discussion

NegativeCognitionsinEmotionalProblemsFollowingRomanticRelationshipBreak-Ups.pdf

N e g a t i v e c o g n i t i o n s i n e m o t i o n a l p r o b l e m s f o l l o w i n g r o m a n t i c r e l a t i o n s h i p b r e a k - u p s

Paul A. Boelen1,*,† and Albert Reijntjes2

1 Department of Clinical and Health Psychology, Utrecht University, Utrecht, The Netherlands 2 Department of Psychosocial Development in Context, Utrecht University, Utrecht, The Netherlands

* Correspondence to: Paul A. Boelen, Department of Clinical and Health Psychology, Utrecht University, PO Box 80140, 3508 TC Utrecht, The Netherlands. † E-mail: [email protected]

S t r e s s a n d H e a l t h Stress and Health 25: 11–19 (2009)

Published online 4 August 2008 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/smi.1219 Received 18 October 2007; Accepted 20 May 2008

Summary This study examined the role of negative cognitions in emotional problems following relationship dissolution. Seventy-nine undergraduate students who experienced a relationship break-up com- pleted measures of break-up related complicated grief, depression and anxiety, together with an adjusted version of the Grief Cognitions Questionnaire (GCQ) that assesses four types of global negative beliefs, negative cognitions about self-blame and the responses of others, and three types of negative interpretations of one’s own grief reactions. Results showed that all cognitive variables tapped by the GCQ were signifi cantly associated with complicated grief, and—except for global beliefs about life—with depression and anxiety. Most of these associations remained signifi cant when controlling for the infl uence of initiator status, variables linked with the ended relationship (e.g. duration) and personality factors (e.g. attachment anxiety, neuroticism). Catastrophic mis- interpretations about one’s own reactions, global negative beliefs about the self and cognitions refl ecting self-blame were the strongest cognitive correlates of break-up related emotional prob- lems. Overall, the fi ndings are in keeping with cognitive models of trauma and loss, and suggest that changing negative cognitions could be a useful intervention for those who fail to recover from a relationship break-up. Copyright © 2008 John Wiley & Sons, Ltd.

Key Words attachment; break-up-of-close-relationships; distress; negative-cognitions; personality

factor for the development of emotional problems (Monroe, Rohde, Seeley, & Lewinsohn, 1999). Yet, as with many other stressful life events (Bonanno, 2004) not every one develops such problems in reaction to this event. In some people, relationship break-ups even give rise to personal growth (Tashiro & Frazier, 2003). What vari- ables are involved in individual differences in reactions to relationship break-up? To date, rela- tively much research in this area has focused on

Introduction

The dissolution of a close relationship can be an extremely upsetting event. It is a prospective risk

P. A. Boelen and A. Reijntjes

static risk factors such as demographic variables (e.g. gender), characteristics of the ended relation- ship (e.g. duration) and personality factors (e.g. attachment style, neuroticism) (Sbarra & Emery, 2005; Sprecher, Felmlee, Metts, Fehr, & Vanni, 1998; Tashiro & Frazier, 2003). Less attention has been paid to the role of potentially modifi able psychological mechanisms involved in recovery from relationship break-up. The present study sought to enhance knowledge about such mecha- nisms, by examining the role of negative cogni- tions in emotional problems following break-up.

Our examination was guided by the view that much can be learned from cognitive models of emotional problems after loss and trauma (Brewin & Holmes, 2003, Dalgleish 2004). Central to these models is the notion that individual variations in cognitions that come to dominate thinking pat- terns after such events (including core beliefs as well as automatic thoughts) account for differ- ences in emotional reactions to such events. Spe- cifi cally, cognitive models of post-traumatic stress disorder (PTSD) postulate that a traumatic event may shatter pre-existing positive core beliefs about the self and the world, or confi rm pre-existing negative ones, as a result of which negative beliefs that ‘the self is worthless’ and ‘the world is danger- ous’ may come to dominate thinking patterns, causing PTSD (Foa & Rothbaum, 1998). Apart from such basic and global core beliefs about the self and the world, other, more superfi cial negative cognitions (also called automatic cognitions) have been associated with PTSD as well, including neg- ative cognitions about self-blame (‘I should have prevented the event’), the responses of others (‘Nobody cares about me’) and catastrophic mis- interpretations of one’s reactions to the event (‘The vivid recollections are signs that I am going mad’) (Ehlers & Clark, 2000). Several studies have con- fi rmed the importance of these different cognitions in PTSD (Brewin & Holmes, 2003).

