- Email: [email protected]
Intensity of positive and negative emotions: Explaining the association between personality and depressive symptoms
Personality and Individual Differences 54 (2013) 214–220
Contents lists available at SciVerse ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
Intensity of positive and negative emotions: Explaining the association between personality and depressive symptoms Annemiek Karreman a,⇑, Marcel A.L.M. van Assen b,1, Marrie H.J. Bekker a,2 a b
Department of Developmental and Clinical Psychology, Tilburg University, Tilburg, The Netherlands Department of Methodology and Statistics, Tilburg University, Tilburg, The Netherlands
a r t i c l e
i n f o
Article history: Received 4 April 2012 Received in revised form 21 August 2012 Accepted 27 August 2012 Available online 23 September 2012 Keywords: Personality Big Five Emotional intensity Positive emotions Negative emotions Depressive symptoms
a b s t r a c t The aim was to examine to what extent emotional intensity accounted for associations between the Big Five personality dimensions and depressive symptoms. Study 1 tested the model cross-sectionally, using survey data of 266 Dutch social science students. Study 2 experimentally examined how personality dimensions were related to emotional reactivity after exposure to various emotional stimuli. Dutch psychology students (N = 130) reported on their personality and viewed an amusing or sad film clip, after which the change in intensity of experienced positive and negative emotions was assessed. Individuals scoring higher on neuroticism generally experienced more intense negative emotions, through which they experienced a higher level of depressive symptoms. Individuals who were more agreeable experienced a lower level of depressive symptoms indirectly through higher general intensity of positive emotions. More agreeable individuals showed stronger increase in negative emotions and stronger decrease in positive emotions, though after exposure to the sad stimulus only. Although replication is needed, our results offer empirical support for a more taylor-made approach in decreasing nonclinical depressive symptoms taking into account both personality characteristics and emotion regulation. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction There is a need to explain individual differences in nonclinical depressive symptoms as these symptoms are common in society (Lee & Guajardo, 2011). Beck’s (2008) influential cognitive model of depression has offered insight into the effects of information processing and cognitive bias on depressive symptoms. This model stresses the importance of individuals’ vulnerability to mild stressful events, expressed in daily cognitive-emotional reactions. However, why certain individuals are more reactive to daily events than others has remained unexplained. The five-factor model (FFM) can contribute to such an explanation, as the biological bases of personality dimensions have been considered to underlie vulnerability to stress and depressive symptoms (Costa, Bagby, Herbst, & McCrae, 2005). ⇑ Corresponding author. Address: Department of Developmental and Clinical Psychology, Tilburg University, P.O. Box 90153, 5000 LE Tilburg, The Netherlands. Tel.: +31 13 466 8280; fax: +31 13 466 2067. E-mail addresses: [email protected] (A. Karreman), [email protected] (M.A.L.M. van Assen), [email protected] (M.H.J. Bekker). 1 Address: Department of Methodology and Statistics, Tilburg University, P.O. Box 90153, 5000 LE Tilburg, The Netherlands. Tel.: +31 13 466 2362; fax: +31 13 466 3002. 2 Address: Department of Developmental and Clinical Psychology, Tilburg University, P.O. Box 90153, 5000 LE Tilburg, The Netherlands. Tel.: +31 13 466 2366/2167; fax: +31 13 466 2067. 0191-8869/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.paid.2012.08.040
Of the Big Five personality dimensions, particularly neuroticism has been associated with depression-proneness (Bagby, Joffe, Parker, Kalemba, & Harkness, 1995; Goodwin & Gotlib, 2004). Lower extraversion, and in some studies lower agreeableness, higher openness and lower conscientiousness, have also been associated with depressive symptoms (e.g., Chioqueta & Stiles, 2005; Harkness, Bagby, Joffe, & Levitt, 2002). The association between personality and depressive symptoms may be explained by the intensity of emotions individuals experience. Studies have found that high intensity of negative emotions and low intensity of positive emotions were linked with depressive symptoms (Lee & Guajardo, 2011). A consistent positive association has been found between neuroticism and intensity of negative emotions, whereas contradictory results (i.e., negative, statistically non-significant, and positive associations) have been reported for the relation between neuroticism and intensity of positive emotions (Bachorowski & Braaten, 1994; Lee & Guajardo, 2011; McFatter, 1998). Individuals high on extraversion have been found to report high intensity of positive emotions, and negative emotions (Bachorowski & Braaten, 1994; McFatter, 1998). Furthermore, McCrae and Costa (1991) found that openness, agreeableness and conscientiousness were moderately related to high intensity of positive emotions. Finally, Lee and Guajardo (2011) found that not affect intensity but difficulty in identifying feelings mediated between neuroticism and depressive symptoms. In sum, most evidence suggests (i) relations
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
between neuroticism and intensity of negative emotions, (ii) relations between the other personality dimensions and intensity of positive emotions, whereas (iii) the extent to which emotional intensity accounts for the association between personality dimensions and depressive symptoms has hardly been examined. To understand emotional experience, also changes in emotional intensity in response to an emotional event, in other words emotional reactivity, should be examined. Larsen and Ketelaar (1991), who examined emotional intensity, showed positive associations between extraversion and intensity of positive emotions after positive affect induction and between neuroticism and intensity of negative emotions after negative affect induction. Experimental studies that examined emotional reactivity after emotion induction in depressive individuals have found the opposite effect; i.e., flattened emotional responding (‘emotion context insensitivity’; see Rottenberg, 2005). Rottenberg (2005) argues more research is needed examining if individuals high on neuroticism show emotion context insensitivity. Two studies were conducted to systematically examine the mechanisms through which personality, emotional intensity, and depressive symptoms are associated. First, we examined if intensity of positive and negative emotions accounted for the associations between personality dimensions and depressive symptoms. Second, we investigated how personality dimensions were related to emotional reactivity after exposure to sad or amusing stimuli. 2. Study 1 In Study 1, we first examined direct associations between personality dimensions and depressive symptoms, expecting depressive symptoms to be positively associated with neuroticism and negatively with extraversion. Second, we studied associations between emotional intensity and depressive symptoms, hypothesizing that intensity of negative and positive emotions would be related to higher and lower levels of depressive symptoms, respectively. Third, we examined if high intensity of negative emotions would account for the positive association between neuroticism and depressive symptoms, and high intensity of positive emotions for the negative association between extraversion and depressive symptoms. We also expected negative indirect effects of agreeableness and conscientiousness on depressive symptoms via more intense positive emotions. 2.1. Methods 2.1.1. Participants and procedure Participants were 266 Dutch social science students (167 women, 99 men). The average age was 21.93 years (SD = 2.32). Of the participants, 13.2% lived with their partner, 24.8% with parents, 36.1% with roommates, and 26% lived alone; 56% was involved in a romantic relationship. Participants voluntarily participated during a course. They read a consent form, explaining the goals and procedure, and completed questionnaires. 2.1.2. Measures The personality dimensions were measured with the Dutch 30item Big Five questionnaire (Gerris et al., 1998). Six markers (e.g., nervous) for each dimension were answered on a 7-point scale, from absolutely agree to absolutely disagree. John and Srivastava (1999) have found good reliability and construct validity for their Big Five scores. In our study, Cronbach’s alphas were .81 (neuroticism), .88 (extraversion), .78 (openness), .78 (agreeableness), and .92 (conscientiousness). The Emotional Intensity Scale-Reduced (EIS-R; Geuens & de Pelsmacker, 2002) was used to measure emotional intensity. Participants described how they usually feel
215
