UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Responsibility in obsessive compulsive disorder: is it worth checking? Lopatka, Cindy Lee 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1994-893494.pdf [ 2.2MB ]
JSON: 831-1.0088153.json
JSON-LD: 831-1.0088153-ld.json
RDF/XML (Pretty): 831-1.0088153-rdf.xml
RDF/JSON: 831-1.0088153-rdf.json
Turtle: 831-1.0088153-turtle.txt
N-Triples: 831-1.0088153-rdf-ntriples.txt
Original Record: 831-1.0088153-source.json
Full Text

Full Text

RESPONSIBILITY IN OBSESSIVE COMPULSIVE DISORDER:IS IT WORTH CHECKING?byCINDY LEE LOPATKAB.A., The University of British Columbia, 1985M.A., The State University of New York at Albany, 1989A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(PSYCHOLOGY)We accept this thesis as conformingto the ry9red standardTHE UNIVERSITY OF BRITISH COLUMBIANovember 1993© Cindy Lee Lopatka, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)__________________________Department of /The University of British ColumbiaVancouver, CanadaDate__________DE-6 (2/88)iiABSTRACTThe purpose of this investigation was to test thehypothesis that perceived responsibility is a majordeterminant of compulsive checking. Thirty participantsrecruited from the community through the local media, whomet criteria for Obsessive Compulsive Disorder, receivedfour conditions. In the low responsibility condition,perceived responsibility for an anticipated negative eventtwas transferred to the experimenter. In contrast, in thehigh responsibility condition, perceived responsibility foran anticipated negative event was given to the participant.The remaining two conditions served as control conditions.Subjects were assessed before and after each experimentalmanipulation.Results suggest a causal connection between decreasesin perceived responsibility and compulsive checking.Decreases in perceived responsibility produced decreases inseveral measures critical to compulsive checking. Resultsfrom increases in perceived responsibility were less clear.However, increases in perceived responsibility lead toincreases in panic and likelihood of anticipated criticism.iiiThere were trends for increases in perceived responsibilityto lead to increases in perceptions of discomfortexperienced, urge to check, and severity of anticipatedcriticism. There was no relationship between variations inperceived responsibility and perceived extent ofcontrollability over an anticipated negative event.Theoretical implications of the results and, inparticular, the value of a cognitive analysis of compulsivechecking, are discussed.Table of ContentsAbstractList of TablesList of FiguresAcknowledgementsPREFACEINTRODUCTIONBiological Approach to OCDCognitive Deficit Approach to OCD.Behavioural Theory of OCDCognitive AnalysesCognitive-Behavioural Analysis ofResponsibility and OCDAttribution ResearchCriticismiiviiviiiix138111424OCD 35485055ivBeliefs of Harm 57Responsibility, Controllability and OCDImmediacy of ThreatBaseratesSummaryHypotheses and PredictionsMETHOD.• .64• .71VExperimental Design and Overview 71Procedure 75Measures 86RESULTS 90DISCUSSION 138Theoretical Implications 145Maintenance 148Cognitive Biases 149Compulsive Checking and Depression 151Compulsive Checking and Criticism 152Limitations of Within Subject Design 155Advantages of Within Subject Design 157Strengths and Weakness of Self-Report Measures 158Bandura’s Social Learning Model 167Kohlberg’s Model 169Implications for Treatment 174Limitations 182Contributions 187Conclusions 188REFERENCES 189APPENDICES 210viAPPENDIX A:APPENDIX B:APPENDIX C:APPENDIX D:APPENDIX E:VERBAL ANALOGUE SCALES 211CONSENT FORM 218STRUCTURED INTERVIEW 220RESPONSIBILITY CONTRACTS 227DEBRIEFING FORM 231viiLIST OF TABLESTABLE 1Subject Characteristics 91TABLE 2Pre Scores, Means and Standard Deviations 96TABLE 2Post Scores, Means and Standard Deviations. . . .97TABLE 4Dunnett Planned Comparisons 105TABLE 5Control ConditionPearson Correlations 122TABLE 6High Responsibility ConditionPearson Correlations 124TABLE 7Low Responsibility ConditionPearson Correlations 126TABLE 8High Responsibility ConditionPerceived responsibility>90 143LIST OF FIGURESFIGURE 1Schematic Representation of Results 128viiiixACKNOWLEDGMENTSI gratefully acknowledge the following people for theircontributions to this study:First, to my advisor, Jack Rachman, for years of hisinspired views, turning the seemingly difficult into play,guidance, and fun; to Charlotte Johnston for her encouragement,commitment, and constructive feedback on methodology; and to myuniversity examiners Ken Craig, and Ishu Ishiyama for theirhelpful comments and insights.I would also like to acknowledge my friends and family whocontinue to make my life richer and