عملکرد نظارتی روحیه جامعیت اتوبیوگرافیک مدیریت شده توسط عزت نفس

This study investigated whether self-esteem (SE) moderates the mood regulation function of autobiographical memory. Music was used to induce a negative mood in participants. Mood was then assessed following one of three tasks: (1) free recall participants were not constrained to retrieve memories of a particular valence; (2) directed recall participants were told to retrieve positive memories; and (3) control participants completed arithmetic problems. It was found that high SE participants recalled more positive memories than did low SE participants in the free recall condition. Consequently, the high SE participants experienced a greater elevation in mood. Memory positivity was equivalent for the high and low SE participants in the directed recall task. Moreover, both high and low SE participants showed an improvement in mood following directed recall of positive memories, although this improvement was less pronounced in the low SE group. No mood change was found in the control condition. The results suggest that individual differences in SE moderate mood regulation by autobiographical recall.

The relationship between mood states and memory processes is at the centre of many cognitive models of affect (e.g. Bower, 1981 and Ellis and Ashbrook, 1989) and of cognitive models of depression in particular (Blaney, 1986). These models assume that the relationship between affective states and memory is reciprocal. Consistent with that assumption is the mood congruency effect whereby affectively valenced material is more likely to be attended to and/or recalled when one is in a mood consistent with that content (Blaney, 1986). Although many studies have reported mood congruency effects (e.g. Bower et al., 1981, Clark and Teasdale, 1982, McDowall, 1984, Teasdale and Talyor, 1981 and Teasdale et al., 1980), several reviews on this topic have noted an asymmetry in the influence of positive and negative moods on memory and behaviour (e.g. Blaney, 1986 and Isen, 1984), such that mood congruency effects have been found more consistently for positive than for negative moods. To account for this asymmetry, Clark and Isen (1982) proposed a mood regulation hypothesis. They suggested that negative moods do not consistently lead to mood congruency effects because individuals may be motivated to regulate aversive mood states. The mood regulation hypothesis assumes that people employ a variety of strategies to regulate their moods, one of which is the recall of mood incongruent information. Recent research has identified several factors that govern whether mood incongruent recall will be effective in attenuating mood states, as well as the likelihood that an individual will engage in mood regulation. These include participants' awareness of the relevance of mood to the experiment (e.g. Parrott & Sabini, 1990), the effortfulness of the task (e.g. Erber and Erber, 1994 and Erber and Tesser, 1992), and personality variables such as self-esteem (e.g. Dodgson and Wood, 1998 and Smith and Petty, 1995). The present study investigated autobiographical memory in individuals with high and low self-esteem (SE) as they experienced a negative mood. The experiment was designed as a conceptual replication of a Smith and Petty (1995) experiment in which participants, after undergoing either a neutral or negative mood induction, were required to recall autobiographical memories. Smith and Petty found that, following the neutral mood induction, high and low SE participants recalled equally positive memories, whereas in the negative mood induction, high SE participants recalled more positive memories than did low SE participants. This interaction was interpreted as evidence for attempts at mood regulation on the part of the high SE participants. However, mood was not assessed following the autobiographical recall task, so the extent to which mood incongruent recall regulated mood was not determined. In the present study a musical mood induction procedure was utilized to induce a negative mood. Music was chosen as the most appropriate method because it is more likely to avoid priming effects that may occur with other induction procedures. For example, viewing videos (Josephson et al., 1996 and Smith and Petty, 1995) may trigger the recall of particular autobiographical events. The task that followed the music mood induction varied across three conditions. In the free recall condition, participants were instructed to recall any three autobiographical memories. Given the results of Smith and Petty (1995), it was expected that high SE participants would recall more positive memories than the low SE participants. In the directed recall condition, participants were asked explicitly to “recall three positive memories” (i.e. memories that are incongruent with the induced mood). This procedure was designed to determine whether the mood regulation effect of mood incongruent recall is dependent on the spontaneous nature of that recall as opposed to merely the content of the recall. Wenzlaff (1993) suggested that simply telling depressed individuals to recall something incongruent with their mood would not result in the attenuation of mood. Rather, individuals must perceive their mood incongruent thoughts to be self-initiated so that they experience a state of dissonance between their present mood and thoughts. This state of dissonance is crucial as an impetus for resolution by a change in mood. However, when an individual is exhorted by outside influences to think incongruent thoughts, the state of dissonance would not be produced because of an available external justification for such thoughts. Therefore, it was hypothesized that, if a state of dissonance is necessary for mood change, then neither the high nor low SE participants would experience an attenuation of mood when directed to recall incongruent memories. A control condition, in which participants solved math problems, was included to determine whether it is the recall of mood incongruent thoughts in particular that leads to an attenuation of mood, or whether mood congruent thoughts are unable to persist because working memory is occupied by other material, regardless of its content (Erber & Tesser, 1992). Erber and Tesser's absorption hypothesis suggests that when people engage in a demanding task they spend less time ruminating on mood congruent thoughts. This has the effect of “absorbing” preexisting moods such that negative moods become less negative and positive moods become less positive.