Boelen, van den Bout and van den Hout (2003) have postulated that the aforementioned negative cognitions are also involved in complicated grief—a debilitating condition that can develop after the death of a close relative (Prigerson & Jacobs, 2001). These authors constructed the Grief Cognitions Questionnaire (GCQ) that taps a wide array of negative loss-related cognitions, including four types of global negative beliefs (similar to core beliefs as defi ned in cognitive therapy) and negative cognitions about self-blame, negative cognitions about the responses of others and three types of negative interpretations of

one’s own grief reactions (similar to automatic thoughts as defi ned in cognitive therapy). Research among bereaved adults confi rmed that endorse- ment of these cognitions was positively associated with more severe and complicated grief, depres- sion and anxiety (Boelen et al., 2003; Boelen & Lensvelt-Mulders, 2005). Global negative beliefs about the self, life and the future, and catastrophic misinterpretations of grief-reactions were found to be the strongest cognitive correlates of these emotional problems. A subsequent prospective study revealed that these four types of cognitions predicted the persistence of grieving problems across 2 years post-loss (Boelen, van den Bout, & van den Hout, 2006).

Building and expanding on cognitive approaches to loss and trauma, the current study was designed to examine negative cognitions in emotional problems after relationship break-ups in under- graduate students. An adjusted version of the GCQ was used to assess negative cognitions. Break-up related complicated grief, depression and anxiety served as the indices of psychological malfunctioning. The assessment of complicated grief represents an extension of earlier studies that typically confi ne their examination of the emotional consequences of relationship break-up to depression, anxiety and general mental health (e.g. Chung et al., 2002; Sprecher et al., 1998).

The aim of the present study, conducted in The Netherlands, was threefold. First, we examined the role of several demographic variables (age, gender, current dating status), variables related to the relationship break-up (duration of the rela- tionship, time since break-up, degree of commit- ment, initiator status and suddenness of break-up) and personality factors (neuroticism, attachment anxiety, attachment avoidance). Based on previ- ous research (e.g. Chung et al., 2002; Davis, Shaver, & Vernon, 2003; Sprecher et al., 1998; Tashiro & Frazier, 2003), it could be expected that current dating status, most of the relation- ship variables, and neuroticism and attachment anxiety would be linked with the severity of break-up related emotional problems. Second, we examined the association between cognitive vari- ables and these emotional problems. Based on research on trauma and loss (Boelen et al., 2003; Ehlers, Mayou, & Bryant, 1998), it was predicted that all cognitive variables tapped by the GCQ would be signifi cantly linked with emotional problems. Moreover, we expected that these cog- nitions would continue to predict variance in emotional problems, over and above the variance

Cognitions in problems after break-ups

explained by demographic, relationship and per- sonality variables. Finally, we sought to deter- mine which cognitive variables explained most variance in emotional problems, after controlling for the infl uence of relevant background variables and the shared variance between the cognitive variables. Based on fi ndings among bereaved indi- viduals (Boelen et al., 2003), it was expected that negative global beliefs about the self, life and the future, and catastrophic misinterpretations of one’s reactions would be among the most impor- tant cognitive correlates of break-up related emotional problems.

Method

Participants and procedures

Data were available from 79 undergraduate stu- dents from Utrecht University. They were all recruited via posters in university buildings and an advertisement on the university Internet site. Solicitations stated that the aim of the study was to learn about emotional problems after a rela- tionship break-up. Criteria for inclusion were: being a student between 18 and 30 years of age and having experienced the break-up of a roman- tic relationship that lasted at least 2 months, within the past 7 years.

The mean age of the sample was 21.5 years [standard deviation (SD) = 2.5], 66 participants (83.5 per cent) were female and 35 participants (44.3 per cent) were dating someone new. On average, the duration of the ended relationship was Mean (M) = 16.6 months (SD = 17.5) and M = 17.5 months (SD = 15.5) had passed since the relationship ended. On a three-point ‘commit- ment’ scale, 11 (13.9 per cent) rated that they had been ‘somewhat’ committed to the previous rela- tionship, 34 (43 per cent) rated ‘strongly’ and 34 (43 per cent) rated ‘very strongly’. Thirty-eight participants (48.1 per cent) had initiated the break-up, 29 (36.7 per cent) were non-initiators and 12 (15.2 per cent) reported that initiation of break-up was mutual. Sixty-one participants (77.2 per cent) experienced the break-up as ‘sudden’, 18 (22.8 per cent) as non-sudden.