more whole: to Karen, forher love, support, comfort, acceptance, and taking me out toplay; to Jen, for her calm, caring perspective and laughter; toCathy, for her “miracle—cure phone calls” and whose constantfaith, and support have been invaluable; to Mel, for his longtermsupport and caring in both practical and spiritual ways; to Jamesand Lorne, for their comfort at important times; and to my motherfor long distance phone calls and always believing in me.1PREFACEMy strong interest in researching and treatingthose experiencing obsessive compulsive problems arosewhile working in a medical center in upstate New York.It was common practice in this facility to give peoplewho were headed for psychosurgery in Boston, a lastattempt at behavioural therapy before surgery wasscheduled. One of my clients was in these direcircumstances. As we worked together I began to learnof the often extreme nature of their difficulties. Ialso became aware of the shortcomings of “my bag oftheories and techniques”. I read widely from varioustheoretical positions and models, and yet failed tofind much solace. The theories and treatmentapproaches stemming from these models seemed to be astart but certainly were not sufficient to deal withthe nature of the problems my clients, and others,grapple with daily.If this type of circumstance was not enough tomotivate me to learn more, my curiosity was piqued bythe seemingly senseless behaviour these individuals2filled their lives with despite their desire not to doso. In many instances, highly articulate andintelligent people would describe their need to performsome tedious, action over and over again. In othercases, they would describe a seemingly bizarre actionor thought and the intense need to literally fill theirlife with it - as if everything depended upon it.As a start, I too wondered why their lives werefilled with performing these senseless tasks and whythey were unable to prevent themselves from carryingout these actions. At this point, I have come fullcircle and now find I am perplexed by why others comingfrom similar circumstances, do n go on to developobsessive compulsive problems.This thesis is an attempt to make senè of thepresumed senselessness of their actions.3INTRODUCTIONOf all the anxiety disorders, obsessive compulsivedisorder (OCD) is often conceptualized as the mostsevere emotional problem. Until recently, OCD wasthought to be very uncommon with prevalence ratesestimated at less than one percent (AmericanPsychiatric Association, 1980). However, thesurprising prevalence rates recently discovered in alarge, multi—centered community study suggestsotherwise. The Epidemiologic Catchment Area (ECA)survey study sponsored by the National Institute ofMental Health, found that for a six-month periodprevalence rates were estimated at approximately 1.3—2% (Myers et al., 1984). The characteristic pattern ofshame and guilt, extreme secrecy and fears of “goingcrazy” often associated with this disorder may resultin a reluctance to self—disclose and enter treatmentand ultimately, to be for accounted for (Rapoport,1990; Talus, 1992). The hundreds of phone calls Ireceived for this study certainly attest to these4prevalence rates. Often their intense fears of “beingfound out” or “going crazy” would result in many notbeing willing to offer even a first name. Many whocalled spoke of their extreme desperation, secrecy andpain but were often unwilling to either take part inour study or accept referrals to community mentalhealth resources.The intensity and severity of this disorder maynot be fully grasped without case illustrations.Barlow (1989) describes a severe example of a patientseeking help that illustrates both the complexity ofthis problem as well as the wide range of behavior thatcan accompany this disorder. He describes the clientas:a nineteen year old, single white male, acollege freshman majoring in philosophy whohad withdrawn from school because ofincapacitating ritualistic behaviors. Thepatient had an 8—year history of severecompulsive rituals. These included excessivehand washing and showering; ceremonialrituals for dressing and studying; compulsiveplacement of any objects handled; grotesquehissing, coughing, and head tossing, whileeating; and shuffling and wiping his feetwhile walking. These rituals interfered withevery aspect of his daily functioning. Thepatient had steadily deteriorated within thepast 2 years, isolating himself from family5and friends, refusing meals, and neglectinghis personal appearance. His hair was verylong, as he had not allowed it to be cut in 5years. He had never shaved or trimmed hisbeard. When he walked, he shuffled, takingsmall steps on his toes while continuallylooking