Obsessive-compulsive symptoms and suppression of personally relevant unwanted thoughts

Pergamon 0191-8869(94)00209-6

Purrou. indi\Id. Diff: Vol. 18. No. 5. pp. 621-625, 1995 Coovrieht 0 1995 Elsevier Science Ltd Printed in Great Britain. All rights reserved Ol9l-8869/95 $9.50 + 0.00

OBSESSIVE-COMPULSIVE SYMPTOMS AND SUPPRESSION OF PERSONALLY RELEVANT UNWANTED THOUGHTS Jakob Smari, Ama Bjork Birgisdottir and Berglind Brynjolfsdottir Faculty of Social Science, University (Received

of Iceland, Reykjavik,

15 August

Iceland

1994)

Summary--Consequences of thought suppression were compared in people characterized by different levels of obsessive-compulsive symptoms. Eighty-five female students were asked to retrieve an uncomfortable thought. Then they were instructed either to think about anything they liked or to do the same with the exception that they were not to think about the uncomfortable thought. The subjects subsequently wrote down their stream of thought. Then all subjects were asked to think without any limitation for a second period. There was a significant interaction between obsessive-compulsive symptoms measured with the Maudsley Obsession-Compulsion Inventory and initial instruction, with regard to the occurrence of the uncomfortable thought during the first period, for two out of three measures. Alternative explanations of the results are discussed.

The large majority of people report that they have intrusive thoughts and use effortful strategies to avoid them (Freeston, Ladouceur, Thibodeau & Gagnon, 1991). An important strategy is the suppression of unwanted thoughts. Wegner recently initiated an original approach to thought suppression (Wegner, Carter, White & Schneider, 1987). He demonstrated that the suppression of innocuous thoughts may lead to an exacerbation of such thoughts. This has recently been discussed by Wegner in terms of what he calls an ironic process (Wegner, 1994). By ironic process Wegner means an automatic process that checks if a mental target that one consciously tries to avoid is in fact being avoided, but that, at the same time, the check serves as a reminder of the target. Thought suppression has been found to have two interesting consequences: initial enhancement of the suppressed thought and rebound. Initial enhancement refers to the paradoxical effect that thought suppression may increase rather than decrease the frequency of target thoughts. Rebound is an increase in thoughts about a target, when suppression instructions are annulled. Initial enhancement and rebound have, however, not been reliably observed in previous research (Clark, Winton & Thynn, 1993; Merckelbach, Muris, van den Hout & deJong, 1991; Muris, Merckelbach, van den Hout & de Jong 1992; Salkovskis & Campell 1994). Research on experimentally induced suppression promises to increase our understanding of disorders involving persistent intrusive thoughts. Above all it seems relevant to obsessive+zompulsive disorders where the person is disturbed by thoughts that he also considers unreasonable. It is thus appropriate to study the behaviour of people high on obsessive-compulsive symptoms in a thought suppression paradigm. People high on obsessive-compulsive symptoms might be expected to try harder to suppress unwanted thoughts and to show rebound, and/or initial enhancement as a consequence. In thought suppression research, suppressed thoughts have generally been ostensively innocuous. However, intrusive thoughts in psychological disorders, such as obsession-compulsion, generally are upsetting, at least to the afflicted person. If the phenomena observed in the paradigm are viable as a model for such disorders, they should occur when emotional material is suppressed. Some studies, however, have failed to find evidence for this when emotional material is used in contrast to non-emotional material (Muris et al., 1992; Smari, Sigurjbnsdottir & Sremundsdbttir, 1994). The reason for this is not clear. Perhaps people are less motivated to suppress emotional material because it is more acceptable not to be able to do so. The relevant factor in this context may rather be whether or not the thoughts are personally relevant. But Kelly and Kahn’s (1994) study indicates that at least rebound does not occur as a consequence of suppression of personally relevant thoughts. The central issue is, however, whether or not these phenomena appear with personally relevant thoughts in people characterized by obsessive-compulsive symptoms. 621

Jakob S&-i ef al.