All participants completed the questionnaires in return for course credits. Participants com- pleted the questionnaires in group-testing ses- sions. None of the 79 participants terminated their participation before they had completed all questionnaires.

Measures

Inventory of complicated grief revised (ICG- R). The ICG-R is a self-report measure con- structed by Prigerson and Jacobs (2001) to assess symptoms of complicated grief. The 29-item Dutch version has good psychometric properties (Boelen, van den Bout, de Keijser, & Hoijtink, 2003). In the present study, break-up related complicated grief was assessed using an adjusted version of the ICG-R. Specifi cally, items were altered, such that references to the death of the respondents’ relative were replaced by the ending of the relationship. For instance, the item ‘I feel I have trouble accepting the death’ was changed into ‘I feel I have trouble accepting that this rela- tionship is over’. For one item (i.e. ‘I feel that it is unfair that I should live when he/she died’), this adjustment could not be made and this item was dropped in the adjusted version. The 28 remain- ing items (α = 0.94) were rated on fi ve-point scales ranging from 1 (not at all) to 5 (very much).

Symptom checklist-90 (SCL-90) depression and anxiety scale. The 16-item depression subscale and 10-item anxiety subscale of the Symptom Checklist-90 (SCL-90) (Derogatis, 1983; Dutch version by Arrindell & Ettema, 2003) were used to assess symptoms of depression and anxiety. Respondents rate how often they experienced the symptoms described in the items during the last week, on fi ve-point scales ranging from 1 (never) to 5 (always). In this sample, the depression and anxiety scales had α’s of 0.89 and 0.83, respectively.

Grief cognitions questionnaire (GCQ). The GCQ, developed by Boelen et al. (2003), is a 38- item measure of negative cognitions involved in emotional problems after loss. It contains nine subscales: four subscales tapping global negative beliefs about the Self (six items), the World (four items), Life (four items) and the Future (fi ve items); a fi fth subscale representing cognitions about Self-Blame over having caused the loss or having done things wrongly in the relationship with the lost person (fi ve items); a sixth subscale assessing negative cognitions about the way Others responded after the loss; and three sub- scales tapping negative cognitions about one’s own reactions to the loss, specifi cally negative cognitions about the Appropriateness of one’s reactions (four items), cognitions refl ecting the

P. A. Boelen and A. Reijntjes

importance of Cherishing the Pain of the loss (three items) and Catastrophic Misinterpretation of one’s feelings as signs of impending mental insanity (fi ve items).

As with the ICG-R, items were altered such that references to the death of a relative were replaced by relationship break-up. For instance, the item ‘Since he/she died, I think I am worthless’ (subscale Self) was changed into ‘Since the rela- tionship ended, I think I am worthless’. Respon- dents rated their agreement with each item on six-point scales ranging from 0 (disagree strongly) to 5 (agree strongly). Psychometric properties of the original (bereavement-related) version are adequate, and confi rmatory factor analysis has shown that the GCQ taps nine distinguishable, but related categories of cognitions (Boelen & Lensvelt-Mulders, 2005). In this study, correla- tions among the GCQ subscales ranged from r = 0.09 (p = 0.45) to r = 0.69 (p < 0.001), which confi rms that they represent relatively indepen- dent constructs.

In the present sample, the α of the GCQ total scale was 0.91. The α’s of the GCQ subscales were moderate to high (Self, α = 0.72; World, α = 0.71; Life, α = 0.62; Future, α = 0.48; Self- Blame, α = 0.62; Others, α = 0.81; Appropriate- ness, α = 0.51, Catastrophic Misinterpretations, α = 0.86), with the exception of the α of the subscale Cherishing the Pain that was low (α = 0.21).

Neuroticism scale from the Eysenck Personal- ity Questionnaire revised and short scale version (EPQ-RRS). The 12-item Neuroticism subscale from the EPQ-RRS (Eysenck & Eysenck, 1991; Dutch version by Sanderman, Arrindell, Ranchor, Eysenck, & Eysenck 1995) was used to assess neuroticism. Respondents indicate their agreement with 12 statements, using a forced- choice response format (yes versus no). The α was 0.73.

Revised experiences in close relationships (ECR-r). The shortened version of the ECR-r, originally developed by Fraley, Waller and Brennan (2000), was used to measure attachment anxiety (i.e. a person’s predisposition towards anxiety and vigilance about rejection and aban- donment) and attachment avoidance (i.e. a per- son’s discomfort with closeness and dependency or a reluctance to be intimate with others). Attach- ment anxiety was tapped by a fi ve-item scale and attachment avoidance by a six-item scale. Respon-

dents rate their agreement with statements on seven-point scales ranging from 0 (strongly dis- agree) to 6 (strongly agree). In the present sample, the attachment anxiety and attachment avoidance scales had α’s of 0.79 and 0.78, respectively.