back, checking and re-checking. Onoccasion he would run quickly in place. Hehad withdrawn his left arm completely fromhis shirt sleeve, as if he were crippled andhis shirt was a sling. Seven weeks prior toadmission, his rituals had become so time—consuming and debilitating that he refused toengage in any personal hygiene for fear thatthe associated rituals would interfere withthe time needed to study. Almost continualshowering became no showering. He stoppedwashing his hair, brushing his teeth, orchanging his clothes. He left his roominfrequently and, to avoid rituals associatedwith the toilet, had began defecating onpaper towels, urinating in paper cups, andstoring the waste in a corner of the closetin his room. His eating habits haddegenerated from eating with the family, toeating in an adjoining room, to eating in hisown room In the two months prior toadmission, he had lost 20 pounds and wouldonly eat late at night when others wereasleep. He felt that eating was “barbaric”;this well described his grotesque eatingrituals, consisting of hissing noises, coughsand hacks, and severe head tossing. His foodintake had been narrowed to ice cream or amixture of peanut butter, sugar, cocoa, milk,and mayonnaise. He considered several foods(e.g., cola, beef, and butter) contaminating,and would not eat these foods. He also had along list of checking rituals associated withthe placement of objects. Excessive time wasspent checking and rechecking to see thatwastebaskets and curtains were in place.These rituals had progressed to tilting ofwastebaskets and twisting of curtains, which6were checked periodically throughout the day.(Barlow, 1989, pp.#599—600)The description of this severe case exampleclearly fits within the definition of OCD in theDiagnostic and Statistical Manual - Revised (DSM-IIIR),which characterizes this disorder as “recurrentobsessions or compulsions sufficiently severe to causemarked distress, be time consuming, or significantlyinterfere with the person’s normal routine,occupational functioning, or with usual socialactivities or relationships with others” (AmericanPsychiatric Association, 1987, p. 245). Obsessions aredefined as “persistent ideas, thoughts, impulses, orimages that are experienced, at least initially, asintrusive and senseless (for example, a parent havingrepeated impulses to kill a love child, or a religiousperson having recurrent blasphemous thoughts)”(American Psychiatric Association 1987, p. 245).Additional criteria and guidelines include that theperson must at least initially, resist, ignore orattempt to suppress these thoughts and that he/shemust be aware that the thought is a product of his/hermind and not externally imposed. Compulsions are7defined as “repetitive, purposeful, and intentionalbehaviors that are performed in response to anobsession, or according to certain rules or in astereotyped fashion..the behavior is designed toneutralize or to prevent discomfort or some dreadedevent or situation; however, either the activity is notconnected in a realistic way with what it is designedto neutralize or prevent, or it is clearly excessive...the person recognizes that his or her behavior isexcessive or unreasonable” (American PsychiatricAssociation, 1987, p.#245).Before turning to the heart of this thesis whichis a cognitive approach to QCD, a brief look at twoalternative explanations, biological and cognitivedeficit, will be discussed. This examination is a verysimplified overview and is not intended to representthe field as a whole. It is provided solely to givereference and note the alternative explanation. Futureresearch may need to consider integrating thesecompeting theories. However, for all intents andpurposes it will be considered outside the scope ofthis thesis.8A Biological Approach to OCDThis conceptualization proposes that OCD is aneurological disorder (Wise & Rapoport 1989). Withinthis framework, OCD is considered to be a motordisorder and the underlying pathology is hypothesizedto be located in the basal ganglia. It links bothethological and neuroanatomic data and proposes aconcept of parallel “loops” interconnecting basalganglia, thalamus, and cerebral cortex (Alexander,DeLong, & Strick, 1988) To vastly simplify, thestriatum acts to trigger ethologically preparedbehaviour and OCD is the result of inappropriatetriggering of genetically stored and learnedbehaviours. Predictions stemming from this positioninclude: inability to inhibit inappropriate responses,perseveration and incorrect planning strategiesGeneral neuropsychological deficits of spatial andshifting ability should be apparent. Rapoport(1991) provides evidence in support of this modelthrough citing a link between OCD and Tourette’sSyndrome, and