Guilt, Distress and Ways of Coping with Guilty Thoughts in a Clinical Sample

A thesis submitted to the University of Manchester for the degree of Doctor of Clinical Psychology (ClinPsyD) in the Faculty of Medical and Human Sciences

2013

Lauren Pugh

School of Psychological Sciences

Division of Clinical Psychology

Table of Contents

List of Tables…………………………………………………………………………6

List of Figures………………………………………………………………………..7

Word Count…………………………………………………………………………..8

Abstract of Thesis…………………………………….………………………………9

Declaration………………………………………………………………………….10

Acknowledgements…………………………………………………………………13

Paper 1: Literature Review……………………………………………………….14

Preface………………………………………………………………………………15

Abstract……………………………………………………………………………..16

Introduction…………………………………………………………………………18

Method………………………………………………………………………………25

Search strategy……………………………………………………………....25

Inclusion and exclusion criteria……………………………………………..26

Results………………………………………………………………………………33

Overview of reviewed studies………………………………………………36

Is guilt related to PTSD symptomology?…………………………………...36

Model 1: Is guilt a causal process driving PTSD symptomology?…………41

Model 2: Is PTSD a causal process driving guilt?………………………….42

Model 3: Are guilt and PTSD symptomology causally unrelated despite their

co-occurrence?………………………………………………………………43

Model 4: Do confounding variables explain the relationship between guilt

and PTSD symptomology?………………………………………………….44

Summary of methodological considerations………………………………..47

Discussion…………………………………………………………………………..51

Implications for future research…………………………………………….53

Implications for clinical practice……………………………………………57

Conclusion…………………………………………………………………..59

References…………………………………………………………………………..60

Part 2: Empirical Study…………………………………………………………...70

Preface………………………………………………………………………………71

Abstract……………………………………………………………………………..72

Introduction…………………………………………………………………………74

Method………………………………………………………………………………80

Participants………………………………………………………………….80

Design……………………………………………………………………….80

Measures…………………………………………………………………….80

Procedure……………………………………………………………………84

Data analysis………………………………………………………………..84

Results………………………………………………………………………………86

Preliminary analysis………………………………………………………...86

Reliability of the GLAMS…………………………………………………..87

Concurrent validity of the GLAMS…………………………………………88

Construct validity of the GLAMS…………………………………………..90

Post-hoc analysis……………………………………………………………92

Discussion…………………………………………………………………………..95

Study limitations and future considerations……………………………….100

Clinical implications……………………………………………………….102

Conclusion…………………………………………………………………103

References…………………………………………………………………………105

Paper 3: Critical Appraisal………………………………………………………111

Overview…………………………………………………………………………..112

Aims of the research……………………………………………………….112

Rationale for Literature Review and Empirical Study…………………………….113

Reflections on Review and Research Process……………………………………..116

Methodological and Ethical Considerations……………………………………….123

Study design……………………………………………………………….123

Measures…………………………………………………………………...124

Recruitment and procedure………………………………………………..128

Sample……………………………………………………………………..130

Ethical issues………………………………………………………………131

Data analysis………………………………………………………………133

Implications for Future Research and Clinical Practice…………………………..136

References…………………………………………………………………………139

Appendix A: Clinical Psychology Review Author Guidance………………….….146

Appendix B: References for PTSD and Guilt Measures Cited (Not Discussed in

Review)……………………………………………………………………………156

Appendix C: Psychological Assessment Author Guidance……………………….158

Appendix D: Study Poster…………………………………………………………166

Appendix E: Eligibility Criteria Form…………………………………………….167

Appendix F: GLAMS……………………………………………………………...168

Appendix G: Content Validity Evaluation Form………………………………….173

Appendix H: Face Validity Evaluation Form……………………………………..177

Appendix I: GI (State Guilt)……..………………………………………………..180

Appendix J: CORE-10 ……………………………………………………………181

Appendix K: BriefCOPE………………………………………………………….182

Appendix L: TCQ…………………………………………………………………184

Appendix M: Demographic Form.………………………………………………..186

Appendix N: Participant Information Sheet………………………………………187

Appendix O: Consent Form……………………………………………………….190

Appendix P: Follow-up Covering Letter………………………………………….191

Appendix Q: Research Ethics Committee Approval Letter……………………....192

List of Tables

Paper 1: Literature Review

Table 1: Study characteristics and findings…………………………………………28

Paper 2: Empirical Study

Table 1: Frequency data for sample demographics…………………………………86

Table 2: Descriptive statistics for scales and subscales…………..…………………87

Table 3: Correlations between GLAMS Guilt Management subscales and

BriefCOPE, TCQ, GI and CORE-10……………….……………………………….91

Table 4: Inter-scale correlations between GLAMS subscales……………………...92

Table 5: Model summary of regression coefficients of GLAMS subscales on distress at follow-up……...………………...………………………………………………..94

List of Figures

Paper 1: Literature Review

Figure 1: Four contrasting conceptualisations of the association between guilt and PTSD symptomology………………………………………………….….24

Figure 2: A flow chart outlining the article identification and selection process…..26

Word Count

Thesis section Text (including References, Total abstract, tables & preface, figures) overview Thesis abstract 456 - 456

Paper 1 (Literature Review) 11,277 3,358 14,635

Paper 2 (Empirical Study) 7,972 2,066 10,038

Paper 3 (Critical Appraisal) 7,260 2,314 9,574

Total 26,965 7,738 34,703* *Excluding Appendices

The University of Manchester

Guilt, Distress and Ways of Coping with Guilty Thoughts in a Clinical Sample

Lauren Pugh Doctor of Clinical Psychology (ClinPsyD) 2013

Abstract of Thesis This thesis explores the role of guilt in post-traumatic stress disorder (PTSD) and ways of coping with guilt-related thoughts in a clinical sample. The thesis is presented as three papers that include a review of the literature, an empirical research study and critical appraisal of the research process. In the first paper, the author provides a systematic review of 27 studies to determine whether an association exists between guilt and symptoms of post-traumatic stress. Guilt remains an associated feature of PTSD; however, how these two constructs might be linked is not fully understood. Therefore the current review further evaluated the evidence for four competing models conceptualising the guilt-PTSD relationship. Overall, trauma-related guilt was positively related to PTSD symptomology even when controlling for depression. Guilt cognitions reflecting self-blame, perceived responsibility and wrongdoing were frequently associated with PTSD symptoms. Few studies found guilt was no longer related to PTSD symptomology when controlling for shame. Future studies ought to control for overlapping or confounding variables and further explore factors that may mediate the guilt-PTSD relationship such as coping. The second paper provided preliminary validation of a newly developed and unique measure of coping with guilty thoughts (GLAMS) in a clinical sample. A total of 67 participants from primary care services completed the GLAMS and measures of distress, guilt, coping and thought control. Eighteen completed the GLAMS and distress measure two weeks later. Overall the GLAMS evidenced moderate to high internal consistency and acceptable to good concurrent validity. Maladaptive subscales were found to be reliable over time. Higher self-punishment was related to greater guilt and distress and more mindful coping was related to a reduction in guilt supporting construct validity. Future research is required to test the stability of the GLAMS factor structure in a larger clinical sample. The GLAMS may have clinical utility in guiding psychological intervention towards more adaptive ways of coping with guilt. It may also provide a suitable outcome measure by monitoring the frequency in which clients engage in maladaptive ways of coping. The final paper provided a critical evaluation and reflection on the research process. Particular reference was made to the research rationale, methodological and ethical issues and considerations were given for future research and clinical practice. Conclusions drawn from this thesis are limited largely by the cross-sectional nature of most of the studies reviewed in paper 1 and insufficient numbers for the empirical study, which due to methodological and service-related constraints, limited further exploration of the data. Factor analysis and subsequent validation of the GLAMS in a larger sample is required to further support inferences drawn.

Declaration

No portion of the work referred to in this thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

i. The author of this thesis (including any appendices and/or schedules to this

thesis) owns certain copyright or related rights in it (the “Copyright”) and

s/he has given The University of Manchester certain rights to use such

ii. Copies of this thesis, either in full or in extracts and whether in hard or

electronic copy, may be made only in accordance with the Copyright,

Designs and Patents Act 1988 (as amended) and regulations issued under it

or, where appropriate, in accordance with licensing agreements which the

University has from time to time. This page must form part of any such

copies made. iii. The ownership of certain Copyright, patents, designs, trademarks and other

intellectual property (the “Intellectual Property”) and any reproductions of

(“Reproductions”), which may be described in this thesis, may not be owned

by the author and may be owned by third parties. Such Intellectual Property

and Reproductions cannot and must not be made available for use without the

prior written permission of the owner(s) of the relevant Intellectual Property

and/or Reproductions. iv. Further information on the conditions under which disclosure, publication

and commercialisation of this thesis, the Copyright and any Intellectual

Property and/or Reproductions described in it may take place is available in

the University IP Policy (see

http://www.campus.manchester.ac.uk/medialibrary/policies/intellectual-

11 property.pdf), in any relevant Thesis restriction declarations deposited in the

University Library, The University Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/regulations) and in The

University’s policy on presentation of Theses

Acknowledgements

I would firstly like to give my sincere thanks to Dr Katherine Berry for her support, patience, time and availability whilst completing this thesis. As a supervisor she provided not only guidance but strong encouragement throughout the supervisory process, all of which have been well received.

I would like to thank the other members of the research team for their input and support that enabled this research project to get off the ground; it has been a pleasure working with you all.

Although not part of the research team, I would like to thank Peter Taylor, a colleague and friend who supported my research and professional development as a scientific practitioner.

I feel it be appropriate to also offer a huge thank you to all those who took part in the study and facilitated the smooth running of this research, with particular reference to the staff team at the main recruitment site with whom I developed a good working relationship with.

Lastly, but by no means least, I would like to give an enormous thank you to my friends and loved ones, particularly my family for their kind words, warmth and belief in me during difficult times. Not once did their support waiver and for that I am forever grateful.

Paper 1: Literature Review

The Role of Guilt in the Development of Post-Traumatic Stress

Disorder: A Systematic Review

Preface

The literature review search and subsequent write-up was carried out between July

2012 and July 2013. Dr Katherine Berry provided support and supervision for this paper. This included reading drafts of the manuscript and supporting preparation for publication. Dr Peter Taylor also significantly contributed to the preparation of this paper for publication. This review was prepared in line with the journal guidelines for “Clinical Psychology Review” and is being submitted for publication. A copy of the author information pack is available in Appendix A. Tables and figures have been incorporated into the main text to aid readability. The authors for this paper will be Lauren Pugh, Dr Peter Taylor and Dr Katherine Berry.

Abstract

Post-traumatic stress disorder (PTSD) can be a debilitating condition associated with a myriad of emotions. Guilt is an important associated feature of PTSD that has received far less recognition than the hallmark symptom clusters. Theorists have suggested that guilt plays a role in PTSD. However, little attention has been given to how these might be linked. To address this, a systematic review of guilt and posttraumatic stress research in adults was carried out. We assess the evidence for four competing models conceptualising the guilt-PTSD relationship: Model 1; guilt is a causal process driving PTSD (Model 2; vice versa); Model 3; guilt and PTSD are unrelated but co-occurring; Model 4; confounds explain the guilt-PTSD link. A systematic search identified 27 articles. Cross-sectional relationships between trauma-related guilt (e.g. self-blame, wrong-doing) and PTSD symptomology was apparent, although, a few intervention studies revealed a lack of correspondence despite their co-occurrence. There was a preliminary indication that guilt mediates the relationship between trauma and PTSD; depression did not better account for this relationship but shame may have partly explained the guilt-PTSD link. The implications for future research, with regards to the further testing of these four models, and relevance for clinical practice is discussed.

Key Words

Guilt, self-blame, responsibility, cognitive appraisals, PTSD, trauma

Highlights

 Overall, guilt was positively related to PTSD symptomology, even when

controlling for negative affect or depression.

 Cognitions reflecting self-blame, the extent of one’s role and perceived

violation of personal standards were frequently associated with PTSD

symptoms and may maintain post-trauma sequelae.

 Some studies found guilt was no longer a significant correlate or predictor of

PTSD symptoms when controlling for the effects of shame.

 Event-related guilt rather than trait guilt was generally associated with PTSD

symptoms suggesting guilt may be trauma-specific than a characteristic

response.

Introduction

Post-traumatic stress disorder (PTSD) represents a prevalent psychological problem that is believed to affect approximately 6.8% of the general population (Kessler,

Chiu, Demler, & Walters, 2005). PTSD is characterised by the re-experiencing of symptoms in the form of intrusive memories, flashbacks and nightmares, avoidance of trauma-related stimuli and a heightened state of physiological arousal secondary to a traumatic event in DSM-IV (American Psychiatric Association [APA], 1994).

Left untreated, PTSD may become a chronic disabling condition associated with increased psychiatric comorbidity and deterioration in physical health (Sareen, Cox,

Clara, & Asmundson, 2005), increased suicidality (Panagioti, Gooding & Tarrier,

2012), and considerable societal costs (Kessler, 2000). Identification of those at risk requires an understanding of the psychological processes underlying the development of PTSD in the wake of a traumatic event.

There is considerable evidence that cognitions play an important part in the development and maintenance of post-traumatic stress symptoms (Ehlers & Clark,

2000; Foa, Ehlers, Clark, Tolin & Orsillo, 1999). Appraisals may be peritraumatic or emerge secondary when trying to make sense of what happened or persisting symptoms (Foa & Rothbaum, 1999; Smith & Bryant, 2000). Maladaptive beliefs distort perception leading to an increase in one’s subjective sense of threat resulting in intense emotional reactions and avoidance (Ehlers & Clark, 2000).

Therapies have primarily evolved to treat the hallmark symptoms of PTSD, typically cognitions and affective states associated with fear and helplessness

(Brewin, Andrews & Rose, 2000; Ozer, Best, Lipsey & Weiss, 2003). Trauma can prompt a myriad of emotional responses other than intense threat such as humiliation, anger, shame and guilt that may even occur more frequently in PTSD

18 than anxiety itself (Hathaway, Boals & Banks, 2010; Power & Fyvie, 2013). The role of guilt in the development of PTSD is increasingly being recognised by clinicians and theorists (Kubany & Watson, 2003; Lee, Scragg & Turner, 2001;

Wilson, Drozdek & Turkovic, 2006). Unlike fear, which is current or future orientated, guilt is considered a retrospective emotion and therefore may be less amenable to change through habitual exposure (Dalgleish, 2004). It has been argued that guilt may impede the emotional processing of fear or be exacerbated by exposure to trauma-related cues that may maintain trauma-related pathology (Ehlers

& Steil, 1995; Pitman et al., 1991). Guilt may also prevent the successful integration of the trauma with prior beliefs, contributing to the use of avoidant coping strategies that maintains PTSD symptomology (Kubany & Manke, 1995; Street, Gibson &

Holohan, 2005). This suggests that guilt, left untreated, may be a barrier to therapeutic change.

The significance of guilt in PTSD is such that it is an associated feature of

PTSD in DSM-IV (APA, 1994), although not part of the hallmark constellation of

PTSD symptoms. The presence of guilt is often referred to as ‘complex PTSD’, frequently seen in those with sustained or multiple traumas (Cloitre et al., 2009).

There have been attempts to formulate the role of guilt in PTSD with a general consensus that the degree to which one experiences guilt may depend on how the traumatic event is appraised and on beliefs held about personal involvement (Kubany

& Manke, 1995; Lee et al., 2001; Wilson et al., 2006). This in turn is thought to predict the severity of post-traumatic stress (Kubany and Watson, 2003). Holding beliefs exaggerating one’s role in a traumatic event, accompanied by distress, are argued to be primary components and determinants of guilt (Kubany et al., 1996).

Such beliefs are associated with perceived responsibility, preventability, lack of

19 justification and violation of personal standards. It is the modification of these distorted ‘guilt cognitions,’ by reframing and reappraisal that is the primary task of cognitive therapy for trauma-related guilt (Kubany, Hill & Owens, 2003). Guilt is also related to cognitions of self-blame, which connotes wrong doing and causal responsibility and which are associated with PTSD (Beck et al., 2004; Foa et al.,

1999; Frijda, 1993).

Evaluating the meaning of a traumatic event is central to the clinical model of guilt-based PTSD proposed by Lee et al. (2001). They suggest that guilt may emerge when a traumatic event is seen to violate personal standards or values leading to guilt charged intrusions and ruminative activity. Where the trauma appears inconsistent with the individual’s self-concept or standards they set for themselves (i.e. schema incongruent), guilt remains circumscribed only to the event meaning their sense of self remains intact. The resulting behaviour is of a reparative nature. In contrast, where a person’s transgression appears to confirm underlying negative beliefs about the self (i.e. schema congruent) this will lead to pervasive feelings of guilt and shame. This model’s emphasis on evaluating one’s role in a traumatic event and its consistency with pre-existing schemas in determining guilt and PTSD is consistent with previous research (Foa & Riggs, 1993; Janoff-Bulman, 1992; McCann &

Pearlman, 1990).

Whilst the various theoretical and clinical models clearly implicate guilt in the formation of PTSD, the precise role of guilt in PTSD and how distress is maintained is far less understood. This may be, in part, due to variation in the definition and measurement of the guilt construct. There is general consensus that guilt possesses both affective and cognitive elements mediated by social and moral standards including beliefs evaluating one’s role in a traumatic event (Baumeister,

Stillwell & Heatherton, 1994; Klass, 1990; Kugler & Jones, 1992; Mosher, 1988).

This is consistent with literature that has associated guilt with intrusive experiences and affective states arising from perceived violation of personally relevant standards for behaviour (Kugler & Jones, 1992; Niler & Beck, 1989; Tangney & Dearing,

2002). Guilt has been further delineated in terms of trait and state-guilt. The dispositional, trait-like tendency to respond with guilt in situations with ambiguous culpability has been described as guilt proneness that is characteristic and chronic

(Harder & Greenwald, 1999; Tangney, Stuewig & Mashek, 2007). This has been differentiated from transient or state-like guilt pertaining to certain situations, which would include traumatic events, or the consequences of specific transgressions

(Tangney, 1992). In an attempt to provide an all-encompassing definition of guilt,

Tilghman-Osborne, Cole and Felton (2010) suggest that guilt ‘involves moral transgressions (real or imagined) in which people believe that their action (or inaction) contributed to negative outcomes’ (p.546).

There also appears a lack of agreement regarding the function and consequences of guilt. Guilt is generally considered to be maladaptive, associated with self-punishment, impaired social behaviour and distress (Burney & Irwin, 2000;

Harder, 1995; Shapiro & Stewart, 2011). However, guilt has also been considered to be adaptive, associated with social adjustment, personality development and empathy (Lutwak, Panish, Ferrari & Razzino, 2001; Tangney, 1990). Guilt may lead to protective outcomes such as atoning for past transgressions or avoiding future ones (Tangney et al., 2007; Williams & Bybee, 1994). Some have even argued that guilt may safeguard an individual against the development of PTSD (Startup,

Makgekgenene & Webster, 2007) challenging the prevailing argument that guilt contributes to PTSD symptomatology.

The inconsistencies reported in the literature regarding the relationship between guilt and psychopathology may be attributed to how guilt is understood and subsequently measured. Concerns have been shared regarding the validity of some guilt measures and the little content overlap between them, which suggests they may be measuring different aspects of the guilt construct (Kim, Thibodeau & Jorgensen,

2011; Tilghman-Osborne et al., 2010). Some have argued that measures of guilt tap into constructs related to but distinct from guilt that may have led to inflated estimates of the relationships with emotional outcomes. Not only does this raise questions about the discriminant validity of these measures but leads to difficulties in distinguishing the psychopathological correlates of guilt and those of other constructs. A primary example would be the conceptual distinction between shame and guilt and their association with distress. Shame involves global condemnation of the self for behaviour perceived as disgraceful reflected in stable appraisals of negative self-worth (Tangney, Wagner & Gramzow, 1992; Wong & Cook, 1992).

Guilt has been associated with self-recrimination for behaviour that arouses a sense of wrong doing. This prompts remorse and tension motivating reparative behaviour and confession (Lee et al., 2001; Singer, 2004). As guilt has been associated with evaluating actions rather than the goodness of the person, the core dimensions of the individual are less affected suggesting guilt may be less psychological damaging

(Tangney, Miller, Flicker & Barlow, 1996; Wilson et al., 2006).

One challenge in determining the relationship between guilt and PTSD is the presence of confounding constructs that overlap with guilt that may better explain this relationship. Guilt has been blurred with the construct of shame and shame has been found to be related to PTSD posing a problem when evaluating psychological correlates of PTSD (Andrews, Brewin, Rose & Kirk, 2000). Excessive guilt also

22 features in major depressive disorder and has been associated with depressive symptomology (Alexander, Brewin, Vearnals, Wolff & Leff, 1999; APA, 1994;

Harder, 1995). This further complicates the relationship between guilt and PTSD because both share a relationship with depressive mood states (Blanchard, Buckley,

Hickling & Taylor, 1998; Breslau, Davis, Andreski & Peterson, 1991).

To our knowledge, the current paper provides the first review of the literature concerning the relationship between guilt and PTSD. The primary aim of this review is to synthesise and evaluate the research evidence in order to determine whether an association exists between guilt and post-traumatic stress symptoms. A second objective is to explore and evaluate the nature of this relationship. Four models conceptualising the relationship between guilt and PTSD symptoms have been proposed, reflecting alternative configurations of the underlying relationships between these constructs (see Figure 1). This relationship may be explained in the following ways: a) guilt is a causal process driving PTSD symptoms and hence can be viewed as a mediator of the effects of trauma on PTSD (Model 1); b) PTSD is a causal process driving guilt (Model 2); c) guilt and PTSD symptoms are unrelated but co-occurring responses to trauma (Model 3); d) confounding variables associated with guilt (e.g. depression and shame) explain the relationship between guilt and

PTSD symptomology (Model 4).

1) Trauma Guilt PTSD

2) Trauma PTSD Guilt

PTSD

3) Trauma

Guilt

Shame 4) Trauma PTSD Depression

Guilt

Figure 1: Four contrasting conceptualisations of the association between guilt and

PTSD symptomology.

Being able to establish which of these models best describes the relationship between guilt and PTSD is essential in determining the clinical importance of this construct and has significant implications for treatment choice and planning. For example, Model 1 suggests that therapeutic interventions aimed specifically at alleviating guilt would be beneficial in treating PTSD by highlighting the content of key appraisals that should be addressed during therapy. Models 2 and 3, however, suggest that treatments focussing on guilt will have little additional benefit and

Model 4 would imply that interventions aimed at guilt would only be beneficial to the extent that they reduce associated constructs such as low mood or shame.

The present review has been organised such that an outline of the review methodology, a summary of research findings evaluated against each model in the above theoretical framework and a critique of the empirical support will be presented. A discussion of the main findings will be summarised and concluded with recommendations regarding future research and clinical practice.

Method

Search strategy

A review of the literature was conducted using selected electronic databases to identify quantitative studies in which the relationship between guilt and PTSD was examined. Databases reviewed were Embase (1974-July 2012), PsycINFO (1806-

July 2012) and Medline (1946-July 2012), searching for relevant articles using the term guilt* in combination with PTSD or associated words (trauma*, post-traumatic stress disorder). To ensure a comprehensive search, the same key search terms and limits were applied to the Web of Science database (1900 – July 2012). All article abstracts were reviewed, after removing duplicates, to establish whether the studies met the threshold criteria for manuscript inclusion for this review. Where this could not be ascertained the full text of the article was retrieved. A search of the reference lists ensured all relevant citations within each full text article were retrieved and read. The numbers of articles identified at each stage of the search process are outlined in Figure 2.

Articles identified through Articles identified

database searching through additional (PsycINFO, Medline, database searching (Web Embase) excluding of Science) duplicates (n = 817) (n =274)

Article abstracts Additional abstracts identified as potentially identified excluding relevant (n = 79) duplicates (n = 9)

Additional articles identified

through citation searches (n = 14)

Total number of full text

articles assessed for eligibility (n = 102) Articles excluded

where inclusion criteria was not met (n =75)

Articles included in

systematic review (n =27)

Figure 2: A flow chart outlining the article identification and selection process.

Inclusion and exclusion criteria

Studies were included that met the following criteria: a) original research articles published in a peer-reviewed journal; b) used adult participants aged 18 years or older; c) were written in English; d) included a measure of PTSD symptoms; and e)

26 included a quantitative measure of guilt. Studies that quantified guilt by coding responses from an interview and assessed guilt exclusively using a single item measure or in the context of depression (i.e. used an inventory for depression) were excluded, as this may not have accurately reflected or quantified the construct of interest. Articles whose focus were to compare the efficacy of treatments for PTSD and associated symptoms such as guilt were included providing both variables were measured as outcomes post-intervention. Studies testing the psychometric properties of guilt scales were included where inclusion criteria were met. Due to the limited number of studies using samples diagnosed with PTSD, studies that assessed PTSD symptoms in populations that may not have met the full diagnostic criteria were still included.

Studies that included samples that were identified as having acute stress disorder were further excluded, as this is a distinctly separate psychological disorder

(APA, 1994), albeit overlapping in symptomology. Studies that grouped together guilt with other constructs into one category of negative affect were excluded to avoid inflating estimates of the relations between guilt and other variables. Similarly, those that combined PTSD symptoms with other forms of global distress were excluded. Applying these criteria, 27 articles were included in this review. Authors, sample, methodology, measures, key findings and the magnitude of effects have been summarised in Table 1. Guilt and PTSD assessment tools discussed as part of this review have been cited in the text accordingly and will appear in the reference section for this article. Corresponding references for all other assessment tools cited but not discussed and not referenced elsewhere in this review, can be found in

Appendix B.