Results

Descriptive statistics

Mean scores of participants on the (28-item) adjusted ICG-R, the SCL depression scale and SCL anxiety scale are shown in Table I. Scores on the SCL depression and anxiety scales were sig- nifi cantly higher than mean scores of a represen- tative non-clinical sample from the Dutch population (Arrindell & Ettema, 2003; see Table I). Table I also shows scores of the current par- ticipants on the GCQ, together with scores of 531 bereaved individuals included in the study of Boelen and Lensvelt-Mulders (2005). Compari- son of these scores using t-tests, outcomes of which are also shown in Table I, showed that mean GCQ subscale and total scores in the current sample were all signifi cantly lower than the scores found in the bereaved sample. The exception was that the current sample scored higher on the GCQ subscale Self-Blame. Scores on the three symptom measures and the GCQ subscales, and GCQ total score were positively skewed and were, therefore, log-transformed in all analyses described below.

Relationship of demographic, relationship and personality variables with symptoms

Relationships of the symptom measures with demographic variables (age, gender, current dating status), relationship variables (duration of relationship, time since break-up, degree of commitment, initiator status and suddenness of break-up) and personality variables (neuroticism, attachment anxiety, attachment avoidance) were examined using Pearson correlations, t-tests and analyses of variance.

Demographic variables age and gender were not related to symptom measures. Current dating status was associated with all three symptom measures. Compared with those who were dating someone new, participants who were still single had higher ICG-R scores [t(77) = 4.00], SCL depression scores [t(77) = 2.34] and SCL anxiety scores [t(77) = 2.04, ps < 0.05].

Cognitions in problems after break-ups

Several relationship variables were associated with complicated grief. Complicated grief was inversely related with time since break-up (r = −0.48, p < 0.001). There was an effect of ‘degree of commitment’ [F(2,78) = 7.15, p < 0.01], such that those who were ‘somewhat’ committed to the relationship had lower ICG-R scores than those who were ‘strongly’ or ‘very strongly’ com- mitted (ps < 0.05). There was an effect of initiator status [F(2,78) = 5.06, p < 0.01], such that ‘ini- tiators’ had lower scores than ‘non-initiators’ (p < 0.05). Moreover, sudden break-ups coin- cided with higher ICG-R scores than non-sudden break-ups [t(77) = 2.42, p < 0.05]. Relationship variables were not associated with depression and anxiety.

With respect to personality variables, compli- cated grief was associated with attachment anxiety (r = 0.35, p < 0.01), depression with attachment anxiety (r = 0.23, p < 0.05) and neuroticism (r = 0.35, p < 0.01), and anxiety with neuroticism (r = 0.29, p < 0.01).

Relationship of cognitive variables with symptoms

Table II shows Pearson correlations between the nine subscales and total score of the GCQ and the three symptom measures. To control for Type I error, we adjusted the alpha-level, using the false discovery rate procedure developed by Benjamini

and Hochberg (1995).1 All GCQ subscales and the GCQ total score were signifi cantly related with break-up related complicated grief. More- over, except for one subscale (i.e. Life), higher scores on all the GCQ subscales and its total score were signifi cantly related to higher levels of depression and anxiety.

Regression analyses were used to examine if the GCQ subscales and total score continued to be asso- ciated with the three symptom measures when con- trolling for the effect of relevant demographic, relationship and personality variables (i.e. those that were related to symptoms as described in the previ- ous section). Again, we adjusted the alpha-level using the procedure of Benjamini and Hochberg (1995).2

1 In controlling the false discovery rate, we reasoned that, for each of the three symptom measures (ICG-R, SCL depression scale and SCL anxiety scale) consecu- tively treated as dependent variables, we tested a series of ten hypotheses (one for each of the nine GCQ sub- scale scores and the GCQ total score). For instance, there were 10 correlations and p values pertaining to the associations of the GCQ subscale and the GCQ total score with the ICG-R. 2 In controlling the false discovery rate, we reasoned that, for each of the three symptom measures (ICG-R, SCL depression scale and SCL anxiety scale) consecu- tively treated as dependent variables, we again tested ten hypotheses (one for each of the nine GCQ subscale scores and the GCQ total score).

Table I. Mean scores on measures of complicated grief, depression, anxiety, and cognitive variables in present sample and reference groups.

