جلوگیری از افکار مانند وسواس در افراد غیربالینی: تاثیر بر فکر فرکانسی، ارزیابی و حالت خلقی

Wegner's (1994, Psychological Review, 101, 34–52) research on the paradoxical effect of thought suppression has been incorporated into contemporary cognitive-behavioural models of obsessive-compulsive disorder. However, findings on the effects of thought suppression on thought frequency have been inconsistent and few studies have actually examined the suppression of thoughts that are obsessional in nature. In the present study 219 nonclinical participants were randomly assigned to suppress or not suppress a neutral, obsessional or positive thought during an initial monitoring interval. In a second thought monitoring interval, all participants received instructions not to suppress their target thought. No paradoxical effect of suppression on frequency was observed for any type of thought, although suppression of obsessional thoughts was associated with greater subsequent discomfort and a more negative mood state than suppression of positive or neutral target thoughts.

Wegner, Schneider, Carter and White's (1987) investigation of the ironic effects of deliberate thought suppression has had a significant impact on cognitive-behavioural models of disorders characterized by the persistent recurrence of unwanted thoughts. Wegner et al. (1987) found that participants instructed to suppress thoughts about “white bears” had more frequent thought occurrences in a later thought expression interval compared to participants who expressed white bear thoughts prior to suppressing them. Other studies have found a close link between suppression and mood, such that negative thoughts are more difficult to suppress when mood is negative (e.g., Wenzlaff, Wegner & Roper, 1988; Wenzlaff, Wegner & Klein, 1991). Wegner and his colleagues proposed that this “rebound” effect of suppression results from the association of the “to-be-suppressed” material with internal and external distracters used to facilitate suppression which become cues for the previously suppressed material during and after suppression efforts have ceased. One obvious application of this research is to the understanding obsessive-compulsive disorder (OCD), which is characterized by the persistent recurrence of unwanted thoughts that are actively resisted. Leading cognitive-behavioural theories of OCD indeed implicate thought suppression as an important factor in thought persistence. Salkovskis, 1985, Salkovskis, 1989, Salkovskis, 1996 and Salkovskis, 1998 argued that obsessional thoughts give rise to wilful suppression efforts because they activate a highly aversive sense that one has become, or may become, responsible for harm to oneself or others. However, such attempts are bound to fail, as per Wegner's model. Furthermore the persistence of thoughts will have a negative impact on mood, which in turn will make the negative obsessional thoughts more accessible, and will prime negative appraisal of the thoughts. Rachman, 1997 and Rachman, 1998 argued that obsessional thoughts escalate because they are interpreted as having catastrophic personal significance. Interpretations of significance lead to greater efforts to control the obsession but such efforts will backfire (again, as per Wegner), resulting in an increase in frequency and negative mood. Negative mood will in turn enhance the negative interpretation of the thought, thereby increasing control efforts. Clark and Purdon (Clark, 1989, Clark & Purdon, 1993 and Purdon & Clark, 1999) have elaborated on these ideas, suggesting that individuals vulnerable to developing obsessional problems may believe that obsessional thoughts are evidence that undesirable personality characteristics exist and that their thoughts can and should be controlled. Failures in thought control are thus experienced as devastating and lead to a more negative mood state which in turn further reduces thought controllability. Thought suppression, then, has been given a central place in cognitive-behavioural models of OCD. However, empirical support for the role of suppression in the persistence of thoughts has been mixed, with some studies replicating Wegner's delayed “rebound” effect, other studies finding an immediate effect of suppression, and other studies finding no effect of suppression on thought frequency at all [see Purdon (1999) and Purdon & Clark (2000) for a comprehensive review]. Furthermore, although the thought suppression paradox has its most obvious application to OCD, very few studies have actually examined the suppression of thoughts that are obsessional in nature. The relationship between obsessionality and suppression of nonobsessional thoughts has been studied (e.g., Smári, Sigurjónsdóttir & Sæmundsdóttir, 1994), but as Salkovskis (1996) has convincingly argued, obsessional problems are not associated with general deficits, but rather with difficulties associated with one or two thoughts in particular. Thus, our understanding of the role of suppression in OCD is likely best advanced by studies of the suppression of obsessional thoughts exclusively. Several studies have examined the suppression of obsessional thoughts, but again, findings have been mixed. For example, Smári, Birgisdóttir and Brynjólfsdóttir (1995) observed no paradoxical effect