Is Timing Everything? Sequential Effects of Rumination and Distraction on Interpersonal Problem Solving
Rumination has been closely linked to risk for depression, whereas distraction has been hypothesized to decrease sad mood and to promote effective problem solving. This study investigates the hypothesis that it is not the use of specific strategies but rather their timing that is critical. Following a negative mood induction, participants were assigned to either immediately ruminate or distract followed by a second set of instructions to either ruminate or distract. Participants who initially engaged in distraction, compared to rumination, generated more effective solutions to interpersonal problems even when they subsequently engaged in rumination immediately prior to the problem solving task. In contrast, participants who engaged in distraction prior to the problem solving task generated less effective solutions when distraction followed a period of rumination. Importantly this effect was not due to differences in current mood state. The results suggest that the timing of the use of emotion regulation strategies is critical.
Although self-focused rumination has been associated with the experience of chronic negative affect (e.g., Mor and Winquist 2002), clearly not all self-reflection is maladaptive. Indeed, there is debate about whether rumination is a detrimental form of emotion regulation that increases the risk of depression or whether it can be thought of as an adaptive coping mechanism that might be part of the problem-solving process (e.g., Nolen-Hoeksema et al. 2008). Martin and Tesser (1996), for example, suggest that rumination can promote problem-focused coping and can contribute to an adjustment of behavior in ways that may eventually lead to the attainment of a desired goal. Despite the fact that some theories on self-reflection suggest that rumination can be adaptive, research has focused primarily on demonstrating the detrimental effects of rumination. In recent years, however, there is a growing trend to differentiate various forms of rumination, such as brooding versus reflection and abstract/analytical rumination versus experiential mindfulness, to differentiate more adaptive forms of rumination from maladaptive ones (e.g., Treynor et al. 2003; Watkins 2004; Watkins and Teasdale 2001).
Rumination is the process of thinking perseveratively about one’s feelings and problems in response to stress (Nolen-Hoeksema et al. 2008). It leads individuals to fixate on the problems and their feelings instead of engaging in active problem solving thereby maintaining or even exacerbating negative affect. Indeed, studies have found that rumination in response to stressful life events is associated with more prolonged periods of depression and an elevated risk for the onset of depression (e.g., Nolen-Hoeksema and Morrow 1991; see Nolen-Hoeksema et al. 2008, for review). According to the response styles theory (Nolen-Hoeksema 1991), rumination is associated with risk for depression through several mechanisms. First, rumination enhances the effects of negative mood on thinking by activating mood-congruent cognitions. Dysphoric individuals who ruminate, for example, predict that negative outcomes are more likely to happen in the future (Lyubomirsky and Nolen-Hoeksema 1995) and evaluate themselves and their current situation in a negative, hopeless manner (Lyubomirsky et al. 1998). Similarly, rumination leads dysphoric individuals to remember more negative events (e.g., Lyubomirsky et al. 1998). Distraction, in contrast, is often considered a more adaptive response to negative affect because it draws one’s attention away from upsetting events and thereby helps mood repair (e.g., Joormann and Siemer 2004; Nolen-Hoeksema 1991).
Importantly, rumination disrupts problem solving (e.g., Nolen-Hoeksema 1991; Watkins and Moulds 2005). Specifically, rumination in conjunction with dysphoric mood interferes with generating good solutions for interpersonal problems (e.g., Lyubomirsky and Nolen-Hoeksema 1995). Dysphoric participants in the rumination condition, for example, were significantly impaired at interpersonal problem solving compared to dysphoric participants who distracted themselves from their mood (e.g., Lyubomirsky and Nolen-Hoeksema 1995; Lyubomirsky et al. 1999). Indeed, the detrimental effects of rumination on problem solving may represent an important link between rumination and depression. Given the high frequency of interpersonal conflicts in everyday life and their relation to the experience of depressive symptoms (Nezu and Ronan 1985), being able to solve these conflicts effectively could be crucial for well-being. Not surprisingly, poor interpersonal problem solving has been considered a core feature of depression (e.g., Nezu 1986, 2004 for a review). These findings demonstrate the importance of gaining a better understanding of the relation between rumination and problem solving specifically in the interpersonal domain.