Present sample M (SD) Reference group M (SD) Difference test t

Symptom measures ICG-R 49.33 (16.46) — — SCL-depression 27.35 (9.39) 21.58 (7.56) 5.46* SCL-anxiety 15.73 (5.39) 12.76 (4.41) 4.90* Cognitive variables Self 3.98 (4.29) 7.25 (7.69) −6.72* World 1.46 (2.53) 5.58 (5.84) −15.31* Life 2.29 (3.02) 5.35 (6.11) −8.94* Future 2.53 (3.01) 7.72 (7.13) −15.23* Self-blame 7.65 (5.07) 5.37 (6.40) 3.98* Others 1.84 (2.60) 5.51 (4.93) −12.43* Appropriateness 2.06 (2.55) 5.01 (5.67) −10.18* Cherish Grief 1.15 (1.70) 3.88 (4.26) −14.18* Catastrophic interpretations 2.99 (4.18) 7.75 (6.36) −10.07* GCQ total score 24.47 (19.94) 53.70 (41.05) −12.94*

* p < 0.001. GCQ = Grief cognitions questionnaire. ICG-R = Inventory of complicated grief revised. SCL = Symptom checklist.

P. A. Boelen and A. Reijntjes

Table II shows the standardized regression coeffi - cients representing the association of each GCQ sub- scale and the GCQ total score with the three symptom measures, controlling relevant background variables. As can be seen, all GCQ subscales and its total score continued to predict variance in complicated grief. Moreover, the GCQ total score and six GCQ sub- scales remained signifi cantly associated with depres- sion and anxiety.

Stepwise regression analyses

Finally, we examined which cognitive variables explained most variance in symptom severity, after controlling for the infl uence of relevant

Table II. Associations of cognitive variables with complicated grief, depression, and anxiety.

Complicated grief

Depression Anxiety

r β r β r β

Self 0.68 0.43 0.43 0.32 0.46 0.38 World 0.41 0.29 0.31 0.22 0.26 0.18 Life 0.46 0.27 0.16 −0.08 0.16 −0.02 Future 0.55 0.36 0.36 0.21 0.27 0.13 Self-blame 0.30 0.21 0.39 0.41 0.36 0.38 Others 0.57 0.42 0.36 0.27 0.31 0.24 Appropriateness 0.55 0.32 0.41 0.38 0.47 0.43 Cherish grief 0.53 0.28 0.31 0.23 0.42 0.37 Catastrophic Interpretations 0.70 0.54 0.49 0.41 0.48 0.42 GCQ total score 0.77 0.56 0.56 0.49 0.54 0.48

Note: Underlined correlations and β’s are signifi cant at p < 0.05 controlling the false discovery rate. GCQ = Grief cognitions questionnaire.

Table III. Summary of hierarchical regression analyses for variables predicting complicated grief, depression, and anxiety.

Variables entered (in order) R2 change F change

Complicated grief Block 1 (enter) Background variables 0.53 9.43*** Block 2 (stepwise) Catastrophic interpretations 0.23 62.01***

Self 0.03 7.31** Depression Block 1 (enter) Background variables 0.17 5.03** Block 2 (stepwise) Self-blame 0.17 18.36***

Catastrophic interpretations 0.10 12.12** Anxiety Block 1 (enter) Background variables 0.12 4.99** Block 2 (stepwise) Catastrophic interpretations 0.16 16.36***

Self-blame 0.09 10.01**

* p < 0.05; ** p < 0.01; *** p < 0.001.

demographic, relationship and personality vari- ables, and the shared variance between the cogni- tive variables. To this end, three hierarchical regression analyses were run, in which compli- cated grief, depression and anxiety were the dependent variables. Relevant demographic, rela- tionship and personality variables were entered in block 1, and the nine GCQ subscales were included in a stepwise analysis in block 2.

Table III summarizes the outcomes. In the anal- ysis with the ICG-R, catastrophic misinterpreta- tions and global negative beliefs about the self explained 26 per cent of variance in complicated grief, over and above the 53 per cent explained by current dating status, time since break-up, commitment, initiator status, suddenness of

Cognitions in problems after break-ups

break-up, and attachment anxiety. The analysis with depression showed that negative cognitions about self-blame and catastrophic misinterpreta- tions explained 27 per cent of variance in depres- sion, over and above the 17 per cent explained by current dating status, attachment anxiety and neuroticism. Finally, the analysis with anxiety showed that catastrophic misinterpretations and negative cognitions about self-blame explained 25 per cent of variance in anxiety over and above the 12 per cent of variance explained by current dating status and neuroticism.3