in situations (5-point scale, from little effect on me to extremely worried), e.g., ‘someone compliments me’ (intensity of positive emotions; 9 items; a = .72) and ‘I think about awful things that might happen’ (intensity of negative emotions; 8 items; a = .78). Geuens and de Pelsmacker (2002) have reported adequate internal consistency (alphas across subsamples ranging from .79 to .89), test–retest stability over a 2-week period (e.g., .88 and .83 for positive and negative emotions), and construct validity of their scores. The Beck Depression Inventory II (Dutch BDI-II; 21 items; van der Does, 2002), measured depressive symptoms. Participants answered how they felt last week on a 4-point scale ranging from not feeling . . . to feeling so . . . that I cannot endure it. Van der Does (2002) reported that the scores in his study were reliable and valid. Cronbach’s alpha in our study was .75. 2.1.3. Statistical analyses We tested models with Mplus 5.0 (Muthén & Muthén, 1998– 2010), using MLR estimation. Bias-corrected (BC) bootstrapping dealt with potential non-normal sampling distributions of the indirect effects. We initially tested if the multiple mediation model (Fig. 1) differed across gender, by performing a multi-group analysis. Because gender had no effect we collapsed all data. Successively, we verified if there was a total effect of each of the personality dimensions on depressive symptoms estimating the model, excluding the mediators. Then, we estimated the saturated multiple mediation model and decomposed the total effect of each Big Five dimension on depressive symptoms into the direct effect and the two indirect effects through the two mediators in explaining depressive symptoms. 2.2. Results and discussion 2.2.1. Preliminary analyses We imputed missing data (0.11%) by estimating the observation using regression analysis with the other scale items as predictors. The variable depressive symptoms was not normally distributed, but skewness and kurtosis values were acceptable (below 2 and 5 respectively, Curran, West, & Finch, 1996). Table 1 displays descriptives and correlations. Neuroticism was consistently positively associated with intensity of negative emotions and depressive symptoms. Also, a positive association with intensity of positive emotions was found. Agreeableness, in contrast with extraversion, was positively associated with intensity of positive emotions and negatively associated with depressive symptoms. Openness was negatively, and conscientiousness was positively, associated with intensity of negative emotions. 2.2.2. Explaining associations with emotional intensity Figure 1 shows the significant direct paths of the final tested model. The personality dimensions explained 9.4% of the variance in depressive symptoms (small-medium effect size; Cohen, 1988). Emotional intensity additionally explained 11.8% (medium-strong effect). Totally, the model explained 21.2% of the depressive symptoms variance. The personality dimensions explained 10.3% of the variance in intensity of positive emotions (medium-strong effect), and 33.7% of the variance in intensity of negative emotions (strong effect). Table 2’s last row depicts the total effects of the personality dimensions on depressive symptoms. Neuroticism and extraversion positively and agreeableness negatively predicted depressive symptoms. Together with the direct (Table 2, first row) and indirect effects (second and third row) the extent to which emotional intensity explains the associations between personality dimensions and depression can be examined. Emotional intensity completely accounted for the positive association between neuroticism and depressive symptoms, mostly
216
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
Neuroticism
.31(.04) ***
.09(.03) **
.36(.30)
.04(.03)
Extraversion
Positive emotions
-2.75(.81) **
.02(.03) .65(.29) ** .01(.03)
Openness
Depressive symptoms
.64(.28)* -.10(.03)
**
-1.25(.55) * 2.91(.57)***
.18(.04) ***
Agreeableness
Negative emotions
.10(.06) .02(.02)
.08(.03) **
-.09(.20)
Conscientiousness Fig. 1. Significant direct paths of the final tested mediation model of Study 1. Note: Results reflect B(SE) values. Covariances were included in the model. ⁄p < .05, ⁄⁄⁄ p < .001.
⁄⁄
p < .01,
Table 1 Correlations and descriptives of Study 1. 1 1. 2. 3. 4. 5. 6. 7. 8.
Neuroticism Extraversion Openness Agreeableness Conscientiousness Positive emotions Negative emotions Depressive symptoms
M SD
2
– .31*** .03 .13* .17** .15* .52*** .22***
3
4
5
6
7
8
–
3.54 .97
.17** .14* .21** .06 .17** .05
– .21** .14* .09 .14* .05
.13* .24*** .03 .18**
– .10 .28*** .00
.33*** .14*
– .29***
–
5.00 .98
4.37 .90
5.62 .55
4.54 1.21
3.66 .41
2.99 .58
5.47 4.31
– –
*
p < .05. p < .01. *** p < .001. **
Table 2 Effects of the final tested mediation model of Study 1. Neuroticism
Direct effect Indirect effect via positive emotions Indirect effect via negative emotions Total effects
Extraversion
Openness
Agreeableness
Conscientiousness
B
SE
B
SE
B
SE
B
SE
B
SE
.36 .23* .91*** 1.04***
.30 .11 .21 .29
.65* -.10 .07 .62*
.29 .08 .10 .32
.64* -.04 .29* .31
.28 .08 .11 .30
1.25* .49* .30 1.43*
.55 .20 .18 .60
.09 .06 .23* .09
.20 .06 .10 .21
Note: Total effects were derived from a model including Big Five and depressive symptoms only. p < .05. p < .01. *** p < .001. *
⁄⁄
intensity of negative emotions, but also intensity of positive emotions. Emotional intensity did not explain the positive association between extraversion and depressive symptoms. The negative effect of agreeableness on depressive symptoms was partly explained by higher intensity of positive emotions. No total effect was observed for openness and conscientiousness. However, inconsistent effects were found for openness; a negative indirect effect via intensity of negative emotions counterbalanced a positive direct effect of openness on depressive symptoms. For conscientiousness we observed an isolated positive indirect effect. We conclude that most associations were found with neuroticism,
extraversion and agreeableness. Associations were as anticipated, except for extraversion, which was related to more depressive symptoms. 3. Study 2 The second study examined associations between personality and emotional reactivity after exposure to sad or amusing stimuli. While Study 1 measured general emotional intensity across different situations (Bachorowski & Braaten, 1994), we now asked for ‘state’ emotions in a specific context. Based on the theory of indi-
217
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
viduals’ vulnerability to stress (Costa et al., 2005), and studies linking neuroticism to negative emotions and the other personality dimensions to positive emotions (e.g., McCrae & Costa, 1991; McFatter, 1998), we hypothesized that neuroticism would be positively related to intensity of negative emotions, and extraversion, openness, agreeableness, and conscientiousness to intensity of positive emotions. We explored if the personality dimensions would be associated with emotional reactivity when viewing a film clip. 3.1. Methods 3.1.1. Participants and procedure Participants were 130 Dutch undergraduate psychology students at Tilburg University (101 women, 29 men), with age varying between 17 and 40 years (M = 20.12, SD = 2.89). As part of a larger research project we randomly assigned participants to one of three conditions—amusement, neutral or sadness—using the amusement (n = 65) and sadness (n = 65) conditions for this study’s purpose. Students voluntarily participated and received course credits. The participant was informed that (s)he would view a film clip and complete questionnaires, and that (s)he would be videotaped. Questions on demography and self-reported emotions (baseline rating) were completed on the computer. The instruction of the film clip was shown (‘‘You will be shown a film clip, just sit down and watch. After the clip we will ask you how you liked it’’), followed by a white screen presenting the word loading for 1 min and the targeted film (Gross & Levenson, 1997). Self-reported emotions during the film were measured (film rating). Next, questionnaires were completed. An informed consent form and a confidentiality agreement were signed. 3.1.2. Film clips The sadness condition involved a clip of the film ‘‘Return to Me’’ (30 34; Rottenberg, Ray, & Gross, 2007). The amusement condition showed the film clip ‘‘Do-It-Yourself Mr. Bean’’ (30 46) from Rowan Atkinson’s Mr. Bean (based on Juckel et al., 2008). 3.1.3. Measures For the assessment of emotional intensity, participants rated to what extent they experienced emotions at the starting moment (baseline) or during the film (Ekman, 1992) on 9-point Likert-type
scales, from 1 (not at all) to 8 (very much). For both baseline and film ratings we created emotional intensity scales by averaging positive and negative emotion scores: (1) intensity of positive emotions (amusement, contentment, happiness, interest, baseline a = .85, sadness condition a = .73, amusement condition a = .88), and (2) intensity of negative emotions (anger, confusion, contempt, disgust, embarrassment, fear, pain, sadness, tension, baseline a = .87, sadness condition a = .70, amusement condition a = .81). Change scores were created by subtracting baseline ratings from film ratings. The NEO-Five Factor Inventory (Dutch NEO-FFI; Hoekstra, Ormel, & De Fruyt, 1996) was used to measure personality dimensions. Sixty items were scored using 5-point Likert-type scales, from 1 (totally disagree) to 5 (totally agree). Alphas were .87 (neuroticism), .83 (extraversion), .71 (openness), .73 (agreeableness), and .83 (conscientiousness). 3.1.4. Statistical analyses We used AMOS 19.0 to test the hypotheses with v2-tests. Four fit-indices were reported for unsaturated models; CFI, GFI, and TLI with values close to .95 indicating superior fit, and RMSEA-values less than .05 indicating good fit (Byrne, 2001). For all models we report the explained variances of the endogenous variables. The analyses were carried out on the combined data of men and women to maintain sufficient statistical power. Three different models were tested. First, we tested if the films induced the expected emotions using a saturated model including the covariance between changes in positive and negative emotions. The second model included effects of condition and the personality dimensions on positive and negative emotions at baseline, and effects of these emotions on changes in emotions (Fig. 2). It also includes covariances between the personality dimensions, of condition with the personality dimensions, and between positive and negative emotions at baseline (not shown in Fig. 2). This model is not saturated since it does not incorporate any effect of personality on emotional change. Two purposes of the second model are testing if personality dimensions are associated with emotions at baseline, and comparing its fit to the saturated model testing if all main effects of personality on the changes in emotions contribute to the prediction of these changes. Finally, the third model tested if the effects of the personality dimensions on changes in emotion would be different be-
Condition -2.28(1.85)
Neuroticism
.00(.02) .04(.01) ***
.06(.02)* .05(.02) *
Extraversion
Positive emotions baseline
.12(.13)
-.63(.09) ***
Positive emotions change
.11(.04) *
.01(.01) -.07(.03) *
Openness
-.01(.01)
.13(.12)
.03(.02) .00(.02)
Agreeableness
.02(.02)
Negative emotions baseline
-.49(.06) ***
Negative emotions change
.04(.02) *
-.02(.01)
Conscientiousness Fig. 2. Significant direct paths of the final tested mediation model of Study 1. Note: Results reflect B(SE) values. Covariances were included in the model. ⁄p < .05, ⁄⁄⁄ p < .001.