of suppression on thought frequency, although there was some indication that participants higher in obsessional symptomatology had more target thoughts during suppression. Kelly and Kahn (1994) used a crossover design to examine suppression vs expression (i.e., active generation) of neutral (“white bear”) thoughts as compared to suppression vs expression of “intrusive”, or normal obsessional thoughts. Suppression of neutral thoughts was associated with a later increase in thought frequency (i.e., a rebound effect), with heightened subjective distress, and a greater perceived loss of control over one's thoughts. However, suppression of intrusive thoughts was not associated with paradoxical effects on frequency and subsequent distress. Rutledge (1998) asked participants to express a target thought during a baseline monitoring interval, suppress it during a second interval, and then express it again during a third interval. No effect of suppression on thought frequency was observed either during or after suppression, relative to the baseline first interval. Obsessionality was associated with greater thought frequency during suppression for women, but not for men. Finally, McNally and Ricciardi (1996) presented participants with a list of examples of thoughts reflecting the various themes of obsessions and asked them to identify one that they had previously experienced. Suppression of obsessional vs neutral thoughts was compared using a crossover design. No effect of suppression was observed although there was a marginally significant tendency for the obsessional thought to occur more often after suppression whereas neutral thoughts tended to occur less frequently after suppression. Although the findings from these studies consistently revealed no suppression effect on frequency of obsessional thoughts, the implications of the findings for understanding OCD may be limited by methodological issues. In the first three studies, it is unclear whether the target thoughts examined actually represented obsessional thoughts because participants simply reported on thoughts that were distressing. One difficulty in assessing obsessional thoughts in nonclinical samples is ensuring that the participants are reporting on thoughts that have actual obsessional content, and are not representative of common worries. Like obsessional thoughts, worries are distressing, but they are far more common and more readily accessible to nonclinical individuals. Indeed, in the examples of target thoughts reported by Kelly and Kahn (1994) and by Rutledge (1998), the thought content appears to be more typical of worry (e.g., thoughts about finances, health). Clark and Purdon (1995) and Clark (1999) have argued that identification of obsessional thoughts requires that the distinguishing features of obsessions (e.g., discrete, spontaneous, intrusive, repugnant, ego-dystonic) be taken into account when eliciting thought content. Use of thought expression as a control condition is also problematic if findings are to be generalized to obsessional problems because obsessions are virtually never actively generated, but rather are actively avoided [see Lavy & van den Hout (1990); Merckelbach, Muris, van den Hout & de Jong, 1991 for a discussion]. The crossover design is also problematic because it confounds suppression instructions with rehearsal of the experimental task [see D.M. Clark, Ball & Pape (1991) and Keppel (1982)]. Purdon and Clark (2000) offer a thorough review and discussion of methodological issues in applying thought suppression research to OCD. Salkovskis offers the most ecologically valid tests of the effects of suppressing obsessional thoughts in analogue samples to date. Salkovskis and Campbell (1994) selected nonclinical participants who reported frequent and distressing obsession-like intrusions and instructed them to either monitor or monitor and suppress their most frequent obsessional thought. During the initial monitoring interval, individuals were given further instructions to either suppress by using one replacement thought as a distracter (which, according to Wegner, should reduce thought rebound compared to other suppression strategies), suppress without using distraction or suppress by completing an attentionally engaging task (a relaxation task). In a second interval, all participants received instructions to simply monitor their thought. All suppression groups except the “attentionally engaging task” group exhibited an immediate and sustained frequency of thoughts relative to “think anything” instructions such that thought frequency was higher in both intervals for those in the “initial suppress” condition. This suggested that the insidious effects of suppression on the frequency of obsessional thoughts may have different mechanisms than those proposed by Wegner. Ratings of discomfort and unacceptability did not vary by group, except that the “attentionally engaging” group reported lower ratings than the other groups. Trinder and Salkovskis (1994) had participants record the occurrence of obsessional thoughts over a 4 day period. One group of participants were instructed to suppress the thought whenever it occurred, a second group were instructed to actively think about the thought (a control for the enhanced salience of the thought