Additional analyses with adjusted symptom measures

The adjusted ICG-R contained six ‘cognitive’ grief-reactions that showed conceptual overlap with items from the GCQ. Similarly, the SCL depression scale included three ‘cognitive’ depres- sion symptoms.4 Hence, it was possible that this overlap in content infl ated associations of the ICG-R and SCL depression scale with the GCQ. To examine this possibility, we reran all analyses using shortened versions of the ICG-R and SCL depression scale from which these cognitive items were removed. Results found with these short- ened measures were almost equivalent to those obtained with the full-length versions. There were, however, two changes. First, the associa- tion of depression with the GCQ subscale World after controlling for relevant demographic, rela- tionship and personality variables that was not

signifi cant initially (see Table II), became signifi - cant when using this shortened depression scale (β = 0.24, p = 0.03). Secondly, in the stepwise regression with depression serving as a dependent variable, the order of variables entering the equa- tion was reversed, with catastrophic misinterpre- tations now entering fi rst (∆R2 = 0.17, ∆F = 18.70, p < 0.001), followed by cognitions about self- blame (∆R2 = 0.10, ∆F = 13.40, p < 0.001).

Discussion

In the current study, we examined the association between negative cognitions and emotional prob- lems after a relationship break-up, in a sample of undergraduate students—also taking into account the role of demographic, relationship and person- ality variables. Notably, mean scores on depres- sion and anxiety measures in the current study group were signifi cantly higher than scores in the normal Dutch population. This is in line with previous work showing that romantic break-ups in adolescence and young adulthood can be accompanied by serious mental health problems (Davis et al., 2003; Monroe et al., 1999).

Examination of the impact of demographic, relationship and personality variables showed that current dating status (but not gender and age) was signifi cantly associated with all three symptom measures. This is consistent with earlier fi ndings that dating someone new is associated with less distress (e.g. Tashiro & Frazier, 2003). Also con- sistent with earlier fi ndings (Sprecher et al., 1998), complicated grief was associated with several rela- tionship variables (e.g. time since break-up, initia- tor status). Somewhat unexpected, was that relationship variables did not affect depression and anxiety. This suggests that these relationship variables have a differential impact on different emotional problems after break-up. Yet, the absence of an association with depression and anxiety symptoms may also be due to the fact that we only assessed the intensity of these symptoms in the preceding week rather than during a more extended period of time in the aftermath of the break-up. With respect to personality factors, attachment anxiety affected complicated grief and depression, and neuroticism was linked with depression and anxiety. These results add to earlier research showing that these variables are involved in break-up related distress (Chung et al., 2002; Davis et al., 2003; Sbarra & Emery, 2005; Spre- cher et al., 1998; Tashiro & Frazier, 2003).

3 As an additional step, we reran all regression analyses, leaving out the GCQ subscale ‘Cherishing the Pain of the Loss’, that had a low internal consistency (α = 0.21). This did not change the outcomes of the regres- sion analyses. 4 The six ‘cognitive’ grief symptoms included in the Dutch ICG-R were ‘Ever since ____ died it is hard for me to trust people’, ‘I feel that life is empty or meaning- less without ____’, ‘I feel like the future holds no purpose or meaning since ____ died’, ‘I feel unable to imagine life being fulfi lling without ____’, ‘I feel that the death of ____ has changed my view of the world’ and ‘I have lost my sense of security, safety or control since the death of ____’. The three ‘cognitive’ depres- sion symptoms from the SCL depression scale were: ‘Blaming yourself for things’, ‘Feeling hopeless about the future’ and ‘Feelings of worthlessness’.

P. A. Boelen and A. Reijntjes

Negative cognitions were assessed using an adjusted version of the GCQ (Boelen & Lensvelt- Mulders, 2005). The GCQ taps a wide range of negative cognitions, including four types of global negative beliefs, negative cognitions about self- blame and the responses of others, and three types of negative interpretations of one’s own grief reactions. As expected, all cognitions tapped by the GCQ were associated with complicated grief and all but global negative beliefs about life with depression and anxiety. Importantly, most of the cognitive variables continued to predict emotional problems after controlling for the effect of relevant demographic, relationship and per- sonality variables. Thus, even when taking into account the impact of being able to reinvest in a new relationship and personality vulnerabilities of attachment anxiety and neuroticism, negative cognitions revolving around the relationship break-up contributed considerably to the distress experienced in its aftermath. These fi ndings link up with previous research showing the pivotal role of negative cognitions in recovery from trauma (e.g. Ehlers et al., 1998) and loss (Boelen et al., 2003, 2006).