Most studies thus far have investigated the effects of rumination and distraction separately. It is unlikely, however, that people use only one strategy when faced with stressful events and when trying to regulate the resulting emotional responses. People may engage in both rumination and distraction at different time points, and the order and timing of these different strategies may be crucial.
To the best of our knowledge, there is only one published study (Trask and Sigmon 1999) that examined sequential effects of rumination and distraction on mood. Participants in the rumination-distraction condition exhibited sustained depressed mood following rumination, but a decrease in depressed mood after engaging in distraction. Participants in the distraction-rumination condition experienced a decrease in depressed mood following distraction. Interestingly, however, these participants did not demonstrate an increase in depressed mood following subsequent rumination. The results, thus, suggest that rumination maintains negative mood only when it is the immediate response to the negative mood induction.
Trask and Sigmon (1999) argued that initially engaging in distraction should not only affect mood but also increase the likelihood of successful problem solving. If encountering a problem induces strong emotion, it may be helpful to immediately engage in distracting activities and thoughts to lift the negative mood before engaging in problem solving (Kross et al. 2005). Otherwise, mood-congruent thoughts can interfere with an objective evaluation of the situation. Eventually, however, individuals may need to engage in reflection to analyze the causes of their depressed mood or problems. As Trask and Sigmon pointed out, initial engagement in distraction can break the cycle of negative automatic thoughts, which can enable individuals to engage in more constructive self-reflection. When individuals engage in this type of constructive rumination, we may not see the typical maladaptive effects of rumination on problem solving. This study, therefore, investigates the proposition that it is not so much the use of specific strategies but rather their timing that is associated with impaired interpersonal problem solving.
The current study was designed to investigate sequential effects of engaging in rumination and distraction on problem solving. Given previous studies on the importance of interpersonal problem solving for depressive disorders (e.g., Nezu 1986) and previous evidence that rumination affects this important ability (Lyubomirsky and Nolen-Hoeksema 1995), we focused on the solving of problems in the interpersonal domain. Following Trask and Sigmon (1999), we hypothesized that the immediate response to a negative mood induction determines the likelihood of successful interpersonal problem solving. Thus, we predicted that participants who engage in distraction immediately following the negative mood induction (i.e., those in the distraction-distraction [DIS/DIS] and distraction-rumination [DIS/RUM] groups) provide more effective solutions to interpersonal problems compared to participants who initially engaged in rumination (i.e., rumination-rumination [RUM/RUM] and rumination-distraction [RUM/DIS] groups).
Fifty-one participants (43 female, 8 male) took part in the experiment in exchange for course credit. Twenty-seven of these participants initially engaged in rumination (17, including 6 male, in RUM/RUM and 10 in RUM/DIS) and 24 participants initially engaged in distraction (14, including 2 male, in DIS/RUM and 10 in DIS/DIS). Participants’ mean age was 22.06 (SD = 5.11).
All participants watched a 10-min film clip taken from the movie Dead Poet’s Society, depicting the suicide of a college student. This film clip has been used in previous research and has been shown to successfully induce sadness in college students (Joormann and Siemer 2004). Participants were asked to get into the feeling of the movie and to take the perspective of the main character (e.g., think how you would feel if you were the main character in the movie).
Participants completed three mood ratings over the course of the study: after the mood induction, after the first rumination/distraction period, and after the second rumination/distraction period (i.e., immediately before the problem solving task). Participants rated their present mood state on several 5-point Likert scales ranging from not at all (0) to very (4). The Likert scales included three mood-related items (i.e., happy, sad, and depressed). The items sad and depressed were averaged to obtain a single measure of negative mood at each time point. These items were interspersed randomly among six filler items (e.g., concentrated; questions about the content of the movie) to help disguise the focus of the study and the repeated measurement of mood state.