afforded by the “suppress” instructions), and a third group simply recorded the thought occurrences. Suppression resulted in greater frequency of thought occurrences and heightened discomfort relative to actively thinking about the thought or simply recording it, whereas discomfort declined for those in the latter group. Furthermore, suppression effort was associated with greater discomfort and thought frequency. Finally, for participants instructed to suppress, discomfort over thought occurrences showed a strong correlation with a symptom measure of anxiety. Taken together, the findings from these latter two studies suggest that, at least in analogue samples, obsessional thoughts may be more difficult to suppress. Furthermore, suppression appeared to be associated with greater thought frequency, discomfort and anxiety. The findings from these two studies support Salkovskis, 1989 and Salkovskis, 1998 contention that thought occurrences, negative appraisal, negative mood, and suppression are components of an insidious feedback loop. Janeck and Calamari (1999) conducted the only study to date that has examined the suppression of obsessional thoughts in a clinical sample of individuals with OCD. Patients and nonclinical controls were administered a validated measure of obsessional intrusions and then instructed to monitor a selected thought from that inventory over three intervals; baseline “think anything”, “suppress” vs “think anything”, and “think anything”. There was no evidence of either a rebound or immediate enhancement effect of suppression on thought frequency overall for either the OCDs or the nonclinical controls, although a greater proportion of the OCD sample experienced a rebound after suppression than did the controls. However, even at that, 75% of the OCD sample did not experience thought rebound. The role of deliberate thought suppression in the development and persistence of obsessional ideation thus remains unclear. As observed earlier, there are methodological problems with some studies that limit their application to understanding obsessional problems and results from studies with strong ecological validity have been inconsistent. At the same time, other issues have remained unresolved in the literature. For example, although there is evidence of a strong link between negative mood and difficulties suppressing negative thoughts (e.g., Wenzlaff et al., 1988) the relationship between mood state and suppression of obsessional thoughts has not been well-examined. Furthermore, very few studies have examined the effects of suppression on variables other than frequency, such as discomfort, unacceptability and unpleasantness. Finally, it would be interesting to examine the effects of suppressing obsessional thoughts in relation to the effects of suppressing emotionally relevant positive vs emotionally neutral thoughts. This would allow for the role of personal relevance of the target thought to be disentangled from the role of emotional valence in the effects of thought suppression. The purpose of the present study was threefold: 1. to contribute to the existing literature on the effects of suppressing exclusively obsessional thoughts on both thought frequency and emotional reaction to the thought whilst observing the methodological recommendations forwarded by Purdon and Clark (2000); 2. to further examine the relationship between suppression of obsessional thoughts and mood state; and 3. to compare the effects of suppressing obsessional thoughts to those of suppressing emotionally relevant positive thoughts and neutral thoughts with no emotional relevance. In this study, 219 nonclinical individuals were randomly assigned to monitor and record either a neutral thought, an emotionally relevant positive thought, or an obsessional thought across two six minute intervals. In the first interval, half of the participants were instructed to suppress their target thought whereas the others were instructed not to suppress any thoughts. In the second interval, all participants received instructions not to suppress any thoughts. The study was guided by three hypotheses based on cognitive models of OCD and the empirical literature. First, it was hypothesized that participants instructed to suppress obsession-like thoughts would report significantly more thought occurrences during the suppression period than participants instructed to suppress neutral or positive thoughts, or controls instructed not to suppress their target thought (i.e., an immediate enhancement effect, as observed by Salkovskis), whereas those instructed to suppress neutral thoughts will experience a resurgence of thoughts after suppression efforts have ceased (i.e., a rebound effect, as observed by Wegner). Second, it was hypothesized that individuals in the suppress/obsessional group would rate their target thought as significantly more emotionally upsetting than participants suppressing neutral or positive thoughts, or the controls instructed not to suppress their target thought. The final hypothesis was that suppression of intrusive thoughts would be associated with a more negative mood state than the suppression of neutral thoughts. No predictions were made about the effects of suppressing positive thoughts due to the exploratory nature of this condition.