Table 1: Study characteristics and findings

Correlation/ Regression Effect Size PTSD PTSD Guilt Effect size (r) Coefficient (d Study Sample Design Diagnostic Tool Measure Measureª Findings (β) ) Kubany et Treatment Cross- MS-Combat, PI, MS- GI (trait) Event-related guilt n = 58; < 0.81** - al. (1995) seeking/ sectional IES; assess Combat, PFQ, correlated with PTSD n = 50; 0.51** community symptoms PI, IES AAGS symptoms in both samples, veterans (n = 58), as did beliefs about role in females from IPV trauma shelters (n = 50)

Kubany et Veterans (n = 74), Cross- MS-Combat; IES, MS- PFQ Trauma-related guilt n = 74; < 0.73* - al. (1996) treatment seeking sectional assess Combat, (trait), GI correlated with PTSD n = 68; < 0.75* female survivors symptoms PCL, (trait), severity in both samples. of IPV (n = 68) MPSS TRGI Beliefs about justification TOSCA not related. Fewer associations between guilt cognitions, intrusions and/or avoidance for IPV sample

Henning & Treatment seeking Cross- CAPS; current MS- GI (trait), Combat specific guilt < 0.49** < 0.63*** Frueh. male veterans sectional diagnosis Combat, CGS correlated with PTSD (1997) (n = 40) CAPS severity, re-experiencing and avoidance and explained much of the variance in PTSD severity independent of exposure and trait guilt

Kubany et Treatment seeking Cross- MS-Combat; MS- PFQ, GI Overall guilt index and total < 0.75*** - al. (1997) (and none sectional assess Combat, (trait), guilt scores significantly treatment seeking) symptoms PCL, IES TRGI, correlated with PTSD veterans (n = 74) STRGS- symptom measures WZ

Correlation/ Regression Effect Size PTSD PTSD Guilt Effect size (r) Coefficient (d Study Sample Design Diagnostic Tool Measure Measureª Findings (β) ) Beckham Treatment seeking Cross- CAPS; current MS- TRGI Trauma-related guilt < 0.50* - et al. veterans sectional diagnosis Combat, correlated with PTSD (1998) (n = 151) DTS severity and symptom clusters. Guilt cognitions did not correlate with intrusions. Perceived wrong doing was consistently related to PTSD symptoms

Keane et Male veteran Cross- SCID for DSM- SCID-III, LPI Higher physiological 0.24ᵇ - al. (1998) patients sectional III; current MS- responses (responders) to (n = 106-120 diagnosis Combat trauma related cues is responders; associated with greater 100-120 non- PTSD severity and guilt in responders) those with current PTSD

Street & Females from Cross- MS-Civilian; MS- TOSCA Guilt was not a significant 0.21 (n/s) 0.18 (n/s) Arias. domestic violence sectional assess Civilian predictor of PTSD (2001) shelters (n = 63) symptoms symptoms, whereas shame was

Leskela et Community Cross- PCL-Military; PCL- TOSCA Guilt proneness did not < 0.11 (n/s) -0.22* al. (2002) residing former sectional assess Military correlate with PTSD veterans (n= 107) symptoms severity; a negative relationship was observed when the effect of shame on PTSD severity was removed

Correlation/ Regression Effect Size PTSD PTSD Guilt Effect size (r) Coefficient (d Study Sample Design Diagnostic Tool Measure Measureª Findings (β) ) Resick et Female survivors Longitudinal CAPS-1; current PSS-I TRGI No difference in proportion - - PTSD: al. (2002) of rape (n = 171) (RCT) diagnosis (global, diagnosed with PTSD or g = < 0.29 cognition) symptom severity between Guilt: CPT and PE groups post- g = < 0.46 treatment. CPT superior to PE in remedying guilt cognitions - - n = 125 Kubany et Female survivors Longitudinal DEQ, CAPS-1; CAPS-1, TRGI, Proportion diagnosed with PTSD: < 1.8 al. (2003), of IPV (n = 37, current DEQ STRGS- PTSD declined with a Guilt: < 2.0 Kubany et n = 125) diagnosis PA, PFQ corresponding reduction in n = 37 al. (2004) PTSD symptoms and guilt PTSD: < 3.3 post-treatment in both Guilt: < 4.2 studies

Lowinger Male drivers Retro- PTSD-I; assess PTSD-I TOSCA, Self-blame correlated with < 0.42** - & convicted of spective symptoms QSG (4 PTSD symptoms post- -0.47* Solomon. reckless driving items) accident. External (2004) (n = 75) attributions of responsibility negatively correlated with PTSD symptoms. Reduction in guilt corresponded with reduction in PTSD symptoms over time

Street et al. Female survivors Cross- PCL-Civilian; PCL- TRGI Trauma-related guilt 0.53** - (2005) of IPV using sectional assess Civilian (global, associated directly and domestic violence (retro- symptoms cognition) indirectly (partially mediated services (n = 63) spective) through avoidant coping) with PTSD symptomology

Correlation/ Regression Effect Size PTSD PTSD Guilt Effect size (r) Coefficient (d Study Sample Design Diagnostic Tool Measure Measureª Findings (β) ) Rusch et Treatment seeking Cross- MINI; current MINI TOSCA-3 Guilt proneness significantly 0.26ᵇ - al. (2007) BPD patients sectional diagnosis higher in women with BPD (n = 23 BPD + and PTSD compared to BPD PTSD; n =37 alone BPD only)

Crisford et Inpatient secure Cross- DAPS (PTS-T); DAPS TRGI Guilt cognitions and al. (2008) unit sample Sectional assess (PTS-T) (cognition) attributions correlated with < 0.43** 0.34** (n = 45) symptoms , Revised offence-related PTSD GBAI symptomology. Cognitions (guilt) explained much of the variance in PTSD symptoms, controlling for negative affect and offence severity

Owens et Treatment seeking Longitudinal PCL-Military; PCL- TRGI Beliefs about wrong doing < 0.34*** 3.27* al. (2008) veteran inpatients current or sub- Military (cognition) significantly correlated with (n = 99) threshold PTSD symptoms pre & post- diagnosis treatment and predicted post-treatment severity. PTSD significantly reduced post-treatment, guilt cognitions did not

Resick et Female PTSD Longitudinal CAPS; current CAPS, TRGI PTSD symptoms and guilt - - PTSD: -1.68 al. (2008) sample (RCT) diagnosis PDS (cognition) cognitions decreased over Guilt: -1.08 (n = 53) time following cognitive processing therapy for PTSD

Correlation/ Regression Effect Size PTSD PTSD Guilt Effect size (r) Coefficient (d Study Sample Design Diagnostic Tool Measure Measureª Findings (β) ) Ginzburg Treatment seeking Longitudinal PCL-Specific; PCL- ARBQ PTSD symptoms did not 0.04 (n/s) 0.03 (n/s) et al. female survivors (RCT) assess Specific (guilt) correlate with guilt related to (2009) of childhood symptoms abuse. Reduction in guilt did sexual abuse not mediate change in PTSD (n = 129) symptoms post-treatment, a reduction in shame did

Owens et Community Cross- PCL-Military; PCL- GI (trait) Guilt correlated with PTSD 0.64** 0.18* al. (2009) residing veterans sectional assess Military symptoms and a significant (n = 174) symptoms predictor of PTSD severity

Marx et al. Combat veterans Cross- SCID for DSM- SCID-III LPI Combat-related guilt 0.73*** - (2010) (n = 1323) sectional III; assess partially mediated (retro- symptoms association between spective) exposure to violence and PTSD symptoms and completely mediated when directly participating

Robinaugh Students and Cross- PCL-Specific; PCL- SSGI State guilt did not predict 0.26** -0.01 (n/s) & community sectional assess Specific (guilt), PTSD symptoms when McNally. sample symptoms TRGI controlling for shame. At (2010) (n = 140) low levels of shame guilt was negatively associated with PTSD, at high levels a positive association was observed

Beck et al. Treatment seeking Cross- CAPS-1; assess CAPS-1 TRGI Guilt-related distress and < 0.33** < 0.41** (2011) females IPV sectional symptoms cognitions were positively (n = 63) associated with PTSD severity

Correlation/ Regression Effect Size PTSD PTSD Guilt Effect size (r) Coefficient (d Study Sample Design Diagnostic Tool Measure Measureª Findings (β) ) Held et al. Treatment seeking Cross- CAPS; current CAPS, TRGI Trauma related guilt was < 0.36* - (2011) veteran inpatients sectional or sub-threshold PCL- (global) positively related to self- (n = 147) diagnosis Specific reported (not clinician reported) PTSD severity, partially mediated through disengagement coping

Pereda et Spanish university Cross- DEQ; assess DEQ TRGI All guilt subscales correlated < 0.68*** - al. (2011) students (n = 650) sectional symptoms significantly with the measure of PTSD symptomology

Semb et al. Victims of violent Cross- HTQ (16 items); HTQ TOSCA, 1 Guilt proneness and event < -0.06 (n/s) - (2011) crime (n = 35) sectional assess item VAS guilt was unrelated to trauma symptoms symptoms when controlling for shame

Browne et Journalist Cross- PCL-Civilian; PCL- TRGI Trauma-related guilt 0.36 - al. (2012) civilians (n = 50) sectional assess Civilian (cognition) cognitions contributed a symptoms significant increase in variance in PTSD symptoms, and mediated the relationship between exposure and PTSD symptomology

Harned et Female BPD Longitudinal PSS-I; current PSS-I TRGI Reduction in PTSD severity, - - PTSD: 1.4 al. (2012) sample (n = 13) diagnosis (cognition) guilt and shame post- Guilt: 1.0 treatment. No significant change in depression

ª The majority of studies listed in Table 1 included more than one measure of guilt. Only those assessment tools used in the statistical evaluation of the guilt-PTSD relationship were discussed as part of this review.

ᵇ Effect size (r) was calculated if the coefficient r was not reported but the authors had reported enough statistics to compute r from r2 = d2 / (4+d2) where d = m1-m2/√sd1+sd2/2 (Cohen, 1988).

Effect sizes reported by the author were included in this review as appropriate to the study design. Cohen’s d statistic was reported unless

Hedges g (unbiased) statistic was otherwise specified.

Note: BPD = Borderline Personality Disorder, IPV = Interpersonal Violence. - denotes authors did not report this information. < denotes there was more than one correlation coefficient given in the analysis but all statistics were less than the number reported at p < 0.05*, p < 0.01**, p <

0.001***. ns indicates not significant. PTSD measures: PTSD Checklist (PCL specific, military and civilian versions; Weathers, Litz, Herman,

Huska & Keane, 1993); Mississippi Scale (MS combat and civilian versions; Keane, Caddell & Taylor, 1988); Clinician Administered PTSD

Scale (CAPS; Blake et al., 1995; Blake et al., 1990); Impact of Events Scale (IES; Horowitz, Wilner & Alvarez, 1979); Structured Clinical

Interview for DSM-III (SCID-III; Spitzer, Williams, Gibson & First, 1989); Distressing Event Questionnaire (DEQ; Kubany, Leisen, Kaplan &

Kelly, 2000); PTSD Symptom Scale-Interview (PSS-I; Foa, Riggs, Dancu & Rothbaum, 1993); Modified PTSD Symptom Scale (MPSS;

Falsetti, Resnick, Resick & Kilpatrick, 1993); Mini International Neuropsychiatric Interview (MINI: Sheehan et al., 1998); Davidson Trauma

Scale (DTS; Davidson et al., 1997); Penn Inventory for PTSD Assessment (PI; Hammarberg, 1992); PTSD-Inventory (PTSD-I; Solomon et al.,

1993); Harvard Trauma Questionnaire (HTQ; Mollica et al., 1992); Posttraumatic Diagnostic Scale (PDS; Foa, Cashman, Jaycox & Perry,

1997); Detailed Assessment of Post-Traumatic Stress (DAPS; Briere, 2001). Guilt measures: Personal Feelings Questionnaire (PFQ; Harder &

Lewis, 1986); Combat Guilt Scale (CGS; Henning & Frueh, 1997); Laufer-Parson Inventory (LPI; Laufer, Yager, Frey-Wouters & Donnellan,

1981); Attitudes About Guilt Survey (AAGS; Kubany & Manke, 1995); Sources of Trauma Related Guilt Survey-War Zone version (STRGS-

WZ; Kubany, Abueg, Kilauano, Manke & Kaplan, 1997) and Partner Abuse version (STRGS-PA; Kubany, Owens & Leigh, 1998); Abuse

Related Beliefs Questionnaire (ARBQ; Ginzburg et al., 2006); State Shame and Guilt Inventory (SSGI; Tangney & Dearing, 2002); Revised

Gudjonsson Blame Attribution Inventory (GBAI; Gudjonsson & Singh, 1989); Questionnaire on Specific Guilt (QSG; Janoff-Bulman, 1989);

Guilt Inventory (GI; Jones, Schratter & Kugler, 2000); Test of Self-Conscious Affect (TOSCA; Tangney, Wagner & Gramzow, 1989); Adapted version of the TOSCA (TOSCA-3; Tangney, Dearing, Wagner & Gramzow, 2000); Trauma-Related Guilt Inventory (TRGI; Kubany et al.,

1996).

Results

Overview of reviewed studies

Sample sizes used in the analysis ranged from n = 13 to n = 1323, with fewer studies recruiting samples with a current or sub-threshold diagnosis of PTSD (n = 10) compared with symptoms consistent with PTSD (n = 17). More than half of the studies recruited an inpatient, treatment or support seeking sample (n = 16), most others recruited a community sample (n = 8). Three articles reported having recruited a mixed sample. The majority of studies were cross-sectional in their design, testing causal relations between guilt and post-traumatic stress (n = 19), two of which compared the level of guilt between two groups that differed in their experience of

PTSD symptomology. The remaining studies used a longitudinal or retrospective design (n = 8), three of which were a randomised control trial (RCT) examining changes in guilt and PTSD symptoms following the delivery of an intervention.

Is guilt related to PTSD symptomology?

A consistent finding across several studies was that trauma-related or combat- specific guilt was significantly associated with post-traumatic stress symptomology

(Beck et al., 2011; Browne, Evangeli & Greenberg, 2012; Crisford, Dare &

Evangeli, 2008; Kubany, Abueg, Kilauano, Manke & Kaplan, 1997; Kubany et al.,

1995; Kubany et al., 1996; Lowinger & Solomon, 2004; Marx et al., 2010; Pereda,

Arch, Pero, Guardia & Forns, 2011; Street et al., 2005). Findings were consistent despite the various different measures and samples used, which included those with a diagnosis of PTSD and those with symptoms consistent with PTSD symptomology.

Interestingly, the association between guilt (global and trauma-related) and

PTSD symptomatology was shown to be mediated by avoidance in two studies

(Held, Owens, Schumm, Chard & Hansel, 2011; Street et al., 2005). In one of these studies this mediation effect was only found when PTSD symptomology was self- reported as opposed to clinician rated. The use of self-report inventories by participants may have stimulated rumination about symptoms increasing the frequency or intensity reported that would be rated as less severe by clinicians

(Carter, Frampton, Mulder, Luty & Joyce, 2010). Nonetheless, these findings suggest that avoiding the experience of guilt may serve to maintain distress.

A number of studies looked at the association between trauma-related or combat-specific guilt and PTSD severity more specifically in samples with a current or sub-threshold diagnosis of PTSD who were also seeking treatment (Beckham,

Feldman & Kirby, 1998; Held et al., 2011; Henning & Frueh, 1997; Owens, Chard &

Cox, 2008). Significant and positive relationships were found confirming that overall findings can be generalized beyond those with symptoms consistent with PTSD to clinical samples. Examining the guilt construct, Owens et al. (2008) found beliefs about wrong doing were a significant predictor of PTSD severity after treatment suggesting this guilt cognition may be particularly relevant in the maintenance of

PTSD. In a study that examined the relationship between guilt and specific PTSD symptom clusters (Henning & Freuh, 1997), combat-specific guilt explained a significant proportion of the variance in symptoms of avoidance and re-experiencing combined, hallmark symptoms of PTSD. Further to this, findings were independent of the effects of exposure and trait guilt, which is particularly relevant as many studies did not control for other variables when assessing the guilt-PTSD relationship. These findings should be interpreted with some caution as combat-

37 related guilt was derived from a measure that was not validated but created for the purpose of this study, the psychometrics of which may be unreliable. Overall, the strength of the relationship between guilt and PTSD symptomology was variable across studies with correlation coefficients ranging between 0.16 and 0.81. This may have been influenced by the use of different measures of guilt and PTSD across studies and the nature or extent of trauma exposure (e.g. acute or chronic).

A relationship between guilt and PTSD was also found in two studies that compared experiences of guilt between groups differing in PTSD symptomatology

(Keane et al., 1998; Rusch et al., 2007). Among treatment seeking females diagnosed with BPD, guilt proneness was found to be significantly higher among those with a comorbid diagnosis of PTSD than those without PTSD (Rusch et al., 2007).

Similarly, guilt was higher in a sample of treatment seeking veterans who demonstrated a heightened physiological response to trauma cues, a core symptom of

PTSD, compared to those not showing this response (Keane et al., 1998).

Several studies looked more closely at the content of guilt-related cognitions or attributions and its relationship with PTSD symptoms (Beckham et al., 1998;

Crisford et al., 2008; Kubany et al., 1995; Kubany et al., 1996; Lowinger &

Solomon, 2004; Owens et al., 2008; Pereda et al., 2011). Beliefs concerning perceived wrong doing, lack of justification, responsibility and preventability were significantly related to PTSD symptomology (Beckham et al., 1998; Kubany et al.,

1995; Kubany et al., 1996; Owens et al., 2008; Pereda et al., 2011). Perceived wrong doing was consistently and significantly related to PTSD symptomology suggesting that perceiving one’s role as violating personal or moral standards may be important in the maintenance of PTSD. Cognitions and attributions relating to themes of regret and remorse were also related to PTSD symptomology among inpatients from a

38 secure unit (Crisford et al., 2008). Guilt significantly predicted PTSD symptomology whilst controlling for offence severity.

Similarly, self-blame, characterised by the belief that one could have prevented the accident and attributing fault to oneself was positively associated with

PTSD symptoms (reported as occurring after the event) among males convicted of causing death by dangerous driving (Lowinger & Solomon, 2004). However these findings were not found to be significant when symptoms were rated at the time of the study. The estimated prevalence of PTSD had declined over time, which corresponded with a reduction in guilt suggesting a link between the two. The reliability of findings may be questioned as responsibility attributions were derived from four items in a scale adapted for use without subsequent validation.

Fewer significant associations were reported between specific guilt cognitions and intrusive experiences and avoidance, particularly appraisals concerning perceived lack of justification (Kubany et al., 1995; Kubany et al., 1996).

This is surprising given that guilt cognitions have been conceptualised as intrusive recollections. Perceived lack of justification may have felt less intrusive thereby arousing less distress and avoidance. Findings may also be limited by the narrow content of the IES (Horowitz, Wilner & Alvarez, 1979), which does not enquire about some symptoms of PTSD (e.g. hyperarousal), which may be related to the construct of guilt. In another study, Beckham et al. (1998) found guilt cognitions did not significantly correlate with the DTS intrusive experiences subscale (Davidson et al., 1997) despite all guilt subscales correlating with the MS-Combat scale (Keane,

Caddell & Taylor, 1988). It may be that guilt cognitions are not experienced as intrusive or are more closely related to other symptoms of PTSD, as measured by the

MS-Combat scale.

Six intervention studies measured guilt and PTSD symptomology as part of their treatment outcome (Harned, Korslund, Foa & Linehan, 2012; Kubany et al.,

2003; Kubany et al., 2004; Owens et al., 2008; Resick, Nishith, Weaver, Astin &

Feuer, 2002; Resick et al., 2008). Preliminary findings from Kubany et al. (2003), later replicated by Kubany et al. (2004) and those reported by Resick et al. (2008) all showed a significant reduction in trauma-related guilt (including guilt cognitions) and PTSD symptomology following the delivery of a trauma-focused intervention.

All three studies described an intervention that addressed cognitive distortions associated with self-blame and other guilt-related beliefs in relation to the trauma.

The finding that an intervention directed at guilt-related cognitions can reduce PTSD symptoms could be taken as proxy evidence of a link between guilt and PTSD, although, an association between guilt and PTSD symptoms was not explicitly tested. Other studies have shown a corresponding decline in both guilt cognitions and PTSD severity among females with BPD following exposure therapy at follow- up, but again this co-variation does not necessarily imply a direct relationship

(Harned et al., 2012). Reduction in trauma-related guilt was significant among the intent to treat sample but not for treatment completers, whose sample size was only seven and was likely to be insufficiently powered to detect a significant effect.

A few studies did not find a significant relationship between guilt and PTSD symptoms (Ginzburg et al., 2009; Leskela et al., 2002; Robinaugh & McNally, 2010;

Semb, Strömsten, Sundbom, Fransson & Henningsson, 2011; Street & Arias, 2001).

Three studies (Leskela et al., 2002; Semb et al., 2011; Street & Arias, 2001) used a measure of trait guilt, the TOSCA (Tangney, Wagner & Gramzow, 1989), which reflects a person’s disposition to respond with guilt to common day to day scenarios.

This may not be an appropriate and reliable measure of guilt in response to a

40 traumatic experience. Those experiencing event-specific guilt may do so independent of a tendency to respond with guilt to everyday situations. Respondents are also providing estimates of how they might feel in hypothetical situations, which may not reflect guilt attached to real life events that are more salient to them. This may have contributed to the lack of significant findings reported, although, positive relationships between trait guilt and PTSD symptomology have been reported and whilst controlling for age and combat exposure (Owens et al., 2009; Rusch et al.,

2007). Some have suggested that the TOSCA and GI (Jones, Schratter & Kugler,

2000) assess different phases of the guilt experience with the TOSCA focusing on adaptive responses to relieve guilt as opposed to measuring unresolved guilt, as indicated by the GI (Ferguson & Crowley, 1997). Consequently, the latter would be considered maladaptive as the person is unable to take appropriate action to deal with their guilt, which would otherwise prevent guilt becoming chronic and distressing. Interestingly, all studies that did not report a significant relationship between guilt and PTSD symptomology reported a significant effect of shame, which may provide an alternative explanation for the guilt-PTSD relationship.

Overall there was evidence of a cross-sectional relationship between guilt and

PTSD across various populations. Specifically, guilt-related distress and cognitions associated with self-blame or wrong-doing were related to PTSD symptomatology.

In contrast cognitions related to lack of justification and trait guilt tended to be less consistently related to PTSD. This may be attributed to the type of measure used, which may not have tapped into the relevant aspects of these constructs.

Model 1: Is guilt a causal process driving PTSD symptomology?

If guilt and PTSD are correlated, a subsequent question concerns the direction and nature of this effect. Model 1 proposes that guilt is a causal process that mediates the

41 relationship between trauma and PTSD symptomatology. Only two studies tested this model, supporting this relationship in both cases. Combat-related guilt partially mediated the association between exposure to combat violence and PTSD symptoms and completely mediated the association when directly participating in violent acts

(Marx et al., 2010). This may be because direct personal involvement denotes greater perceived responsibility and sense of wrong doing. This study had a large sample size and used path analysis as opposed to traditional mediational analyses to enable the exploration of direct and indirect effects simultaneously. However, these findings should be interpreted with caution as this study relied on retrospective self-reports of exposure, which may have been subject to memory biases over time.

Further support is derived from Browne et al. (2012) who found trauma exposure, as a predictor of PTSD symptoms, to be non-significant when guilt cognitions were entered into the regression model. Further analyses revealed a marginal yet significant indirect effect of guilt cognitions, which partially mediated the relationship between exposure and PTSD symptoms. Therefore beliefs about one’s role in a traumatic event may, in part, explain the association between exposure and PTSD symptoms suggesting a causal role of guilt in PTSD. This study used a moderate sample size, comparable in age and gender with a larger population of journalists suggesting the sample was representative. Overall, these results increase the plausibility of Model 1; however, as these studies were cross-sectional they limit inferences regarding the direction of effect.

Model 2: Is PTSD a causal process driving guilt?

There were no studies included in this review that tested the conceptualisation that guilt emerged as a result of developing PTSD symptoms. Furthermore, no

42 longitudinal studies could verify the direction of the causal relationship between guilt and PTSD. Hence the validity of Model 2 remains unknown.

Model 3: Are guilt and PTSD symptomology causally unrelated despite their co- occurrence?

It is possible that the relationship between guilt and PTSD is artefactual, reflecting the fact that both are co-occurring products of trauma. Hence, whilst several studies have suggested those who experience guilt also experience PTSD symptomology, there may be no underlying causal link. Indirect evidence for this model comes from intervention studies where there appears to be a lack of synchrony between guilt and

PTSD symptoms over time. Groups receiving cognitive processing therapy (CPT) or prolonged exposure (Resick et al., 2002) both demonstrated a similar reduction in

PTSD symptom severity and the proportion diagnosed with PTSD after treatment.

This non-significant group difference remained despite a significant reduction in guilt cognitions, namely hindsight bias and perceived lack of justification, among those receiving CPT. If guilt were a causal process underlying PTSD, a shift in guilt would be expected to correspond with a shift in PTSD symptomatology. Similarly,

Owens et al. (2008) found a reduction in PTSD symptoms that did not correspond with a reduction in guilt cognitions post-treatment. Both studies suggest guilt and

PTSD may have co-occurred among both trauma groups yet may be unrelated. Guilt cognitions may not have been expected to change if the intervention did not target such beliefs, which may explain this finding. It is also possible that another mechanism of change may be underlying the improvement in PTSD symptomology other than remedying guilt cognitions. For example a change in coping strategy or increase in one’s perceived competence to cope with trauma-related stressors, which

43 has been argued may better account for a reduction in PTSD symptomology

(Benotsch, Brailey, Vasterling & Uddo, 2000; Foa & Rauch, 2004).