Rumination Response Scale (Nolen-Hoeksema and Morrow 1991)
The 21-item Rumination Response Scale (RRS) is a subscale of the Response Style Questionnaire (RSQ). The RRS is designed to measure the extent to which individuals respond to depressed mood by focusing on the self, symptoms and on the causes and consequences of their mood (i.e., trait rumination). The rumination subscale of the RRS can be further divided into the brooding and the reflection subcomponents, each with five items (Treynor et al. 2003). Brooding has been considered to be more strongly related to depression and, thus, a more maladaptive form of rumination than reflection. The items are rated on a 4-point Likert-type scale. The reliability of the rumination scale for the current sample was Cronbach’s α = .79. The reliabilities of the brooding and reflection subscales were Cronbach’s α = .60 and .42, respectively.
Center for Epidemiological Studies Depression Scale (Radloff 1977)
The Center for Epidemiological Studies Depression Scale (CES-D) is a commonly used self-report measure to assess levels of depressive symptoms. The 20 items in this scale are rated on a 4-point Likert-type scale. The reliability of the scale for the current sample was Cronbach’s α = .85.
The task was adapted from the work by Nolen-Hoeksema and her colleagues (e.g., Morrow and Nolen-Hoeksema 1990; Lyubomirsky and Nolen-Hoeksema 1993) and was successfully used in previous studies (e.g., Ciesla and Roberts 2007; Joormann and Siemer 2004). To influence thought content, participants were asked to focus their attention on a series of items developed by Lyubomirsky and Nolen-Hoeksema (1993). In the rumination condition, participants focused their attention on thoughts that were emotion-focused, symptom-focused, and self-focused. Participants, however, were not instructed to think specifically about negative emotions or negative personal attributes. Participants, for example, were asked to think about “your feelings right now and why you are feeling this way,” or “your character and who you strived to be.” In the distraction condition, participants focused their attention on thoughts that were externally oriented and not related to symptoms, emotions, or self (e.g., imagine looking at the shiny surface of a trumpet). During rumination/distraction, participants spent exactly 8 min focusing on eight items (self-paced).
Means-Ends Problem Solving Procedure
The Means-Ends Problem Solving (MEPS; Platt and Spivack 1972) measures interpersonal problem solving ability by assessing how the person conceptualizes the means of moving towards a pre-defined goal (i.e., a positive resolution). Participants are presented with a series of stories describing hypothetical real life interpersonal problems or conflicts. For each MEPS situation, participants are presented with the beginning and the end of a problem and are instructed to list the means that will get them to the stated ending. Following the precedent of previous studies that investigate the effects of rumination on interpersonal problem solving (e.g., Lyubomirsky and Nolen-Hoeksema 1995), we used the following four interpersonal problem solving situations: (a) you realize that a friend is avoiding you, (b) your boyfriend/girlfriend tells you that he or she is very angry with you, (c) your professor writes to you that you may fail a class, and (d) you realize that a committee’s suggestions will not work. Participants were asked to imagine experiencing the situations, and to write what they would do in each situation that would lead to the specified positive ending.
Two raters who were unaware of participants’ group assignment scored all responses. Following Lyubomirsky and Nolen-Hoeksema (1995), two measures of problem-solving effectiveness were obtained. First, each response was given a global rating of problem-solving effectiveness, on a 7-point Likert-type scale (1 = not at all effective to 7 = extremely effective). For this rating, raters considered the entire set of solutions or strategies offered in each response. Second, we used the percentage of all solutions offered by the participants that were considered “model” responses that Lyubomirsky and Nolen-Hoeksema (1995) compiled from the ratings of eight independent judges. Interrater agreement was good; the interrater correlations for the four situations ranged from .73 to .92 (M = .78) for the global effectiveness ratings and from .69 to .85 (M = .78) for the percentage of model solutions.