⁄⁄
p < .01,
218
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
Table 3 Correlations and descriptives of Study 2. 1 1. Neuroticism
2
–
3
.50 **
***
4
.13
5
.07
6
.00
7
.14 *
8
9
M (SD)
.15
.07
.40
.10
33.72 (8.32) 42.20 (6.28) 39.62 (5.71) 43.11 (5.88) 41.12 (6.78) 5.61 (1.31)
.04
.55***
1.81 (.84)
.15 .59***
– .19
.15 –
.24 (1.38) .27 (.55)
1.94 (1.09)
.81 (1.88)
1.04 (1.10)
–
2. Extraversion
.38
–
.02
.03
.05
.25
.18
.05
.08
3. Openness
.15
.09
–
.05
.03
.20
.00
.17
.10
4. Agreeableness
-.02
.43***
.22
–
.25*
.17
.15
.01
.05
5. Conscientiousness
.15
.31*
.04
.31*
–
.09
.10
.09
**
*
*
6. Positive emotions baseline 7. Negative emotions baseline 8. Positive emotions change 9. Negative emotions change
.13
.39
.25
.42***
.25*
.09
.13
.15 .23
.26* .34**
.20 .10
M (SD)
34.17 (7.56)
40.78 (6.85)
39.14 (6.56)
.29
.28
*
**
–
.05
.39
.17
.00
–
.34** .29*
.17 .20
.66*** .29*
43.62 (5.54)
40.28 (6.60)
5.07 (1.42)
.01 **
Note: Results for the amusement condition are presented above the diagonal, results for the sadness condition are presented below the diagonal. p < .05. p < .01. *** p < .001. *
**
tween the two conditions (interaction effect between condition and personality dimension). To limit effects of multicollinearity, interactions were tested individually. Covariances between the interaction and personality dimensions and condition were estimated as well. 3.2. Results and discussion 3.2.1. Preliminary analyses There were no missing values, and no outliers were detected. Table 3 displays descriptives and correlations. Personality dimensions and baseline intensity of negative emotions did not differ between conditions, but baseline intensity of positive emotions was higher in the amusement than in the sadness condition, t(128) = 2.25, p = .03, d = .40. Increase in positive emotions was significantly higher in the amusement (M = .24, SD = 1.38) than in the sadness condition (M = .81, SD = 1.88), t(117.45) = 3.62, p < .001, d = .67. Increase in negative emotions was higher in the sadness (M = 1.04, SD = 1.10) than in the amusement condition (M = .27, SD = .55), t(94.76) = 8.64, p < .001, d = 1.78. Strong negative correlations (from .39 to .66) were observed between baseline ratings and change of emotions (both conditions), corresponding with an expected ‘regression to the mean’ effect. Change in positive emotions was not significantly correlated with change in negative emotions. In the amusement condition, neuroticism was negatively associated with change in negative emotions; in the sadness condition, extraversion and agreeableness were negatively associated with change in positive emotions, and positively associated with change in negative emotions.
3.2.3. Effect of Big Five on emotions at baseline The model in Fig. 2 fitted the data very well, GFI = .99, CFI = 1.00, TLI = 1.03, RMSEA = 0.00, and did not fit significantly worse than the saturated model, v2(10) = 8.62, p = .57, meaning that personality dimensions were not associated with emotional changes, averaged over both conditions. Table 4 provides the estimates of the personality dimensions on the baseline emotions. The personality dimensions explained 18.7% and 11.1% of the variances of positive and negative baseline emotions, respectively. At baseline, openness and extraversion had a statistically significant positive effect on positive emotions, and neuroticism on negative emotions. The model explained 37.3% and 59.9% of changes in positive and negative emotions, respectively.
3.2.4. Interaction: effect of Big Five on change in emotions in each condition To estimate and test the interaction effect between condition and one personality dimension, we added four effects to the model represented in Fig. 2; effects of this dimension on changes in both positive and negative emotions, for both conditions. Allowing for these four effects did not lead to a statistically significant improvement of fit for effects of conscientiousness, v2(4) = 1.38, p = .85, extraversion, v2(4) = 5.66, p = .23, neuroticism, v2(4) = 2.52, p = .64, and openness, v2(4) = .83, p = .94. Thus, for these dimensions neither
Table 4 Effect of Big Five on baseline emotions. Baseline emotions
3.2.2. Effect of condition on change in emotions As expected, the change in positive emotions was higher in the amusement than in the sadness condition, B = 1.05, CR = 3.63, p < .001, and the change in negative emotions was lower in the amusement than in the sadness condition, B = 1.32, CR = 8.67, p < .001. The covariance between the errors of the changes in emotions was not statistically significant, cov = .13, CR = 1.08, p = .28. Condition explained 9.3% and 36.8% of the variance of change in positive and negative emotions, respectively, representing a medium to strong and a very strong effect.
Positive
Neuroticism Extraversion Openness Agreeableness Conscientiousness *
p < .05. p < .01. *** p < .001. **
Negative
B
SE
B
SE
.002 .060 .045 .025 .023
.016 .019** . 018* .021 .017
.040 .013 .010 .003 .020
.011*** .014 .013 .015 .013
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
an interaction effect nor an effect on change of emotions appeared in the amusement and sadness conditions. We did observe an interaction effect for agreeableness, however, v2(4) = 12.00, p = .02. A positive effect was found on change of negative emotions in the sadness condition only, CR = 2.36, p = .02. Highly agreeable individuals increased their negative emotions stronger when viewing the sad clip. We found a negative effect on change in positive emotions in the sadness condition only, CR = 2.13, p = .03. Highly agreeable individuals also decreased their positive emotions stronger when viewing the sad clip. In sum, neuroticism had the anticipated positive effect on baseline intensity of negative emotions, and extraversion and openness on baseline intensity of positive emotions. Unexpectedly, agreeableness and conscientiousness had no effect on baseline intensity of positive emotions; their effects may be present only when emotions are related to events. With respect to emotional reactivity we found that individuals high on neuroticism did not show reduced emotional responding, possibly indicating experience of prolonged negative emotions. However, the more agreeable, the more likely were respondents emotionally reactive, but only in the sadness condition. Future research should find out if effects are stronger for negative than for positive stimuli.