We also investigated which cognitions were most important (i.e. showed the strongest links with emotional problems) when taking into account relevant demographic, relationship and personality variables, and the shared variance among cognitions. Consistent with our expecta- tions and studies on loss (Boelen et al., 2006) and trauma (Ehlers et al., 1998), catastrophic misin- terpretations about one’s own reactions to the break-up were among the strongest cognitive cor- relates of complicated grief, depression and anxiety. Also as expected, negative beliefs about the self were key cognitive correlates of compli- cated grief. This suggests that having diffi culties sustaining a positive sense of self is an important variable impeding recovery from a relationship break-up. The expectation that global negative beliefs about life and the future would prove to be central to complicated grief was not confi rmed. Notably, cognitions about self-blame, but none of the global negative beliefs, emerged as key cognitive correlates of depression and anxiety. This suggests that it is the tendency to blame oneself for one’s role in the break-up or one’s behaviour during the relationship more than it is the presence of global, core negative beliefs about the self, life or the future that put persons at risk for depression and anxiety following a break-up.

Several limitations of the present study deserve mention. First, the relatively small sample size may have reduced the statistical power necessary to detect some signifi cant associations. Second, the cross-sectional design of this study does not allow drawing conclusions about the direction of causality between cognitive variables and emo- tional problems following break-up. Prospective longitudinal studies are needed to gain more insight into the direction of causality between these variables. It would be interesting for these studies to use methods based in structural equa- tion modelling rather than regression analysis, to allow for a better understanding of the direct and indirect relationships of cognitive and background variables with symptoms after break-up. Third, generalization of the current fi ndings to other groups should be done cautiously, given that par- ticipants were all undergraduate students, were all young, and likely had fewer, shorter and less intense relationships than older victims of break- ups most likely would have had. Fourth, as this study relied on self-report measures, we can not rule out that shared method variance affected some of the associations between dependent and independent variables. Finally, adaptation for relationship break-up of the GCQ and ICG-R may have compromised psychometric properties of these scales. Indeed, internal consistencies of the adjusted GCQ subscales were lower than those of the original bereavement-related version (Boelen & Lensvelt-Mulders, 2005). Thus, it is important for future studies to replicate the current fi ndings using measures specifi cally designed and validated for the assessment of break-up related cognitions and emotions.

Notwithstanding these limitations, the current fi ndings may have both theoretical and clinical implications. Theoretically, the fi ndings provide evidence that focusing on maladaptive cognitions is a fruitful approach in research on relationship break-up. More generally, fi ndings suggest that a common core process of negative thinking runs through the process of coping with all kinds of stressors. With regard to clinical implications, the present results suggest that cognitive therapeutic interventions that have been successfully applied in the treatment of PTSD (Ehlers, Clark, Hack- mann, McManus, & Fennell, 2005) and compli- cated grief (Boelen, de Keijser, van den Hout, & van den Bout, 2007) may prove useful in counselling or therapy for those suffering from psychological problems after relationship break-up.

Cognitions in problems after break-ups

Acknowledgments

This research was supported by an Innovative Research Incentive Veni Grant (451-06-011) from the Nether- lands Organization for Scientifi c Research (NWO) awarded to the fi rst author. The authors thank Corine Geurts, Coby de Vries, Marlinde Stam and Marije Ver- meulen for their help in collecting data for this study.

References

Arrindell, W.A., & Ettema, J. (2003). Symptom Checklist. Handleiding bij een multidimensionele psychopathologie- indicator [Symptom Checklist. Manual for a multidimen- sional indicator of psychopathology]. Lisse: Swets & Zeitlinger.

Benjamini, Y., & Hochberg, Y. (1995). Controlling the False Discovery Rate: a practical and powerful approach to mul- tiple testing. Journal of the Royal Statistical Society, 57, 289–300. Series B

Boelen, P.A., van den Bout, J., & van den Hout, M.A. (2003). The role of cognitive variables in psychological functioning after the death of a fi rst degree relative. Behaviour Research and Therapy, 41, 1123–1136.

Boelen, P.A., van den Bout, J., & van den Hout, M.A. (2006). Negative cognitions and avoidance in emotional problems after bereavement: A prospective study. Behaviour Research and Therapy, 44, 1657–1672.

Boelen, P.A., van den Bout, J., de Keijser, J., & Hoijtink, H. (2003). Reliability and validity of the Dutch version of the Inventory of Traumatic Grief. Death Studies, 27, 227– 249.