3.1. Preliminary analyses There were no significant group differences on the ROII, ATQ-P or T1 Mood Scale, indicating that random group assignment had been achieved. The overall means for the ROII and ATQ-P were within the expected range for a university sample. There was significant variability in the range of intrusions identified by participants as being their most upsetting, with the maximum endorsement for any one thought being ca. 10%. The four most frequently reported upsetting intrusive thoughts were: 1. unwanted, unacceptable thoughts of engaging in a sexual act that goes against one's sexual preference; 2. unwanted, unacceptable thoughts of self-harm; 3. unwanted, unacceptable thoughts of having left the stove, heat or lights on and thereby caused an accident; and 4. unwanted, unacceptable thoughts of having sex with someone who is repugnant. The top four most pleasant positive thoughts were: 1. There are many people who care about me; 2. I have friends who support me; 3. I will be successful; and 4. I'm fun to be with. The rates of endorsement of these top four ranged from 27% to 5.5%. Analyses of variance were conducted on each thought appraisal item administered after the visualization task, as well as the two items assessing the effectiveness of the prime (vividness and difficulty imagining the thought). Bonferroni adjustment for Type I error resulted in an alpha level of 0.01, and post hoc tests were conducted using Student Newman–Keuls tests. These analyses revealed that the obsessive intrusive group rated their thought as more discomforting [F(2, 215)=153.16, p<0.001], more unpleasant [F(2, 215)=17.02, p<0.001], more unacceptable [F(2, 215)=72.91, p<0.001] and more difficult to imagine [F(2, 215)=15.91, p<0.001] than did the neutral or positive groups. Those in the positive group rated their target thought as more pleasant than the neutral and obsessive intrusive groups [F(2, 215)=17.02, p<0.001]. These data indicated that, as intended, the assigned target thoughts represented positive, neutral and negative stimuli to participants. Participants in the obsessive intrusive group reported having less vivid images of their target thought during the priming task than the other two groups [F(2, 216)=4.34, p<0.01]. The next analysis was conducted to determine whether participants complied with the experimental instructions to suppress or not suppress. Means and standard deviations of the suppression effort scores across group and experimental interval are presented in Table 1. A 2 (Suppression Group)×3 (Target Thought Group)×2 (Interval, within subjects factor) repeated measures analysis of variance on suppression effort ratings was then conducted. There was a significant Target Thought Group×Suppression Group×Interval interaction [F(2, 213)=4.63, p<0.011]. Post hoc analyses following the principles of Fisher's protected tests were then conducted. In this procedure, tests of simple main effects are conducted using the same alpha level as used to determine the significance of the main effect or interaction (see Cohen & Cohen, 1983). Within interval 1, there was a significant interaction of target thought group×suppression group [F(2, 213)=7.13, p<0.001]. Post hoc analyses revealed that the suppress group had significantly higher suppression effort ratings than the nonsuppress group [F(1, 217)=66.59, p<0.001]. However, suppression effort scores within the nonsuppress groups varied according to target thought, which accounted for the two-way interaction within this interval. Analysis of simple main effects revealed that the nonsuppress/obsessional group had significantly higher suppression effort ratings than those in the nonsuppress groups assigned a neutral target thought [F(1, 213)=14.08, p<0.001] or a positive target thought [F(1, 213)=33.06, p<0.001]. Nonetheless, the suppression effort scores for the nonsuppress/obsessional group were significantly lower than those of their suppress/obsessional group counterparts [F(1, 213)=3.93, p<0.05]. In Interval 2 there was no target thought group×suppression group interaction and no significant main effect of suppression group. Thus, suppression effort ratings in Interval 2 were equivalent across the suppress and nonsuppress groups. However, there was a main effect of target thought group [F(2, 213)=7.19, p<0.001] such that the obsessional thought group had higher scores than those in the neutral