4. General discussion We systematically examined to what extent emotional intensity accounted for the associations between the Big Five personality dimensions and nonclinical depressive symptoms. Study 1 examined the full model using self-report measures. Study 2 examined specific emotional processes. As experimentally testing effects on depressive symptoms—reflecting enduring feelings—was impossible, Study 2 focused on the first part of the model: how personality is related to emotional reactivity after exposure to positive and negative stimuli. Intensity of negative emotions could explain the association between neuroticism and depressive symptoms. Individuals high on neuroticism generally experienced high intensity of negative emotions, which was not changed by a specific negative stimulus. Emotion context insensitivity may be found in clinically depressed individuals only (Rottenberg, 2005). Neuroticism was related to more intense positive emotions too, which may reduce nonclinical depressive symptoms by compensating negative feelings. Agreeableness predicted lower levels of depressive symptoms, which could be partly explained by higher general intensity of positive emotions. Individuals higher on agreeableness also showed a higher increase in negative emotions and a stronger decrease in positive emotions when exposed to a negative stimulus. However, we did not find associations between agreeableness and baseline ‘state’ positive emotions, neither with change in positive emotions after exposure to a positive stimulus. Although more research is needed, highly agreeable individuals may share specific negative emotions with other people, and have fewer depressive symptoms because they generally experience more positive emotions (McCrae & Costa, 1991). Additional effects were found of extraversion, openness and conscientiousness. Extraversion unexpectedly directly predicted a higher level of depressive symptoms. Individuals higher on extraversion neither consistently reported higher intensity of positive emotions, nor were found to be more emotionally reactive. One explanation may be that extraversion affects the extent to which one reports different affective states. That is, two persons having similar depressive symptoms may differ in their reported symptoms because the more extraverted person more likely reports the symptoms. Another explanation is that the energetic approach towards the world, an aspect of extraversion (McCrae & Costa,
219
1995), may be associated with a higher level of depressive symptoms because this approach increases the likelihood to experience negative events. Low extraversion may also be related to the clinical state of depression only (Bagby et al., 1995). Inconsistent results were found for openness, suggesting that aspects of openness have different effects on depressive symptoms. We found a positive effect of conscientiousness on depressive symptoms through higher intensity of negative emotions, but these findings should be interpreted carefully as replication is needed. Several study limitations should be noted. First, our samples contained college students only. We did not examine other groups, neither examined how demographics like education, SES, or ethnic and cultural background may affect our results, which could be the focus of further research. Second, since participants were assigned to either the sadness or amusement condition, it was not possible to examine within-person difference in the effects of sadness and amusement. Third, we were not able to experimentally examine effects on depressive symptoms. Ideally, emotional reactivity scores are used to predict effects on depressive symptoms prospectively (Rottenberg, Kasch, Gross, & Gotlib, 2002). Despite these limitations, our results offer empirical support for a more taylor-made approach in decreasing nonclinical depressive symptoms taking into account both personality characteristics and emotion regulation. For example, our findings indicate that by strengthening general positive emotional intensity, one may decrease depressogenic aspects, specifically in those with higher neuroticism and lower agreeableness. References Bachorowski, J., & Braaten, E. B. (1994). Emotional intensity: Measurement and theoretical implications. Personality and Individual Differences, 17, 191–199. http://dx.doi.org/10.1016/0191-8869(94)90025-6. Bagby, R. M., Joffe, R. T., Parker, J., Kalemba, V., & Harkness, K. L. (1995). Major depression and the five-factor model of personality. Journal of Personality Disorders, 9, 224–234. Beck, A. T. (2008). The evolution of the cognitive model of depression and its neurobiological correlates. The American Journal of Psychiatry, 165, 969–977. http://dx.doi.org/10.1176/appi.ajp. 2008.08050721. Byrne, B. M. (2001). Structural equation modeling with AMOS: Basic concepts, applications, and programming. Mahwah, NJ: Lawrence Erlbaum Associates. Chioqueta, A. P., & Stiles, T. C. (2005). Personality traits and the development of depression, hopelessness, and suicide ideation. Personality and Individual Differences, 38, 1283–1291. http://dx.doi.org/10.1016/j.paid.2004.08.010. Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum Associates. Costa, P. T., Jr., Bagby, R. M., Herbst, J. H., & McCrae, R. R. (2005). Personality selfreports are concurrently reliable and valid during acute depressive episodes. Journal of Affective Disorders, 89, 45–55. http://dx.doi.org/10.1016/ j.jad.2005.06.010. Curran, P. J., West, S. G., & Finch, J. F. (1996). The robustness of test statistics to nonnormality and specification error in confirmatory factor analysis. Psychological Methods, 1, 16–29. http://dx.doi.org/10.1037/1082-989X.1.1.16. Ekman, P. (1992). Are there basic emotions? Psychological Review, 99, 550–553. http://dx.doi.org/10.1037/0033-295X.99.3.550. Gerris, J. R. M., Houtmans, M. J. M., Kwaaitaal-Roosen, E. M. G., Schipper, J. C., Vermulst, A. A., & Janssens, J. M. A. M. (1998). Parents, adolescents and young adults in Dutch families: A longitudinal study. Nijmegen, NL: Institute of Family Studies. Geuens, M., & de Pelsmacker, P. (2002). Validity and reliability of scores on the reduced Emotional Intensity Scale. Educational and Psychological Measurement, 62, 299–315. http://dx.doi.org/10.1177/0013164402062002007. Goodwin, R. D., & Gotlib, I. H. (2004). Gender differences in depression: The role of personality factors. Psychiatry Research, 126, 135–142. http://dx.doi.org/ 10.1016/j.psychres.2003.12.024. Gross, J. J., & Levenson, R. W. (1997). Hiding feeling: The acute effects of inhibiting negative and positive emotion. Journal of Abnormal Psychology, 106, 95–103. http://dx.doi.org/10.1037/0021-843X.106.1.95. Harkness, K. L., Bagby, M., Joffe, R. T., & Levitt, A. (2002). Major depression, chronic minor depression, and the Five-Factor Model of personality. European Journal of Personality, 16, 271–281. http://dx.doi.org/10.1002/per.441. Hoekstra, H. A., Ormel, J., & De Fruyt, F. (1996). NEO Persoonlijkheidsvragenlijsten: NEO-PI-R: NEO-FFI. Lisse, NL: Swets Test Services. John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (2nd ed., pp. 102–138). New York, NY: Guilford.
220
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
Juckel, G., Mergl, R., Prässl, A., Mavrogiorgou, P., Witthaus, H., Möller, H. J., et al. (2008). Kinematic analysis of facial behaviour in patients with schizophrenia under emotional stimulation by films with ‘‘Mr. Bean’’. European Archives of Psychiatry and Clinical Neuroscience, 258, 186–191. http://dx.doi.org/10.1007/ s00406-007-0778-3. Larsen, R. J., & Ketelaar, T. (1991). Personality and susceptibility to positive and negative emotional states. Journal of Personality and Social Psychology, 61, 132–140. http://dx.doi.org/10.1037/0022-3514.61.1.132. Lee, S. A., & Guajardo, N. R. (2011). Affect intensity and alexithymia differentially influence the relationship between neuroticism and depressive symptomatology among college students. Personality and Individual Differences, 50, 646–650. http://dx.doi.org/10.1016/j.paid.2010.11.036. McCrae, R. R., & Costa, P. T. Jr., (1991). Adding liebe und arbeit: The full five-factor model and well-being. Personality and Social Psychology Bulletin, 17, 227–232. http://dx.doi.org/10.1177/014616729101700217. McCrae, R. R., & Costa, P. T. Jr., (1995). Trait explanations in personality psychology. European Journal of Personality, 9, 231–252. http://dx.doi.org/10.1002/ per.2410090402. McFatter, R. M. (1998). Emotional intensity: Some components and their relations to extraversion and neuroticism. Personality and Individual Differences, 24, 747–758. http://dx.doi.org/10.1016/S0191-8869(97)00220-1.
Muthén, L. K., & Muthén, B. O. (1998–2010). Mplus user’s guide (6th ed.). Los Angeles, CA: Muthén & Muthén. Rottenberg, J. (2005). Mood and emotion in major depression. Current Directions in Psychological Science, 14, 167–170. http://dx.doi.org/10.1111/j.09637214.2005.00354.x. Rottenberg, J., Kasch, K. L., Gross, J. J., & Gotlib, I. H. (2002). Sadness and amusement reactivity differentially predict concurrent and prospective functioning in major depressive disorder. Emotion, 2, 135–146. http://dx.doi.org/10.1037/15283542.2.2.135. Rottenberg, J., Ray, R. R., & Gross, J. J. (2007). Emotion elicitation using films. In J. A. Coan & J. J. B. Allen (Eds.), The handbook of emotion elicitation and assessment (pp. 9–28). New York, NY: Oxford University Press. van der Does, A. J. W. (2002). Handleiding bij de Nederlandse versie van Beck Depression Inventory (BDI-II-NL) (2nd ed.). San Antonio, TX/Lisse, NL: The Psychological Corporation/Swets Test Publisher [Manual of the Dutch version of the BDI-II].