Boelen, P.A., de Keijser, J., van den Hout, M.A., & van den Bout, J. (2007). Treatment of Complicated Grief: A com- parison between cognitive behavioral therapy and support- ive counseling. Journal of Consulting and Clinical Psychology, 75, 277–284.

Boelen, P.A., & Lensvelt-Mulders, G.J.L.M. (2005). Psycho- metric properties of the Grief Cognitions Questionnaire (GCQ). Journal of Psychopathology and Behavioral Assess- ment, 27, 291–303.

Bonanno, G.A. (2004). Loss, trauma, and human resilience. Have we underestimated the human capacity to thrive after extremely aversive events? American Psychologist, 59, 20– 28.

Brewin, C.R., & Holmes, E.A. (2003). Psychological theories of posttraumatic stress disorder. Clinical Psychology Review, 23, 339–376.

Chung, M.C., Farmer, S., Grant, K., Newton, R., Payne, S., Perry, M., Saunders, J., Smith, C., & Stone, N. (2002). Self-esteem, personality and posttraumatic stress symptoms following the dissolution of a dating relationship. Stress and Health, 18, 83–90.

Dalgleish, T. (2004). Cognitive approaches to Posttraumatic Stress Disorder: The evolution of multirepresentational theorising. Psychological Bulletin, 130, 228–260.

Davis, D., Shaver, P.R., & Vernon, M.V. (2003). Physical, emotional, and behavioral reactions to breaking up. Personality and Social Psychology Bulletin, 29, 871– 884.

Derogatis, L.R. (1983). SCL-90-R: Administration, Scoring and procedures. Manual II (Second edition). Towson, Maryland: Clinical Psychometric Research.

Ehlers, A., & Clark, D.M. (2000). A cognitive model of post- traumatic stress disorder. Behaviour Research and Therapy, 38, 319–345.

Ehlers, A., Clark, D.M., Hackmann, A., McManus, F., & Fennell, M. (2005). Cognitive therapy for PTSD: Develop- ment and evaluation. Behaviour Research and Therapy, 43, 413–431.

Ehlers, A., Mayou, R.A., & Bryant, B. (1998). Psychological predictors of chronic Posttraumatic Stress Disorder after motor vehicle accidents. Journal of Abnormal Psychology, 107, 508–519.

Eysenck, H.J., & Eysenck, S.B.G. (1991). Manual of the Eysenck Personality scales (EPS Adult). London: Hodder & Stoughton.

Foa, E.B., & Rothbaum, B.O. (1998). Treating the trauma of rape: Cognitive behavioral therapy for PTSD. New York: Guilford Press.

Fraley, R.C., Waller, N.G., & Brennan, K.A. (2000). An item response theory analysis of self-report measures of adult attachment. Journal of Personality and Social Psychology, 78, 350–365.

Monroe, S.M., Rohde, P., Seeley, J.R., & Lewinsohn, P.M. (1999). Life events and depression in adolescence: Relation- ship loss as a prospective risk factor for fi rst onset of major depressive disorder. Journal of Abnormal Psychology, 108, 606–614.

Prigerson, H.G., & Jacobs, S.C. (2001). Traumatic grief as a distinct disorder: a rationale, consensus criteria, and a pre- liminary empirical test. In M.S. Stroebe, R.O. Hansson, W. Stroebe, & H.A.W. Schut (Eds), Handbook of bereavement research. Consequences, coping, and care (pp. 613–647). Washington DC: APA Press.

Sanderman, R., Arrindell, W.A., Ranchor, A., Eysenck, H.J., & Eysenck, S.B.G. (1995). Het meten van persoonli- jkheidskenmerken met de Eysenck Personality Question- naire:Een handleiding [Measuring personality characteristics with the Eysenck Personality Questionnaire: A Manual]. Groningen: Noordelijk Centrum voor Gezondheids- vraagstukken.

Sbarra, D.A., & Emery, R.E. (2005). The emotional sequelae of nonmarital relationship dissolution: Analysis of change and intraindividual variability over time. Personal Relation- ships, 12, 213–232.

Sprecher, S., Felmlee, D., Metts, S., Fehr, B., & Vanni, D. (1998). Factors associated with distress following the breakup of a close relationship. Journal of Social and Per- sonal Relationships, 15, 791–809.

Tashiro, T., & Frazier, P. (2003). I’ll never be in a relationship like that again: Personal growth following romantic relationship breakups. Personal Relationships, 10, 113– 128.