and the positive thought groups [F(1, 213)=14.38, p<0.001]. These findings indicated that, in general, participants complied with the experimental instructions to suppress or not to suppress their target thought. Table 1. Means and standard deviations of suppression effort across experimental groups and intervalsa Interval 1 Interval 2 Group M SD M SD Suppress b Neutral 74.73 23.69 26.41 28.48 Positive 71.47 30.00 23.67 23.36 Intrusive 75.24 23.46 38.71 31.33 Nonsuppress b Neutral 38.63 30.70 39.77 32.89 Positive 26.53 28.62 25.00 28.10 Intrusive 62.74 24.38 46.49 30.10 a Note: n per group ranges from 35 to 38. b “Suppress” and “Nonsuppress” refer to Interval 1 instructions; in Interval 2 all participants received the “Do not suppress” instructions. Table options 3.2. Effects of suppression on thought frequency The first hypothesis predicted that participants in the suppress/obsessional group would experience an immediate and sustained resurgence of target thoughts relative to all other groups (i.e., an immediate enhancement effect), whereas those in the suppress/neutral group were expected to have fewer thoughts during the first interval and more target thought occurrences than the nonsuppress groups in the second interval (i.e., a rebound effect). Table 2 presents the means and standard deviations on the frequency of target thoughts during the first and second monitoring intervals for the six experimental groups. In order to understand the effect of the experimental manipulation on thought frequency across intervals a 2 (suppression group)×3 (target thought group)×2 (interval, within Ss factor) repeated measures analysis of variance was conducted on thought frequency. A significant target thought group×suppression group×interval interaction was observed [F(2, 212)=5.69, p<0.004]. This interaction is presented in Fig. 1. Table 2. Means and standard deviations of thought frequency by Interval and Experimental groupsa Interval 1 Interval 2 Group M SD M SD Suppress b Neutral 9.14 9.37 10.68 11.91 Positive 7.72 5.84 10.44 10.13 Intrusive 6.47 5.54 4.61 6.01 Nonsuppress b Neutral 20.46 15.33 12.91 10.00 Positive 13.55 8.23 8.97 7.08 Intrusive 5.88 3.46 2.71 2.36 a Note: n per group ranges from 34 to 38. b “Suppress” and “Nonsuppress” refer to Interval 1 instructions; all participants received the “Do Not Suppress” instructions in Interval 2. Table options Full-size image (8 K) Fig. 1. Number of thought occurrences by experimental group and interval. Figure options Analysis of Interval 1 thought frequency revealed a significant target thought group×suppression group interaction [F(2, 212)=8.27, p<0.001]. Analyses of simple main effects revealed that the nonsuppress/neutral group had significantly more thoughts than the suppress/neutral group [F(1, 212)=29.84, p<0.001]. Similarly, those in the nonsuppress/positive group reported significantly more thoughts than those in the suppress/positive group [F(1, 212)=8.13, p<0.01]. When intrusive thought frequency was compared across the suppress and nonsuppress obsessional groups, however, no significant difference was observed, thereby accounting for the interaction. Examination of changes in reported thought frequency across interval revealed a significant decline for all nonsuppress groups from interval 1 to interval 2 [F(1, 212)=39.77, p<0.001, F(1, 212)=15.91, p<0.001 and F(1, 212)=6.85, p<0.01 for the neutral, positive and obsessional groups, respectively]. The suppress groups showed no significant change in thought frequency from interval 1 to interval 2. Examination of reported thought frequency within Interval 2 revealed a main effect of target thought group only [F(2, 212)=18.98, p<0.001)]. Thus, thought frequency within target thought groups did not vary according to suppression group. Analyses of the main effect of target thought group revealed that the obsessional group had significantly fewer thoughts in interval 2 (M=3.66, SD=4.18) than either the positive group (M=9.40, SD=7.50, F(1, 212)=18.14, p<0.001) or the neutral group (M=11.79, SD=10.96, F(1, 212)=35.892, p<0.001) 1. These analyses did not support the first hypothesis. There was no evidence of an immediate enhancement effect or rebound effect for any type of thought, although prior suppression of target thoughts did result in sustained levels of the target thought during the subsequent monitoring interval for all suppress groups (whereas frequency declined for all nonsuppress groups). 