Contents lists available at SciVerse ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
Intensity of positive and negative emotions: Explaining the association between personality and depressive symptoms Annemiek Karreman a,⇑, Marcel A.L.M. van Assen b,1, Marrie H.J. Bekker a,2 a b
Department of Developmental and Clinical Psychology, Tilburg University, Tilburg, The Netherlands Department of Methodology and Statistics, Tilburg University, Tilburg, The Netherlands
a r t i c l e
i n f o
Article history: Received 4 April 2012 Received in revised form 21 August 2012 Accepted 27 August 2012 Available online 23 September 2012 Keywords: Personality Big Five Emotional intensity Positive emotions Negative emotions Depressive symptoms
a b s t r a c t The aim was to examine to what extent emotional intensity accounted for associations between the Big Five personality dimensions and depressive symptoms. Study 1 tested the model cross-sectionally, using survey data of 266 Dutch social science students. Study 2 experimentally examined how personality dimensions were related to emotional reactivity after exposure to various emotional stimuli. Dutch psychology students (N = 130) reported on their personality and viewed an amusing or sad film clip, after which the change in intensity of experienced positive and negative emotions was assessed. Individuals scoring higher on neuroticism generally experienced more intense negative emotions, through which they experienced a higher level of depressive symptoms. Individuals who were more agreeable experienced a lower level of depressive symptoms indirectly through higher general intensity of positive emotions. More agreeable individuals showed stronger increase in negative emotions and stronger decrease in positive emotions, though after exposure to the sad stimulus only. Although replication is needed, our results offer empirical support for a more taylor-made approach in decreasing nonclinical depressive symptoms taking into account both personality characteristics and emotion regulation. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction There is a need to explain individual differences in nonclinical depressive symptoms as these symptoms are common in society (Lee & Guajardo, 2011). Beck’s (2008) influential cognitive model of depression has offered insight into the effects of information processing and cognitive bias on depressive symptoms. This model stresses the importance of individuals’ vulnerability to mild stressful events, expressed in daily cognitive-emotional reactions. However, why certain individuals are more reactive to daily events than others has remained unexplained. The five-factor model (FFM) can contribute to such an explanation, as the biological bases of personality dimensions have been considered to underlie vulnerability to stress and depressive symptoms (Costa, Bagby, Herbst, & McCrae, 2005). ⇑ Corresponding author. Address: Department of Developmental and Clinical Psychology, Tilburg University, P.O. Box 90153, 5000 LE Tilburg, The Netherlands. Tel.: +31 13 466 8280; fax: +31 13 466 2067. E-mail addresses: [email protected] (A. Karreman), [email protected] (M.A.L.M. van Assen), [email protected] (M.H.J. Bekker). 1 Address: Department of Methodology and Statistics, Tilburg University, P.O. Box 90153, 5000 LE Tilburg, The Netherlands. Tel.: +31 13 466 2362; fax: +31 13 466 3002. 2 Address: Department of Developmental and Clinical Psychology, Tilburg University, P.O. Box 90153, 5000 LE Tilburg, The Netherlands. Tel.: +31 13 466 2366/2167; fax: +31 13 466 2067. 0191-8869/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.paid.2012.08.040
Of the Big Five personality dimensions, particularly neuroticism has been associated with depression-proneness (Bagby, Joffe, Parker, Kalemba, & Harkness, 1995; Goodwin & Gotlib, 2004). Lower extraversion, and in some studies lower agreeableness, higher openness and lower conscientiousness, have also been associated with depressive symptoms (e.g., Chioqueta & Stiles, 2005; Harkness, Bagby, Joffe, & Levitt, 2002). The association between personality and depressive symptoms may be explained by the intensity of emotions individuals experience. Studies have found that high intensity of negative emotions and low intensity of positive emotions were linked with depressive symptoms (Lee & Guajardo, 2011). A consistent positive association has been found between neuroticism and intensity of negative emotions, whereas contradictory results (i.e., negative, statistically non-significant, and positive associations) have been reported for the relation between neuroticism and intensity of positive emotions (Bachorowski & Braaten, 1994; Lee & Guajardo, 2011; McFatter, 1998). Individuals high on extraversion have been found to report high intensity of positive emotions, and negative emotions (Bachorowski & Braaten, 1994; McFatter, 1998). Furthermore, McCrae and Costa (1991) found that openness, agreeableness and conscientiousness were moderately related to high intensity of positive emotions. Finally, Lee and Guajardo (2011) found that not affect intensity but difficulty in identifying feelings mediated between neuroticism and depressive symptoms. In sum, most evidence suggests (i) relations
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
between neuroticism and intensity of negative emotions, (ii) relations between the other personality dimensions and intensity of positive emotions, whereas (iii) the extent to which emotional intensity accounts for the association between personality dimensions and depressive symptoms has hardly been examined. To understand emotional experience, also changes in emotional intensity in response to an emotional event, in other words emotional reactivity, should be examined. Larsen and Ketelaar (1991), who examined emotional intensity, showed positive associations between extraversion and intensity of positive emotions after positive affect induction and between neuroticism and intensity of negative emotions after negative affect induction. Experimental studies that examined emotional reactivity after emotion induction in depressive individuals have found the opposite effect; i.e., flattened emotional responding (‘emotion context insensitivity’; see Rottenberg, 2005). Rottenberg (2005) argues more research is needed examining if individuals high on neuroticism show emotion context insensitivity. Two studies were conducted to systematically examine the mechanisms through which personality, emotional intensity, and depressive symptoms are associated. First, we examined if intensity of positive and negative emotions accounted for the associations between personality dimensions and depressive symptoms. Second, we investigated how personality dimensions were related to emotional reactivity after exposure to sad or amusing stimuli. 2. Study 1 In Study 1, we first examined direct associations between personality dimensions and depressive symptoms, expecting depressive symptoms to be positively associated with neuroticism and negatively with extraversion. Second, we studied associations between emotional intensity and depressive symptoms, hypothesizing that intensity of negative and positive emotions would be related to higher and lower levels of depressive symptoms, respectively. Third, we examined if high intensity of negative emotions would account for the positive association between neuroticism and depressive symptoms, and high intensity of positive emotions for the negative association between extraversion and depressive symptoms. We also expected negative indirect effects of agreeableness and conscientiousness on depressive symptoms via more intense positive emotions. 2.1. Methods 2.1.1. Participants and procedure Participants were 266 Dutch social science students (167 women, 99 men). The average age was 21.93 years (SD = 2.32). Of the participants, 13.2% lived with their partner, 24.8% with parents, 36.1% with roommates, and 26% lived alone; 56% was involved in a romantic relationship. Participants voluntarily participated during a course. They read a consent form, explaining the goals and procedure, and completed questionnaires. 2.1.2. Measures The personality dimensions were measured with the Dutch 30item Big Five questionnaire (Gerris et al., 1998). Six markers (e.g., nervous) for each dimension were answered on a 7-point scale, from absolutely agree to absolutely disagree. John and Srivastava (1999) have found good reliability and construct validity for their Big Five scores. In our study, Cronbach’s alphas were .81 (neuroticism), .88 (extraversion), .78 (openness), .78 (agreeableness), and .92 (conscientiousness). The Emotional Intensity Scale-Reduced (EIS-R; Geuens & de Pelsmacker, 2002) was used to measure emotional intensity. Participants described how they usually feel
215