622

In the present study we address the moderating role of obsessive-compulsive symptoms in the thought suppression paradigm. We hypothesize an interaction between instructions (suppression-control) and the level of obsessive-compulsive symptoms with regard to the frequency of personally relevant, unwanted thoughts. We expect, in a first period, people characterized by obsessive-compulsive symptoms to show initial enhancement when they try to suppress unwanted, personally relevant thoughts, to a greater extent than people less so characterized. For rebound the expectations are less clear, but we tentatively include a second control period for all subjects, for exploratory purposes.

METHOD Subjects

Eighty-five female students of education, psychology and music served as Ss in this study. Their age varied between 19 and 39 years (m = 23.99, SD = 4,17). Procedure

The study was presented to Ss as relating to different kinds of thoughts and feelings. They were asked to fill in three different inventories, one of which was the Icelandic translation (Smari, Bjarnason & Thorleifsson, 1994) of the Maudsley Obsessive-Compulsive Inventory (MOCI) (Rachman & Hodgson, 1980). After filling in the inventories, they were asked to retrieve a thought that bothered them in daily life, write a description of this thought on a piece of paper, and then tear it up. The Ss were given randomly either a suppression or a control instruction. In the suppression instruction they were asked to write down on sheets of paper their stream of thought for a few minutes. They were told to think about anything they liked with one exception; they were not to think about the thought they had just written down. Anytime that thought occurred to them, they were to mark an X at the right margin of the sheets. The number of these marks was used as a measure of current target thoughts. A similar measure is used by Wenzlaff, Wegner and Roper (1988). The control instruction was exactly the same, apart from that the Ss were told they could think about anything they liked, including the thought they had written down previously. After a period of 5 min all the Ss were given a control instruction and told to write their thoughts as before. After a second period of 5 min the Ss turned to a booklet with a series of questions relevant to their behaviour during the two periods. First came a set of questions related to the first, and then a set of questions related to the second period. On the first measure the Ss rated on a 7-point scale, from 1 = not at all to 7 = very much, how much they had tried not to think about the target thought during the period. This measure also served as a manipulation check as to how well the Ss had understood and tried to follow the initial instructions. Then there were two measures of the frequency of occurrence of the target thoughts during the period. The first measure was a rating on a 7-point scale, of how much the S had thought about the target thought, from 1 = not at all to 7 = constantly. The second measure was a rating of the percentage of thinking time devoted to target thoughts, on a scale from 0 to 100%. These are, together with the X marks, supposed to reflect the frequency of occurrence of the target thoughts. Also the Ss rated, on two different 7-point scales: (a) how much control of their thoughts they had experienced during the period: and (b) how much control they estimated they should have had over their thoughts during the period. Finally, in order to assess other potentially relevant characteristics of the target thoughts, Ss were asked to rate also on two 7-point Likert scales how much the thought used to disturb them in daily life, and how often the thought used to occur in daily life.

RESULTS Obsessive-compulsive

symptoms

The mean and standard deviation on the Maudsley Obsessive-Compulsive Inventory (MOCI) were 7.82 and 4.27, respectively. Ss were divided into those high and low on obsessive-compulsive symptoms based on whether their scores were below or above the median. The median value was 7

Obsessive-compulsive

623

Table I. Initial instruction, obsessive%ompulsive symptoms fMOCf) and means for: (a) x marks for target thoughts; fb) amount of target thoughts; (c) percentage time devoted to target thoughts. The first period High on MOCI

Low INI MOCI

3.0 3.1

3.0 1.6

4. I 4.6

4.2

Control

51

Suppression

57

42 24

(0)