3.3. Suppression and thought appraisal To determine whether suppression would interfere with emotional response to thoughts, a 2 (suppression group)×3 (target thought group)×2 (T1 vs T2) repeated measures analysis of variance was conducted on discomfort ratings. Participants who did not report at least one target thought occurrence during interval 1 or interval 2 did not make appraisal ratings after those intervals (n=27), and so were excluded from the analysis, as were others who had missing data (n=2). These participants were evenly distributed across groups. Two multivariate outliers were also excluded, leaving a final total N=188 2. Means and standard deviations of the discomfort ratings across experimental groups are presented in Table 3. Significant main effects of target thought group [F(2, 182)=28.22, p<0.001] and interval were observed [F(1, 182)=21.54, p<0.001], but there was also a significant interaction of interval with target thought group [F(2, 182)=7.05, p<0.001], and an interaction of suppression group with interval [F(1, 182)=5.43, p<0.02] 3. These interactions were qualified by a three-way interaction of target thought group, suppression group and interval [F(2, 182)=9.13, p<0.001]. Post hoc analyses revealed that discomfort scores for the suppress/obsessional group were not significantly different from the nonsuppress/obsessional group after the first interval [F(1, 182)=0.93, p<0.34], but that there was a significant decline in discomfort scores for both the suppress/obsessional [F(1, 182)=4.02, p<0.05] and nonsuppress obsessional groups [F(1, 182)=25.95, p<0.001] from interval 1 to interval 2. However, after the second interval, the suppress/obsessional group ratings of discomfort were significantly higher at interval 2 than those of their nonsuppress/obsessional counterparts [F(1, 182)=9.87, p<0.002]. Table 3. Means and standard deviations of “Discomfort” ratings across Experimental groupsa Group Interval 1 Discomfort Interval 2 Discomfort M SD M SD Suppress Neutral 19.66 23.48 9.97 16.42 Positive 25.70 28.09 10.23 14.54 Intrusive 52.54 29.97 43.89 31.09 Nonsuppress Neutral 11.17 17.19 21.14 25.94 Positive 18.05 22.67 18.05 23.35 Intrusive 46.04 27.35 24.86 19.42 a Note: N=188, with exclusion of two outliers and participants who did not report a single thought occurrence during either interval 1 or interval 2 and subsequently not able to give “Discomfort” ratings. Higher ratings reflect higher discomfort. Cell ns range from 26 to 37. Table options The discomfort scores in the suppress/neutral group and the suppress/positive group were not different from their respective nonsuppress counterparts after interval 1, and declined significantly from interval 1 to interval 2 [F(1, 182)=14.83, p<0.001 and F(1, 182)=6.21, p<0.01, respectively], whereas discomfort scores for the nonsuppress/neutral scores increased [F(1, 182)=7.19, p<0.01]. After interval 2, discomfort scores for the suppress/neutral group were significantly lower than those for the nonsuppress/neutral group [F(1, 182)=4.08, p<0.05]. Discomfort scores for the nonsuppress/positive group were not significantly different from their suppress/positive counterparts after either interval. A 2 (suppression group)×3 (target thought group)×2 (interval; within Ss factor) analysis was then conducted on participants' ratings of unpleasantness associated with the target thought at intervals 1 and 2. The total N for this analysis was 190. A significant main effect of target thought group emerged [F(2, 184)=57.54, p<0.001], but no other main effects were observed, nor were there any interactions. Post hoc analyses revealed that overall, the obsessional group rated their target thought as being more unpleasant (M=66.82, SD=21.96) than did the positive group [M=29.13, SD=27.34; F(1, 184)=59.62, p<0.001] and the neutral group [M=40.04, SD=25.69; (F(1, 184)=49.30, p<0.001]. Unacceptability ratings were similarly examined (N=190). A main effect of target thought group was observed [F(2, 184)=62.09, p<0.001]. Post hoc analyses revealed that those in the obsessional thought group had significantly higher unacceptability ratings (M=61.67, SD=28.09) than the neutral [M=22.00, SD=25.88; F(1, 184)=84.73, p<0.001] or the positive group [M=19.83, SD=27.83; F(1, 184)=100.76, p<0.001]. A main effect of interval was also observed [F(1, 184)=5.32, p<0.02], but