in situations (5-point scale, from little effect on me to extremely worried), e.g., ‘someone compliments me’ (intensity of positive emotions; 9 items; a = .72) and ‘I think about awful things that might happen’ (intensity of negative emotions; 8 items; a = .78). Geuens and de Pelsmacker (2002) have reported adequate internal consistency (alphas across subsamples ranging from .79 to .89), test–retest stability over a 2-week period (e.g., .88 and .83 for positive and negative emotions), and construct validity of their scores. The Beck Depression Inventory II (Dutch BDI-II; 21 items; van der Does, 2002), measured depressive symptoms. Participants answered how they felt last week on a 4-point scale ranging from not feeling . . . to feeling so . . . that I cannot endure it. Van der Does (2002) reported that the scores in his study were reliable and valid. Cronbach’s alpha in our study was .75. 2.1.3. Statistical analyses We tested models with Mplus 5.0 (Muthén & Muthén, 1998– 2010), using MLR estimation. Bias-corrected (BC) bootstrapping dealt with potential non-normal sampling distributions of the indirect effects. We initially tested if the multiple mediation model (Fig. 1) differed across gender, by performing a multi-group analysis. Because gender had no effect we collapsed all data. Successively, we verified if there was a total effect of each of the personality dimensions on depressive symptoms estimating the model, excluding the mediators. Then, we estimated the saturated multiple mediation model and decomposed the total effect of each Big Five dimension on depressive symptoms into the direct effect and the two indirect effects through the two mediators in explaining depressive symptoms. 2.2. Results and discussion 2.2.1. Preliminary analyses We imputed missing data (0.11%) by estimating the observation using regression analysis with the other scale items as predictors. The variable depressive symptoms was not normally distributed, but skewness and kurtosis values were acceptable (below 2 and 5 respectively, Curran, West, & Finch, 1996). Table 1 displays descriptives and correlations. Neuroticism was consistently positively associated with intensity of negative emotions and depressive symptoms. Also, a positive association with intensity of positive emotions was found. Agreeableness, in contrast with extraversion, was positively associated with intensity of positive emotions and negatively associated with depressive symptoms. Openness was negatively, and conscientiousness was positively, associated with intensity of negative emotions. 2.2.2. Explaining associations with emotional intensity Figure 1 shows the significant direct paths of the final tested model. The personality dimensions explained 9.4% of the variance in depressive symptoms (small-medium effect size; Cohen, 1988). Emotional intensity additionally explained 11.8% (medium-strong effect). Totally, the model explained 21.2% of the depressive symptoms variance. The personality dimensions explained 10.3% of the variance in intensity of positive emotions (medium-strong effect), and 33.7% of the variance in intensity of negative emotions (strong effect). Table 2’s last row depicts the total effects of the personality dimensions on depressive symptoms. Neuroticism and extraversion positively and agreeableness negatively predicted depressive symptoms. Together with the direct (Table 2, first row) and indirect effects (second and third row) the extent to which emotional intensity explains the associations between personality dimensions and depression can be examined. Emotional intensity completely accounted for the positive association between neuroticism and depressive symptoms, mostly
216
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
Neuroticism
.31(.04) ***
.09(.03) **
.36(.30)
.04(.03)
Extraversion
Positive emotions
-2.75(.81) **
.02(.03) .65(.29) ** .01(.03)
Openness
Depressive symptoms
.64(.28)* -.10(.03)
**
-1.25(.55) * 2.91(.57)***
.18(.04) ***
Agreeableness
Negative emotions
.10(.06) .02(.02)
.08(.03) **
-.09(.20)
Conscientiousness Fig. 1. Significant direct paths of the final tested mediation model of Study 1. Note: Results reflect B(SE) values. Covariances were included in the model. ⁄p < .05, ⁄⁄⁄ p < .001.
⁄⁄
p < .01,
Table 1 Correlations and descriptives of Study 1. 1 1. 2. 3. 4. 5. 6. 7. 8.
Neuroticism Extraversion Openness Agreeableness Conscientiousness Positive emotions Negative emotions Depressive symptoms
M SD
2
– .31*** .03 .13* .17** .15* .52*** .22***
3
4
5
6
7
8
–
3.54 .97
.17** .14* .21** .06 .17** .05
– .21** .14* .09 .14* .05
.13* .24*** .03 .18**
– .10 .28*** .00
.33*** .14*
– .29***
–
5.00 .98
4.37 .90
5.62 .55
4.54 1.21
3.66 .41
2.99 .58
5.47 4.31
– –
*
p < .05. p < .01. *** p < .001. **
Table 2 Effects of the final tested mediation model of Study 1. Neuroticism
Direct effect Indirect effect via positive emotions Indirect effect via negative emotions Total effects
Extraversion
Openness
Agreeableness
Conscientiousness
B
SE
B
SE
B
SE
B
SE
B
SE
.36 .23* .91*** 1.04***
.30 .11 .21 .29
.65* -.10 .07 .62*
.29 .08 .10 .32
.64* -.04 .29* .31
.28 .08 .11 .30
1.25* .49* .30 1.43*
.55 .20 .18 .60
.09 .06 .23* .09
.20 .06 .10 .21
Note: Total effects were derived from a model including Big Five and depressive symptoms only. p < .05. p < .01. *** p < .001. *
⁄⁄
intensity of negative emotions, but also intensity of positive emotions. Emotional intensity did not explain the positive association between extraversion and depressive symptoms. The negative effect of agreeableness on depressive symptoms was partly explained by higher intensity of positive emotions. No total effect was observed for openness and conscientiousness. However, inconsistent effects were found for openness; a negative indirect effect via intensity of negative emotions counterbalanced a positive direct effect of openness on depressive symptoms. For conscientiousness we observed an isolated positive indirect effect. We conclude that most associations were found with neuroticism,
extraversion and agreeableness. Associations were as anticipated, except for extraversion, which was related to more depressive symptoms. 3. Study 2 The second study examined associations between personality and emotional reactivity after exposure to sad or amusing stimuli. While Study 1 measured general emotional intensity across different situations (Bachorowski & Braaten, 1994), we now asked for ‘state’ emotions in a specific context. Based on the theory of indi-
217
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
viduals’ vulnerability to stress (Costa et al., 2005), and studies linking neuroticism to negative emotions and the other personality dimensions to positive emotions (e.g., McCrae & Costa, 1991; McFatter, 1998), we hypothesized that neuroticism would be positively related to intensity of negative emotions, and extraversion, openness, agreeableness, and conscientiousness to intensity of positive emotions. We explored if the personality dimensions would be associated with emotional reactivity when viewing a film clip. 3.1. Methods 3.1.1. Participants and procedure Participants were 130 Dutch undergraduate psychology students at Tilburg University (101 women, 29 men), with age varying between 17 and 40 years (M = 20.12, SD = 2.89). As part of a larger research project we randomly assigned participants to one of three conditions—amusement, neutral or sadness—using the amusement (n = 65) and sadness (n = 65) conditions for this study’s purpose. Students voluntarily participated and received course credits. The participant was informed that (s)he would view a film clip and complete questionnaires, and that (s)he would be videotaped. Questions on demography and self-reported emotions (baseline rating) were completed on the computer. The instruction of the film clip was shown (‘‘You will be shown a film clip, just sit down and watch. After the clip we will ask you how you liked it’’), followed by a white screen presenting the word loading for 1 min and the targeted film (Gross & Levenson, 1997). Self-reported emotions during the film were measured (film rating). Next, questionnaires were completed. An informed consent form and a confidentiality agreement were signed. 3.1.2. Film clips The sadness condition involved a clip of the film ‘‘Return to Me’’ (30 34; Rottenberg, Ray, & Gross, 2007). The amusement condition showed the film clip ‘‘Do-It-Yourself Mr. Bean’’ (30 46) from Rowan Atkinson’s Mr. Bean (based on Juckel et al., 2008). 3.1.3. Measures For the assessment of emotional intensity, participants rated to what extent they experienced emotions at the starting moment (baseline) or during the film (Ekman, 1992) on 9-point Likert-type
scales, from 1 (not at all) to 8 (very much). For both baseline and film ratings we created emotional intensity scales by averaging positive and negative emotion scores: (1) intensity of positive emotions (amusement, contentment, happiness, interest, baseline a = .85, sadness condition a = .73, amusement condition a = .88), and (2) intensity of negative emotions (anger, confusion, contempt, disgust, embarrassment, fear, pain, sadness, tension, baseline a = .87, sadness condition a = .70, amusement condition a = .81). Change scores were created by subtracting baseline ratings from film ratings. The NEO-Five Factor Inventory (Dutch NEO-FFI; Hoekstra, Ormel, & De Fruyt, 1996) was used to measure personality dimensions. Sixty items were scored using 5-point Likert-type scales, from 1 (totally disagree) to 5 (totally agree). Alphas were .87 (neuroticism), .83 (extraversion), .71 (openness), .73 (agreeableness), and .83 (conscientiousness). 3.1.4. Statistical analyses We used AMOS 19.0 to test the hypotheses with v2-tests. Four fit-indices were reported for unsaturated models; CFI, GFI, and TLI with values close to .95 indicating superior fit, and RMSEA-values less than .05 indicating good fit (Byrne, 2001). For all models we report the explained variances of the endogenous variables. The analyses were carried out on the combined data of men and women to maintain sufficient statistical power. Three different models were tested. First, we tested if the films induced the expected emotions using a saturated model including the covariance between changes in positive and negative emotions. The second model included effects of condition and the personality dimensions on positive and negative emotions at baseline, and effects of these emotions on changes in emotions (Fig. 2). It also includes covariances between the personality dimensions, of condition with the personality dimensions, and between positive and negative emotions at baseline (not shown in Fig. 2). This model is not saturated since it does not incorporate any effect of personality on emotional change. Two purposes of the second model are testing if personality dimensions are associated with emotions at baseline, and comparing its fit to the saturated model testing if all main effects of personality on the changes in emotions contribute to the prediction of these changes. Finally, the third model tested if the effects of the personality dimensions on changes in emotion would be different be-
Condition -2.28(1.85)
Neuroticism
.00(.02) .04(.01) ***
.06(.02)* .05(.02) *
Extraversion
Positive emotions baseline
.12(.13)
-.63(.09) ***
Positive emotions change
.11(.04) *
.01(.01) -.07(.03) *
Openness
-.01(.01)
.13(.12)
.03(.02) .00(.02)
Agreeableness
.02(.02)
Negative emotions baseline
-.49(.06) ***
Negative emotions change
.04(.02) *
-.02(.01)
Conscientiousness Fig. 2. Significant direct paths of the final tested mediation model of Study 1. Note: Results reflect B(SE) values. Covariances were included in the model. ⁄p < .05, ⁄⁄⁄ p < .001.