In summary, support for Model 3 has been largely derived from intervention studies that indicate a lack of correspondence between guilt and PTSD symptoms despite their co-occurrence. Guilt and PTSD symptoms are both consequences of trauma but may remain independent of each other. However, interpretation of these results requires caution as these studies were not designed to specifically test the guilt-PTSD link. In contrast, mediation studies discussed previously did test this link and provide credible counter-evidence to Model 3 by demonstrating that guilt and

PTSD symptoms are related over and above their shared relationship with trauma.

Model 4: Do confounding variables explain the relationship between guilt and

PTSD symptomology?

It is possible that the observed relationship between guilt and PTSD symptomatology may result from the overlap between guilt and other constructs. The majority of studies did not control for the effects of low mood despite several studies showing significant correlations between depression and aspects of the guilt construct and between depression and PTSD symptomology (Kubany et al., 1997; Kubany et al.,

1995; Kubany et al., 1996; Owens et al., 2008; Owens et al., 2009; Pereda et al.,

2011; Robinaugh & McNally, 2010; Street & Arias, 2001). Depression may therefore explain the relationship between guilt and PTSD symptomology. A change in mood may have influenced findings from treatment outcome studies as a reduction in depressive symptomology or the number that met criteria for diagnosis post- intervention was reported (Resick et al., 2002; Resick et al., 2008; Kubany et al.,

2003; Kubany et al., 2004). Significant group differences in mood may also explain

44 the findings from Keane et al. (1998). In this study, veterans showing greater physiological responses to trauma-related cues reported greater guilt and higher levels of depression than non-responders. It remains unclear if the significant group difference in guilt would remain after controlling for depression, whose diagnostic criteria features guilt.

Some findings, however, challenge the suggestion that depression underlies or inflates the relationship between guilt and PTSD symptomology. Crisford et al.

(2008) found that trauma-related guilt cognitions independently predicted PTSD symptoms whilst controlling for negative affect among secure unit inpatients. This counters the suggestion that negative affect or low mood explains the relationship between guilt and PTSD symptoms. Likewise, Marx et al. (2010) found a relationship between guilt and PTSD among veterans whilst controlling for the effect of guilt on depressive symptomatology within a path model. Another study found that depression did not significantly reduce among a sample of females with BPD who were receiving treatment for PTSD although a reduction in guilt and PTSD symptomology was observed (Harned et al., 2012). This suggests guilt was not present in the context of low mood but may have been trauma-specific. Findings by

Rusch et al. (2007) support this conclusion having found guilt-proneness to be significantly higher among females with a comorbid diagnosis of PTSD than those without PTSD whilst there were no significant group differences in depression.

Although, the presence of PTSD was identified using an assessment tool, which is not typically used in establishing caseness for PTSD.

The majority of studies reviewed did not control for the effects of shame despite several studies showing significant correlations between shame and guilt and between shame and PTSD symptomology (Beck et al., 2011; Kubany et al., 1997;

Kubany et al., 1995; Kubany et al., 1996; Leskela et al., 2002; Lowinger & Solomon,

2004; Robinaugh & McNally, 2010; Street & Arias, 2001). Shame was also a significant treatment outcome that was seen to reduce over time in studies discussed previously (Harned et al., 2012; Kubany et al., 2003; Kubany et al., 2004). Like depression, it is difficult to disentangle guilt from shame and be confident that guilt is a driver of PTSD symptoms. Interestingly, studies that did not report a relationship between guilt and PTSD symptomology did report a significant effect of shame suggesting shame may better explain the guilt-PTSD relationship (Ginzburg et al.,

2009; Leskela et al., 2002; Robinaugh & McNally, 2010; Semb et al., 2011; Street &

Arias, 2001).

Ginzburg et al. (2009) found that guilt, in relation to childhood sexual abuse, was not significantly associated with PTSD symptoms. A reduction in shame, not guilt, explained a change in PTSD symptoms post-treatment. Four other studies found either no relationship between guilt and PTSD symptoms (Robinaugh &

McNally, 2010; Semb et al., 2011; Street & Arias, 2001), or a negative relationship when controlling for the effects of shame (Leskela et al., 2002). Robinaugh &

McNally (2010) also reported an interaction whereby state guilt was positively related to PTSD symptoms among those experiencing high shame and negatively associated at low levels of shame. These findings support claims that guilt is related to distress when fused with shame (Tangney et al., 2007).

Some of the above studies may be limited by the use of measures with poor psychometric properties. The MS-Civilian version used by Street and Arias (2001) has shown to correlate more strongly with measures of depression and anxiety than

PTSD (Lauterbach, Vrana, King & King, 1997; Vreven, Gudanowksi, King & King,

1995). Factor analytic studies involving the HTQ, used by Semb et al. (2011), have

46 also revealed that items associated with social withdrawal reflected avoidance consistent with depressive symptomology than trauma raising similar issues regarding discriminant validity (Smith-Fawsi et al., 1997). The HTQ (Mollica et al.,

1992) has been validated as a cross-cultural instrument; however, only 16 trauma- specific items were used by Semb et al. (2011), the reliability of which is unknown.

The same study used a non-standardised one-item measure to assess guilt and shame, which would insufficiently encapsulate either construct, which has implications for construct validity.

Overall, there is evidence to suggest that the relationship between guilt and

PTSD symptomology may be confounded by overlapping constructs. Although the majority of studies reviewed did not control for negative affect or depression, those that did continued to report a significant relationship between guilt and PTSD symptomology. Unlike depression, the relationship between guilt and PTSD symptoms was not maintained among studies controlling for the effects of shame. It is therefore possible that shame better explains the guilt-PTSD relationship.

Nonetheless, only a minority of studies controlled for co-existing shame, further research is required to explore the contributions of both these constructs to PTSD, independently and in relation to one another.

Summary of methodological considerations

Consideration has been given to the overall methodological limitations of the studies discussed in this review that may affect interpretation of the findings. The primary limitation was the lack of control over significant covariates such as depression, even when found to correlate with guilt and PTSD (e.g. Robinaugh & McNally, 2010). As guilt has shown to be more closely identified with depression (Bennice, Grubaugh &

Resick, 2001; Nishith, Nixon & Resick, 2005), and depression remains a significant

47 comorbidity of PTSD, this may have inflated the association between guilt and

PTSD. A decline in guilt after treatment may also be associated with a reduction in depressive rather than PTSD symptomology, where the effects of low mood were not controlled (e.g. Kubany et al., 2003). Nevertheless, positive relationships between guilt and PTSD symptomology remained significant whilst statistically controlling for negative affect or whilst there were no significant differences between groups regarding depression scores (e.g. Crisford et al., 2008; Rusch et al., 2007).

Similarly, the effects of shame were rarely controlled for when investigating the guilt-PTSD link (e.g. Beck et al., 2011). Few studies directly tested the association between shame and PTSD symptoms and those that did found a significant relationship (e.g. Ginzburg et al., 2009; Street & Arias, 2001). It was shame, not guilt that was found to be a significant correlate or predictor of PTSD symptomology. The role of guilt remains difficult to determine in the absence of controlling for covariates significantly related to PTSD.

The diverse conceptualisation of guilt and its overlapping with related constructs may have led to inconsistent findings regarding the relationship between guilt and psychopathology. One concern is the appropriateness of measures used with trauma samples. The TOSCA was developed from college and community samples and reflects a person’s tendency to respond with guilt to everyday situations

(Tangney, Wagner, Fletcher & Gramzow, 1992). This may not tap into trauma- specific guilt in response to a traumatic event, which may account for a lack of a positive relationship between guilt and PTSD symptoms. The TOSCA also measures specific behavioural responses associated with emotional states such that guilt is framed with pro-social behaviour and shame with avoidance or negative self- evaluations, which has relevant implications for this review. Firstly, the TOSCA

48 guilt scale may reflect a measure of adaptive behaviour than a measure of guilt affect

(Silfver, 2007). This may have contributed to the lack of significant relationships reported between guilt and PTSD symptomology when using the TOSCA (Leskela et al., 2002; Semb et al., 2011; Street & Arias, 2001). Secondly, shame may be expected to correlate more positively with distress when maladaptive aspects of the construct are assessed (Luyten, Fontaine & Corveleyn, 2002). Discrepant findings may also suggest other measures have not adequately distinguished between guilt and shame (Tangney, Burggraf and Wagner, 1995). The guilt scale used by Crisford et al. (2008) included items referencing ‘self-hate’ and ‘ashamed,’ therefore it is possible that in its assessment of guilt this measure tapped into the construct of shame.

Additional limitations include the use of scales, shortened or translated into a different language without subsequent validation (e.g. Kubany et al., 1995; Lowinger

& Solomon, 2004; Rusch et al., 2007; Semb et al., 2011), which may have affected the integrity of findings. Other studies described translated measures with adequate psychometrics (Pereda et al., 2011; Strömsten et al., 2009). Two studies (Kubany et al., 1995; Kubany et al., 1997) assessed guilt derived from four or less items that may not have fully captured the construct of guilt, limiting content validity. The factor structure and internal consistency of these scales also remain unknown. The psychometrics of other measures may also be unreliable (Henning & Freuh, 1997).

Inadequate psychometrics for measures of PTSD were apparent, having included items that appear to tap into constructs reflecting psychopathology other than PTSD (e.g. Semb et al., 2011; Street & Arias, 2001). Some studies used measures that reported PTSD symptom severity based on whether respondents were

‘bothered’ by symptoms (e.g. Semb et al., 2011). Those experiencing chronic PTSD

49 may have adjusted to living with symptoms and are therefore less distressed by these, albeit still being present. Therefore there may be substantial variation in the reporting of PTSD symptomology that may affect the integrity of findings. Other considerations include the potential for factors to have inflated the relationship between guilt and PTSD symptoms. This may have been the case for studies using measures of PTSD symptomology that included guilt-related items in its assessment

(e.g. Pereda et al., 2011). Similarly, positive relationships between guilt-related distress and PTSD symptom scores could be expected given they both reflect distress. Research has suggested it may just tap into the higher level of overall distress experienced by trauma survivors (Marshall, Schell & Miles, 2010).

With regards to study design, five studies reported a sample size less than 50

(Crisford et al., 2008; Harned et al., 2012; Henning & Frueh, 1997; Kubany et al.,

2003; Semb et al., 2011). The majority continued to report significant findings suggesting effects were generally strong. Across three intervention studies, drop-out rates were reviewed as moderate and ranged from 15% by Ginzburg et al. (2009) up to 26.8% by Resick et al. (2002), which may have resulted in findings being reported from a biased sample by excluding those who did not tolerate the intervention. Two studies used data collected from the same sample (Keane et al., 1998; Marx et al.,

2010), however, sample sizes varied greatly between each study, as did PTSD symptomology, and the guilt-PTSD relationship was examined using an additional measure of PTSD in the study by Keane et al. (1998).

Lastly, all studies included a largely Caucasian sample, which has implications for generalising findings as response to trauma is likely to vary between cultures.

Discussion

One aim of this present review was to synthesise and evaluate the research evidence that may suggest an association between guilt and PTSD symptoms. A positive relationship between guilt and PTSD symptoms was generally observed across different samples. Guilt cognitions and attributions associated with perceived lack of justification, responsibility or preventability were associated with PSTD symptoms.

Beliefs about wrong doing were consistently related to PTSD symptomology suggesting that perceiving one’s role as violating personal or moral standards may be important in PTSD, which supports existing clinical models of guilt-based PTSD

(Kubany et al., 2003; Lee et al., 2001). Perceived lack of justification was associated less frequently with intrusive symptoms and avoidance suggesting this guilt cognition may be less distressing, which would necessitate fewer avoidant responses.

Treatments that aimed to modify cognitive distortions found a reduction in trauma- related guilt that corresponded with a reduction in PTSD symptoms. The lack of synchrony between guilt and PTSD symptoms across a few studies may suggest a mechanism of change, other than remedying guilt cognitions, is underlying the improvement in PTSD symptomology. Alternatively, specific guilt-related beliefs may not have been targeted as part of the treatment delivered.

A second objective was to evaluate the nature of the guilt-PTSD link by summarising the findings that support each of the four models conceptualising the relationship (see Figure 1). Mediation studies support the relationship proposed in

Model 1, whereby guilt acts as a meditational process underlying the development of

PTSD following trauma, which may have an intuitive appeal to clinicians as it clearly highlights the value of interventions directed at guilt. However, support for this model is limited by the lack of longitudinal data, which makes it impossible to

51 draw conclusions about the direction of effect. Consequently, Model 2, whereby guilt is a product of emerging PTSD, is equally plausible, although no studies tested this model directly.

The suggestion in Model 3 that Guilt and PTSD may have no meaningful relationship but are artefactually correlated due to their shared relationship with trauma received mixed support. The lack of correspondence between guilt and

PTSD reported by some intervention studies could be taken as support for this model, although there was no direct test of this relationship, which reduces the credibility of these findings. Mediation studies exploring whether guilt acts as a mediator between trauma and PTSD provide counter evidence for this model. These studies have shown that guilt and PTSD are linked over and above their shared relationship with trauma whilst controlling for the effect of guilt on depressive symptomatology within a path model in one study. Overall, there is currently little basis to support Model 3.

There was more convincing evidence to suggest the guilt-PTSD relationship may be better explained by shame than depressive symptomology, consistent with

Model 4. Guilt remained positively and significantly related to PTSD symptoms when controlling for the influence of negative affect suggesting guilt may be trauma- specific rather than a feature of low mood. The same could not be said when controlling for the effects of shame, a significant correlate and predictor of PTSD symptomology, suggesting shame may better explain the guilt-PTSD relationship.

Interestingly, guilt was negatively associated with PTSD symptomology when the effects of shame were removed or exhibited at low levels in two studies, which further supports the argument that guilt is associated with distress when fused with shame (Tangney et al., 2007). The lack of a significant relationship between guilt

52 and PTSD symptoms could be explained by the use of trait guilt measures that do not reflect event-specific guilt experienced post-trauma that may be driving symptoms of

PTSD.

Further to the main results, it was suggested that avoidance or disengagement may mediate the relationship between trauma-related guilt and PTSD symptomology. Consistent with this, prolonged exposure that serves to challenge avoidant responses, was found to reduce guilt cognitions and PTSD severity (Harned et al., 2012). These findings may explain why studies have found a correlational relationship between guilt and PTSD symptoms but why interventions have not consistently reported a corresponding reduction in both outcomes, as this may depend on various mediating processes, which would be important targets for intervention. Avoidance, as an explanatory link between guilt and PTSD is not consistent with the literature that states avoidance is typical in response to shame whereas guilt motivates reparative action (Tangney & Dearing, 2002). It could be that where this is not possible, guilt results in avoidant coping in an attempt to minimise acute distress, which ultimately maintains symptoms of post-traumatic stress longer term. Avoidant coping and disengagement has been claimed to interfere with emotional processing of a traumatic event and has been associated with PTSD severity and distress (Gil, 2005; Littleton, Horsley, John & Nelson, 2007).

Implications for future research

The methodological limitations that restrict our understanding of the guilt-PTSD relationship have important implications for future investigations. Several recommendations, considered relevant to the future testing of the four models presented in this review, have been made.

Longitudinal and experimental design

The majority of studies utilised a cross-sectional design making it difficult to determine causation and the directionality of effects or the influence of third variables that may mediate or moderate the relationship between guilt and PTSD.

Chronic PTSD may generate guilt over time (Model 2) or guilt-related peritraumatic appraisals may increase the likelihood of PTSD development (Model 1). This review does not allow for distinguishing between these two competing explanations.

Similarly, no causal inferences can be drawn from intervention studies reporting a reduction in guilt and PTSD symptomology post-treatment, as there may be alternative mechanisms of change that can account for both findings. Longitudinal studies exploring whether differences in baseline levels of guilt can predict subsequent changes in PTSD symptomatology (and vice versa) are necessary to make inferences about the direction of effect and to start to distinguish between

Model 1 and Model 2. Experimental research may be necessary to make firmer conclusions regarding causality but such research is difficult due to the ethical issues of experimentally manipulating aversive states such as guilt. Nonetheless, there may be a role for research in non-clinical analogue populations that could help to support basic principles underlying the putative guilt – PTSD relationship. Such studies could explore if short-term induced states of guilt can interfere with memory or produce PTSD-like phenomena such as intrusive imagery at a non-clinical level.

Experience sampling methodology, a novel approach of using moment-by-moment assessments taking place within the context of an individual’s daily life, could also be beneficial in establishing the temporal relationships between guilt and PTSD

(Palmier-Claus et al., 2011).

Assessing guilt and PTSD symptoms

The lack of conceptual clarity surrounding guilt has led to various measures assessing different aspects of the construct or process, which makes it difficult to fully explore the relationship between guilt and post-trauma sequelae. This review has highlighted the need for a unified measure of guilt encompassing cognitive, affective and behavioural components of this multi-faceted construct. Perceived wrong-doing, responsibility and self-blame may be particularly relevant to the concept of guilt. It is also important to distinguish between those who experience event-related guilt and those with a tendency to respond to everyday events with guilt. When investigating guilt post-trauma, a trauma-related guilt measure would be more appropriate where psychometric data has been developed from appropriate samples and reported cross-culturally.

The majority of studies assessed symptoms consistent with PTSD that were not always consistent with recent diagnostic criteria (e.g. Rusch et al., 2007) therefore findings may not be fully generalizable to those with a current diagnosis of

PTSD. The ‘gold standard’ structured diagnostic interview such as the CAPS (Blake et al., 1995) is recommended to reliably assess the relationship between guilt and

PTSD to ensure coverage of all relevant symptom clusters.

Confounding variables

In considering the link between PTSD and guilt, it is necessary to differentiate guilt from closely related affective experiences such as shame and control for any subsequent extraneous effects. This is required in order to further evaluate the validity of Model 4. Therefore a reliable measure of shame should be incorporated in future studies to help tease apart the effects of what appears to be two overlapping

55 constructs. As this review suggested, shame is a significant correlate of both guilt and PTSD symptomology, an area for future research could be to synthesise the quantitative evidence linking shame to PTSD to enable researchers to compare and contrast a shame-based model of PTSD with that of guilt. As guilt is also a feature of major depressive episodes and depression is highly comorbid with PTSD, the effects of mood should also be controlled for when testing the relationship between guilt and PTSD. Researchers may wish to consider including a general measure of negative affect, as some have argued it is the disposition to experiencing negative mood states that is important than guilt per se (Clark & Watson, 1995).

Behavioural responses to guilt

The findings of this review suggest avoidant coping is a mediator in the guilt-PTSD relationship for those likely to have endured on-going trauma. The use of disengagement coping styles during trauma of a longer duration has shown to be positively related to distress (Littleton et al., 2007). To understand how symptoms of

PTSD are maintained, future investigations of how respondents cope with guilt in the context of PTSD may be necessary, further refining existing clinical models of guilt- based PTSD. Several measures of coping have been developed (see Skinner, edge,

Altman & Sherwood, 2003, for a review); however, no existing measures assess coping with guilt specifically. This remains an area of development for future research, as does the relationship between guilt and behavioural markers of distress associated with PTSD. This includes suicide ideation and anger, which have shown to be related to PTSD and would have significant implications for managing risk in clinical practice (Glover, 1985; Panagioti et al., 2012).

Implications for clinical practice

The findings of this review suggest evaluation of one’s role in a traumatic event may contribute to post-traumatic stress for some trauma groups, which has important implications for clinical practice with respect to the application of cognitive therapies. It is recommended that clinicians be vigilant to the content of appraisals to identify themes consistent with guilt that may indicate the presence of a more guilt- based PTSD. Where appropriate, treatment interventions should target guilt based cognitions in addition to treating those associated with fear to produce effective treatment outcomes. As avoidant coping may mediate the relationship between guilt and PTSD symptoms, increased therapeutic focus on addressing unhelpful coping strategies that serves as a barrier to addressing underlying guilt may lead to a reduction in overall distress.

Intervention studies reviewed suggest cognitive therapy, whose focus was to remedy faulty guilt cognitions, may be more effective in reducing guilt than traditional exposure techniques that serve to break the avoidance-intrusion cycle that perpetuates anxiety. This suggests that those presenting with a guilt-based PTSD may benefit more from cognitive treatment. Findings in this review suggested exposure therapy was less efficacious than cognitive therapy in reducing guilt cognitions but was just as effective in reducing PTSD symptomology (e.g. Resick et al., 2002). This suggests there may be factors other than cognitive distortions maintaining distress that are amenable to exposure work. That said, unresolved guilt may not only maintain residual distress but leave the person vulnerable to re- traumatisation if exposed to trauma cues that elicit peritraumatic cognitions and guilt-related affect. Findings from this review found greater levels of guilt and PTSD symptoms among those exhibiting arousal in response to trauma-related material. It

57 may therefore be more appropriate to apply trauma-focused cognitive therapy in conjunction with exposure therapy. The application of cognitive trauma therapy for the treatment of guilt in PTSD is a manualized approach showing promising findings but requires further evaluation using more robust RCT methodology (Kubany &

Manke, 1995).

Findings suggest that having more direct involvement in a traumatic event, which may elicit greater guilt, is more strongly associated with the development of

PTSD. Therefore it could be argued that guilt may be a more prominent feature of

PTSD among specific trauma groups such as veterans who have engaged in frequent violent acts of commission or omission. This may underpin findings that show participating in abusive violence is predictive of PTSD above and beyond that explained by combat exposure alone (Beckham et al., 1998; Breslau & Davis, 1987).

For therapy to produce clinically meaningful outcomes in this trauma group, a detailed history of atrocities exposure that delineates one’s role and the delivery of treatment that targets multiple sources of war-related guilt may be necessary. There has been some support in the literature that the association between combat experience and PTSD remains after controlling for level of exposure (Fontana &

Rosenheck, 1993; Yehuda, Southwick & Giller, 1992). This suggests frequency or severity of exposure may not be an important factor. Instead, the findings of this review suggest it is the extent and evaluation of one’s role in a traumatic event that appears to influence pathology.

The present review highlighted the on-going issue regarding the conceptual overlap between shame and guilt, which both feature in PTSD (Wilson et al., 2006).

Few studies reviewed supported the notion that guilt may be related to PTSD symptomology only when combined with shame, which has been long argued to be a

58 stronger correlate and predictor of distress and psychopathology overall (Tangney et al., 2007). A distinction ought to be made between the two when clients present to services, as it may necessitate a different psychological approach to treatment.

Historically, shame has been differentiated from guilt, in part, based on behavioural responses such that shame prompts concealment and avoidance. However, findings from this review suggest guilt may be associated with avoidant coping that not only makes identifying this affective experience more difficult but may highlight a barrier to treatment. Similarly, it may be important to clinically distinguish between trauma- related guilt and guilt experienced in the context of depression. The use of avoidant coping in response to guilt would imply the experience of guilt was intrusive and threatening in some way, which may closely resemble the affective experience of anxiety than depression.

Conclusion

This review found evidence for a relationship between guilt, particularly cognitions such as perceived wrong doing and self-blame, and PTSD symptoms. The direction of causality remains difficult to ascertain. The guilt-PTSD relationship may have been confounded by overlapping constructs often not controlled for (e.g. shame).

Nonetheless, the present review represents a significant step forward in clarifying how guilt and PTSD may interact, and in providing a framework, in terms of four competing models, for guiding future experimental research in this area. Intervention studies should incorporate treatments that aim to reduce both guilt cognitions and

PTSD symptomology and consider whether addressing mediating variables (e.g. coping style) produces more successful outcomes.

References

Alexander, B., Brewin, C. R., Vearnals, S., Wolff, G., & Leff, J. (1999). An investigation of shame and guilt in a depressed sample. British Journal of Medical Psychology, 72, 323-338.

American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders (4th ed.). Washington, DC: APA.

Andrews, B., Brewin, C. R., Rose, S., & Kirk, M. (2000). Predicting PTSD symptoms in victims of violent crime: The role of shame, anger and childhood abuse. Journal of Abnormal Psychology, 109, 69-73.

Baumeister, R. F., Stillwell, A. M., & Heatherton, T. F. (1994). Guilt: An interpersonal approach. Psychological Bulletin, 115, 243-267.

Beck, G. J., Coffey, S. F., Palyo, S. A., Gudmundsdottir, B., Miller, L. M., & Colder, C. R. (2004). Psychometric properties of the posttraumatic cognitions inventory (PTCI): A replication with motor vehicle accident survivors. Psychological Assessment, 16, 289-298.

Beck, J. G., McNiff, J., Clapp, J. D., Olsen, S. A., Avery, M. L., & Hagewood, J. H. (2011). Exploring negative emotion in women experiencing intimate partner violence: Shame, guilt and PTSD. Behavior Therapy, 42, 740-750.

Beckham, J. C., Feldman, M. E., & Kirby, A. C. (1998). Atrocities exposure in Vietnam combat veterans with chronic posttraumatic stress disorder: Relationship to combat exposure, symptom severity, guilt, and interpersonal violence. Journal of Traumatic Stress, 11(4), 777–785.

Bennice, J. A., Grubaugh, J. L., & Resick, P. A. (2001). Guilt, depression and PTSD among female rape victims. Poster presented at the 17th annual meeting of the International Society for Traumatic Stress Studies. New Orleans, LA.

Benotsch, E. G., Brailey, K., Vasterling, J. J., & Uddo, M. (2000). War zone stress, personal and environmental resources, and PTSD symptoms in Gulf War veterans: A longitudinal perspective. Journal of Abnormal Psychology, 109, 205–213.