was qualified by the emergence of a two-way interaction of suppression group by interval [F(1, 184)=7.81, p<0.01]. Post hoc analyses revealed that unacceptability ratings for the suppress group were significantly higher at interval 1 [M=40.63, SD=34.51; F(1, 188)=5.59, p<0.02] than for the nonsuppress group (M=29.56, SD=32.14), and declined from interval 1 to interval 2 [M suppress T2=29.97, SD=31.79; F(1, 188)=12.77, p<0.001] whereas those of the nonsuppress group did not change across intervals (M nonsuppress T2=30.92, SD=32.34). Unacceptability ratings after interval 2 were the same across groups, thus accounting for the interaction. Finally, only a main effect of interval was observed on responsibility ratings [F(1, 184)=6.97, p<0.01], such that ratings after interval 2 were significantly lower (M=49.114, SD=30.22) than those completed after interval 1 (M=54.24, SD=29.67). The results of these analyses offer partial support for the second hypothesis. Suppression of obsessional, but not positive or neutral thoughts resulted in higher levels of self-reported discomfort associated with target thought occurrences relative to the nonsuppress/obsessional thought group during the second monitoring interval. Furthermore, target thoughts that occurred during the first interval for the suppress group (i.e., thoughts occurring whilst participants were supposed to be suppressing) were rated as more unacceptable than thoughts occurring during the first interval for the nonsuppress group (i.e., when there were instructions not to suppress). Suppression did not have a significant effect on the pleasantness or responsibility appraisal ratings of intrusive, positive or negative thoughts. 3.4. Suppression and mood state It was predicted that participants in the suppress/obsessional group would report a significantly more negative mood state after thought suppression than individuals who suppressed neutral or positive thoughts, or participants in the no suppress conditions. Mood state was measured prior to the visualization task and then again after the second monitoring interval. A 2 (suppression group)×3 (target thought group) analysis of variance was performed on T2 mood scores, with T1 mood Scores entered as a continuous independent predictor4. Total number of thoughts across intervals 1 and 2 were entered as a second continuous predictor. The main effects were entered sequentially according to chronological order of events. This allowed for examination of the unique contribution of each predictor after the events preceding it had been removed. Results indicated a significant main effect of T1 mood scores [F(1, 200)=249.31, p<0.001], such that higher scores at T1 predicted higher scores at T2. A main effect of target thought group was also observed [F(2, 200)=6.76, p<0.001]. This was qualified by a significant Target Thought Group×T1 mood interaction [F(2, 200)=5.03, p<0.01]. In order to interpret this interaction the zero-order correlations between T1 and T2 mood scores were examined within each level of target thought group. The correlations between T1 and T2 mood were quite high in the neutral and positive groups (r=0.77, p<0.001 and r=0.86, p<0.001, respectively). However, the correlation within the obsessive intrusive group was lower (r=0.48, p<0.001). Post hoc analyses revealed that T2 mood scores were significantly lower (i.e., more negative) for those in the obsessional thought group than for those in the neutral [F(1, 200)=7.12, p<0.01] or positive thoughts [F(1, 200)=12.44, p<0.001] groups. A significant Target Thought Group×Thought Frequency interaction was also found [F(2, 200)=3.99, p<0.02]. It was qualified by a significant Target Thought Group×Suppression Group×Thought Frequency interaction [F(2, 200)=3.18, p<0.04]. The zero-order correlations between thought frequency and T2 mood within each of the six experimental groups were examined. Thought frequency had no relationship with T2 mood in any group (i.e., rs=approximately 0, all ns except for the suppress/obsessional group r=−0.40, p<0.01), such that higher thought frequency was associated with more negative T2 mood. This analysis supported the third hypothesis. More frequent thought occurrences were associated with more negative mood state for individuals instructed to suppress obsessional thoughts. Suppression of neutral or positive target thoughts did not have a significant impact on mood state