3. Results 3.1. Self-esteem A median value of 30 was used to assign participants to high and low SE groups. The mean SE scores for the high and low SE groups across tasks are presented in Table 1. A 2 (SE: high vs. low) × 3 (tasks) ANOVA of the SE scores indicated that SE was equivalent across task groups, F (2, 82) < 1, and that the high and low SE values did not differ across tasks, F (2, 82) < 1. Table 1. Mean SEa scores for high and low SE groups across tasks Critical tasks Free recall Directed recall Control High SE M 33.64 34.43 33.94 S.D. 2.24 2.31 2.26 Low SE M 26.23 26.43 26.00 S.D. 3.22 2.90 2.88 a SE=Self esteem Table options 3.2. Sex The distribution of males and females across tasks and SE groups was not significantly different, χ2 (2, n=88)=0.41, χ2 (1, n=88)=2.54, P>0.05, respectively. The following analyses were initially conducted using sex as an independent variable. However, because sex did not yield a significant main effect or enter into significant interactions, it was excluded as a factor in the analyses. 3.3. PA and NA scores The PA and NA scores were recorded on three separate occasions: before the mood induction (BMI), after the mood induction (AMI), and after the critical tasks (ACT). Initial analyses indicated that the NA scores, although moving in the expected direction, did not yield significant results. Therefore, an attempt was made to gain a more sensitive overall index of the participants' mood, by using the ratio of PA to NA (PA/NA). The result was three ratio scores (BMI, AMI, ACT) for each participant, with larger scores representing a more positive mood (see Table 2). Table 2. Mean ratio Scores (PA/NA) at three occasions for high and low SE groups across tasksa Tasks Mood rating occasions n BMI AMI ACT High SE Free recall M 14 2.44 1.95 2.39 S.D. 0.65 0.62 0.73 Directed recall M 14 2.50 1.99 2.78 S.D. 0.72 0.79 0.51 Control M 18 2.39 1.96 2.12 S.D. 0.87 1.00 1.12 Low SE Free recall M 13 1.46 1.19 1.22 S.D. 0.48 0.69 0.60 Directed recall M 14 1.38 1.00 1.43 S.D. 0.69 0.62 0.75 Control M 15 1.84 1.54 1.64 S.D. 0.95 0.54 0.68 a Higher means indicate more positive moods. SE= Self esteem; BMI=before mood induction; AMI=after mood induction; ACT=after critical task. Table options 3.4. Mood manipulation check To determine whether the music induced the expected decrease in mood, a 2 (SE: high vs. low) X 3 (tasks) X 2 (time: BMI vs. AMI) mixed ANOVA, with time as the repeated measures variable, was performed. The main effect of time, F (1, 82)=32.37, P< 0.001, indicated that mood decreased significantly following the musical mood induction (M BMI=2.02 vs. M AMI=1.63). Time did not interact with SE, F (1, 82)=1.35, P>0.05, or task, F (2, 82) < 1, indicating that the mood induction procedure was equally effective in decreasing mood for all groups. 3.5. Mood following the critical tasks The change in mood following the critical tasks was examined using a 2 (SE: high vs. low) X 3 (tasks) X 2 (time: AMI vs. ACT) mixed ANOVA with time as the repeated measures variable. A main effect of SE, F (1, 82)=31.64, P< 0.001, revealed that low SE participants were significantly lower in mood than were the high SE participants. The effect of SE did not interact with task, F (2, 82)=2.11, P>0.05. Most importantly, the analysis indicated that time had a significant effect, F (1, 82)=35.46, P< 0.001, such that mood was higher following the critical tasks (M AMI=1.63 vs. M ACT=1.94). The significant interaction between time and task, F (2, 82)=7.25, P< 0.01, was explored using univariate ANOVAs. The increase in mood following the critical tasks was significant in the free recall task, F (1, 26)=4.49, P< 0.05, and in the directed recall task, F (1, 27)=34.97, P< 0.001, but not in the control condition, F (1, 32) =3.02, P>0.05. There was also a significant interaction between time and SE, F (1, 82)=6.66, P< 0.05. Univariate ANOVAs indicated that the change in mood was significant for both the low SE participants, F (1, 41)=8.99, P< 0.01, and the high SE participants, F (1, 45)=20.95, P< 0.001. However, the interaction suggests that the effect of time had a differential impact on the two SE groups, such that the change in mood was more pronounced in the high SE participants than in the low SE participants. 