Control Suppression

(bt Control Suppression

3.0

Cc)

and Ss scoring at the median (10) were excluded in the following analyses that compare Ss high and low on obsessive-compulsive symptoms. The number of remaining S’swas 36, in the high group and 39, in the low group. The means and standard deviations of the two groups were (m = 11.49, SD = 3.67) for those high, and (m = 4.28, SD = 1.26) for the group that was low, with respect to obsessive~ompulsive symptoms. Manipulution check In order to check that the control-suppression instructions had been properly understood, a one-way analysis of variance was performed on scores on the try not to think about the target variable, during period I, with initial instruction as the independent variable. The difference between conditions was significant in the expected direction, F( 1,83) = 6.01, P = 0.02. Target thoughts during the first period In order to test the principal hypothesis of an interaction between initial instruction and level of obsessive-compulsive symptoms with regard to target thoughts, three two-way analyses of variance were performed, one for each measure of target thoughts, with inst~ction (suppression-control) and level of obsessive-compulsive symptoms (high-low) as independent variables. The interactions between instruction and obsessive-compulsive symptoms were all significant or marginally significant, F( 1,7 1) = 3.7 1, P = 0.06 for X marks, F( 1,7 1) = 4.15, P = 0.04 for percentage of time and F( 1,7 1) = 4.09, P = 0.05 for amount of thought. An inspection of the means revealed that Ss low on obsessive-compulsive symptoms showed less target thoughts with the suppression instruction, whereas the reverse is true for Ss high on obsessive-compulsive symptoms (see Table 1). The main effect for obsessive-compulsive symptoms was significant for two of the dependent variables, X marks, F( I ,7 1) = 4.05, P = 0.05 and percentage of time, F( I,7 1) = 12.08, P = 0.001, and marginally significant for amount of thought, F( 1,71) = 3.16, P = 0.08. The main effect for instruction was non-significant in all three analyses (P > 0.10). Thus there was no general initial enhancement effect. It is possible that the pattern of results obtained is due to the fact that target thoughts of people high on obsessive-compulsive symptoms are particularly emotional or well rehearsed. In fact two one-way analyses of variance with level of obsessive-compulsive symptoms as the independent variable, and the variables “how much do you think about the target thought in daily life”, and “how disturbing is the target thought in your daily life”, revealed significant effects of obsessive~ompulsive symptoms for both variables, F( 1,7 1) = 6.75, P = 0.01 for the former and F( 1,70) = 5.02, P = 0.03 for the latter. Ss high on obsessive-compulsive symptoms feel that the unwanted thought occurs more often and is more disturbing to them in their daily life. Thus three two-way analyses of variance were performed on the three target thought variables with obsessive~ompulsive symptoms and initial insertion, as independent variables, and the “how much do you think about the target in your daily life”, and “how disturbing is the target thought in your daily life” measures, as covariates. Three Ss were excluded because of missing values on either one of the two covariates. The interactions between obsessive~ompulsive symptoms and inst~ction were no longer significant, but there were still

Jakob Smari et al.

624 Table

2.

Means

target during:

for reports

of trying

(a) the first; High

((1) Control

not to think

about the

and (b) the second period

on MOCI

Low

on MOCI

3.21

3.29

Suppression

4.94

3.10

Control

2.63

2.88

Suppression

3.64

2.63

(h)

tendencies in the same direction as before [F( 1,66) = 2.79, P = 0.10 for X marks, F( 1,66) = 3.50, P = 0.07 for amount of thought and F( 1,66) = 2.72, P = 0.10 for percentage of time]. The main effect of obsessive-compulsive symptoms remained significant for percentage of time, F( 1,66) = 7.29, P = 0.01, marginally significant for X marks, F( 1,66) = 3.73, P = 0.06, but non-significant for amount of thought, F( 1,66) = 1.70, P = 0.20. These results are difficult to interpret unambiguously, but they indicate that the interaction between instruction and obsessive-compulsive symptoms, and the difference between subjects high and low on obsessive-compulsive symptoms with regard to target thoughts are not wholly due to the characteristics of the target thoughts as they are measured with the two scales. Target thoughts during the second period Three two-way analyses of variance were performed with level of obsessive-compulsive symptoms and initial instruction as independent variables and the three target thought measures as dependent variables. The interactions between the independent variables were non-significant, P > 0.10 except for a non-significant trend (P = 0.08) in the same direction as in the first period for percentage of thought. The main effects for instruction were non-significant for all target thought measures (P > 0.10). There was thus no trace of a rebound effect. For one target thought measure (percentage of time) there was a significant main effect for obsessive-compulsive symptoms in the same direction as in the first period, F(1,71) = 5.14, P = 0.03. Trying not to think about the target Above it was noted that Ss in the suppression condition reported more attempts at trying not to think about the target than the Ss of the control condition. But how did these attempts relate to level of obsessive-compulsive symptoms ? Across periods a two-way analysis of variance yielded an interaction between initial instruction and obsessive-compulsive symptoms, F( 1,7 1) = 6.11, P = 0.02. The main effect of obsessive-compulsive symptoms was not significant (P > 0.10). The interaction reflects the fact that Ss high in obsessive-compulsive symptoms make much more effort, not to think about the target, given suppression instructions, whereas Ss low in obsessive-compulsive symptoms do not (see Table 2).

