health measures have a stronger correlation with at least one of the Big-Five traits, while 8 of the 11 (73%) Time 2 health measures have a stronger correlation with at least one of the Big-Five than with any of the EI variables (see Table 4). To examine this issue further we conducted tests comparing the magnitude of the correlations. In none of the cases described above, in which one of the EI variables was more strongly correlated with health than the Big-Five was the correlation significantly stronger for EI. Finally, in Study 1, 10 of 11 health measures (91%) had a significantly (p < .05) stronger correlation between health and at least one of the Big-Five measures than between health and at least on EI measure. In Study 2, 9 of the 11 (82%) Time 1 and 10 of the 11 (91%) of the Time 2 had at least one Big-Five measure significantly more strongly correlated with health than at least one EI measure. In sum these findings fail to support the possibility that dispositional EI is linked with health (independent of the Big-Five) simply due to greater measurement overlap with the health measures. Rather, the associations with health seem in fact to be stronger with the Big-Five than with dispositional EI, providing further weight to our findings that dispositional EI predicts health independent of personality.

5. Perception also marginally predicted several other features (e.g., autonomy, see Table 5). However, as these effects did not emerge in Study 1 (even marginally) they are not discussed further.

6. Of course this statement only applies to personality as conceived of by the five-factor model. It is possible that EI might show stronger relationships with other components of personality and that those associations might qualify somewhat the conclusions of the current research.

7. Another finding that is difficult to interpret is that while in Study 1 understanding and regulation predicted lower emotional distress (after controlling for personality), these relationships were not observed in Study 2. The magnitude of the correlations between understanding, regulation and distress were consistent from Study 1 to Study 2, however, after controlling for Time 1 distress and the Big-Five in Study 2, the relationships were no longer present. One possible explanation is that, after controlling for Time 1 distress, there was simply not enough free variance in Time 2 distress to allow for predicting from understanding and regulation.

Additional information

Notes on contributors

Tirza E. Shulman

This research was supported by a University Small Research Grant from Kansas State University. Thanks go to all the students who helped with this research: Brandi Ditch, Rondi Erickson, Nikki Fisher, Colleen Green, Jessica Koran, Patricia Marsh, Lana McCarthy, Joy Murphy, Nicki Roe and Jolyn Whitaker.