All participants were run individually. The experimenter informed them that the study was about “individual differences in imagination and memory.” Throughout the study, filler tasks were administered to reinforce the cover story and to reduce possible demand effects. Participants’ responses during debriefing demonstrated that the cover story was successful. After providing informed consent, participants completed a filler task, which was described as an imagination task and was used to reinforce the cover story. The experimenter then introduced the mood induction which was described as an imagination task, and participants were asked to “try to get into the feeling of the movie” and “to take the actors’ perspective.” They were also told to try to imagine the events as vividly as possible by imagining how they would feel in the situation. Participants then watched the movie for 10 min, after which they completed a “movie questionnaire” consisting of mood ratings and filler items concerning the movie (e.g., “How easy was it to understand the plot of the movie?”). Participants then engaged in either rumination or distraction, which was introduced as another imagination task, based on the group assignment. Participants then completed a second mood rating. After completing the second mood rating, another rumination or distraction task was administered followed by another mood rating. Participants then completed the MEPS followed by a packet of questionnaires mentioned above.
Participants in the four groups did not differ significantly in age, F(3,47) < 1. CES-D and RRS scores are missing for two participants. For the entire sample, the average CES-D score was M = 13.92 (SD = 7.8) and the average RRS score was M = 51.02 (SD = 8.13). The mean brooding score was M = 12.42 (SD = 2.50) and the mean reflection score was M = 11.55 (SD = 2.60). No significant group differences were obtained for RRS or CES-D scores, both Fs < 1, ns. Likewise, no significant group differences were found for brooding or reflection scores, both Fs < 1.1
Interpersonal Problem Solving
To examine our central hypothesis regarding the sequential effects of rumination and distraction on problem solving, one-way ANOVAs were conducted on the global effectiveness of the solutions provided by the participants and the percentage of model solutions endorsed on the MEPS. The results revealed that there were significant differences among groups for the percentage of model solutions provided, F(3, 47) = 3.11, p < .05, and for the effectiveness of generated solutions, F(3, 47) = 3.62, p < .05.
Research on rumination has focused primarily on maladaptive aspects of this form of self-reflection. Rumination, however, does not always exert detrimental effects (e.g., Lyubomirsky et al. 1999) and some researchers have argued that whether rumination is maladaptive or not depends on the context (e.g., Martin and Tesser 1996). Along this line, researchers have started to investigate factors that make rumination adaptive (e.g., experiential focus) versus maladaptive (e.g., analytical focus; Watkins 2004) and have started to differentiate adaptive (reflection) from maladaptive (brooding) forms of rumination (Treynor et al. 2003). Similarly, distraction, which is often viewed as a positive strategy to lift negative mood (e.g., Nolen-Hoeksema 1991), is not always helpful and might even be detrimental (e.g., Kross and Ayduk 2008). Previous research investigating the effects of rumination and distraction considered these emotion regulation strategies independently. In everyday life, however, individuals are likely to engage in diverse strategies when responding to negative affect. Thus, it is possible that it is not necessarily the use of a specific strategy but rather their timing that is crucial in determining whether a response is adaptive or not.
We investigated the effects of the timing of rumination and distraction on interpersonal problem solving, and we predicted that it is the initial response to the mood induction that exerts effects on problem solving. Consistent with previous findings (Trask and Sigmon 1999), the current study demonstrated that as long as participants distracted themselves immediately after a negative mood induction, later engagement in rumination did not lead to increased negative mood.
The current study extends previous research on sequential effects of rumination and distraction by demonstrating that the timing with which participants engage in these different emotion regulation strategies affects not only mood but also the quality of interpersonal problem solving. Regardless of whether participants further engaged in distraction or rumination, those who responded to the negative mood induction with immediate distraction generated more effective solutions to interpersonal problems compared to participants who responded to the negative mood induction with immediate rumination. Rumination had no detrimental effects on problem solving if the person initially engaged in distraction. An initial period of distraction, thus, seems to protect individuals from the detrimental effects of rumination on problem solving. Conversely, the current results demonstrate that individuals are less likely to generate effective solutions if they ruminate as an immediate response to negative mood, regardless of whether they subsequently engage in rumination or distraction. That is, distraction after rumination did not help participants to generate more effective solutions compared to the participants who continued to ruminate.