3.6. Positivity of memories Two independent judges, who were unaware of the participants' level of SE or the task to which they had been assigned, rated the positivity of the autobiographical memories on a seven-point scale (1=extremely negative to 7=extremely positive). The raters achieved a high degree of reliability for each memory [r (53)=0.95, for the first memory recalled; r (53)=0.96, for the second memory; r (53)=0.89, for the final memory]. Accordingly, the two sets of ratings were combined to determine mean positivity scores for each of the three memories and for the three memories overall (see Table 3). Table 3. Mean positivity ratings of memories in order of recall by high and low SEa participants Tasks Memories recalled First Second Third Overall High SE Free recall M 4.29 4.00 5.46 4.58 S.D. 2.49 2.38 2.01 1.64 Directed recall M 6.61 6.21 6.36 6.39 S.D. 0.40 1.01 0.46 0.49 Low SE Free recall M 3.19 3.69 3.12 3.33 S.D. 2.16 2.10 1.56 1.41 Directed recall M 6.36 6.29 6.11 6.25 S.D. 0.63 0.54 0.98 0.53 a Higher means indicate more positive memories. SE= Self esteem. Table options In view of the finding that mood-related effects may be more pronounced for the first memory generated, the order in which memories were written by participants was considered as a factor in the analyses (Parrott & Sabini, 1990). The positivity ratings for the memories were submitted to a 2 (SE: high vs. low) X 2 (task: free vs. directed) X 3 (memory output order) mixed ANOVA, with output order as the repeated measures variable. The positivity ratings of the memories did not differ across output order, F (2, 102) < 1, nor were the interactions between output order and task or SE significant, F (2, 102)=1.2, P>0.05, F (2, 102)=2.59, P>0.05, respectively. Memories were more positive in the directed recall task (M=6.32) than in the free recall task (M=3.98), F (1, 51)=59.92, P< 0.001. Ratings of positivity were also higher for the high SE participants (M=5.49) than for the low SE participants (M=4.85), F (1, 51)=5.20, P< 0.05. Although the interaction between task and SE was not significant, F (1, 51)=3.29, P=0.076, planned comparisons carried out to test the effect of SE on memory positivity in the free recall task revealed that high SE participants recalled significantly more positive memories (M=4.58) than did low SE participants (M=3.33), F (1, 25)=4.50, P< 0.05. In the directed recall task, the positivity ratings of the memories generated by the high SE participants (M=6.39) did not differ from those of the low SE participants (M=6.25), F (1, 26) < 1. 3.7. Relationship between SE and recall To provide an additional test of the extent to which SE was related to the positivity of the autobiographical memories, correlations were performed separately for each task. In the free recall task, a significant positive correlation was revealed between SE scores and memory positivity, r (25)=0.54, P< 0.01. The correlation between SE and memory positivity in the directed recall task was not significant, r (26)=−0.02, P> 0.05. 3.8. Relationship between mood and memory recall To assess the extent to which participants' mood scores were related to their recall, within-cell correlations between mood, both before and after the recall tasks, and memory positivity were calculated (see Table 4). The absence of significant correlations in the directed recall task is not surprising as the range of memory positivity scores was restricted. However, the absence of a significant correlation for the low SE participants in the free recall task suggests that their mood had little impact on memory recall, r (13)=0.16, P> 0.05. In addition, memory positivity had little impact on the low SE participants mood following recall, r (13)=0.47, P> 0.05. In contrast, the high SE participants' in the free recall task exhibited significant positive correlations between mood, both before and after the recall tasks, and memory positivity, r (14)=0.57, P< 0.05, r (14)=0.67, P< 0.01, respectively. Table 4. Correlations between mood before recall and mood following recall and memory positivity scores for high and low SEa participants across recall tasks Correlated variables Tasks Free recall Directed recall Low SE High SE Low SE High SE Mood before recall and memorypositivity 0.16 0.57∗ 0.06 −0.23 Memory positivity and mood following recall 0.47 0.67∗∗ 0.22 −0.02 a SE=Self esteem ∗ P< 0.05 ∗∗ P< 0.01