Blake, D. D., Weathers, F. W., Nagy, L. M., Kaloupek, D. G., Gusman, F. D., Charney, D. S., & Keane, T. M. (1995). The development of a clinician- administered PTSD scale. Journal of Traumatic Stress, 8, 75-90.

Blanchard, E. B., Buckley, T. C., Hickling, E. J., & Taylor, A. E. (1998). Posttraumatic stress disorder and comorbid major depression: Is the correlation an illusion? Journal of Anxiety Disorders, 12, 21-37.

Breslau, N., & Davis, G.C. (1987). Posttraumatic stress disorder: The etiologic specificity of wartime stressors. American Journal of Psychiatry, 144, 578- 583.

Breslau, N., Davis, G. C., Andreski, P., & Peterson, E. (1991). Traumatic events and posttraumatic stress disorder in an urban population of young adults. Archives of General Psychiatry, 48(3), 216-222.

Brewin, C., Andrews, B., & Rose, S. (2000). Fear, helplessness and horror in posttraumatic stress disorder: Investigating DSM-IV criterion 2 in victims of violent crime. Journal of Traumatic Stress, 13(3), 499-509.

Browne, T., Evangeli, M., & Greenberg, N. (2012). Trauma-related guilt and posttraumatic stress among journalists. Journal of Traumatic Stress, 25, 207- 210.

Burney, J., & Irwin, H. J. (2000). Shame and guilt in women with eating disorder symptomology. Journal of Clinical Psychology, 56, 51-61.

Carter, J. D., Frampton, C. M., Mulder, R. T., Luty, S. E., & Joyce, P. R. (2010). The relationship of demographic, clinical, cognitive and personality variables to the discrepancy between self and clinician rated depression. Journal of Affective Disorders, 124, 202-206.

Clark, L. A., & Watson, D. B. (1995). Constructing validity: Basic issues in objective scale development. Psychological Assessment, 7, 309-319.

Cloitre, M., Stolbach, B. C., Herman, J. L., Van der Kolk, B., Pynoos, R., Wang, J., & Petkova, E. (2009). A developmental approach to complex PTSD: Childhood and adulthood cumulative trauma as predictors of symptom complexity. Journal of Traumatic Stress, 22, 399-408.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Earlbaum Associates.

Crisford, H., Dare, H., & Evangeli, M. (2008). Offence-related posttraumatic stress disorder (PTSD) symptomatology and guilt in mentally disordered violent and sexual offenders. The Journal of Forensic Psychiatry & Psychology, 19(1), 86-107.

Dalgleish, T. (2004). Cognitive approaches to posttraumatic stress disorder: The evolution of multirepresentational theorizing. Psychological Bulletin, 130(20), 228-260.

Davidson, J. R. T., Book, S. W., Colket, J. T., Tupler, L. A., Roth, S., David, D.,…Feldman, M. E. (1997). Assessment of a new self-rating scale for posttraumatic stress disorder: The Davidson Trauma Scale. Psychological Medicine, 27, 153-160.

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

Ehlers, A., & Steil, R. (1995). Maintenance of intrusive memories in posttraumatic stress disorder: A cognitive approach. Behavioural Cognitive and Psychotherapy, 23, 217-249.

Ferguson, T. J., & Crowley, S. L. (1997). Measure for measure: A multitrait- multimethod analysis of guilt and shame. Journal of Personality Assessment, 69, 425-441.

Foa, E. B., Ehlers, A., Clark, D. M., Tolin, D. F., & Orsillo, S. M. (1999). The Posttraumatic Cognitions Inventory (PTCI): Development and validation. Psychological Assessment, 11, 303-314.

Foa, E. B., & Rauch, S. A. M. (2004). Cognitive changes during prolonged exposure verses prolonged exposure plus cognitive restructuring in female assault survivors with posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 72, 879-884.

Foa, E. B., & Riggs, D. S. (1993). Posttraumatic stress disorder in rape victims. In J. Oldham., M. B. Riba., & A. Tasman (Eds.), American psychiatric press review of psychiatry (vol. 12, pp. 273-303). Washington, DC: American Psychiatric Press.

Foa, E. B., & Rothbaum, B. O. (1999). Treating the trauma of rape: Cognitive behavioural therapy for PTSD. New York, NY: Guilford Press.

Fontana, A., & Rosenheck, R. (1993). A causal model of the etiology of war related PTSD. Journal of Traumatic Stress, 6, 475-500.

Frijda, N. H. (1993). The place of appraisal in emotion. Cognition and Emotion, 7, 357-387.

Gil, S. (2005). Coping style in predicting posttraumatic stress disorder among Israeli students. Anxiety, Stress & Coping: An International Journal, 18, 351-359.

Ginzburg, K., Butler, L. D., Giese-Davis, J., Cavanaugh, C. E., Neri, E., Koopman, C.,…Spiegel, D. (2009). Shame, guilt, and posttraumatic stress disorder in adult survivors of childhood sexual abuse at risk for human immunodeficiency virus: Outcomes of a randomized clinical trial of group psychotherapy treatment. The Journal of Nervous and Mental Disease, 197(7), 536-42.

Glover, H. (1985). Guilt and aggression in Vietnam veterans. The American Journal of Social Psychiatry, 1, 15-18.

Harder, D. W. (1995). Shame and guilt assessment, and relationships of shame-and guilt-proneness to psychopathology. In J. P. Tangney & K. W. Fischer (Eds.), Self-conscious emotions: The psychology of shame, guilt, embarrassment and pride (pp. 368-392). New York, NY: Guildford Press.

Harder, D. W., & Greenwald, D. F. (1999). Further validation of the shame and guilt scales of the harder personal feelings questionnaire-2. Psychological Reports, 85, 271-281.

Harned, M. S., Korslund, K. E., Foa, E. B., & Linehan, M. M. (2012). Treating PTSD in suicidal and self-injuring women with borderline personality disorder: Development and preliminary evaluation of a Dialectical Behavior Therapy Prolonged Exposure Protocol. Behaviour Research and Therapy, 50(6), 381-386.

Hathaway, L. M., Boals, A., & Banks, J. B. (2010). PTSD symptoms and dominant emotional response to a traumatic event: An examination of DSM-IV Criterion A2. Anxiety, Stress & Coping, 23(1), 119-126.

Held, P., Owens, G. P., Schumm, J. A., Chard, K. M., & Hansel, J. E. (2011). Disengagement coping as a mediator between trauma-related guilt and PTSD severity. Journal of Traumatic Stress, 24(6), 708-715.

Henning, K. R., & Frueh, B. C. (1997). Combat guilt and its relationship to PTSD symptoms. Journal of Clinical Psychology, 53(8), 801-808.

Horowitz, M. J., Wilner, N., & Alvarez, W. (1979). Impact of Event Scale: A measure of subjective stress. Psychosomatic Medicine, 41, 209-218.

Janoff-Bulman, R. (1992). Shattered assumptions: Towards a new psychology of trauma. New York, NY: Free Press.

Jones, W. H., Schratter, A. K., & Kugler, K. (2000). The Guilt Inventory. Psychological Reports, 87, 1039-1042.

Keane, T. M., Caddell, J. M., & Taylor, K. L. (1988). Mississippi Scale for Combat- Related Posttraumatic Stress Disorder: Three studies in reliability and validity. Journal of Consulting and Clinical Psychology, 56, 85-90.

Keane, T. M., Kolb, L. C., Kaloupek, D. G., Orr, S. P., Blanchard, E. B., Thomas, R. G.,…Lavori, P.W. (1998). Utility of psychophysiological measurement in the diagnosis of posttraumatic stress disorder: Results from a Department of Veterans Affairs Cooperative Study. Journal of Consulting and Clinical Psychology, 66(6), 914-923.

Kessler, R. (2000). Post-traumatic stress disorder: The burden to the individual and to society. Journal of Clinical Psychiatry, 61, 4-12.

Kessler, R. C., Chiu, W. T., Demler, O., & Walters, E. E. (2005). Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 62, 617- 627.

Kim, S., Thibodeau, R., & Jorgensen, R. S. (2011). Shame, guilt, and depressive symptoms: A meta-analytic review. Psychological Bulletin, 137, 68-96.

Klass, E. T. (1990). Guilt, shame and embarrassment: Cognitive-behavioural approaches. In H. Leiten-berg (Ed.), Handbook of social and evaluation anxiety (pp. 385-414). New York, NY: Plenum.

Kubany, E. S., Abueg, F. R., Kilauano, W. L., Manke, F. P., & Kaplan, A. S. (1997). Development and validation of the Sources of Trauma-Related Guilt Survey – War-Zone Version. Journal of Traumatic Stress, 10(2), 235-258.

Kubany, E. S., Abueg, F. R., Owens, J. A., Brennan, J. M., Kaplan, A. S., & Watson, S. B. (1995). Initial examination of a multidimensional model of trauma- related guilt: Applications to combat veterans and battered women. Journal of Psychopathology and Behavioural Assessment, 17(4), 353-375.

Kubany, E. S., Haynes, S. N., Abueg, F. R., Manke, F. P., Brennan, J. M., & Stahura, C. (1996). Development and validation of the Trauma-Related Guilt Inventory (TRGI). Psychological Assessment, 8(4), 428-444.

Kubany, E. S., Hill, E. E., & Owens, J. A. (2003). Cognitive trauma therapy for battered women with PTSD: Preliminary findings. Journal of Traumatic Stress, 16(1), 81–91.

Kubany, E. S., Hill, E. E., Owens, J. A., Iannce-Spencer, C., McCaig, M. A., Tremayne, K. J., & Williams, P. L. (2004). Cognitive trauma therapy for battered women with PTSD (CTT-BW). Journal of Consulting and Clinical Psychology, 72(1), 3-18.

Kubany, E. S., & Manke, F. P. (1995). Cognitive therapy for trauma-related guilt: Conceptual bases and treatment outlines. Cognitive and Behavioral Practice, 2, 23-61.

Kubany, E. S., & Watson, S. B. (2003). Guilt: Elaboration of multidimensional model. The Psychological Record, 53, 51-90.

Kugler, K., & Jones, W. H. (1992). On conceptualizing and assessing guilt. Journal of Personality and Social Psychology, 62, 318-327.

Lauterbach, D., Vrana, S., King, D. W., & King, L. A. (1997). Psychometric properties of the Civilian Version of the PTSD Mississippi Scale. Journal of Traumatic Stress, 10(3), 499-513.

Lee, D. A., Scragg, P., & Turner, S. (2001). The role of shame and guilt in traumatic events: A clinical model of shame-based and guilt-based PTSD. British Journal of Medical Psychology, 74, 451-466.

Leskela, J., Dieperink, M., & Thuras, P. (2002). Shame and posttraumatic stress disorder. Journal of Traumatic Stress, 15(3), 223-226.

Littleton, H., Horsley, S., John, S., & Nelson, D. V. (2007). Trauma coping strategies and psychological distress: A meta-analysis. Journal of Traumatic Stress, 20, 977-988.

Lowinger, T., & Solomon, Z. (2004). PTSD, guilt, and shame among reckless drivers. Journal of Loss and Trauma, 9, 327-344.

Lutwak, N., Panish, J. B., Ferrari, J. R., & Razzino, B. E. (2001). Shame and guilt and their relationship to positive expectations and anger expressiveness. Adolescence, 36(144), 641-654.

Luyten, P., Fontaine, J. R. J. & Corveleyn, J. (2002). Does the Test of Self- Conscious Affect (TOSCA) measure maladaptive aspects of guilt and adaptive aspects of shame? An empirical investigation. Personality and Individual Differences, 33(8), 1373-1387.

Marshall, G. N., Schell, T. L., & Miles, J. N. V. (2010). All PTSD symptoms are highly associated with general distress: Ramifications for the Dysphoria Symptom Cluster. Journal of Abnormal Psychology, 119, 126-135.

Marx, B. P., Foley, K. M., Feinstein, B. A., Wolf, E. J., Kaloupek, D. G., & Keane, T. M. (2010). Combat-related guilt mediates the relations between exposure to combat-related abusive violence and psychiatric diagnoses. Depression and Anxiety, 27, 287-293.

McCann, I. L., & Pearlman, L. A. (1990). Psychological trauma and the adult Survivor: Theory, therapy and transformation. New York, NY: Brunner and Mazel.

Mollica, R. F., Caspi-Yavin, Y., Bollini, P., Truong, T., Tor, S., & Lavelle, J. (1992). Validating a cross-cultural instrument for measuring torture, trauma, and posttraumatic stress disorder in Indochinese refugees. Journal of Nervous and Mental Disease,180, 111-116.

Mosher, D. L. (1988). Revised Mosher Guilt Inventory. In C. M. Davis & W. L. Yarber (Eds.), Sexuality-related measures: A compendium. Lake Mills, IA: Graphic Publications.

Niler, E. R., & Beck, S. J. (1989). The relationship among guilt, dysphoria, anxiety and obsessions in a normal population. Behaviour Research and Therapy, 27, 213-220.

Nishith, P., Nixon, R., & Resick, P. (2005). Resolution of trauma-related guilt following treatment of PTSD in female rape victims: A result of cognitive processing therapy targeting comorbid depression? Journal of Affective Disorders, 86, 259-265.

Owens, G. P., Chard, K. M., & Cox, T. A. (2008). The relationship between maladaptive cognitions, anger expression, and posttraumatic stress disorder among veterans in residential treatment. Journal of Aggression, Maltreatment & Trauma, 17(4), 439-452.

Owens, G. P., Steger, M. F., Whitesell, A. A., & Herrera, C. J. (2009). Posttraumatic stress disorder, guilt, depression, and meaning in life among military veterans. Journal of Traumatic Stress, 22(6), 654-657.

Ozer, E. J., Best, S. R., Lipsey, T. L., & Weiss, D. R. (2003). Predictors of posttraumatic stress disorder and symptoms in adults: A meta-analysis. Psychological Bulletin, 129, 52-73.

Palmier-Claus, J. E., Myin-Germeys, I., Barkus, E., Bentley, L., Udachina, A, Delespaul, P. A.,…Dunn, G. (2011). Experience sampling research in individuals with mental illness: Reflections and guidance. Acta Psychiatrica Scandinavica, 123(1), 12-20.

Panagioti, M., Gooding, P. A., & Tarrier, N. (2012). A meta-analysis of the association between posttraumatic stress disorder and suicidality: The role of comorbid depression. Comprehensive Psychiatry, 53, 915-930.

Pereda, N., Arch, M., Peró, M., Guàrdia, J., & Forns, M. (2011). Assessing guilt after traumatic events: The Spanish adaptation of the Trauma-Related Guilt Inventory. European Journal of Psychological Assessment, 27(4), 251–257.

Pitman, R. K., Altman, B., Greenwald, E., Longpre, R. E., Macklin, M. M., Poire, R. E., & Steketee, G. S. (1991). Psychiatric complications during flooding therapy for posttraumatic stress disorder. Journal of Clinical Psychiatry, 52(1), 17-20.

Power, M. J., & Fyvie, C. (2013). The role of emotion in PTSD: Two preliminary studies. Behavioural and Cognitive Psychotherapy, 41(2), 162-172.

Resick, P. A., Nishith, P., Weaver, T. L., Astin, M. C., & Feuer, C. A. (2002). A comparison of cognitive-processing therapy with prolonged exposure and a waiting condition for the treatment of chronic posttraumatic stress disorder in female rape victims. Journal of Consulting and Clinical Psychology, 70(4), 867-879.

Resick, P. A., Uhlmansiek, M. O. B., Clum, G. A., Galovski, T. E., Scher, C. D., & Young-Xu, Y. (2008). A randomized clinical control trial to dismantle components of cognitive processing therapy for posttraumatic stress disorder in female victims of interpersonal violence. Journal of Consulting and Clinical Psychology, 76(2), 243-258.

Robinaugh, D. J., & McNally, R. J. (2010). Autobiographical memory for shame or guilt provoking events: Association with psychological symptoms. Behaviour Research and Therapy, 48, 646-652.

Rüsch, N., Corrigan, P. W., Bohus, M., Kühler, T., Jacob, G. A., & Lieb, K. (2007). The impact of posttraumatic stress disorder on dysfunctional implicit and explicit emotions among women with borderline personality disorder. The Journal of Nervous and Mental Disease, 195(6), 537-539.

Sareen, J., Cox, B. J., Goodwin, R. D., & Asmundson, G. J. G. (2005). Co- occurrence of posttraumatic stress disorder with positive psychotic symptoms in a nationally representative sample. Journal of Traumatic Stress, 18, 313- 322.

Semb, O., Strömsten, L. J., Sundbom, E., Fransson, P., & Henningsson, M. (2011). Distress after a single violent crime: How shame-proneness and event-related shame work together as risk factors for post-victimization symptoms. Psychological Reports, 109(1), 3-23.

Shapiro, L. J., & Stewart, S. E. (2011). Pathological guilt: A persistent yet overlooked treatment factor in obsessive-compulsive disorder. Annals of Clinical Psychiatry, 23(1), 63-70.

Silfver, M. (2007). Coping with guilt and shame: A narrative approach. Journal of Moral Education, 36(2), 169-183.

Singer, M. (2004). Shame, guilt, self-hatred and remorse in the psychotherapy of Vietnam combat veterans who committed atrocities. American Journal of Psychotherapy, 58, 377-385.

Skinner, E. A., Edge, K., Altman, J., & Sherwood, H. (2003). Searching for the structure of coping: A review and critique of category systems for classifying ways of coping. Psychological Bulletin, 129, 216-269.

Smith, K., & Bryant, R. A. (2000). The generality of cognitive bias in acute stress disorder. Behaviour Research and Therapy, 38(7), 709-715.

Smith-Fawsi, M. C., Pham, T., Lin, L., Nguyen, T. V., Ngo, D., Murphy, E., & Mollica, R. F. (1997). The validity of posttraumatic stress disorder among Vietnamese refugees. Journal of Traumatic Stress, 10, 101-108.

Startup, M., Makgekgenene, L., & Webster, R. (2007). The role of self-blame for trauma as assessed by the Posttraumatic Cognitions Inventory (PTCI): A self- protective cognition? Behaviour Research and Therapy, 45(2), 395-403.

Street, A. E., & Arias, I. (2001). Psychological abuse and posttraumatic stress disorder in battered women: Examining the roles of shame and guilt. Violence and Victims, 16(1), 65-78.

Street, A. E., Gibson, L. E., & Holohan, D. R. (2005). Impact of childhood traumatic events, trauma-related guilt, and avoidant coping strategies on PTSD symptoms in female survivors of domestic violence. Journal of Traumatic Stress, 18(3), 245-252.

Strömsten, L. M. J., Henningsson, M., Holm, U., & Sundbom, E. (2009). Assessment of self-conscious emotions: A Swedish psychometric and structure evaluation of the Test of Self-Conscious Affect (TOSCA). Scandinavian Journal of Psychology, 49, 71-77.

Tangney, J. P. (1990). Assessing individual differences in proneness to shame and guilt: Development of the Self Conscious Affect and Attribution Inventory. Journal of Personality and Social Psychology, 59, 102-111.

Tangney, J. P. (1992). Situational determinants of shame and guilt in young adulthood. Personality and Social Psychology Bulletin, 18, 199-206.

Tangney, J. P., Burggraf, S. A., & Wagner, P. E. (1995). Shame-proneness, guilt- proneness, and psychological symptoms. In J. P. Tangney & K. W. Fischer (Eds.), Self-conscious emotions: The psychology of shame, guilt, embarrassment, and pride (pp. 343-367). New York, NY: Guilford Press.

Tangney, J. P., & Dearing, R. L. (2002). Shame and guilt. New York, NY: Guilford Press.

Tangney, J. P., Miller, R. S., Flicker, L., & Barlow, D. H. (1996). Are shame, guilt and embarrassment distinct emotions? Journal of Personality and Social Psychology, 70, 1256-1269.

Tangney, J. P., Stuewig, J., & Mashek, D. (2007). Moral emotions and moral behaviour. Annual Review of Psychology, 58, 345-372.

Tangney, J. P., Wagner, P., Fletcher, C., & Gramzow, R. (1992). Shamed into anger? The relation of shame and guilt to anger and self-reported aggression. Journal of Personality and Social Psychology, 62, 669-675.

Tangney, J. P., Wagner, P., & Gramzow, R. (1989). The Test of Self-Conscious Affect. Fairfax, VA: George Mason University.

Tangney, J. P., Wagner, P., & Gramzow, R. (1992). Proneness to shame, proneness to guilt, and psychopathology. Journal of Abnormal Psychology, 101, 469- 478.

Tilghman-Osborne, C., Cole, A. D., & Felton, J. W. (2010). Definition and measurement of guilt: Implications for clinical research and practice. Clinical Psychology Review, 30, 536-546.

Vreven, D. L., Gudanowski, D. M., King, L. A., & King, D. W. (1995). The Civilian Version of the Mississippi PTSD Scale: A psychometric evaluation. Journal of Traumatic Stress, 8, 91-109.

Williams, C., & Bybee, J. (1994). What do children feel guilty about? Development and gender differences. Developmental Psychology, 30, 617-623.

Wilson, J. P., Drozdek, B., & Turkovic, S. (2006). Posttraumatic shame and guilt. Trauma, Violence and Abuse, 7, 122-141.

Wong, M. R., & Cook, D. (1992). Shame and its contribution to PTSD. Journal of Traumatic Stress, 5, 557-562.

Yehuda, R., Southwick, S. M., & Giller, E. L. (1992). Exposure to atrocities and severity of chronic posttraumatic stress disorder in Vietnam combat veterans. American Journal of Psychiatry, 149, 333-336.

Paper 2: Empirical Study

Validating a Measure of Coping with Guilty Thoughts in a

Clinical Sample

Preface

Research for the empirical study and the subsequent write-up took place between

August 2012 and June 2013. There were a number of collaborators that made significant contributions towards this research. Dr Katherine Berry provided academic support and supervision for the study including reading drafts of the manuscript. Dr Katherine McIvor developed the GLAMS and supported the organisation of the study as field supervisor. Amy Degnan further supported the conceptual development of the GLAMS and provided theoretical support to the current study having provided validation of the GLAMS in a student sample prior.

The research paper was prepared in line with the journal guidelines for

“Psychological Assessment”. A copy of the author information pack is available in

Appendix C. Tables have been left in the main body of the text to aid reading.

Abstract

Over the last few years there has been a remarkable growth of empirical studies investigating the definition, function and measurement of guilt but research on coping with guilt remains scarce. Although guilt has been associated with prosocial behaviour and positive outcomes, guilt, when chronic or unresolved, may lead to psychopathology. Coping response may mediate this relationship, as some coping strategies are more adaptive than others in response to distress. There exists to date, no measure of coping with guilt. The objective of this study was to provide preliminary validation of a newly developed Guilt Level and Management Scale

(GLAMS) in a clinical sample. Sixty-seven participants attending psychological services completed the GLAMS, measures of thought control, coping, guilt and distress. Eighteen participants completed the GLAMS and distress measure again two weeks later. Moderate to high internal consistency was reported for all but one

GLAMS subscale ( = 0.70-0.89). Test re-test reliability was variable (r = 0.02-

0.82), with more maladaptive subscales demonstrating greater stability over time.

Using a correlational analysis, acceptable to good standards of concurrent validity were evidenced with theoretically similar subscales relating to thought control and coping. In support of construct validity, self-punishment was associated with greater guilt and distress and mindfulness-based coping was related to a reduction in guilt.

Study limitations, implications for research and clinical practice are discussed.

Preliminary evidence is given that the GLAMS is a valid and reliable measure with promising clinical utility in guiding intervention towards adaptive ways of coping with guilt-related distress.

Key Words

Guilt, coping, intrusive thoughts, psychometric, scale

Introduction

Guilt is considered a multifaceted construct, possessing emotional and cognitive components mediated by social and moral standards of behaviour (Kugler & Jones,

1992), which when violated result in negative consequences (Tilghman-Osborne,

Cole & Felton, 2010). Guilt has been construed as an intrusion into awareness or affective response (Baumeister, Stillwell & Heatherton, 1994; Niler & Beck, 1989), it may be trait-like and enduring (Tangney, Stuewig & Mashek, 2007) or state-like, pertaining to the consequences of specific transgressions. Although guilt carries negative connotations there are discrepant findings regarding its function and consequences. Guilt has been referred to as adaptive associated with personality development, empathy and reconciliatory actions promoting social contact and healthy relationships (Fontaine et al., 2006; Tangney, 1990; Zahn-Waxler &

Kochanska, 1990). Guilt has also been considered maladaptive associated with regret and self-punishment resulting in impaired social behaviour and distress (Kugler &

Jones, 1992; Tangney, 1990).

It is therefore not surprising the relationship between guilt and psychopathology is inconsistent. Guilt has shown to be positively related to various indices of psychopathology (Alexander, Brewin, Vearnals, Wolff & Leff, 1999;

Burney & Irwin, 2000; Shapiro & Stewart, 2011). However, inverse relationships have too been highlighted (Sanftner, Barlow, Marschall & Tangney, 1995; Tangney,

1991). The guilt-distress relationship is highly variable depending on how guilt has been conceptualised and measured (Silfver, 2007; Tangney & Dearing, 2002). It is also possible that guilt left unresolved becomes chronic and maladaptive, which may depend on the coping strategy employed.

The coping with distress literature remains vast, whereas existing literature on coping with guilt is still in its infancy. Coping is a multidimensional process involving a plethora of cognitions and behaviours used to deal with stressful encounters (Folkman & Moskowitz, 2004). There exists a lack of consensus regarding how the construct should be conceptualised and much diversity in what coping measures assess making it difficult to aggregate findings across studies (see

Skinner, Edge, Altman & Sherwood, 2003 for a review). Researchers have clustered together coping responses using theory-based higher order categories. One example is the theoretical distinction made between problem-focused coping (PFC) that modifies the problem source and emotion-focused coping (EFC), which ameliorates distress (Lazarus and Folkman, 1984). These types of coping have often been found to co-occur due to changing demands and resources (Tennen, 2000).