⁄⁄
p < .01,
218
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
Table 3 Correlations and descriptives of Study 2. 1 1. Neuroticism
2
–
3
.50 **
***
4
.13
5
.07
6
.00
7
.14 *
8
9
M (SD)
.15
.07
.40
.10
33.72 (8.32) 42.20 (6.28) 39.62 (5.71) 43.11 (5.88) 41.12 (6.78) 5.61 (1.31)
.04
.55***
1.81 (.84)
.15 .59***
– .19
.15 –
.24 (1.38) .27 (.55)
1.94 (1.09)
.81 (1.88)
1.04 (1.10)
–
2. Extraversion
.38
–
.02
.03
.05
.25
.18
.05
.08
3. Openness
.15
.09
–
.05
.03
.20
.00
.17
.10
4. Agreeableness
-.02
.43***
.22
–
.25*
.17
.15
.01
.05
5. Conscientiousness
.15
.31*
.04
.31*
–
.09
.10
.09
**
*
*
6. Positive emotions baseline 7. Negative emotions baseline 8. Positive emotions change 9. Negative emotions change
.13
.39
.25
.42***
.25*
.09
.13
.15 .23
.26* .34**
.20 .10
M (SD)
34.17 (7.56)
40.78 (6.85)
39.14 (6.56)
.29
.28
*
**
–
.05
.39
.17
.00
–
.34** .29*
.17 .20
.66*** .29*
43.62 (5.54)
40.28 (6.60)
5.07 (1.42)
.01 **
Note: Results for the amusement condition are presented above the diagonal, results for the sadness condition are presented below the diagonal. p < .05. p < .01. *** p < .001. *
**
tween the two conditions (interaction effect between condition and personality dimension). To limit effects of multicollinearity, interactions were tested individually. Covariances between the interaction and personality dimensions and condition were estimated as well. 3.2. Results and discussion 3.2.1. Preliminary analyses There were no missing values, and no outliers were detected. Table 3 displays descriptives and correlations. Personality dimensions and baseline intensity of negative emotions did not differ between conditions, but baseline intensity of positive emotions was higher in the amusement than in the sadness condition, t(128) = 2.25, p = .03, d = .40. Increase in positive emotions was significantly higher in the amusement (M = .24, SD = 1.38) than in the sadness condition (M = .81, SD = 1.88), t(117.45) = 3.62, p < .001, d = .67. Increase in negative emotions was higher in the sadness (M = 1.04, SD = 1.10) than in the amusement condition (M = .27, SD = .55), t(94.76) = 8.64, p < .001, d = 1.78. Strong negative correlations (from .39 to .66) were observed between baseline ratings and change of emotions (both conditions), corresponding with an expected ‘regression to the mean’ effect. Change in positive emotions was not significantly correlated with change in negative emotions. In the amusement condition, neuroticism was negatively associated with change in negative emotions; in the sadness condition, extraversion and agreeableness were negatively associated with change in positive emotions, and positively associated with change in negative emotions.
3.2.3. Effect of Big Five on emotions at baseline The model in Fig. 2 fitted the data very well, GFI = .99, CFI = 1.00, TLI = 1.03, RMSEA = 0.00, and did not fit significantly worse than the saturated model, v2(10) = 8.62, p = .57, meaning that personality dimensions were not associated with emotional changes, averaged over both conditions. Table 4 provides the estimates of the personality dimensions on the baseline emotions. The personality dimensions explained 18.7% and 11.1% of the variances of positive and negative baseline emotions, respectively. At baseline, openness and extraversion had a statistically significant positive effect on positive emotions, and neuroticism on negative emotions. The model explained 37.3% and 59.9% of changes in positive and negative emotions, respectively.
3.2.4. Interaction: effect of Big Five on change in emotions in each condition To estimate and test the interaction effect between condition and one personality dimension, we added four effects to the model represented in Fig. 2; effects of this dimension on changes in both positive and negative emotions, for both conditions. Allowing for these four effects did not lead to a statistically significant improvement of fit for effects of conscientiousness, v2(4) = 1.38, p = .85, extraversion, v2(4) = 5.66, p = .23, neuroticism, v2(4) = 2.52, p = .64, and openness, v2(4) = .83, p = .94. Thus, for these dimensions neither
Table 4 Effect of Big Five on baseline emotions. Baseline emotions
3.2.2. Effect of condition on change in emotions As expected, the change in positive emotions was higher in the amusement than in the sadness condition, B = 1.05, CR = 3.63, p < .001, and the change in negative emotions was lower in the amusement than in the sadness condition, B = 1.32, CR = 8.67, p < .001. The covariance between the errors of the changes in emotions was not statistically significant, cov = .13, CR = 1.08, p = .28. Condition explained 9.3% and 36.8% of the variance of change in positive and negative emotions, respectively, representing a medium to strong and a very strong effect.
Positive
Neuroticism Extraversion Openness Agreeableness Conscientiousness *
p < .05. p < .01. *** p < .001. **
Negative
B
SE
B
SE
.002 .060 .045 .025 .023
.016 .019** . 018* .021 .017
.040 .013 .010 .003 .020
.011*** .014 .013 .015 .013
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
an interaction effect nor an effect on change of emotions appeared in the amusement and sadness conditions. We did observe an interaction effect for agreeableness, however, v2(4) = 12.00, p = .02. A positive effect was found on change of negative emotions in the sadness condition only, CR = 2.36, p = .02. Highly agreeable individuals increased their negative emotions stronger when viewing the sad clip. We found a negative effect on change in positive emotions in the sadness condition only, CR = 2.13, p = .03. Highly agreeable individuals also decreased their positive emotions stronger when viewing the sad clip. In sum, neuroticism had the anticipated positive effect on baseline intensity of negative emotions, and extraversion and openness on baseline intensity of positive emotions. Unexpectedly, agreeableness and conscientiousness had no effect on baseline intensity of positive emotions; their effects may be present only when emotions are related to events. With respect to emotional reactivity we found that individuals high on neuroticism did not show reduced emotional responding, possibly indicating experience of prolonged negative emotions. However, the more agreeable, the more likely were respondents emotionally reactive, but only in the sadness condition. Future research should find out if effects are stronger for negative than for positive stimuli.