Why does rumination not hurt problem solving, once participants had a chance to distract? Distancing oneself from negative mood can reduce arousal that typically accompanies negative affect (Kross et al. 2005), and thus may change the nature of subsequent rumination. That is, rumination following distraction might be similar to a self-distanced analysis, in which an individual analyzes the event in “cool” cognitive terms (Kross et al. 2005). Individuals, therefore, would be able to engage in generating and simulating alternative options rather than being preoccupied with their negative mood and/or the current situation that initiated the problem solving. Ruminating as an immediate response to negative mood, however, might result in a self-immersed analysis which activates “hot” representations (Kross et al. 2005). Clearly, future research is needed to investigate underlying mechanisms of the observed effects.
Interestingly, our results also suggest that once a person ruminates in response to the negative mood induction, subsequent engagement in distraction does not help problem solving. It is possible that distraction after rumination is difficult to initiate or limited in its effectiveness, given the high accessibility of negative cognitions and memories. Interestingly, however, the RUM/DIS group, after engaging in distraction, showed equivalent levels of negative mood to those of the DIS/DIS group (in fact, they were the same!). The mood ratings, thus, indicate that distraction after rumination was effective; indeed, the RUM/DIS group was the only group that exhibited a significant decline in their negative mood from time 2 to time 3. Ineffective distraction following rumination, thus, cannot be a viable explanation for the current results.
The current results could be interpreted as another indication that rumination has negative effects when it occurs in the context of an already existing negative mood state. It is worth noting, however, that the sequential effects of rumination and distraction on problem solving in our study cannot be fully explained by group differences in negative mood. That is, the RUM/DIS group was not as effective as the DIS/DIS group in their problem-solving despite reporting comparable levels of negative emotion at the time when they engaged in problem-solving. In addition, the RUM/DIS group did worse than the DIS/RUM group who reported higher levels of negative mood. These results are consistent with previous findings demonstrating that individuals with depression in a state-oriented rumination condition were worse at problem solving despite experiencing similar levels of despondent mood (Watkins and Baracaia 2002). Therefore, improvements in negative mood before engaging in problem solving do not necessarily lead to effective problem solving. It seems that the sequential effects of rumination and distraction on problem solving ability are somewhat independent of an individuals’ current mood state.
The current results suggest that an initial engagement in distraction before thinking through a problem could be beneficial for problem solving. It remains unclear, however, how long individuals need to distract to experience these beneficial effects. It is also possible that prolonged rumination after distraction leads to detrimental effects on problem solving. We also do not know whether the insulating effects of distraction can be generalized to personally relevant, real-life problems. Future studies should further investigate whether the obtained results generalize to clinical samples. It is possible that rumination after distraction does not affect college students but would have detrimental effects in people diagnosed with depression or at risk for depression.
It should also be noted that the reliabilities for the RRS subscales, particularly the reflection subscale, were low. Previous studies have reported low reliabilities for RRS subscales (e.g., Armey et al. 2009; Treynor et al. 2003) and these low reliabilities may be due to the small number of items that compose each scale. We should point out, however, that the RRS was administered in this study only to demonstrate that the different experimental groups did not exhibit preexisting differences in their tendency to ruminate. Therefore, despite casting doubt on the utility of the reflection subscale, the low reliabilities found in this study do not affect the interpretations of our experimental findings.