The stability of the coping construct has received mixed reviews. Lazarus and

Folkman (1984) suggested coping responses are situation specific and a dynamic process that may change over time. Theoretically, there ought to be some stability in coping as a result of personality attributes or static environmental factors. Research findings have highlighted personal and situational factors that may predict consistency or instability in coping across events and time (Carver, Scheier &

Weintraub, 1989; Terry, 1994). Identifying ways of coping that are more or less effective in reducing distress remains a perplexing issue in coping research. Adaptive value is thought to depend on the dynamic and situational context (Lazarus &

Folkman, 1984). What is considered effective coping at the offset may be less effective over time. Similarly, coping may be effective in one situation and not another. PFC is considered adaptive when an event is appraised as controllable and less so when unchangeable, in which case EFC would be preferable (Terry & Hynes,

1998; Vitaliano, DeWolfe, Maiuro, Russo & Katon, 1990). This may be relevant to guilt as reparation for past transgressions may not be possible, which may facilitate the use of EFC.

A wealth of literature has identified, however, specific coping responses to be more adaptive than others, many of which may be relevant to guilt-related distress.

Adaptive coping has been associated with PFC such as problem-solving and seeking instrumental social support (Thoits, 1995). Social support may also be emotionally beneficial by providing opportunity to receive sympathy (Semmer et al., 2008) or to enable confession. Problem-solving of a reparative nature may be unique to guilt and prevent chronicity by channelling guilt into positive behaviour (Bybee & Quiles,

1998; Fontaine et al., 2006). Adaptive cognitive strategies have too been highlighted such as positive reappraisal where adversity is construed as benign or beneficial

(Charlton & Thompson, 2011; Garnefski et al., 2002). On the contrary, maladaptive reappraisal is related to increased susceptibility to affective disorders (Beck, 1991;

Martin & Dahlen, 2005), the restructuring of which is central to the efficacy of cognitive and behavioural therapy (CBT; Beck, 1991). This may be particularly relevant to guilt, as appraisals relating to self-blame and wrong doing may maintain guilt-related pathology (Kubany & Manke, 1995).

Several other EFC strategies have consistently shown to be maladaptive in response to distress. Self-punishment has been associated with psychopathology, particularly where compensation for past wrongdoings is not possible (Nelissen &

Zeelenberg, 2009; Warda & Bryant, 1998). Self-punishment in response to guilt has been indicated where guilt was found to be inversely related to self-forgiveness

(Strelan, 2007). Self-punitive feelings of guilt have been associated with poorer mental health (Bybee & Quiles, 1998). Religiosity, however, has produced mixed

76 findings such that positive religious coping has been associated with improved and poorer psychological outcomes (Pargament, 1997).

Ways of coping commonly cited as maladaptive often reflect thought control processes such as rumination and worry that have been associated with psychopathology and poor recovery (Nolen-Hoeksema, 2000; Reynolds & Wells,

1999; Warda & Bryant, 1998). Studies have suggested the role of perseverative processes and thought suppression in elevated perceived moral wrongness and guilt

(Fontaine et al., 2006; Lee, Scragg & Turner, 2001; Marcks & Woods, 2007).

Thought suppression may be related to cognitive avoidance, which has been implicated in various psychological disorders (Aldao, Nolen-Hoeksema &

Schweizer, 2010). Avoidance or disengagement are argued to be forms of dysfunctional coping (DC), which have shown to be related to increased distress and was found to mediate the relationship between guilt and pathology (Carver et al.,

1989; Held, Owens, Schumm, Chard & Hansel, 2011). Acceptance, which underlies mindfulness-based practice (Segal, Williams & Teasdale, 2002) has been viewed as an adaptive alternative to experiential avoidance (Marcks & Woods, 2005) and may be relevant in guilt management.

Although mindfulness has sat at the heart of Buddhist traditions for centuries, it has only gained attention in psychological literature more recently. Mindfulness is a non-judgemental, present-centred awareness where experiences are accepted as they are in a detached way (Kabat-Zinn, 1990; Segal et al., 2002). It is metacognitive in the sense it involves a metalevel of awareness that monitors cognition whilst reflecting back on the process. This shift in cognitive set is postulated to reduce distress by allowing one to step back from mental experiences, treating them as momentary events than a reflection of the self or reality (Segal et al., 2002).

Few coping measures have included mindfulness and others are argued to be narrow in focus or a product of inadequate sampling of coping (Carver et al., 1989).

Few have incorporated acceptance yet remain limited in their assessment of meta- awareness and cognition in response to intrusive experiences. Measures of thought control (e.g. Wells & Davies, 1994) have indicated that certain strategies are maladaptive in response to intrusive thoughts, although not guilt-specific. Currently, there exists no validated measure of coping with guilt. Existing guilt measures are limited in their over-reliance on responders’ interpretation of guilt and demonstrate little content overlap highlighting the importance of a theoretically-informed definition of guilt (Kim, Jorgensen & Thibodeau, 2011; Tilghman-Osborne et al.,

2010). Many coping measures are also empirically not theoretically derived and contain ambiguous items or confounds that may have inflated the coping-distress relationship (Carver et al., 1989; Stanton, Danoff-Burg, Cameron & Ellis, 1994). In the absence of suitable instruments, understanding how coping mediates the relationship between distressing events and guilt remains elusive.

A measure of coping with guilty thoughts (GLAMS; Guilt Level and

Management Scale) was developed, which unlike previous measures of coping, was theoretically derived and has incorporated mindfulness-based coping that reflects metacognitive awareness and acceptance. It is also inclusive of several other coping responses thought to be relevant to guilt management. A recent study examining the factor structure and psychometrics of this scale identified a six-factor solution.

Reparative action, external amelioration relating to social support, spirituality and metacognitive thinking were identified as adaptive in response to guilt, whereas self- punishment and avoidance were indicative of maladaptive coping (Degnan, 2013).

The validity of the GLAMS to date had been tested in a student population, the

78 clinical relevance of which remains unknown. Therefore the primary aim of the present study was to provide validation to the GLAMS in a clinical sample. The following predictions were made:

Reliability

The GLAMS scale and Guilt Management subscales would be internally consistent and subscales would demonstrate stability over 2-4 weeks.

Concurrent validity

With regards to coping, GLAMS External Amelioration, Metacognition and

Spirituality would be positively related to EFC; GLAMS Reparation and External

Amelioration would be positively related to PFC; GLAMS Self-Punishment and

Avoidance would be positively related to DC. With regards to thought control,

GLAMS Self-Punishment would be positively related to Worry and Punishment;

GLAMS External Amelioration would be positively related to Social Control;

GLAMS Metacognition would be positively related to Reappraisal; GLAMS

Avoidance would be positively related to Distraction.

Construct validity

GLAMS Self-Punishment, Avoidance and Spirituality would be positively related to state guilt and distress; GLAMS Reparation, External Amelioration, Metacognition would be negatively associated with state guilt and distress.

Method

Participants

A total of 67 participants were recruited between August 2012 and April 2013 from

5 primary care psychology services in the North West of England using opportunistic sampling. Posters of the study were placed in waiting rooms at recruitment sites (Appendix D). Only those attending for an initial assessment and not treatment were approached. Inclusion criteria included those a) aged 18 years or over; b) able to speak and read English; and c) able to offer informed consent.

Exclusion criteria included a) organic cognitive impairment; b) substance dependence (i.e. in receipt of support from a drug and/or alcohol team); and c) high risk (e.g. current suicidal ideation). Eligibility criteria were given to screening clinicians who informed the researcher of those unsuitable (Appendix E).

Design

A cross-sectional study design was employed to enable within-subject comparisons between the GLAMS and other established measures. To assess reliability over time a longitudinal design was used to compare participant responses over a 2-4 week interval. Between-subject comparisons were made to compare relevant group differences in guilt and coping.

Measures

Guilt Level and Management Scale (GLAMS; McIvor & Bettney, 2011)

The GLAMS is a self-report measure of coping with guilty thoughts (Appendix F).

The Guilt Management scale used in the present study is comprised of 31 items asking respondents to rate how frequently they endorsed various coping strategies in

80 response to guilty thoughts using a four-point Likert scale (1 = ‘never’, 4 = ‘almost always’). Six subscales have been identified; Self-Punishment, Reparation,

Spirituality, External Amelioration, Metacognition and Avoidance, scored by adding up item scores comprising each subscale. This proposed factor structure was derived from a study that evaluated the GLAMS psychometrics in a student sample (Degnan,

2013). Although the GLAMS initially incorporated 5 items measuring guilt level these were not reported in the main findings for the current study.

A critical first step in the development of the GLAMS was to conceptualise the target construct and its theoretical context and develop an initial item pool. Guilt was appraised as an intrusive thought and conceptualised as ‘a real or imagined violation of personal morals in which people believe that their action (or inaction) contributed to negative outcomes,’ based on a review of existing guilt measures by

Tilghman-Osborne et al. (2010, p. 546). The conceptual framework and item pool was developed by McIvor & Bettney based on a review of the guilt and coping literature and clinical experience. Coping strategies identified were grouped into conceptually distinct categories. To assess content validity, 6 lay persons and 4 experts in the field were contacted by email and asked to independently rate the relevance of each GLAMS item to the conceptual framework provided (Appendix

G), using a four-point Likert scale (1 = ‘not relevant’, 4 = ‘highly relevant’). Using the Content Validity Index (CVI; Lynn, 1986), all items reported a CVI of 0.80 or above and were retained. Feedback led to the inclusion of a further 6 coping items reflecting confession, relaxation, self-punishment and rationalisation. Face validity was examined by 6 lay persons and 4 members of a Service User and Carer Forum who attended a focus group. Qualitative feedback was obtained from semi-structured

81 interviews to ensure the GLAMS was user-friendly and the layout and language was appropriate (Appendix H).

Guilt Inventory (GI; Jones, Schratter & Kugler, 2000)

The GI is a 45-item self-report measure consisting of three subscales. Only the State

Guilt subscale (10 items) was used in this study to ensure current level of guilt was captured rather than guilt as a dispositional tendency (Appendix I). It is one of few brief measures of state guilt defined by Kugler & Jones (1992) as present guilt, based on recent transgressions. Responses were rendered on a five-point Likert scale (1 =

‘Strongly Agree’, 5 = ‘Strongly Disagree’). Higher scores reflected higher levels of guilt. Minimum possible score was 10 and the maximum was 50. This scale has demonstrated good discriminant and convergent validity, high internal consistency and acceptable test re-test reliability (Jones et al., 2000; Kugler & Jones, 1992). The alpha coefficient for this scale in the current study was 0.84.

Clinical Outcomes in Routine Evaluation-10 (CORE-10; Connell & Barkham, 2007)

The CORE-10 is a brief psychological screening tool derived from the 34-item

CORE-Outcome Measure (CORE-OM; Evans et al., 2000). The 10-item version was selected (Appendix J), as it is routinely used in primary care. It has excellent internal reliability and has correlated highly with CORE-OM scores in clinical samples

(Barkham et al., 2013; Connell & Barkham, 2007). The CORE-10 assessed global distress over the last week using a four-point Likert scale (0 = ‘not at all’, 4 = ‘most or all of the time’). The maximum possible score was 40. A score of 11 or above was used as a reliable indicator of clinically significant distress (Barkham et al., 2013). In the present study the alpha coefficient was 0.86.

Brief Coping Orientation to Problems Experienced (BriefCOPE; Carver, 1997)

The BriefCOPE is a 28-item self-report measure of coping in response to distress

(Appendix K), an abbreviated version of the COPE-Inventory (Carver et al., 1989).

The BriefCOPE encompasses 14 subscales, each with 2 items reflecting distinct responses that fall under broader domains of coping; emotion-focused coping (EFC; emotional support, positive reframing, religion, humour and acceptance), problem focused coping (PFC; active coping, planning and instrumental support) and dysfunctional coping (DC; denial, venting, self-blame, self-distraction, behavioural disengagement and substance use). The BriefCOPE has reported high internal consistencies, test re-test reliability of up to a year and concurrent validity (Cooper,

Katona & Livingston, 2008; Yusoff, Low & Yip, 2010). Responses were given on a four-point Likert scale (1 = ‘I usually don’t do this at all’, 4 = ‘I usually do this a lot’). The possible range of scores varied between subscales: EFC; 10-40, PFC; 6-24 and DF; 12-48. Alpha coefficients in this study ranged between 0.65 and 0.85.

Thought Control Questionnaire (TCQ; Wells & Davies, 1994)

The TCQ is a 30-item self-report measure of coping strategies used to control unwanted thoughts (Appendix L). It is comprised of 5 subscales, each with 6 items that assessed Distraction, Social Control, Worry, Punishment and Re-appraisal. The

TCQ has a stable factor structure, acceptable to good standards of internal consistency and reliability over time (Reynolds & Wells, 1999; Wells & Davies,

1994). Responses were rated on a Likert scale (1 = ‘never’, 4 = ‘almost always’).

Subscale scores ranged from 6 to 24. Higher subscale scores indicated greater use of that method of thought control. Alpha coefficients in the present study ranged between 0.71 and 0.82.

Demographic Form

Age, gender, ethnicity, age at which left education, occupational and marital status was routinely collected (Appendix M). Reason for referral to psychological services was accessed from clinical records where prior consent was obtained to do so.

Procedure

Participants were approached by the researcher or clinician before or after their appointment. Those eligible were given a Participant Information Sheet (PIS;

Appendix N) and involvement, anonymity and confidentiality were discussed.

Informed consent was obtained (Appendix O). Participants were asked to complete the demographic form, CORE-10, GI, GLAMS, TCQ and BriefCOPE with a stamped addressed envelope provided to return by post. Participants that had not returned questionnaires after 2 weeks were contacted by telephone, as agreed. Those that returned questionnaires and consented to follow-up were sent the CORE-10 and

GLAMS after 2 weeks. A covering letter reminding participants of the second study phase was enclosed with the PIS (Appendix P). It was emphasised that participation was voluntary and participants could withdraw at any time at no detriment to themselves. Identified risk was shared with the screening clinician and GP, where appropriate. The study received ethical approval from the North West NHS Research

Ethics Committee (study reference: 12/NW/0395) and each locality’s Research and

Development (R&D) department.

Data Analysis

Descriptive statistics were reported; the Shapiro-Wilks test highlighted a number of variables that significantly deviated from normality resulting in transformations

84 being carried out. Successfully transformed variables were included in the analysis.

Spirituality remained significantly positively skewed therefore non-parametric tests were used for this variable.

Cronbach’s alpha values evaluated internal consistency. A coefficient of 0.6 was considered the lower limit of acceptability (Sim & Wright, 2000). Intra-class correlation coefficients (ICCs) using a mixed effects model assessed test-retest reliability. ICCs of 0.6 and above have been considered acceptable (Sim & Wright,

2000). Pearson’s and Spearman’s correlation coefficient were used to evaluate

GLAMS concurrent and construct validity and inter-scale relationships between coping subscales. Independent samples t-test and its non-parametric equivalent tested for significant group differences in guilt management for demographic variables; correlations were used for age.

Missing data was imputed by pro-rating the scores. Where more than 20% of items from each scale were missing data was excluded from the analysis. Where two or more answers were given for the same item, the first answer was taken. All available data was analysed using SPSS 20.0 for Windows. Due to the large number of statistical analyses, a more conservative significance criterion was elected and therefore findings are not discussed in relation to the main hypotheses unless they are significant at the p ≤ 0.01 level.

Results

Preliminary analysis

Sample characteristics

The frequencies of categorical demographic variables have been summarised (Table 1).

The mean age of the complete sample (n = 67) was 39.93 years (SD = 11.56), ranging from 20-73 years. The mean age at which participants left education was 17.78 (SD =

4.52, n = 64).

Table 1. Frequency data for sample demographics Demographic Frequency Demographic Frequency (n) (%) (n) (%) Gender (n = 67) Marital Status (n = 64) Female 41 (61.2) Single 40 (62.5) Male 26 (38.8) Married/Separated 17 (26.6) Ethnicity (n = 67) Divorced 6 (9.4) White British 62 (92.5) Widowed 1 (1.6) Other White 2 (3) Reason for Referral (n = 64) Other Asian 2 (3) Depression/Anxiety 19 (29.7) White and Black African 1 (1.5) Anxiety 15 (23.4) Occupation (n = 65) Depression 11 (17.2) Unemployed 46 (70.8) Other* 15 (23.4) Full time 8 (12.3) No consent given 4 (6.3) Part time 7 (10.8) Retired/Unable to work 4 (6.1) Note: * Included anger, physical ill-health, psychosis and low self-esteem; where anxiety or depression was comorbid but not the primary reason for referral.

Descriptive statistics

Scale mean, standard deviation (SD) and range have been summarised (Table 2). Of the GLAMS subscales, the greatest mean score was reported for Self-Punishment

(17.83 ± 5.84) and the lowest for Spirituality (5.2 ± 2.72). Of the TCQ subscales, the highest mean score was observed for Worry (13.03) and the lowest for Social

Control (11.08 ± 3.74). With regards to the BriefCOPE, a greater mean score was reported for DC (27.19 ± 5.58) and the lowest score reported for PFC (13.34 ± 4.31).

Table 2. Descriptive statistics for scales and subscales

Measure n Mean SD Range (minimum- maximum) CORE-10* 66 25.62 - 39 (0-39) GI State Guilt* 67 37.58 - 40 (10-50) GLAMS Self Punishment 66 17.83 5.84 21 (7-28) GLAMS Reparation 66 12.29 3.2 14 (6-20) GLAMS Spirituality 65 5.2 2.72 9 (3-12) GLAMS External Amelioration 65 7.57 2.47 9 (4-13) GLAMS Metacognition* 64 13.11 - 16 (7-23) GLAMS Avoidance 66 7.86 2.05 9 (3-12) TCQ Distraction 64 12.31 3.07 15 (6-21) TCQ Social Control 64 11.08 3.74 13 (6-19) TCQ Worry* 64 13.03 - 18 (6-24) TCQ Punishment 65 12.26 4.09 15 (6-21) TCQ Reappraisal* 63 12.80 - 15 (7-22) BriefCOPE PFC 63 13.34 4.31 16 (6-22) BriefCOPE EFC 63 20.24 5.15 22 (10-32) BriefCOPE DC 63 27.19 5.58 27 (14-41) Note: * Geometric means (and no SD) were reported for transformed scales.

Demographic and clinical variables

Age was not found to be significantly associated with any of the GLAMS coping subscales (at p > 0.05). There were no significant differences in GLAMS subscales or state guilt scores between males and females or between those with anxiety and depression (at p > 0.05).

Reliability of the GLAMS

Internal consistency

Cronbach’s alpha for the GLAMS 31 item scale was 0.85 and individual subscales ranged from 0.70 to 0.89 reflecting good to excellent internal reliability with the exception of the External Amelioration subscale (a = 0.56). Dropping item 3 (‘I

87 think of how other people have done worse things’) improved the reliability of this subscale with a revised alpha value of 0.63.

Test re-test reliability

The stability of the GLAMS subscales over a 2-4 week interval was evaluated using

ICCs. Avoidance (ICC = .65, n = 18, 95% confidence interval (CI) = 0.29 to 0.85, p

= 0.001), Spirituality (ICC = .76, n = 17, 95% CI = 0.45 to 0.91, p = 0.001) and Self-

Punishment (ICC = .82, n = 18, 95% CI = 0.59 to 0.93, p = 0.001) demonstrated good levels of reliability as all ICCs were above 0.6. Those associated with more adaptive coping; Reparation (ICC = .39, n = 18, 95% CI = -0.06 to 0.72, p = 0.05),

Metacognition (ICC = .02, n = 18, 95% CI = -0.02 to 0.10, p = 0.21) and External

Amelioration (ICC = .22, n = 17, 95% CI = -0.13 to 0.58, p = 0.09) revealed poorer test re-test reliability, particularly Metacognition, suggesting consistency in scores was not maintained over time.

Concurrent validity of the GLAMS

Correlations between GLAMS subscales and BriefCOPE subscales

Hypotheses: 1) GLAMS External Amelioration, Metacognition and Spirituality would be positively related to BriefCOPE EFC; 2) GLAMS Reparation and External

Amelioration would be positively related to BriefCOPE PFC; 3) GLAMS Self-

Punishment and Avoidance would be positively related to BriefCOPE DC.

As predicted, there was a positive significant relationship between External

Amelioration, Metacognition and the BriefCOPE EFC. Unexpectedly, Reparation was also found to be related to BriefCOPE EFC. Reparation and External

Amelioration were positively related to BriefCOPE PFC, as predicted, as was

Metacognition, which was not predicted. The BriefCOPE DC demonstrated a strong positive correlation with Self-Punishment and a smaller yet significant positive association with Avoidance consistent with predictions. However, a positive correlation was also observed between BriefCOPE DC and External Amelioration

(Table 3).

Correlations between GLAMS subscales and TCQ subscales

Hypotheses: 1) GLAMS Self-Punishment would be positively related to TCQ Worry and Punishment; 2) GLAMS External Amelioration would be positively related to

TCQ Social Control; 3) GLAMS Metacognition would be positively related to TCQ

Reappraisal; 4) GLAMS Avoidance would be positively related to TCQ Distraction.

Self-Punishment correlated strongly with TCQ subscales Punishment and Worry supporting our hypothesis. Interestingly, a significant negative relationship was observed between Self-Punishment and TCQ Social Control. External Amelioration correlated with TCQ Social Control, as predicted, as well as with Worry, which was not expected. Metacognition correlated significantly with TCQ Reappraisal, as predicted. Avoidance correlated with TCQ Worry and Punishment, however, no relationship was found with TCQ Distraction. Rather, Reparation and Metacognition were significantly associated with TCQ Distraction. Other findings included significant correlations between Reparation and Metacognition with TCQ Social and between Reparation and TCQ Reappraisal. Interestingly Spirituality was significantly related to TCQ Punishment and Worry (Table 3).

Construct validity of the GLAMS

Correlations between GLAMS subscales and GI State Guilt scale.

Hypotheses: 1) GLAMS Self-Punishment, Avoidance and Spirituality would be positively related to GI State Guilt; 2) GLAMS Reparation, External Amelioration and Metacognition would be negatively associated with GI State Guilt.

A significant correlation was found between GLAMS Self-Punishment and GI State

Guilt, as predicted, however, no relationships were found between Avoidance or

Spirituality and State Guilt. A significant negative association was evidenced between Metacognition and State Guilt, as predicted, however, no relationships were found between State Guilt and Reparation or External Amelioration (at p > 0.01,

Table 3).

Correlations between GLAMS subscales and CORE-10 (distress)

Hypotheses: 1) GLAMS Self-Punishment, Avoidance and Spirituality would be positively associated with CORE-10; 3) GLAMS Reparation, External Amelioration,

Metacognition would be negatively associated with CORE-10.

Self-Punishment was positively related to CORE-10, as predicted. A trend emerged between Avoidance and CORE-1O, however, only reached significance at p ≤ 0.05.

Unexpectedly, no significant relationships were found between Reparation, External

Amelioration, Metacognition, Spirituality and CORE-10 (at p > 0.01, Table 3).

Table 3. Correlations between GLAMS Guilt Management subscales and BriefCOPE, TCQ, GI and CORE-10

Subscales Brief- Brief- Brief- TCQ TCQ TCQ TCQ TCQ GI State CORE-10 COPE COPE COPE Distraction Social Worry Punishment Reappraisal Guilt EFC PFC DC Control 1. Self- -.07 -.12 .64** -.15 -.35* .40** .69** .00 .62** .56** Punishment p = 0.59 p = 0.37 p = 0.001 p = 0.24 p = 0.01 p = 0.001 p = 0.001 p = 1.00 p = 0.001 p = 0.001

2. Reparation .50** .61** .17 .56** .43** .05 .09 .53** -.18 -.22 p = 0.001 p = 0.001 p = 0.18 p = 0.001 p = 0.001 p = 0.70 p = 0.46 p = 0.001 p = 0.14 p = 0.08

3. Spirituality .26 .00 .27 .16 -.07 .36* .40** .23 .24 .21 p = 0.04 p = 0.98 p = 0.03 p = 0.21 p = 0.56 p = 0.001 p = 0.001 p = 0.08 p = 0.06 p = 0.10

4. External .45** .43** .43** .24 .46** .38* .28 .21 -.06 .00 Amelioration p = 0.001 p = 0.001 p = 0.001 p = 0.06 p = 0.001 p = 0.001 p = 0.02 p = 0.10 p = 0.66 p = 0.98

5. Meta- .35* .46** -.07 .48** .32* .13 -.03 .55** -.31* -.14 cognition p = 0.01 p = 0.001 p = 0.60 p = 0.001 p = 0.01 p = 0.29 p = 0.83 p = 0.001 p = 0.01 p = 0.29

6. Avoidance -.07 -.14 .34* .03 -.10 .41** .31* .18 .14 .24 p = 0.59 p = 0.29 p = 0.01 p = 0.83 p = 0.44 p = 0.001 p = 0.01 p = 0.17 p = 0.26 p = 0.05

Note: * p ≤ 0.01, ** p ≤ 0.001 (n = 63 – 66)

Post-hoc analysis

Inter-scale correlations between GLAMS Guilt Management subscales have been displayed (Table 4). Overall, subscales were not highly inter-related and those that were appeared to reflect either adaptive or maladaptive coping. Self-Punishment significantly correlated with Avoidance and Spirituality. External Amelioration significantly correlated with Metacognition and Reparation. Avoidance and External

Amelioration were also positively related, a possibly unexpected finding.