4. General discussion We systematically examined to what extent emotional intensity accounted for the associations between the Big Five personality dimensions and nonclinical depressive symptoms. Study 1 examined the full model using self-report measures. Study 2 examined specific emotional processes. As experimentally testing effects on depressive symptoms—reflecting enduring feelings—was impossible, Study 2 focused on the first part of the model: how personality is related to emotional reactivity after exposure to positive and negative stimuli. Intensity of negative emotions could explain the association between neuroticism and depressive symptoms. Individuals high on neuroticism generally experienced high intensity of negative emotions, which was not changed by a specific negative stimulus. Emotion context insensitivity may be found in clinically depressed individuals only (Rottenberg, 2005). Neuroticism was related to more intense positive emotions too, which may reduce nonclinical depressive symptoms by compensating negative feelings. Agreeableness predicted lower levels of depressive symptoms, which could be partly explained by higher general intensity of positive emotions. Individuals higher on agreeableness also showed a higher increase in negative emotions and a stronger decrease in positive emotions when exposed to a negative stimulus. However, we did not find associations between agreeableness and baseline ‘state’ positive emotions, neither with change in positive emotions after exposure to a positive stimulus. Although more research is needed, highly agreeable individuals may share specific negative emotions with other people, and have fewer depressive symptoms because they generally experience more positive emotions (McCrae & Costa, 1991). Additional effects were found of extraversion, openness and conscientiousness. Extraversion unexpectedly directly predicted a higher level of depressive symptoms. Individuals higher on extraversion neither consistently reported higher intensity of positive emotions, nor were found to be more emotionally reactive. One explanation may be that extraversion affects the extent to which one reports different affective states. That is, two persons having similar depressive symptoms may differ in their reported symptoms because the more extraverted person more likely reports the symptoms. Another explanation is that the energetic approach towards the world, an aspect of extraversion (McCrae & Costa,
219
1995), may be associated with a higher level of depressive symptoms because this approach increases the likelihood to experience negative events. Low extraversion may also be related to the clinical state of depression only (Bagby et al., 1995). Inconsistent results were found for openness, suggesting that aspects of openness have different effects on depressive symptoms. We found a positive effect of conscientiousness on depressive symptoms through higher intensity of negative emotions, but these findings should be interpreted carefully as replication is needed. Several study limitations should be noted. First, our samples contained college students only. We did not examine other groups, neither examined how demographics like education, SES, or ethnic and cultural background may affect our results, which could be the focus of further research. Second, since participants were assigned to either the sadness or amusement condition, it was not possible to examine within-person difference in the effects of sadness and amusement. Third, we were not able to experimentally examine effects on depressive symptoms. Ideally, emotional reactivity scores are used to predict effects on depressive symptoms prospectively (Rottenberg, Kasch, Gross, & Gotlib, 2002). Despite these limitations, our results offer empirical support for a more taylor-made approach in decreasing nonclinical depressive symptoms taking into account both personality characteristics and emotion regulation. For example, our findings indicate that by strengthening general positive emotional intensity, one may decrease depressogenic aspects, specifically in those with higher neuroticism and lower agreeableness. References Bachorowski, J., & Braaten, E. B. (1994). Emotional intensity: Measurement and theoretical implications. Personality and Individual Differences, 17, 191–199. http://dx.doi.org/10.1016/0191-8869(94)90025-6. Bagby, R. M., Joffe, R. T., Parker, J., Kalemba, V., & Harkness, K. L. (1995). Major depression and the five-factor model of personality. Journal of Personality Disorders, 9, 224–234. Beck, A. T. (2008). The evolution of the cognitive model of depression and its neurobiological correlates. The American Journal of Psychiatry, 165, 969–977. http://dx.doi.org/10.1176/appi.ajp. 2008.08050721. Byrne, B. M. (2001). Structural equation modeling with AMOS: Basic concepts, applications, and programming. Mahwah, NJ: Lawrence Erlbaum Associates. Chioqueta, A. P., & Stiles, T. C. (2005). Personality traits and the development of depression, hopelessness, and suicide ideation. Personality and Individual Differences, 38, 1283–1291. http://dx.doi.org/10.1016/j.paid.2004.08.010. Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum Associates. Costa, P. T., Jr., Bagby, R. M., Herbst, J. H., & McCrae, R. R. (2005). Personality selfreports are concurrently reliable and valid during acute depressive episodes. Journal of Affective Disorders, 89, 45–55. http://dx.doi.org/10.1016/ j.jad.2005.06.010. Curran, P. J., West, S. G., & Finch, J. F. (1996). The robustness of test statistics to nonnormality and specification error in confirmatory factor analysis. Psychological Methods, 1, 16–29. http://dx.doi.org/10.1037/1082-989X.1.1.16. Ekman, P. (1992). Are there basic emotions? Psychological Review, 99, 550–553. http://dx.doi.org/10.1037/0033-295X.99.3.550. Gerris, J. R. M., Houtmans, M. J. M., Kwaaitaal-Roosen, E. M. G., Schipper, J. C., Vermulst, A. A., & Janssens, J. M. A. M. (1998). Parents, adolescents and young adults in Dutch families: A longitudinal study. Nijmegen, NL: Institute of Family Studies. Geuens, M., & de Pelsmacker, P. (2002). Validity and reliability of scores on the reduced Emotional Intensity Scale. Educational and Psychological Measurement, 62, 299–315. http://dx.doi.org/10.1177/0013164402062002007. Goodwin, R. D., & Gotlib, I. H. (2004). Gender differences in depression: The role of personality factors. Psychiatry Research, 126, 135–142. http://dx.doi.org/ 10.1016/j.psychres.2003.12.024. Gross, J. J., & Levenson, R. W. (1997). Hiding feeling: The acute effects of inhibiting negative and positive emotion. Journal of Abnormal Psychology, 106, 95–103. http://dx.doi.org/10.1037/0021-843X.106.1.95. Harkness, K. L., Bagby, M., Joffe, R. T., & Levitt, A. (2002). Major depression, chronic minor depression, and the Five-Factor Model of personality. European Journal of Personality, 16, 271–281. http://dx.doi.org/10.1002/per.441. Hoekstra, H. A., Ormel, J., & De Fruyt, F. (1996). NEO Persoonlijkheidsvragenlijsten: NEO-PI-R: NEO-FFI. Lisse, NL: Swets Test Services. John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (2nd ed., pp. 102–138). New York, NY: Guilford.
220
A. Karreman et al. / Personality and Individual Differences 54 (2013) 214–220
Juckel, G., Mergl, R., Prässl, A., Mavrogiorgou, P., Witthaus, H., Möller, H. J., et al. (2008). Kinematic analysis of facial behaviour in patients with schizophrenia under emotional stimulation by films with ‘‘Mr. Bean’’. European Archives of Psychiatry and Clinical Neuroscience, 258, 186–191. http://dx.doi.org/10.1007/ s00406-007-0778-3. Larsen, R. J., & Ketelaar, T. (1991). Personality and susceptibility to positive and negative emotional states. Journal of Personality and Social Psychology, 61, 132–140. http://dx.doi.org/10.1037/0022-3514.61.1.132. Lee, S. A., & Guajardo, N. R. (2011). Affect intensity and alexithymia differentially influence the relationship between neuroticism and depressive symptomatology among college students. Personality and Individual Differences, 50, 646–650. http://dx.doi.org/10.1016/j.paid.2010.11.036. McCrae, R. R., & Costa, P. T. Jr., (1991). Adding liebe und arbeit: The full five-factor model and well-being. Personality and Social Psychology Bulletin, 17, 227–232. http://dx.doi.org/10.1177/014616729101700217. McCrae, R. R., & Costa, P. T. Jr., (1995). Trait explanations in personality psychology. European Journal of Personality, 9, 231–252. http://dx.doi.org/10.1002/ per.2410090402. McFatter, R. M. (1998). Emotional intensity: Some components and their relations to extraversion and neuroticism. Personality and Individual Differences, 24, 747–758. http://dx.doi.org/10.1016/S0191-8869(97)00220-1.
Muthén, L. K., & Muthén, B. O. (1998–2010). Mplus user’s guide (6th ed.). Los Angeles, CA: Muthén & Muthén. Rottenberg, J. (2005). Mood and emotion in major depression. Current Directions in Psychological Science, 14, 167–170. http://dx.doi.org/10.1111/j.09637214.2005.00354.x. Rottenberg, J., Kasch, K. L., Gross, J. J., & Gotlib, I. H. (2002). Sadness and amusement reactivity differentially predict concurrent and prospective functioning in major depressive disorder. Emotion, 2, 135–146. http://dx.doi.org/10.1037/15283542.2.2.135. Rottenberg, J., Ray, R. R., & Gross, J. J. (2007). Emotion elicitation using films. In J. A. Coan & J. J. B. Allen (Eds.), The handbook of emotion elicitation and assessment (pp. 9–28). New York, NY: Oxford University Press. van der Does, A. J. W. (2002). Handleiding bij de Nederlandse versie van Beck Depression Inventory (BDI-II-NL) (2nd ed.). San Antonio, TX/Lisse, NL: The Psychological Corporation/Swets Test Publisher [Manual of the Dutch version of the BDI-II].