Another important limitation of our study is the group difference in sample size. Our main hypothesis compared people who ruminated immediately after the negative mood induction to people who distracted, and this comparison relied on 24 versus 27 participants and yielded significant findings with effect sizes greater than .80. Some of the follow-up tests, however, were done using smaller subsamples (e.g., the comparison of RUM/RUM vs. RUM/DIS and the comparison of DIS/RUM vs. DIS/DIS) and should be regarded as preliminary and in need of replication. For example, Fig. 1 suggests that there is a difference in mood ratings between the RUM/RUM and the RUM/DIS group at time 2, even though both groups underwent the identical rumination induction. This difference is not significant and we can only speculate that it reflects random differences between our groups that would not replicate in future studies. Clearly, a replication of our data with a larger sample size is needed. It is also possible that we failed to find further differences among these subsamples due to a lack of power. A closer inspection of the data, however, suggests that this is unlikely given the negligible effect sizes when comparing RUM/RUM versus RUM/DIS and DIS/RUM versus DIS/DIS (Cohen’s d ranging from .03 to .14). In addition, we obtained significant findings comparing sub-groups (see Footnote 2), indicating that power might not be an issue despite the small sample size. Follow-up studies using larger samples, however, are needed.
Finally, the effects of mood inductions generally attenuate over time. Our participants completed two blocks of 8 min of rumination/distraction after the mood induction. Indeed, negative mood ratings dropped between time 1 (i.e., mood induction) and time 2 (i.e., first rumination/distraction period) in all four conditions even if participants engaged in rumination. Given the attenuation of the effects of mood induction over time, our results may simply be regarded as a replication of previous findings that the effects of rumination on problem solving occur only when individuals are in a dysphoric mood (e.g., Lyubomirsky and Nolen-Hoeksema 1995). Our analysis, however, show that the sequential effects of rumination on problem solving cannot be entirely accounted for by the levels of dysphoria. As discussed earlier, the RUM/DIS group produced less effective solutions than the DIS/RUM group despite the fact the DIS/RUM group reported greater negative mood.
To conclude, this is the first study, to our knowledge, to demonstrate that the detrimental effects of rumination on problem solving depend on the timing with which an individual engages in rumination or distraction. After distraction, later engagement in rumination no longer exerts negative effects on problem solving. Distraction, however, does not help problem solving if an individual initially ruminates in response to negative mood. These findings might provide further insight into the relation between rumination and depression (e.g., Nolen-Hoeksema and Morrow 1991). When faced with interpersonal conflicts that are likely to elicit negative mood, one’s tendency to immediately ruminate could result in ineffective problem solving in the interpersonal domain. Indeed, studies have reported that depressed participants are less likely to use distraction even though they state that it could help them alleviate their negative affect (Lyubomirsky and Nolen-Hoeksema 1993). Poor interpersonal problem solving can lead to higher levels of interpersonal stress, which then may increase risk for the onset of a depressive episode (e.g., Davila et al. 1995). Implications for treatments could include to not only target problem solving strategies directly, an effective intervention when treating depression (e.g., Nezu 1986), but to also help individuals engage in distraction in a timely manner.
There were no significant differences between male and female participants on any of our variables. Furthermore, when we included sex of the participants as a factor, none of the effects involving sex were significant. The results from the analyses examining the effects of gender are available upon request from the first author.
Although we matched the analyses closely to our a priori hypothesis, we also conducted conventional post-hoc analyses to follow-up on the significant ANOVA results. As expected, post-hoc LSD analyses revealed that the DIS/DIS group provided a higher percentage of model solutions and their solutions were rated as more effective compared to the RUM/RUM group (both ps < .05). More importantly, the DIS/RUM group provided a significantly higher percentage of model solutions and their solutions were rated as more effective compared to the RUM/RUM group (both ps < .05). Furthermore, the solutions generated by the RUM/DIS group were rated as less effective than the ones provided by the DIS/RUM and the DIS/DIS groups (both ps < .05). Thus, more conventional analyses also support our hypotheses despite small sample sizes in each group.
The results did not change, when the mood rating after the second rumination/distraction (i.e., right before completing the MEPS) was entered as a covariate. That is, the solutions generated by the participants who first engaged in distraction (DIS/RUM and DIS/DIS) were still rated as more effective than those who first engaged in rumination (RUM/RUM and RUM/DIS), and the difference between the two groups with regards to the amount of model solutions remained significant after controlling for individual differences in negative mood.
We thank Sonja Lyubomirksy for sharing her materials with us.