Table 4. Inter-scale correlations between GLAMS subscales

Reparation Spirituality External Meta- Avoidance Amelioration cognition Self- .05 .28* .24 -.22 .28* Punishment p = 706 p = 0.02 p = 0.06 p = 0.08 p = 0.02 Reparation .13 .47*** .53*** .07 p = 0.29 p = 0.001 p = 0.001 p = 0.59 Spirituality .08 -.04 -.13 p = 0.51 p = 0.74 p = 0.32 External .30* .31** Amelioration p = 0.02 p = 0.01 Meta- .17 cognition p = 0.18 Note: * p < 0.05, ** p ≤ 0.01, *** p ≤ 0.001 (n = 64 – 66)

Further exploratory analysis examined the relationship between state guilt, as measured by the GI and guilt level, as measured by the GLAMS Guilt Level scale

(items 3, 4 and 5). GLAMS guilt frequency (r = 0.62), guilt-related distress (r =

0.65) and negative self-perception caused by guilt (r = 0.56) in the last week all demonstrated moderate positive correlations with GI State Guilt (at p = 0.001).

At follow-up (n = 18), the mean score for CORE-10 distress (22.56 ± 5.82) was slightly less than the baseline levels of distress initially reported (23.63 ± 7.59) by

92 this sample indicating a marginal reduction over time, on average. ICCs indicated that distress was not maintained over time (ICC = .01, n = 18, 95% CI = -0.07 to

0.17, p = 0.46), which may be related to the lack of consistency observed for some

GLAMS subscales during this time period. To explore this further, a series of three separate linear hierarchical regression models investigated whether initial use of more adaptive coping strategies (External Amelioration, Metacognition and

Reparation) predicted a change in CORE-10 distress at follow-up, whilst controlling for baseline distress.

The Durbin-Watson statistic for Metacognition (2.78), Reparation (2.66) and

External Amelioration (1.85) are close to the recommended value of 2 suggesting no meaningful autocorrelation in the residuals at the 1% level of significance. Therefore the assumption of independent errors is likely to have been met for the regression analyses. The variance inflation factor (VIF) for Metacognition (1.02), Reparation

(1.05) and External Amelioration (1.00) are very close to the value of 1 indicating no linear relationship between predictor variables. Therefore multicollinearity is unlikely to be biasing the regression model, which also meets the necessary assumptions for predictor variables in these regression models.

As can be seen in Table 5, Reparation was not found to be a significant predictor of distress and did not explain a significant proportion of the variance in distress at follow-up (adjusted R2 = .60, F(2, 15) = 13.53, p < 0.001). However,

External Amelioration significantly predicted distress scores and explained a significant proportion of the variance in distress at follow-up (adjusted R2 = .87, F(2,

15) = 56.92, p < 0.001), as did Metacognition (adjusted R2 = .67, F(2, 15) = 18.57, p

< 0.001). These findings suggest that greater use of External Amelioration and

Metacognitive coping strategies at baseline predicted greater distress at follow-up.

Table 5. Model summary of regression coefficients of GLAMS subscales on distress at follow-up

Modal Summary Standardised t (df) p 95% Confidence β Interval for β Lower Upper Regression Model 1 - - - - CORE-10ª .78 - - - External Amelioration .52 5.90 (15) p = 0.001* .78 1.67 Regression Model 2 - - - - CORE-10 .83 - - - Metacognition .32 2.27 (15) p = 0.04* .25 7.88 Regression Model 3 - - - - CORE-10 .82 - - - Reparation .18 1.12 (15) p = 0.28 -.29 .93 Note: Predictor variables: External Amelioration, Metacognition and Reparation.

Dependent variable: CORE-10 distress at follow up. ª Baseline CORE-10 distress controlled. * p ≤ 0.05.

Discussion

The study examined the psychometric properties of the GLAMS with a view to establishing preliminary reliability and validity of this measure in a clinical sample.

The study confirmed that the GLAMS scale and subscales were internally reliable, as predicted, having reported good to excellent internal consistency ( = 0.70-0.89), with the exception of External Amelioration ( = 0.56). There were mixed findings regarding subscale reliability over 2-4 weeks (r = 0.02-0.82). Those considered maladaptive were consistent over time unlike those considered adaptive, partially supporting predictions. The study confirmed good standards of concurrent validity, although Avoidance was not related to Distraction. Construct validity was demonstrated with Self-Punishment being strongly related to state guilt and distress and Metacognition was negatively related to guilt. However, contrary to predictions,

Avoidance, Reparation, External Amelioration and Spirituality were not related to guilt or distress levels. Further analysis revealed External Amelioration and

Metacognition were significant predictors of distress at follow-up.

In the present study, high mean levels of state guilt (37.58) and distress

(25.62) were much greater than those previously reported in non-clinical populations

(GI; 27.10 ± 6.98, CORE-10; 17.1 ± 8.6), although, the range was the same for the

GI (Barkham et al., 2013; Jones et al., 2000). Approximately 97% of respondents were above the clinical cut off for psychological distress, which was also higher than

84.5% reported in other primary care settings (Barkham et al., 2013). These findings are consistent with the more frequently reported use of maladaptive coping in this sample, as seen in the higher mean scores on Self-Punishment, Worry and DC subscales.

Internally reliable subscales were reported, however, External Amelioration fell below the lower limit of acceptability. Dropping item 3 (‘I think of how other people have done worst things’) improved reliability ( = 0.63) suggesting this item may not have been as relevant to the construct of interest. Item 3 reflected a cognitive process rather than seeking support. However, values below 0.7 for psychological constructs can realistically be expected given the diversity in the construct measured (Kline, 1999). Item 3 may not be redundant but may load differently in a factor analysis and should be retained until this can be established.

Overall, reliabilities reported were not so high they could be considered redundant, ideally they were related but each item can be argued to be contributing unique information as well.

Avoidance, Spirituality and Self-Punishment demonstrated acceptable to good standards of reliability over time unlike Metacognition, External Amelioration and Reparation, inconsistent with previous research that cited values above 0.6 for conceptually related subscales (Carver et al., 1989; Cooper et al., 2008). A pattern can be seen whereby ways of coping considered adaptive were less stable over time than those considered maladaptive. One might expect this sample, high in guilt and distress, to have consistently used maladaptive coping that consequently maintained distress. Respondents had not commenced therapy where maladaptive coping would be challenged and therefore was not expected to change over time. Avoidance and

Self-Punishment may be more relevant to guilt as reparative action for past transgressions is not always possible thus these EFC strategies may be preferable as the stressor is unchangeable (Terry & Hynes, 1998; Vitaliano et al., 1990).

Spirituality remained stable, as expected, as religious affiliation and faith are based on stable belief systems that were unlikely to have changed over a short time. One

96 might have expected greater consistency in adaptive coping responses; however, coping is a shifting process sensitive to the changing demands and resources of the environment (Lazarus and Folkman, 1984). With regards to External Amelioration, access to social support as a coping resource may have changed over time, for example, as it relies on availability of others.

Further exploration of the data revealed that distress levels were also not consistent over time and greater use of External Amelioration and Metacognition actually predicted greater distress at follow-up. This is an interesting and unexpected finding as seeking social support and mindful coping were hypothesised as being adaptive and associated with reduced distress. These findings could be a product of using a small sample size (n = 18), which may not have been representative of the study’s overall sample or the population from which they were derived. For this sample, their usual support seeking behaviour may have exacerbated distress if approaching others resulted in negative outcomes or others were increasingly unavailable during the follow-up interval. Alternatively, seeking reassurance from others may have reduced distress in the short term but increased distress over time as seen with many clinically anxious presentations. More mindful thinking may have exacerbated distress for those who intermittently tried to resist engaging with their intrusive thoughts and did so unsuccessfully. Having not entered into therapy it is feasible that the sample had little understanding of how to consistently apply mindfulness in a way that was therapeutic. Therefore, engaging in what has generally been considered as adaptive coping may simply have led to negative outcomes or increased distress for this subset of individuals. Replicating these findings in a much larger sample would be necessary to avoid attributing findings to random sampling error.

Although the GLAMS Guilt Level scale was primarily included in the overall measure for clinical purposes, exploratory analysis of the data collected revealed significant positive correlations between items measuring guilt frequency, associated distress and negative self-perception in the last week and the GI State Guilt scale. As these GLAMS items purport to measure state guilt, this can be taken as further support for the concurrent validity of the GLAMS scale. More interestingly, the

GLAMS coping subscales was strongly related to theoretically similar subscales of the TCQ and BriefCOPE thereby indicating good concurrent validity as a measure of coping. However, GLAMS Avoidance was found to be related to TCQ Worry and

Punishment and not Distraction. Worry is argued to perpetuate avoidance (Wells &

Matthews, 1994), which may explain their association. Avoidance and Distraction subscales may have reflected different constructs as avoidance serves to maintain psychological problems and distraction has shown to reduce distress (Aldao et al.,

2010; Reynolds & Wells, 1999; Wells & Carter, 2009).

More generally, unpredicted associations may have emerged as ways of coping are not mutually exclusive or one-dimensional. Coping strategies reflect both emotional and problem-focused coping enabling overlap between constructs

(Skinner et al., 2003). Metacognition may be related to EFC and PFC, as a mindful- state leads to increased emotional insight and awareness but may also provide a foundation for PFC by allowing one to choose more constructively further functions of coping (Tharaldsen & Bru, 2011). One coping strategy may also increase the use of another hence the predicted relationships found between Metacognition and

Reappraisal, as mindful decentring allows for new appraisals to be made (Garland,

2007).

Associations also likely reflect different conceptualisations given to coping constructs that are highly variable across the literature (Skinner et al., 2003). The association between External Amelioration, TCQ Worry and DFC contrasts with the view that seeking social support is adaptive (Thoits, 1995). However, conceptual links may be explained where seeking support refers to venting or reassurance seeking, which has been associated with negative outcomes (Joiner, Metalsky, Katz

& Beach, 1999). Negative relationships (e.g. between GLAMS Self-Punishment and

TCQ Social Control) may be explained by one having a mitigating effect on the other (e.g. social support may provide corrective feedback regarding a past transgression reducing the need to torture oneself).

As findings and the literature suggest an overlap between ways of coping, inter-scale correlations between the GLAMS coping subscales were explored as a post-hoc analysis. The findings suggested that the subscales were not highly interrelated and those that were, were related in conceptually meaningful ways such that adaptive subscales tended to cluster together as did those considered to be maladaptive. This can be expected given that coping is not a one-dimensional construct. Overlap between coping constructs is to be assumed, particularly as ways of coping have been shown to cluster together in various yet meaningful ways in the literature (Skinner et al., 2003). It can be argued that the GLAMS 6 subscales are relatively distinct and measure different yet associated constructs. Further analysis of the GLAMS underlying factor structure in a clinical sample is required to determine whether the identified coping subscales are in fact distinct and make unique contributions to the overall coping with guilt scale.

In support of the GLAMS construct validity, greater use of Self-Punishment was associated with greater guilt and distress, consistent with the literature (Amir,

Cashman & Foa, 1997; Bybee & Quiles, 1998). Avoidance did not reach significance in contrast to the literature (Aldao et al., 2010; Degnan, 2013); however, this may have reflected the study’s lack of power as it had reached significance at p

≤ 0.05 in its association with distress. Alternatively, avoidance may be more closely related to anxiety management than guilt. Higher use of Metacognition was related to lower levels of guilt consistent with the psychological benefits of mindful coping and acceptance-based practice (Baer, 2003; Kabat-Zinn, 1990; Marcks & Woods,

2005), further supporting the construct validity of the GLAMS. External

Amelioration, Reparation and Spirituality were not found to correlate significantly with guilt or distress suggesting these were neither adaptive nor maladaptive, although purported to serve a function in their relationship with distress (Bybee &

Quiles, 1998; Pargament, 1997; Thoits, 1995). This may suggest constructs are beneficial for some and not for others in alleviating distress.

Study limitations and future considerations

There is preliminary evidence to suggest the GLAMS is a valid tool for use in research and clinical practice, however, limitations have been acknowledged. Cross- sectional data does not allow for inferences to be made regarding causation or the influence of meditating or moderating variables. Potential limitations include the lack of control over variables (e.g. perceived control and self-efficacy) known to influence the coping-distress relationship (Chwalisz, Altmaier & Russell, 1992;

Thoits, 1995; Vitaliano et al., 1990). It is not known if these are relevant to guilt thus further exploration is warranted, as these should not be dissociated from the assessment of coping (Lazarus, 2006). Changes in adaptive coping may be explained by a change in guilt, which should be assessed in future studies at follow-up.

Researchers may also consider measuring positive outcomes of coping with guilt

100

(e.g. improved self-esteem and relationships). Failure to find adequate test re-test reliability across the whole measure may have reflected the small sample at follow- up or the variable test re-test interval, which may have affected the integrity of findings as questionnaires were not completed under the same conditions. The possibility of inflated Type 1 error should be acknowledged, as a result of computing multiple comparisons, although, significant correlations in support of predictions were reasonably strong and only reported at the p ≤ 0.01 significance level.

The majority presented with depression as a primary complaint or as a comorbidity, which may explain the high guilt observed; a feature of major depression (American Psychiatric Association, 1994). This is considered representative of the population studied as depression is the most commonly cited problem presenting to primary care services across the adult age range and is most frequently comorbid with anxiety (Clark et al., 2009; Watts et al., 2002). The sample size was too small to make meaningful group comparisons between specific problems, a future consideration as guilt features across a range of disorders where coping responses may vary.

The sample was largely unemployed and white Caucasian therefore findings may not generalize beyond this, as evidence suggests guilt varies cross-culturally

(Fessler, 2004). As Spirituality was associated with greater Punishment and Worry, future research should collect data on religious affiliation, as a common belief cross- culturally is that suffering brings relief from sin (Glucklich, 2001) and some groups, as seen in Catholicism, have demonstrated greater guilt compared to non-religious groups (Walinga, Corveleyn & van Saane, 2005).

101

Further to this, research should focus on exploratory factor analysis of a larger clinical data set to test the stability of the proposed factor structure and remove redundant items. This would provide greater fidelity to the conceptualisation of coping with guilt upon which this measure is based. It would also enable the exploration of substance use, for which data was collected for in this study but not analysed having failed to load onto factors previously. Substance use is highly prevalent in clinical populations and constituting a form of escapism, may be relevant to coping with guilt (Jane-Llopis & Matytsina, 2006).

Clinical implications

The GLAMS has future clinical utility in assessment and guiding intervention towards more effective guilt management. Guilt and distress may be reduced by generating adaptive responses, such as mindful coping and minimising those maladaptive (e.g. self-punishment). Current literature has focused on targeting maladaptive guilty thoughts that are amenable to change and targeted in cognitive therapy (Kubany & Manke, 1995). Identifying coping strategies that maintain distress has equally important therapeutic value in that they too are modifiable and remain an integral component and target of CBT (Beck, 1991). Emerging evidence has suggested coping mediates the relationship between guilt and distress and therefore is important to identify early on (Held et al., 2011).

The GLAMS has identified potentially unhelpful coping responses that may be amenable to CBT, however, where self-punishment is high; individuals may do less well with standard therapies. Compassionate mind training might be preferable for those whose focus is self-criticism and self-condemnation (Gilbert & Irons,

2005), which are reflected in the GLAMS Self-Punishment subscale. Findings

102 suggest the application of mindful coping, where acceptance and metacognitive awareness are central, would be beneficial in the treatment of guilty thoughts.

Acceptance underlies the efficacy of acceptance and commitment therapy (Hayes,

Strosahl & Wilson, 1999) and mindfulness-based practice (Baer, 2003; Segal et al.,

2002), which might demonstrate good therapeutic outcomes for guilt-related distress.

The GLAMS could provide a suitable measure of treatment outcome for more maladaptive coping responses that have shown stability over time. However, test re-rest reliability over a longer period of time is required initially. The instability of adaptive coping responses suggests the GLAMS may be more clinically useful at the assessment stage. The use of the GLAMS in clinical practice may help to draw attention to the importance of guilt as a clinical problem across disorders, the resolution of which may facilitate recovery. This research also highlighted the importance of cultural factors in formulating guilt-related distress. It is understandable that guilt would be viewed as aversive and a target for treatment; however, clinicians must not lose sight of the function of guilt for the individual and consider the extent to which guilt is intertwined with religious values and practices.

Conclusion

Overall, the GLAMS scale and subscales demonstrated good to excellent internal consistency. Those considered maladaptive remained stable over time whereas those arguably adaptive were less consistent. Acceptable to good standards of concurrent validity were evidenced with scales relating to state guilt, coping and thought control. Self-Punishment was associated with greater guilt and distress and

Metacognition, in relation to mindfulness-based coping, was associated with a reduction in guilt supporting the construct validity of this tool. Future research is

103 required to test the stability of the GLAMS factor structure in a larger clinical sample. Preliminary evidence suggests that the GLAMS is a valid and reliable measure with promising clinical utility in guiding psychological intervention towards more adaptive ways of coping with guilt.

104

References

Aldao, A., Nolen-Hoeksema, S., & Schweizer, S. (2010). Emotion-regulation strategies across psychopathology: A meta-analytic review. Clinical psychology review, 30(2), 217-37.

Alexander, B., Brewin, C. R., Vearnals, S., Wolff, G., & Leff, J. (1999). An investigation of shame and guilt in a depressed sample. British Journal of Medical Psychology, 72, 323-338.

American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders (4th ed.). Washington, DC: APA.

Amir, N., Cashman, L., & Foa, E. B. (1997). Strategies of thought control in obsessive-compulsive disorder. Behaviour research and therapy, 35(8), 775- 7.

Baer, R. A. (2003). Mindfulness training as a clinical intervention. Psychological Science, 10(2), 125-143.

Barkham, M., Bewick, B., Mullin, T., Gilbody, S., Connell, J., Cahill, J.,…Evans, C. (2013). The CORE-10: A short measure of psychological distress for routine use in psychological therapies. Counselling and Psychotherapy Research, 13(1), 3-13.

Baumeister, R. F., Stillwell, A. M., & Heatherton, T. (1994). Guilt: An interpersonal approach. Psychological Bulletin, 115(2), 243-267.

Beck, A. T. (1991). Cognitive therapy: A 30 year retrospective. American Psychologist, 46(4), 368-75.

Burney, J., & Irwin, H. J. (2000). Shame and guilt in women with eating-disorder symptomatology. Journal of Clinical Psychology, 56, 51-61.

Bybee, J., & Quiles, Z. N. (1998). Guilt and mental health. In J. Bybee (Ed.), Guilt and children (pp. 269-291). San Diego, CA: Academic Press.

Carver, C. S. (1997). You want to measure coping but your protocol’s too long: Consider the Brief COPE. International Journal of Behavioral Medicine, 4, 92-100.

Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing coping strategies: A theoretically based approach. Journal of Personality and Social Psychology, 56, 267-283.

Charlton, P. F. C., & Thompson, J. A. (2011). Ways of coping with psychological distress after trauma. British of Journal of Clinical Psychology, 35(4), 517- 530.

105

Chwalisz, K., Altmaier, E. M., & Russell, D. W. (1992). Causal attributions, self- efficacy cognitions, and coping with stress. Journal of Social and Clinical Psychology, 11(4), 377-400.

Clark, D. M., Layard, R., Smithies, R., Richards, D. A., Suckling, R., & Wright, B. (2009). Improving access to psychological therapy: Initial evaluation of two UK demonstration sites. Behaviour Research and Therapy, 47(11), 910-20.

Connell, J., & Barkham, M. (2007). CORE-10 User Manual, Version 1.1. Bath, UK: CORE System Trust & CORE Information Management Systems Ltd.

Cooper, C., Katona, C., & Livingston, G. (2008). Validity and reliability of the Brief COPE in carers of people with dementia: The LASER-AD study. Journal of Nervous and Mental Disease, 196, 838-843.

Degnan, A. (2013). Development and validation of the GLAMS to measure guilt level and management in a non-clinical sample. (Unpublished dissertation). University of Manchester, Greater Manchester.

Evans, C., Mellor-Clark, J., Margison, F., Barkham, M., Audin, K., Connell, J., & McGrath, G. (2000). CORE: Clinical outcomes in routine evaluation. Journal of Mental Health, 9(3), 247-255.

Fessler, D. M. (2004). Shame in two cultures: Implications for evolutionary approaches. Journal of Cognition and Culture, 4, 207-262.

Folkman. S., & Moskowitz, J. T. (2004). Coping: Pitfalls and promise. Annual Review of Psychology, 55, 745-774.

Fontaine, J. R. J., Luyten, P., De Boeck, P., Corveleyn, J., Fernandez, M., Herrera, D.,…& Tomcsa´nyi, T. (2006). Untying the Gordian knot of guilt and shame: The structure of guilt and shame reactions based on situation and person variation in Belgium, Hungary and Peru. Journal of Cross-Cultural Psychology, 37(3), 273-292.

Garland, E. L. (2007). The meaning of mindfulness: A second-order cybernetics of stress, metacognition, and coping. Complementary Health Practice Review, 12(1), 15-30.

Garnefski, N., Legerstee, J., Kraaij, V., Van Den Kommer, T., & Teerds, J. (2002). Cognitive coping strategies and symptoms of depression and anxiety: A comparison between adolescents and adults. Journal of Adolescence, 25(6), 603-611.

Gilbert, P., & Irons, C. (2005). Focused therapies and compassionate mind training for shame and self-attacking. In P. Gilbert (Ed.), Compassion: Conceptualisations, research and use in psychotherapy (pp. 263-325). London, UK: Routledge.

106

Glucklich, A. (2001). Sacred pain: Hurting the body for the sake of the soul. New York, NY: Oxford University Press.

Hayes, S. C., Strosahl, K. D., & Wilson, K. G. (1999). Acceptance and commitment therapy: An experiential approach to behaviour change. New York: Guilford Press.

Jané-Llopis, E., & Matytsina, I. (2006). Mental health and alcohol, drugs and tobacco: A review of the comorbidity between mental disorders and the use of alcohol, tobacco and illicit drugs. Drug and Alcohol Review, 25(6), 515- 36.

Joiner, T. E., Jr., Metalsky, G. I., Katz, J., & Beach, S. R. H. (1999). Depression and excessive reassurance-seeking. Psychological Inquiry, 10, 269-278.

Jones, W. H., Schratter, A. K., & Kugler, K. (2000). The Guilt Inventory. Psychological Reports, 87, 1039-1042.

Kabat-Zinn, J. (1990). Full catastrophe living: Using the wisdom of your body and mind to face stress, pain and illness. New York, NY: Delacorte.

Kim, S., Jorgensen, R. S., & Thibodeau, R. (2011). Shame, guilt and depressive symptoms: A meta-analytic review. Psychological Bulletin, 137, 68-96.

Kline, P. (1999). The handbook of psychological testing (2nd ed.). London, UK: Routeledge.

Kubany, E. S., & Manke, F. P. (1995). Cognitive therapy for trauma-related guilt: Conceptual bases and treatment outlines. Cognitive and Behavioral Practice, 2, 23-61.

Kugler, K., & Jones, W. H. (1992). On conceptualizing and assessing guilt. Journal of Personality and Social Psychology, 62(2), 318-327.

Lazarus, R. S. (2006). Emotions and interpersonal relationships: Toward a person- centered conceptualization of emotions and coping. Journal of Personality, 74(1), 9-46.

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal and coping. New York, NY: Springer Publishing Company, Inc.

Lee, D. A., Scragg, P., & Turner, S. (2001). The role of shame and guilt in traumatic events: A clinical model of shame-based and guilt-based PTSD. British Journal of Medical Psychology, 74, 451-466.

107

Lynn, M. R. (1986). Determination and quantification of content validity. Nursing Research, 35, 382-385.

Marcks, B. A., & Woods, D. W. (2005). A comparison of thought suppression to an acceptance-based technique in the management of personal intrusive thoughts: A controlled evaluation. Behaviour Research and Therapy, 43(4), 433-45.

Marcks, B. A., & Woods, D. W. (2007). Role of thought-related beliefs and coping strategies in the escalation of intrusive thoughts: An analog to obsessive- compulsive disorder. Behaviour Research and Therapy, 45, 2640-2651.

Martin, R. C., & Dahlen, E. R. (2005). Cognitive emotion regulation in the prediction of depression, anxiety, stress, and anger. Personality and Individual Differences, 39(7), 1249-1260.

Nelissen, R. M. A., & Zeelenberg, M. (2009). When guilt evokes self-punishment: Evidence for the existence of a Dobby effect. Emotion, 9, 118-122.

Niler, E. R., & Beck, S. J. (1989). The relationship among guilt, dysphoria, anxiety and obsessions in a normal population. Behaviour Research and Therapy, 27, 213-220.

Nolen-Hoeksema, S. (2000). The role of rumination in depressive disorders and mixed anxiety/depressive symptoms. Journal of Abnormal Psychology, 109, 504-511.

Pargament, K. I. (1997). The psychology of religion and coping: Theory, research, and practice. New York, NY: Guilford.

Reynolds, M., & Wells, A. (1999). The Thought Control Questionnaire- Psychometric properties in a clinical sample, and relationships with PTSD and depression. Psychological Medicine, 29(5), 1089-99.

Sanftner, J. L., Barlow, D. H., Marschall, D. E., & Tangney, J. P. (1995). The relation of shame and guilt to eating disorder symptomology. Journal of Social and Clinical Psychology, 14, 315-324.

Segal, Z. V., Williams, J. M. G., & Teasdale, J. D. (2002). Mindfulness-based cognitive therapy for depression: A new approach to preventing relapse. New York, NY: Guilford Press.