DISCUSSION

No general effects of initial enhancement or rebound were observed. On the other hand the hypothesis of an interaction between obsessive-compulsive symptoms and initial instruction received support in the first period. There was thus a tendency of people high on obsessive-compulsive symptoms to show more thoughts about the target when they were to suppress such thoughts, whereas the reverse was true for people low on obsessive+ompulsive symptoms. There are at least two different explanations of why consequences of thought suppression vary, as a function of obsessive-compulsive symptoms. Primarily, the thoughts people high on obsessivecompulsive symptoms try to suppress may be higher in emotional intensity or better rehearsed, and secondly these people’s attempts to get rid of thoughts may be less efficient. The present study does not allow to decide between these interpretations of the results. However, the interaction between initial instruction and level of obsessive-compulsive symptoms, even though weakened, did not

625

Obsessive-compulsive

disappear when thought characteristics were taken into account. This indicates, that differences between thoughts might at least not tell the whole story. Instead differences in suppression strategies between people high and low on obsessivexompulsive symptoms may be at least partly responsible for the results. Ss high on obsessive-compulsive symptoms were found to try harder not to think about the target, when they received the suppression instruction. This coincides with an increase in target thoughts in these Ss and gives support to the notion that it is the effort put into suppressing thoughts that makes them more intrusive. In order to come to grips with such issues, it is necessary in future research on thought suppression and obsessive-compulsive symptoms to address more directly the Ss’ perception of the situation and the precise strategies they implement to tackle intrusive thoughts. Ackno~ledRrments-This research was supported of the University of Iceland to the first author.

by grants from the Icelandic

Science Foundation

and the Research

Fund

REFERENCES Clark, D. ., Winton, E. & Thynn, L. (1993) A further experimental investigation of thought suppression. Behariour Research and Therapy, 31, 207-210. Freeston, M. H., Ladouceur, R., Thibodeau, N. & Gagnon, F. (1991) Cognitive Intrusions in a non-clinical population. I. Response style, subjective experience, and appraisal. Behaviour Research and Therapy, 29, 585-597. Kelly, A. E. & Kahn, J. H. (1994) Effects of suppression of personal intrusive thoughts. Juurnal c?fPersonalifF and Social P sychologv, 66, 998- 1006. Merckelbach, H., Muris, P., van den Hout, M. & de Jong, P. (1991) Rebound effects of thought suppression: Instruction dependent. Behavioural Psychotherapy, 18, 25 I-258. Muris, P. Merckelbach, H., van den Hout, M. & de Jong, P. (1992) Suppression of emotional and neutral material. Behnviou~ Research and Therap.v, 6, 639-642. Rachman S. & Hodgson, R. J. (I 980) Obsessions and compulsions. Englewood Cliffs, NJ: Prentice-Hall, Inc. Salkovskis, P. M. & Campell, P. (1994) Thought suppression induces intrusion in naturally occurring negative intrusive thoughts. Behaviour Research and Therapy, 32, 1-8. Smari, J., Bjarnason, B. & Thorleifsson, K. (in preparation) Psychometric assessment of two obsession-compulsion inventories. Smari, J., Sigurjonsdbttir, H. & S2emundsdottir. H. (I 994) Thought suppression and obsession