Semmer, N. K., Elfering, A., Jacobshagen, N., Perrot, T., Beehr, T. A., & Boos, N. (2008). The emotional meaning of instrumental social support. International Journal of Stress Management, 15(3), 235-251.

Shapiro, L. J., & Stewart, E. S. (2011). Pathological guilt: A persistent yet overlooked treatment factor in obsessive-compulsive disorder. Annals of Clinical Psychiatry, 23(1), 63-70.

108

Silfver, M. (2007). Coping with guilt and shame: A narrative approach. Journal of Moral Education, 36(2), 169-183.

Sim, J., & Wright, C. (2000). Research in health care: concepts, designs and methods. Cheltenham, UK: Stanley Thornes Ltd.

Stanton A. L., Danoff-Burg, S., Cameron C. L., & Ellis, A. P. (1994). Coping through emotional approach: Problems of conceptualization and confounding. Journal of Personality and Social Psychology, 66, 350-62.

Strelan, P. (2007). The prosocial, adaptive qualities of just world beliefs: Implications for the relationship between justice and forgiveness. Personality and Individual Differences, 43(4), 881-890.

Tangney, J. P. (1991). Moral affect: The good, the bad, and the ugly. Journal of Personality and Social Psychology, 61, 598-607.

Tangney, J. P., & Dearing, R. L. (2002). Shame and guilt. New York, NY: Guilford Press.

Tangney, J. P., Stuewig, J., & Mashek, D. (2007). Moral emotions and moral behaviour. Annual Review of Psychology, 58, 345-372.

Tennen, H., Affleck, G., Armeli, S., Carney, M. A. (2000). A daily process approach to coping: Linking theory research and practice. American Psychologist, 55(6), 626-36.

Terry, D. (1994). Determinants of coping: The role of stable and situational factors. Personality Processes and Individual Differences, 66(5), 895-910.

Terry D. J., & Hynes G. J., (1998). Adjustment to a low-control situation: Reexamining the role of coping responses. Journal of Personality and Social Psychology, 74(4), 1078-92.

Tharaldsen, K. B. & Bru, E. (2011). Validation of the mindful coping scale. Emotional and Behavioural Difficulties, 16, 87-103.

Thoits, P. A. (1995). Stress, coping, and social support processes: Where are we? What next? Journal of Health and Social Behaviour, 35, 53-79.

109

Tilghman-Osborne, C., Cole, D. A., & Felton, J. W. (2010). Definition and measurement of guilt: Implications for clinical research and practice. Clinical Psychology Review, 30(5), 536-46.

Vitaliano, P. P., DeWolfe, D. J., Maiuro, R. D., Russo, J., & Katon, W. (1990). Appraised changeability of a stressor as a modifier of the relationship between coping and depression: A test of the hypothesis of fit. Journal of Personality and Social Psychology, 59(3), 582-592.

Walinga, P., Corveleyn, J., & van Saane, J. (2005). Guilt and religion: The influence of orthodox Protestant and orthodox Catholic conceptions of guilt on guilt- experience. Archive for the Psychology of Religion, 27(1), 113-136.

Warda, G., & Bryant, R. A. (1998). Thought control strategies in acute stress disorder. Behaviour Research and Therapy, 36(12), 1171-1175.

Watts, S. C., Bhutani, G. E., Stout, I. H., Ducker, G. M., Cleator, P. J., McGarry, J., & Day, M. (2002). Mental health in older adult recipients of primary care services: Is depression the key issue? Identification, treatment and the general practitioner. International Journal of Geriatric Psychiatry, 17(5), 427-37.

Wells, A., & Carter, K. (2009). Maladaptive thought control strategies in generalized anxiety disorder, major depressive disorder, and non-patient groups and relationships with trait anxiety. International Journal of Cognitive Therapy, 2, 224-234.

Wells, A., & Davies, M. (1994). The Thought Control Questionnaire: A measure of individual differences in the control of unwanted thoughts. Behaviour Research and Therapy, 32, 871-878.

Wells, A., & Matthews, G. (1994). Attention and emotion: A clinical perspective. Hove, UK: Earlbaum.

Yusoff, N., Low, W. Y., & Yip C. H. (2010). Reliability and validity of the Brief COPE Scale (English version) among women with breast cancer undergoing treatment of adjuvant chemotherapy: A Malaysian study. Medical Journal Malaysia, 65, 41-44.

Zahn-Waxler, C., & Kochanska, G. (1990). The origins of guilt. In R. Thomson (Ed.), 36th annual Nebraska symposium on motivation: Socio-emotional development (pp. 183-258). Lincoln, NE: University of Nebraska Press.

110

Paper 3: Critical Appraisal

111

Overview

This paper provides a critical evaluation and reflection on the research carried out. It begins initially with the researcher’s rationale for conducting the literature review and empirical study, followed by some of the researcher’s reflections on the review process. A more detailed critique of the research methodology and a discussion of ethical issues are given for the empirical study whilst reflecting on the research process throughout. The implications for future research and clinical practice were further explored with regards to the research study findings and subsequent recommendations were made.

Aims of the research

The primary aims of the literature review was to synthesise and evaluate the research evidence to determine whether an association exists between guilt and symptoms of post-traumatic stress, with a view to conceptualising the underlying relationship between these two constructs. Findings from the review suggested that coping may mediate the relationship between guilt and distress. In order to explore this further, research is required that assesses ways of coping with guilt. Currently there exists no broad measure of coping with guilt specifically.

The overall objective of the empirical study was to provide validation to a unique and newly developed measure of coping with guilty thoughts in a clinical sample. This enabled exploration of the adaptive and maladaptive value of ways of coping with guilt by investigating the relationships between coping constructs, guilt and global psychological distress.

112

Rationale for Literature Review and Empirical Study

The construct of guilt is not an emerging concept; its origin has been well researched across developmental studies, as has its function and consequences (Bybee, 1998;

Kochanska, Gross, Lin & Nichols, 2002). Although guilt has been portrayed as a prosocial and protective emotion (Tangney, 1990; Williams & Bybee, 1994; Zahn-

Waxler & Kochanska, 1990), guilt has also been implicated in various psychological disorders including PTSD (Burney & Irwin, 2000; Harder, 1995; Kubany & Manke,

1995; Shafran, Watkins & Charman, 1996). The role of guilt in depression and OCD has been explored in recent reviews (Kim, Thibodeau & Jorgensen, 2011; Shapiro &

Stewart, 2011) and there have been a growing number of studies looking at guilt in

PTSD (Kubany & Manke, 1995; Lee, Scragg & Turner, 2001; Wilson, Drozdek

&Turkovic, 2006). This is an important area of research given the increasing number of veterans returning from conflict zones with symptoms consistent with a guilt- based PTSD that can be costly and debilitating (Kubany et al., 1996; Marx et al.,

2010).

The role of guilt in PTSD is such that it has been recognised as a secondary feature in current diagnostic criteria (DSM-IV; American Psychiatric Association,

1994). The evidence base suggests trauma-focused interventions demonstrate superiority over other therapeutic approaches (Bisson et al., 2007; Ehlers et al.,

2010) and should be offered in the treatment of PTSD (National Institute for Health and Care Excellence, 2005). However, the focus of these therapies is on treating hallmark symptoms driven by high arousal and anxiety (Brewin, Andrews & Rose,

2000; Ozer, Best, Lipsey & Weiss, 2003). As a retrospective emotion guilt may not respond to or may even impede current therapeutic approaches (Hendin & Haas,

113

1991; Pitman et al., 1991). The present review aimed to evaluate and conceptualise the guilt-PTSD relationship to enhance our understanding of the role of guilt in

PTSD. It is also considered a relevant contribution to the on-going deliberation about the position of guilt, as an associated feature within current diagnostic criteria, which has led to guilt receiving far less recognition than the hallmark symptom clusters.

The outcomes of the present review identified several guilt cognitions relating to PTSD symptomology, which lends itself to a cognitive model of guilt- based PTSD. This model could be further substantiated by looking at the relationship between guilt and behavioural responses, as maladaptive responses formulate a large part of cognitive and behavioural therapy (CBT; Beck, 1991). Evidence suggests avoidant coping or disengagement may be a mediator in the guilt-PTSD relationship

(Held, Owens, Schumm, Chard & Hansel, 2011; Street, Gibson & Holohan, 2005).

This contrasts with existing clinical models of guilt-based PTSD and research suggesting guilt leads to ruminative activity and reparative action than avoidance, which has shown to be more closely linked with shame (Lee et al., 2001; Tangney,

Wagner & Gramzow, 1992). To further understand how guilt-related distress is maintained, ways of coping with guilt need to be explored. The current review highlighted a lack of research in this domain therefore the empirical study aimed to contribute to this area by investigating guilt management and its association with distress.

The study objective was to validate a newly developed measure of coping with guilty thoughts (GLAMS) by correlating ways of coping with guilt with theoretically similar subscales, state guilt and global distress. This not only sought to validate an innovative and unique measure but enabled exploration of the adaptive or

114 maladaptive value of cognitive and behavioural responses to guilt-related distress.

Guilt may have adaptive origins and remain functional up until a point in its frequency or severity, whereby it then becomes dysfunctional. Guilt left unresolved may become chronic and maladaptive as a result, which may be determined by how guilt is managed.

The empirical study extended previous research as the GLAMS is the first measure of its kind to examine ways of coping with guilt-specific intrusive thoughts.

Unlike previous measures of coping (see Skinner, Edge, Altman & Sherwood, 2003, for a review) the GLAMS incorporates facets of mindfulness-based coping such as acceptance and meta-awareness with regards to managing unwanted thoughts. The

GLAMS also encompasses a broad range of coping methods thought to be related to guilt and is theoretically derived. Based on a recent review of the definition and measurement of the guilt construct (Tilghman-Osborne, Cole & Felton, 2010) the

GLAMS established a conceptual framework that encompassed a theoretically- informed understanding of guilt. This was important as existing measures of guilt have demonstrated little content overlap due to the diverse conceptualisation of the construct and have been critiqued as being highly inferential having relied on respondents’ interpretation of guilt (Kim et al., 2011; O’Connor, Berry, Weiss, Bush

& Sampson 1997). It was decided that validating a measure of coping with guilt- related distress in a clinical sample, whose inclusion criteria was not restricted to those with a diagnosis of PTSD, would have greater clinical value. It was also considered more feasible given the anticipated recruitment difficulties of a small subgroup in the time available.

115

Reflections on Review and Research Process

One of the aims of the review was to explore the link between guilt and PTSD using an interpretive framework, in terms of four competing models, within which the existing evidence base could be assessed and understood. These hypotheses are plausible in their explanation of the guilt-PTSD link, however, have been difficult to test based on existing research. For the most part, the research indicates a relationship between guilt and PTSD; however, despite a fair number of studies, the nature of this relationship remains largely undetermined and constrained by study design limitations. Therefore the evidence remains limited in speaking to the validity of these four models, although, evidence supporting some models may be stronger perhaps than others. The framework has its advantages in that it can readily identify where the literature falls short in being able to explain the guilt-PTSD link and can make recommendations regarding future studies that are necessary to further explore the conceptualisations presented. In particular, reference is made to experimental research that is required to make firmer conclusions regarding causality such as inducing guilt in a non-clinical analogue population to see if guilt produces PTSD phenomena at a non-clinical level. Alternatively, experience sampling methodology could investigate the temporal relationship between guilt and PTSD. The ethical constraints and logistics of such methodology in this field may limit their application.

With regards to exploring the validity of Model 1, research that used mediational analyses can be argued to be the most appropriate as we already know a relationship between trauma and PTSD exists. Mediation models help us to understand a known relationship by exploring the underlying mechanism by which

116 an independent variable (e.g. trauma) influences the dependent variable (e.g. PTSD) via the inclusion of a third explanatory variable (e.g. guilt). Guilt was found to mediate the relationship between trauma and PTSD suggesting a causal role of guilt.

However, partial mediation suggests that guilt accounts for some but not all of the relationship between trauma and PTSD suggesting the role of other factors.

Unfortunately, few studies controlled for or investigated the effect of potential mediating confounds or moderating variables that may have better explained the relationship. That said, one study found a significant effect of guilt on PTSD above and beyond that explained by depression that was explored in a path model. To confidently report mediation, a significant reduction in variance explained by the independent variable must also have occurred as a trivial amount of variance may be explained by the third variable and yet be significant. This was the case for one of two studies that reported a marginal yet significant indirect effect of guilt on PTSD.

Although mediation analyses can be a powerful statistical test, a common problem is that cross-sectional and therefore correlational designs are often used making it difficult to establish the direction of effect. Further to that, a third variable independent from the proposed mediator may explain the link. In order to address this, future studies need to establish temporal precedence whereby the independent variable precedes the dependent variable in time using a longitudinal design, to suggest a directional link. It is acknowledged that even with longitudinal studies, a direction of effect doesn’t equate with causality. However, this would enable one to explore whether changes in guilt at baseline predict subsequent changes in PTSD symptomology. One also needs to prove that potential confounds do not influence the relationship between the independent and dependent variable to be able to build a stronger argument for the mediation effect. Ideally, future studies should measure

117 both PTSD and guilt and any other known associated variables at two time points and explore which relationships are significant in a path model. This would support the exploration of Models 1 and 2, as the relationship between guilt and PTSD and vice versa can be explored. Currently, there is some evidence to suggest the validity of Model 1; however, further exploration is limited by the lack of longitudinal research designs.

Little direct support was found for Model 3 and conclusions drawn from intervention studies were tentative at best, as the link between guilt and PTSD was not explicitly tested. Two therapies that demonstrated a similar reduction in PTSD severity post-treatment and only one that targeted and reduced guilt could indicate a lack of synchrony between guilt and PTSD. However, this evidence is weak as various other factors may have led to a decline in guilt and/or PTSD. In some instances it was not clear what element of the intervention was expected to reduce guilt or how interventions differed and there was no evidence to suggest this had been reliably tested. No causal inferences can be drawn from intervention studies as there may be alternative mechanisms of change. Due to a lack of longitudinal research, the evidence assessing the validity of this model remains inconclusive.

For Model 3 to suggest that guilt and PTSD co-exist but are unrelated, there has to a variable that accounts for both of these (e.g. trauma) to explain why they co- occur. In order to test whether they are truly unrelated, one must account for likely common causes of both guilt and PTSD such as trauma for example. If the relationship between guilt and PTSD suddenly disappears by doing so, one could argue guilt and PTSD are not related but are driven by a shared common denominator. This would require a large sample and in-depth knowledge of common factors that may influence the guilt-PTSD relationship. Further to this, mediation

118 studies that support Model 1 also provide some evidence against Model 3, which can be used in exploring the validity of this model also.

Research that controls for common causes of both guilt and PTSD would also lend itself to exploring the validity of Model 4, which suggests that the guilt-PTSD link is accounted for by other mediating and related constructs such as shame and depression. Few studies in the current review controlled for such confounds making it difficult to tease apart the effects of two clearly overlapping constructs. The five studies that controlled for the effects of shame, thereby testing Model 4, reported that guilt was no longer a significant correlate or predictor of PTSD. This suggests that shame may better explain the relationship between trauma and PTSD and needs to be controlled for in future studies.

One limitation of the review is the narrow search criteria used with regards to guilt. The review highlighted that guilt may be associated with regret and remorse and perceived responsibility in relation to self-blame appeared a key aspect of the construct. None of these terms were included in the database search, as it was initially felt that such terms reflected consequences of guilt than aspects of the guilt construct. This may have resulted in relevant studies being missed that may have influenced the conclusions drawn. The search did not include self-blame as this has been argued to conceptually overlap with shame, both of which focus on internal, global and stable attributions, which may have inflated the guilt-PTSD relationship

(Andrews & Brewin, 1990). A strong association has been demonstrated between characterological self-blame and shame-proneness (Tangney et al., 1992). Self- blame has also been associated with dysphoria, possibly due to the attribution of internal causation to negative events that characterises self-blame (Clements &

Sawhney, 2000). However, in hindsight, self-blame may have been related to either

119 perceived responsibility, as seen in guilt, or self-loathing, as seen in shame, the latter of which could have formulated exclusion criteria if this search term was used.

The review was limited to peer-reviewed articles therefore book chapters and reviews, editorial presentations, letters, proceedings papers and meeting abstracts were excluded. This is particularly important, as studies with positive findings tend to be published and therefore would have been included in this review. Psycho- analytic and qualitative studies were also not included, which may have brought about further meaning to the role of guilt in PTSD. Studies were not excluded where a reliable diagnosis of PTSD could not be obtained as this would have led to an overly narrow review and missed important findings. More than half the studies reviewed included samples of war veterans and victims of interpersonal violence, both of which are associated with chronic trauma featuring violent acts of commission or omission where guilt may strongly feature. This has implications for generalising findings to other trauma groups.

The methodological quality of studies included in the review critiqued throughout where appropriate and summarised in the methodological constraints section of the paper. Study quality was not weighted using a quality assessment tool

(QAT) for several reasons. There exists much disagreement regarding how to conceptualize and measure study quality and there is currently no consensus on the ideal assessment tool (Petticrew & Roberts, 2006). Existing tools have generally been created to grade the quality of RCT’s, cohort and observational studies and diagnostic accuracy and treatment effect studies, which would not be appropriate for the majority of studies included in the current review. The conclusions drawn from large reviews are that QAT’s have generally not been tested using standard scale development techniques and information is lacking with regards to their validity and

120 reliability (Moher et al., 1995; Deeks et al., 2003). Further to this, different tools evaluating the same study have shown to produce divergent results (Herbison, Hay-

Smith & Gillespie, 2006). Each study is unique to some extent and a generic quality checklist therefore may not be appropriate.

That’s not to say QAT’s don’t contribute added value as they enable the standardizing of decisions regarding the quality of studies. A lack of adherence to criteria that help define the validity of studies may explain why studies reporting on the same subject provide different results. In the absence of a QAT, unidentified methodologically flawed studies may lead to inaccurate interpretation of findings and conclusions drawn. A QAT can guide interpretation and the strength of inferences made by helping to identify potential for bias. Although more relevant to intervention studies and meta-analyses there is also empirical support for the use of a

QAT such that studies with lower methodological quality have reported greater treatment effects (Moher et al., 1998). That said, others have reported no reliable association between quality score and effect size (Balk et al., 2002). It is good practice to consider the quality of studies in relation to findings as conclusions drawn can guide the direction of future research and clinical practice. If the ‘raw’ material is flawed then the outcome from systematic reviews may be questioned. In the current review, weighting the quality of studies may have helped identify more clearly those studies most reliable and therefore those that best support conclusions drawn. A QAT may also have highlighted methodological inadequacies more fairly and clearly, which may have helped explain different findings across studies. Further to this, a QAT may have allowed for more concrete recommendations to be made regarding the quality of studies that are necessary in this area in future research.

121

With regards to the research process for the empirical study, the recruitment stage proved the most stressful aspect. It aroused much anxiety due to the pressure of obtaining a high participant to variable ratio and often obstacles were outside the researcher’s control (e.g. service-related changes, attendance rates). Expanding recruitment to beyond 5 sites proved difficult for both logistic and practical reasons.

Services across the North West were contacted, however, none of whom could accommodate a research project. Upon reflection, the researcher may have been naive to only consider those eligible and not those unsuitable (e.g. acutely distressed), which may have reflected a lack of experience in complex primary care.

The researcher maintained regular contact with R&D offices and in return received positive and prompt support that helped to ensure the smooth running of the study.

On occasion, the researcher received a mixed reception from teams that were undergoing service-related changes that had caused anxiety among the staff. The researcher was sensitive to this and proposed that individual staff could volunteer to support the study than attempting to gain support from the whole team. This flexible approach was well received and support was obtained from recently qualified

Clinical Psychologists that could relate to the difficulties encountered by a Trainee recruiting in the NHS. Overall, the researcher enjoyed exploring guilt and its management, which has informed their clinical practice.

The researcher can more readily identify guilt, should this present, by listening to the language used by clients consistent with perceived responsibility, wrong doing, lack of justification and hindsight bias (Kubany & Watson, 2003). The study has guided the selection of appropriate clinical tools for assessing guilt, particularly those relevant to specific trauma groups such as veterans whose history

122 is compiled of multiple war time events, which the researcher had previously found difficult and overwhelming to navigate through. The researcher incorporated into clinical practice the available evidence-base regarding the treatment of a more guilt- based PTSD where appropriate. This complemented more conventional trauma- focused therapy in order to maximise therapeutic outcomes.

The study also enabled the researcher to reflect on past experiences of guilt and ways of coping, which have not always been effective. The researcher thought about their coping style and contemplated using alternative methods that may hold greater therapeutic value. Having previously identified with guilt as a negative and destructive emotion, the researcher was able to think more broadly and reflect on the functional and adaptive aspects of guilt in its acute form. Drawing parallels with anxiety, the researcher recognised that like anything that is unresolved and chronic, guilt may become maladaptive longer term if not appropriately managed.

Methodological and Ethical Considerations

Study design

The cross-sectional design was appropriate for this questionnaire-based study to enable within subject comparisons between the GLAMS and other subscales. This design has the advantage of reducing error variance associated with individual difference. A longitudinal design was appropriate for assessing reliability over time, however, this was difficult to implement at times due to service related and methodological constraints. Services with a high turnover at screening clinics sometimes directed participants to treatment before follow-up data could be collected therefore the design of this study was not suitable within some primary care mental

123 health teams. There is also an argument to be made that the first assessment session carries therapeutic benefit (Finn & Tonsager, 1997), which may have affected distress levels at follow-up.

A follow-up interval of at least two weeks is argued to be a reasonable compromise between memory recall bias and unwanted clinical change (Marx,

Menezes, Horovitz, Jones & Warren, 2003). The precise interval time for reliability could not be ascertained but was estimated to be 2-4 weeks in the current study.

Although this remains an inevitable consequence of using postal follow-ups and a practical reality for other studies of reliability, it does enable findings to be generalized to other studies using the same methodology. In hindsight, administrative staff could have recorded the date questionnaires were returned, as envelopes were unique to the study and easily identifiable, which may have provided a better approximation of interval time.

Measures

The design of the GLAMS was given careful consideration. An even numbered four- point Likert scale was selected to avoid excessive use of the mid-point value

(Garland, 1991; Johns, 2005). Providing further response alternatives may have reduced scale validity if respondents were unable to make the more subtle distinctions between each anchored point that would be required (Clark & Watson,

1995). A limitation was that some questions could be construed as double-barrelled, although, they appear to be measuring the same construct.

Another limitation was that the GLAMS initial item pool was not developed using a structured systematic approach by the authors. Therefore the assessment of

124 content validity was particularly relevant to determine which items were retained based on their relevance to the conceptual framework as rated by reviewers of the

GLAMS. A total of 10 reviewers were included based on recommendations of 5 or more, with more than 10 being deemed unnecessary (Haynes, Richard & Kubany,

1995; Lynn, 1986). Using the Content Validity Index (CVI) those reporting a CVI of

0.80 or above were retained, as recommended by Davis (1992). The CVI has been reviewed as having many advantages over alternative methods, particularly with regards to its focus on agreement of relevance than agreement per se (Polit, Beck &

Owen, 2007). The GLAMS was considered appropriate for clinical samples having been developed in collaboration with service users. Suitable modifications were made to the structure, language, length and use-ability to establish face validity

(DeVon et al., 2007).

Initially, the GLAMS proposed a Guilt Level scale (5 items) that assessed guilt frequency, distress and situations that may have elicited guilty thoughts in the past week. Data collected was generally not reported in the main findings as they did not contribute anything further to the study or existing literature. The Guilt Level scale items were included in the overall GLAMS measure for clinical purposes, it was not intended that these items would be incorporated into the factor structure of the coping scale. As the Guilt Level items measured different aspects of the guilt construct, computing a 5 item total score was not considered appropriate as it would not reflect an internally consistent scale. In further support of the concurrent validity of the GLAMS, the GI State Guilt scale was correlated with only those GLAMS

Guilt Level items relating to guilt in the last week (items 3, 4 and 5). Pearson’s correlation coefficient was used as all variables were considered normally distributed. Item 1 was not included the analysis as it reflected a measure of trait

125 guilt by asking respondents if they had been affected by guilt throughout their life.

Item 2 was excluded as it assessed situations that elicited guilt. Although created for clinical purposes, this scale did reveal some interesting information. Guilt was most frequently triggered by ‘memories of past events’, endorsed by 79% of the sample, followed by ‘thinking bad things,’ experienced by just over 50%, which lends itself to a cognitive model of guilt-related distress.

The GI (Jones, Schratter & Kugler, 2000) provided a separate index of unresolved state guilt that was short in length for administration ease. The GI is a widely used valid and reliable measure, although, scale psychometrics were derived from non-clinical populations (Jones et al., 2000; Kugler & Jones, 1992). The GI was still considered suitable based on a recent review that highlighted a number of conceptual and methodological flaws across existing guilt measures (Tilghman-

Osborne et al., 2010). The GI did not focus on self-denigration or items relating to shame. The GI was preferable to the TOSCA (Tangney, Wagner & Gramzow, 1989) that appeared to measure prosocial behavioural responses to guilt than the guilt construct (Ferguson & Crowley, 1997; Silfver, 2007). There remains considerable heterogeneity in the measurement of guilt therefore it is possible that a different measure would produce different findings.

The CORE-10 was selected as a global measure of distress routinely used in primary care (Connell & Barkham, 2007). Although the CORE-10 has reported excellent internal reliability and has a reliable cut-off for clinical caseness, other psychometric properties of this scale have not been reported (Barkham et al., 2013).

The CORE-OM (34-item) has demonstrated high internal consistency, acceptable reliability and construct validity (Barkham, Gilbert, Connell, Marshall & Twigg,

126

2005; Evans et al., 2002) but was considered too long to administer given the number of questionnaires in the present study. However, CORE-10 scores have shown to correlate highly with CORE-OM scores in clinical samples (Barkham et al., 2013).

The BriefCOPE (Carver, 1997) was selected as it was a valid and reliable tool that provided a broad assessment of coping that may be relevant to guilt management. This abbreviated version of the original 60-item COPE-Inventory

(Carver, Scheier & Weintraub, 1989) provided a good balance between breadth and time taken to complete the measure. However, the three broad categories of coping used in the present study may not have been the best way of conceptualising the coping construct. Problem and emotion-focused higher-order categories may not be useful distinctions as any given way of coping is likely to serve many functions

(Skinner et al., 2003). A lack of consensus regarding what constitutes different types of coping makes it difficult to establish concurrent validity for any new measure of coping.

The TCQ (Wells & Davies, 1994) was selected as a valid, reliable and widely cited measure of thought control. Subscales were conceptually relevant to those featured in the GLAMS and therefore were appropriate for investigating concurrent validity. Whilst an overlap in subscale content between the measures was acknowledged, due to the inclusion of similar items that may have inflated relationships, correlations were not so strong findings could be considered redundant. The GLAMS was created as the TCQ (and BriefCOPE) are not guilt- specific; they do not reflect all strategies unique to or associated with guilt (Fontaine

127 et al., 2006) and neither have incorporated mindfulness, which is arguably different to reappraisal (Garland, 2007).

Recruitment and procedure

Recruitment took place at 1 main site and across 4 secondary sites in primary care where clinicians took an active role supporting the project. Recruitment was expanded across multiple locations in an attempt to safeguard against lower numbers, however, recruitment proved more difficult than predicted due to several factors. Of the 67 participants, 59 were recruited from a screening clinic at the main site averaging 2-3 participants a week. An unexpected change in the referral pathway

12 weeks before the end of recruitment meant that the number of clinic days significantly reduced until the clinic was no longer available. Further to this, the

DNA rate across 26 weeks ranged from 17% to 58%, of which 22 weeks had a DNA rate greater than 30%. This is much higher than that reported by other primary care psychology services (Abrahams & Udwins, 2002; Greasley & Small, 2005). The number of screening clinicians each week varied between 1 and 6; however 77% of the time there were 5 or more. The number of patients that attended ranged between

4 and 24, with an average of 17 attending weekly. Thus it was the lack of patient attendance than clinician availability that was the greatest constraint on the number recruited.

A total of 119 patients (43% of the total that attended the screening clinic) were not suitable or able to take part. Upon reflection, this is a surprisingly high number as only a quarter of these had met the exclusion criteria. The largest number

(31%) could not be approached prior to their appointment having attended late or were missed due to a higher volume of patients attending that day. A further 15%

128 refused, 12% were acutely distressed upon arrival and 7% had a history of violence posing a potential risk. The remaining 12% could not take part for other reasons.

Given the complex needs of this sample and that 69% were currently unemployed, which may have affected attendance due to the cost of travel; it may not be surprising that recruitment numbers were low. In an attempt to boost return rate, those that had not returned questionnaires after 2 weeks were contacted. Of those given questionnaires, 58% were non-contactable (i.e. no consent given or invalid number) and of those that were, the proportion that answered the phone was the same as those unavailable. Only 18% of those that answered the phone and agreed to return the questionnaires actually did suggesting this was not perhaps the most effective method but none the less prompted few returns.

Recruitment at secondary sites did not prove fruitful even though 2 of the 4 sites were operating several screening clinics a week. With approximately 20 clinicians assessing 4-6 new referrals a day and much enthusiasm from both teams, the researcher had hoped on recruiting more than 7 here. It is likely that a lack of presence on the researcher’s behalf influenced this significantly. A main barrier cited by clinicians centred on back to back 30 minute appointments that left little time to discuss the study without over running. Each clinician agreed to approach one person each day and gave a participant information sheet (PIS) to all those eligible so the study may be considered in their own time. Other barriers included clinicians forgetting to take study material to community appointments thus the study became an agenda item for team meetings, which clinicians found helpful for keeping the study in mind.

129

An unexpected shortage of clinicians at another site increased the clinical workload for other team members. As a result, fewer new referrals were taken on reducing the overall turn-over and number eligible for the study. Due to the competing number of research projects that were already on-going the present study could not be prioritised. There were also concerns from some clinicians that asking patients, who had previously been given a PIS, if they would consider taking part might appear coercive and did not feel it to be appropriate to ask. The researcher maintained frequent contact and attended team meetings to collaborate with clinicians regarding these issues and possible solutions.

Sample

Reason for referral provided a suitable description of the clinical sample in this study. This is particularly relevant as guilt is a significant feature of various clinical presentations, as discussed previously. One limitation is that presenting problem was obtained from the GP referral letter; however, GP’s rarely use standardized interviews focusing on diagnostic criteria and therefore this may not have been the most reliable indicator of mental health status (Clark et al., 2009).

With regards to sample bias, it is possible those high in guilt were more likely to respond with a view to learning more about coping with guilt or to avoid feeling guilty if they did not return questionnaires after having agreed to take part.

Every effort was made to communicate that participation was voluntary and they could withdraw at any time. Given that 97% exceeded the CORE-10 clinical cut off for distress, the sample does not appear biased by missing out those too distressed to take part. It is difficult to comment on how representative the sample was as no data was collected for those who declined to take part. Data was collected for non-

130 responders who consented to take part and for referral information to be accessed. Of the non-responders, 41% were referred primarily for depression, which was a higher percentage than those that had responded (17%) and may have affected return rates.

A similar proportion of non-responders and responders were referred for anxiety

(20% & 23% respectively).

Response rate for the main site of recruitment was 38% for Time 1 and 47% for Time 2. A much smaller sample size was obtained for follow-up data (n = 18).

Small sample sizes can lead to loss of study power and insignificant findings or the over-estimation of the magnitude of an association (Field, 2005). A review has shown that questionnaires were more likely to be returned if short, letters were personalised and stamped returned envelopes were included (Edwards et al., 2009), all of which were considered in the current study. Questionnaires of a sensitive nature have shown poorer response rates (Edwards et al., 2002), which may have been a relevant factor in the current study.

Ethical issues

The study was approved by the University Of Manchester’s Division of Clinical

Psychology and ethical approval was given by the North West NHS Research Ethics

Committee (Appendix Q). One issue initially raised by the committee was the use of a prize draw for a gift voucher as an incentive to take part, which they felt may be misconstrued and be unethical if every participant could not be compensated for their time. The prize draw was therefore withdrawn, which likely hampered recruitment.

Research has shown that the odds of questionnaires being returned are doubled when a financial incentive is involved (Edwards et al., 2002).

131

Participants identified as at risk of self-harm by clinicians were not included in the study, although this was difficult to control at times. Clinicians relied on referral information or past history to indicate risk, which was often brief or absent if a new referral. Given that high variability in suicidal ideation has shown to be the most potent predictor of past attempts and may confer future risk (Witte, Fitzpatrick,

Joiner & Schmidt, 2005), it was important to enquire about current risk status as determined by clinicians as part of their standard clinical assessment. Participants indicating high risk in response to item 6 on the CORE-10 were not followed up as to not exacerbate distress. Clinicians were subsequently informed and the GP where appropriate. Although identifying risk was acknowledged by the ethics committee as difficult to control and possible risks, though minimal, were highlighted on the PIS, the researcher was familiar with a more thorough exploration of risk given their clinical training. Separating the role of researcher from Trainee Clinical Psychologist was an important task.

The researcher did, however, encounter ethical issues that required acute clinical management. The partner of a participant contacted the researcher detailing personal distress unrelated to the study. The researcher sought supervision and suggested they attend their GP who could offer options of support. The researcher was able to use their clinical training to handle this sensitively and professionally whilst maintaining appropriate boundaries. This highlighted the importance of ensuring all researchers working with vulnerable persons have received the appropriate training and supervision to manage these issues.

Informed consent was sought at all levels. Consent to being contacted for the purpose of the study, for access to referral information or to being posted a summary

132 of study findings was optional. This allowed for participants to enter into the first phase even if they did not wish to be followed up to boost recruitment. The study protocol was carried out in accordance with guidelines governing data protection and confidentiality (Data Protection Act, 1998; Department of Health, 2003). Participants were identified by a number to preserve anonymity. Returned data and consent forms were stored securely and independently from each other in a lockable facility on

Trust premises. Only the researcher could match consent forms to data and only non- identifiable data, fully encrypted, was transferred to personal computers for analysis.

Data analysis

One of the original aims was to conduct an exploratory factor analysis (EFA) on the

GLAMS in a clinical sample. Based on the recruitment of 5 persons per variable item, as recommended (Munroe, 2005), it was hoped that 155 would take part, however, a much smaller participant to variable ratio was obtained and therefore

EFA was not possible. A 6 factor solution in a student sample suggested distinct dimensions of coping, which allowed for theoretical predictions to be made. Post- hoc analysis in the current study suggested the GLAMS subscales were not highly inter-correlated and those that were clustered together in conceptually meaningful ways by reflecting adaptive or maladaptive coping consistent with the literature. For example, Self-Punishment significantly correlated with Avoidance and Spirituality and Reparation significantly correlated with External Amelioration and

Metacognition. The GLAMS coping subscales are considered to be distinct yet overlapping with each other, as to be expected based on previous research.

One unexpected association was found between External Amelioration and

Avoidance as these coping constructs seemed unrelated to one another from a

133 theoretical stand point. Seeking social support is considered adaptive and an approach-based form of coping unlike avoidance that is often construed to be unhelpful and refers to disengagement and evading the problem. It may be that recruiting support from others to resolve a problem may be a form of avoidance by not dealing with the problem yourself. For example, relying on others for reassurance, as seen in many anxious presentations, may relieve distress in the short term but may be construed as avoiding the underlying issue in the long term.

The target sample size may have been unrealistic as this did not account for missing data. Due to low participant numbers, it was agreed that missing items would be prorated to minimise lost data and increase the power of the study. Data was excluded when more than 20% of items were missing, which was consistently used across all subscales. Although the CORE-10 must have no more than 10% of items missing for the score to be prorated and reliable (Leach et al., 2006), this assumption had continued to be met in the current study with only two respondents providing 1 or 3 missing values; the latter was excluded from analyses.

The Shapiro-Wilks test identified several variables non-normally distributed, however, due to the potential over-sensitivity of this test to small departures from normality; variables with significant findings in the absence of skewness and kurtosis were accepted as representative of the population. Log, root, square and square root transformations were applied. Following this, skewness and kurtosis values ranged between 0.01- 2.31 and 0.01-1.88, respectively. Due to the moderate to small sample size, values below 2.58 were considered normally distributed and therefore it was appropriate to use parametric tests (Field, 2005). The GLAMS

Spirituality subscale remained significantly positively skewed (Z-skewness; 4.29) thus non-parametric tests were applied.

134

Internal consistency of the GLAMS scale and subscales was reported using

Cronbach’s alpha coefficients (0.56-0.89). Coefficients were generally consistent with those cited for beliefs and values (0.70-0.82) in behavioural research (Peterson,

1994). Although this is the most frequently cited internal reliability statistic (DeVon et al., 2007), Cronbach’s alpha has attained high values even when a set of items measures several latent variables therefore an alternative index of reliability could have been considered (Revelle & Zinbarg, 2009). Reliability over time was assessed using intra-class correlation coefficients (ICCs), deemed appropriate when examining the reliability of a new instrument (Yen & Lo, 2002). ICCs of 0.6 and above have been considered acceptable (Chinn, 1991).

To ensure the necessary assumptions for predictor variables in a linear regression model were met, additional statistics were provided. The Durbin Watson test statistic, which looks at the independence of residuals, can vary between 0 and 4 with a value of 2 meaning residuals are uncorrelated (Field, 2005). The observed values were greater than 2, which may have indicated a slight negative autocorrelation, however, values were evaluated against the upper and lower acceptable limits provided in the original paper by Durbin & Watson (1951), which concluded in favour of zero autocorrelation. This method of evaluation was selected for accuracy, as the Durbin Watson statistic varies depending on the number of observations and predictor variables. A second assumption is lack of multicollinearity between predictor variables, the magnitude of which was analysed by looking at the size of the variance inflation factors (VIF). The cut-off at which one can argue that multicollinearity is high has varied among researchers. The more conservative approach recommended by Bowerman and O’Connell (1990) was taken that suggests values greater than 1 are a cause for concern. In the current study, the

135 necessary assumptions for carrying out a series of regression models were met satisfactorily.

Implications for Future Research and Clinical Practice

An overview of the implications of findings has been considered in the empirical study (see paper 2). The author has selected a few discussion points to further reflect on in this paper. The empirical study provided preliminary evidence in support of the validity and reliability of a unique measure of coping with guilty thoughts. The stability of the proposed factor structure in a clinical population is yet to be established and warrants further investigation in a larger sample. Further refinement of the GLAMS may be necessary to increase the homogeneity of subscales, particularly with regard to External Amelioration; however, this should be considered pending the outcome of an EFA. Future research should also consider confirmatory factor analysis where sample size permits.

The findings of this study suggest guilt-related distress may be amenable to change using mindfulness-based practice, which should be explored in future research using randomised controlled trials. This serves to expand the existing evidence base that currently promotes trauma-focused and cognitive-based interventions that centre on modifying guilt-based appraisals. As guilt features across various psychological disorders future research may wish to explore if ways of coping with guilt varies as a function of clinical presentation. Although the present study found no difference in guilt or its management between groups anxious or depressed, this may have reflected the small sample sizes (< 15). There may be within group differences (e.g. between OCD and PTSD samples) that have been

136 overlooked, as the number in each group were too small to make meaningful comparisons.

Guilt is also argued to be qualitatively different across non-westernised cultures and there may also be cross-cultural variation in the correlates and consequences of guilt (Wong & Tsai, 2007). Given that the GLAMS was validated in a sample predominantly White British it may be less culturally relevant. Further research is required to better our understanding of the cultural variations of this construct and how it is managed. This is important if we are to avoid marginalising the needs of vulnerable groups associated with various sub-cultures.

Findings suggest the GLAMS may not function completely as a reliable outcome measure when assessing coping pre and post intervention as more adaptive ways of coping do not appear to be stable over time. Given the stability of maladaptive coping responses the GLAMS may be helpful in monitoring change by measuring the frequency in which respondents engage in destructive ways of coping with guilt over time. The clinical utility of the GLAMS may be found at the assessment stage whereby maladaptive guilt management may be indicated necessitating a specific treatment that targets guilt-related thoughts. For therapy to produce clinically meaningful outcomes effective guilt management may need to complement more traditional therapies in the future where guilt is identified.

Different ways of coping with guilt may also guide clinicians towards alternative forms of therapy such as mindfulness or compassionate mind training.

GLAMS Spirituality was found to be significantly related to Dysfunctional

Coping (p < 0.05), Worry and Punishment subscales (p ≤ 0.001) and demonstrated a positive trend with State Guilt (no relationship was found with global distress). This

137 was surprising given that items for this subscale reflected positive aspects of religious coping relating to forgiveness and faith as opposed to negative aspects (e.g.

God is punishing me). This suggests that religious coping may not be adaptive and may exacerbate guilt-related distress. That is not to say religious coping does not serve to ameliorate other types of distress or serve other functions and may create somewhat of an impasse in therapy. Addressing this issue with a view to reducing religious or spiritual coping could be inappropriate and culturally insensitive and a delicate matter that would need to be explored with the client.

138

References

Abrahams, S., & Udwin, O. (2002). An evaluation of a primary care-based child clinical psychology service. Child and Adolescent Mental Health, 7(3), 107- 113.

American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders (4th ed.). Washington, DC: APA.

Andrews, B., & Brewin, C. R. (1990). Attributions for martial violence: A study of antecedents and consequences. Journal of Marriage and the Family, 52, 757- 767.

Balk, E. M., Bonis, P. A. L., Moskowitz, H., Schmid, C. H., Ioannidis, J. P. A., Wang, C., Lau, J. (2002). Correlation of quality measures with estimates of treatment effect in meta-analyses of randomized controlled trials. JAMA, 287(22), 2973-2982.

Barkham, M., Gilbert, N., Connell, J., Marshall, C., & Twigg, E. (2005). Suitability and utility of the CORE-OM and CORE-A for assessing severity of presenting problems in psychological therapy services based in primary and secondary care settings. British Journal of Psychiatry, 186, 239-46.

Beck, A. T. (1991). Cognitive therapy: A 30 year retrospective. American Psychologist, 46(4), 368-75.

Bisson, J. I., Ehlers, A., Matthews, R., Pilling, S., Richards, D., & Turner, S. (2007). Psychological treatments for chronic post-traumatic stress disorder: Systematic review and meta-analysis. British Journal of Psychiatry, 190, 97- 104.

Bowerman, B. L., & O’Connell, R. T. (1990). Linear statistical models: An applied approach (2nd ed.). Belmont, CA: Duxbury.

Burney, J., & Irwin, H. J. (2000). Shame and guilt in women with eating-disorder symptomatology. Journal of Clinical Psychology, 56, 51-61.

Bybee, J. (1998). Guilt and children. San Diego, CA: Academic Press.

139

Carver, C. S. (1997). You want to measure coping but your protocol’s too long: Consider the Brief COPE. International Journal of Behavioral Medicine, 4, 92-100.

Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing coping strategies: A theoretically based approach. Journal of Personality and Social Psychology, 56, 267-283.

Chinn, S (1991). Repeatability and method comparison. Thorax, 46, 454-456.

Clark, L. A., & Watson, D. (1995). Constructing validity: Basic issues in objective scale development. Psychological Assessment, 7(3), 309-319.

Clements, C. M., & Sawhney, D. K. (2000). Coping with domestic violence: Control attributions, dysphoria, and hopelessness. Journal of Traumatic Stress, 13(2), 219-240.

Connell, J., & Barkham, M. (2007). CORE-10 User Manual, Version 1.1. Bath, UK: CORE System Trust & CORE Information Management Systems Ltd.

Data Protection Act (1998). Retrieved from http://www.legislation.gov.uk/ukpga/1998/29/contents/enacted

Davis, L. L. (1992). Instrument review: Getting the most from your panel of experts. Applied Nursing Research, 5, 194-197.

Deeks, J. J., Dinnes, J., D’Amico, R., Sowden, A. J., Sakarovitch, C., Song, F.,…Petticrew, M. (2003). Evaluating non-randomised intervention studies. Health Technology Assessment, 7(27).

Department of Health (2003). NHS confidentiality: Code of practice. London, UK: The Stationary Office.

DeVon, H. A., Block, M. E., Moyle-Wright, P., Ernst, D. M., Hayden, S. J., Lazzara, D. J., & Kostas-Polston, E. (2007). A psychometric toolbox for testing validity and reliability. Journal of Nursing Scholarship, 39, 155-164.

Durbin, J., & Watson, G. S. (1951). Testing for serial correlation in least squares regression, II. Biometrika, 30, 159-178.

Edwards, P. J., Roberts, I., Clarke, M., DiGuiseppi, C., Pratap, S., Wentz, R., & Kwan, I. (2002). Increasing response rates to postal questionnaires: Systematic review. British Medical Journal, 324, 1183-1185.

140

Edwards, P. J., Roberts I., Clarke, M. J., DiGuiseppi, C., Wentz, R., Kwan, I.,…Pratap S. (2009). Methods to increase response to postal and electronic questionnaires. Cochrane Database of Systematic Reviews, 8(3), MR000008.

Ehlers, A., Bisson, J., Clark, D. M., Creamer, M., Pilling, S., Richards, D.,…Yule, W. (2010). Do all psychological treatments really work the same in posttraumatic stress disorder? Clinical Psychology Review, 30, 269-276.

Evans C., Connell J., Barkham M., Margison F., Mcgrath G., Mellor-Clark, J., & Audin, K. (2002). Towards a standardised brief outcome measure: Psychometric properties and utility of the CORE-OM. British Journal of Psychiatry, 180, 51-60.

Ferguson, T. J., & Crowley, S. (1997). Measure for measure: A multitrait- multimethod analysis of guilt and shame. Journal of Personality Assessment, 69(2), 425-441.

Field, A. P. (2005). Discovering statistics using SPSS. London, UK: Sage.

Finn, S. E., & Tonsager, M. E. (1997). Information-gathering and therapeutic models of assessment: Complementary paradigms. Psychological Assessment, 9(4), 374-385.

Garland, R. (1991). The mid-point on a rating scale: Is it desirable? Marketing Bulletin, 2, 66-70.

Garland, E. L. (2007). The meaning of mindfulness: A second-order cybernetics of stress, metacognition, and coping. Complementary Health Practice Review, 12(1), 15-30.

Greasley, P., & Small, N. (2005). Evaluating a primary care counselling service: Outcomes and issues. Primary Health Care Research and Development, 6(2), 125-136.

Harder, D. W. (1995). Shame and guilt assessment, and relationships of shame- and guilt-proneness to psychopathology. In J. P. Tangney & K. W. Fischer (Eds.), Self-conscious emotions: The psychology of shame, guilt, embarrassment and pride (pp. 368-392). New York, NY: Guildford Press.

Haynes, S. N., Richard, D. C. S., & Kubany, E. S. (1995). Content validity in psychological assessment: A functional approach to concepts and methods. Psychological Assessment, 7, 238-247.

141

Hendin, H., & Haas, A. P. (1991). Suicide and guilt as manifestations of PTSD in Vietnam combat veterans. American Journal of Psychiatry, 148, 586-591.

Herbison, P., Hay-Smith, J., & Gillespie, W. J. (2006). Adjustment of meta-analyses on the basis of quality scores should be abandoned. Journal of Clinical Epidemiology, 59, 1249-1256.

Johns, R. (2005). One size doesn’t fit all: Selecting response scales for attitude items. Journal of Elections, Public Opinion & Parties, 15(2), 237‐64.

Jones, W. H., Schratter, A. K., & Kugler, K. (2000). The Guilt Inventory. Psychological Reports, 87, 1039-1042.

Kim, S., Thibodeau, R., & Jorgensen, R. S. (2011). Shame, guilt, and depressive symptoms: A meta-analytic review. Psychological Bulletin, 137, 68-96

Kochanska, G., Gross, J. N., Lin, M., & Nichols, K. E. (2002). Guilt in young children: Development, determinants, and relations with a broader system of standards. Child Development, 73, 461-482.

Kubany, E. S., & Manke, F. P. (1995). Cognitive therapy for trauma-related guilt: Conceptual bases and treatment outlines. Cognitive and Behavioral Practice, 2, 23-61.

Kubany, E. S., & Watson, S. B. (2003). Guilt: Elaboration of multidimensional model. The Psychological Record, 53, 51-90.

Kugler, K., & Jones, W. H. (1992). On conceptualizing and assessing guilt. Journal of Personality and Social Psychology, 62(2), 318-327.

Leach, C., Lucock, M., Barkham, M., Stiles, W. B., Noble, R., & Iveson, S. (2006). Transforming between Beck Depression Inventory and CORE-OM scores in routine clinical practice. British Journal of Clinical Psychology, 45, 153-166.

Lee, D. A., Scragg, P., & Turner, S. (2001). The role of shame and guilt in traumatic events: A clinical model of shame-based and guilt-based PTSD. British Journal of Medical Psychology, 74, 451-466.

Lynn, M. R. (1986). Determination and quantification of content validity. Nursing Research, 35, 382-385.

142

Marx, A. G., Menezes, A., Horovitz, L., Jones, E. C., & Warren, R. F. (2003). A comparison of two time intervals for test re-test reliability of health status instruments. Journal of Clinical Epidemiology, 56, 730-735.

Moher, D., Jadad, A., Nichol, G., Penman, M., Tugwell, P., & Walsh, S. (1995). Assessing the quality of randomized controlled trials: An annotated bibliography of scales and checklists. Controlled Clinical Trials, 16, 62-73.

Moher, D., Pham, B., Jones, A., Cook, D. J., Jadad, A. R., Moher, M.,…Tugwell, P. (1998). Does quality of reports of randomised trials affect estimates of intervention efficacy reported in meta-analyses? Lancet, 352, 609-613.

Munro, B. H. (2005). Statistical methods for health care research. Philadelphia, PA: Lippincott, Williams & Wilkins.

National Institute for Health and Care Excellence (2005). Post-traumatic stress disorder (PTSD). The management of PTSD in adults and children in primary and secondary care, CG26. London, UK: National Institute for Health and Care Excellence.

O’Connor, L. E., Berry, J. W., Weiss, J., Bush, M., & Sampson, H. (1997). Interpersonal guilt: The development of a new measure. Journal of Clinical Psychology, 53, 73-89.

Ozer, E. J., Best, S. R., Lipsey, T. L., & Weiss, D. R. (2003). Predictors of posttraumatic stress disorder and symptoms in adults: A meta-analysis. Psychological Bulletin, 129, 52-73.

Peterson, R. A. (2013). Meta-analysis of Cronbach’s coefficient alpha. Journal of Consumer Research, 21(2), 381-391.

Petticrew, M., & Roberts, H. (2006). Systematic reviews in the social sciences. A practical guide. Malden, MA: Blackwell.

Polit, D. F., Beck, C. T., & Owen, S. V. (2007). Is the CVI an acceptable indicator of content validity? Appraisal and recommendations. Research in Nursing & Health, 30(4), 459-467.

Revelle, W., & Zinbarg, R. (2009). Coefficients alpha, beta, omega, and the glb: Comments on Sijtsma. Psychometrika, 74(1), 145-154.

143

Shafran, R., Watkins, E., & Charman, T. (1996). Guilt in obsessive–compulsive disorder. Journal of Anxiety Disorders, 10, 509-516.