other apparent associated basal gangliadysfunction disorders such as Sydenham’s chorea, and9postencephalitic Parkinsonism (see Wise & Rapoport 1989for review). Prevalence rates of the co—occurrence ofOCD and these disorders are used to substantiate thishypothesis.In addition, brain imaging studies have beenconducted to further test this basal gangliadysfunction hypothesis. Structural as well asfunctional abnormalities in OCD patients have beenreported using computed tomography (CT) and positronemission tomography (PET) (see Rapoport, 1991 for areview). However, the evidence is scanty at best andlacks reliability. That is, areas that showedpathology in one study failed to be replicated inanother. Overall, as Rapoport (1991) reports none ofthe twelve studies conducted is without significantproblems and tighter methodology must be employed toadequately test this hypothesis. This area of researchis being pursued vigorously at present.The second wave of evidence in support of abiological approach to OCD comes from drug treatmentresponse. Individuals with OCD have been found to10respond to serotonin (5-hydroxytryptamine (5-HT) uptakeblocking drugs clomipramine, fluoxetine, andfluvoxamine (see Jenike, Baer, Minichiello, 1990 for areview). There is presumed to be a specific 5—HTeffect because noradrenergically selectiveantidepressants such as desipramine do not work(Leonard, Swedo, Rapoport, Koby, Lenane, Cheslow, &Hamburger, 1989), and clomipramine has not been shownto be effective with other anxiety disorders such aspanic disorder (Zohar, 1989). In addition, individualswith OCD do not respond to provocative agents (ie.sodium lactate, carbon dioxide) that have been shown toincrease symptoms in other anxiety disorders (Gorman,Liebowitz, Fyer, Dillon, Davies, Stein, & Klein, 1988;Hollander, DeCaria & Liebowitz, 1989). However, theuse of provocation agents to indicate biologicalvulnerability or as a means of classifying anxietydisorders has been seriously challenged (for reviewssee Margraf, Ehlers, & Roth ; Rachman & Maser, 1988)Overall, this “serotonin dysfunction” positionhas been acknowledged by many within this field as toosimplistic and lacks in direct evidence of serotonin11pathology (see Rapoport, 1991). For example, there islittle evidence to suggest that individuals with OCDhave more or less serotonin production when compared tocontrols and apparent “selective drugs” are often foundwith time to have other actions. Further, to implycausation from the effect of a treatment is logicallyfaulty. A variety of different treatments may beeffective for a particular disorder but it tells uslittle about causation.To summarize, the evidence in support of abiological approach to OCD appears to be weak. Giventhe active research being conducted from thisperspective, we should be much clearer in the nearfuture about the promise of the basal gangliadysfunction hypothesis.Cognitive Deficit Approach To OCDCognitive deficit approaches to OCD, asexemplified by Reed (1977) see OCD as a breakdown ofcognitive control processes or a type of memorydysfunction. This approach draws from an informationprocessing model and hypothesizes that the cognitive12processes of those with obsessive—compulsive problemsare characterized by “overstructuring, involving closeattention to the details of input, over—specificity,search for information and the deferral of decision asto completion.”(Reed, 1977, p. 1984) This hypothesisis in keeping with the general characteristic thatthose with OCD appear excessively concerned withdetails that many may find irrelevant.Studies providing support for this model includeReed (l969a) which found that those with OCDsymptomatology selected fewer alternatives than acontrol group in a card sorting task and Reed (l969b)which found that those with OCD generated morecategories than did a psychiatric control group. +Although these results were provocative they werenot specific in identifying deficits which may berelated to compulsive checking. More recently, Personsand Foa (1984) hypothesized that rather than a generalcognitive deficit as Reed (1977) posits, those with OCDmay have a more complex definition of concepts butsolely for items that are fear-related. Results did13not confirm this hypothesis but pointed to a generalmemory deficit.. Others such has Sher, Frost and Otto(1983) also attempted to identify specific cognitivedeficits which could be linked back to compulsivebehaviour. They suggest that those with OCD have apoorer memory for actions but not an overall memorydeficit.Current research continues to address theseissues. The focus of late has been to use collegestudents who score high on the Maudsley Obsessional—compulsive Inventory (MOCI) in analogue typeexperiments. The thrust of this research is to lookfor specific memory deficits related to compulsivebehaviour. At present it is still unclear whetherthere is a more general slowness of processinginformation or whether there is are specific deficitsthat can be linked back to compulsive behaviour. Thisarea of research is currently being pursued vigorouslyby Sher, Frost and colleagues as well as others andresults may begin to give us greater insight into thecognitive processes of those with compulsive problems.14The focus of this thesis will be on compulsivechecking and the benefit of a cognitive analysis.However, before turning to these specifics, it may befruitful to examine the forces that led to theemergence of this theoretical position. First, a briefexamination of the behavioural theory of OCD and howthe gaps in this model encouraged the emergence of amore encompassing approach to OCD will be discussed.Second, it will be argued that the spirit of the times,or zeitgeist, has been to re—examine emotional problemsthrough the lens of cognitive theory and thereformulation of OCD is in keeping with this wave ofchange.Behavioural Theory of OCDAs a start, it must be stated that prior to thedevelopment of the behavioural theory of OCD, there wasno successful treatment for this disorder. Overall,traditional psychotherapy had not proven to besuccessful in treatment of OCD (Black, 1974). In fact,many considered OCD to be an intractable problem. Aslate as 1960, Breitner stated that “most of us are15agreed that the treatment of obsessional states is oneof the most difficult tasks confronting thepsychiatrist and many of us consider it hopeless” (p.32).The roots of the behavioural theory of OCD liewith Mowrer’s (1939) two—factor theory of fear. Mowrerproposed that the onset and maintenance of fear andavoidance behaviour could be explained through simpleconditioning theory. He predicted that when a neutralstimulus is paired with a unconditioned stimulus (e.g.it innately produces fear) the neutral stimulus willeventually take on the ability to evoke fear. Thisneutral stimulus may include physical objects as wellas cognitions and images. The second stage of thismodel posits that the individual uses passive or activeavoidance as a way of dealing with the intrinsicallyaversive anxiety state. Passive avoidance is difficultto maintain with obsessive—compulsive problems giventhe massive extent of generalization. Thus, activeavoidance patterns, in the form of checking and washingoften follow.16In more recent years, this theory has beengenerally viewed as insufficient for explaining theonset of fear (e.g. Rachman & Wilson, 1980). Althoughthe beginnings of this disorder do often occur after astressor, there is little evidence to suggest atraumatic onset or co—occurrence of symptoms and “hard—wired” fear reaction. Rachman (1981) suggested thatheightened arousal may lead to sensitization ofthoughts which are associated with this state.Similarly, Eysenck (1979) stated that individuals whohave been exposed to aversive events which are aboveindividual limits will lead to sensitization of thesecues. Boulougouris (1977) provided experimentalevidence in support of these views when demonstratingthat individuals with OCD were more physiologicallyreactive when anticipating an aversive event than werean experimental control group.According to behavioural theory, the maintenanceof the disorder can be explained more readily. Thereis support for the notion that obsessions lead to anincrease in anxiety whereas compulsions lead to adecrease in anxiety. In learning terms, an obsessive17thought produces anxiety because it is associated withan unconditioned stimulus eliciting anxiety.Similarly, compulsions arise because initially theserituals result in short—term anxiety reduction.However, in the long—run the decrease in anxiety servesto reinforce the compulsion. Boulougouris, Rabavilas,and Stefanis (1977) found that obsessive thoughtsincreased heart rate and skin conductance when comparedto neutral thoughts. Similarly, exposure to fearedobjects led to increases in heart rate and subjectivelevels of anxiety (Hodgson & Rachman, 1972) and skinconductance levels (Hornsveld, Kraaimaat, & Van daml—baggen, 1979).There is considerable evidence suggestingritualistic behaviours result in anxiety reduction. Apioneering series of investigations with individualsexperiencing washing and checking compulsions foundthat after intentional provocation of the urge toritualize, the pattern of behaviour led to a decreasein subjective as well as physiological measures ofanxiety (Hodgson & Rachman, 1972; Hornsveld et al.181979, Roper, Rachman & Hodgson, 1973; Roper & Rachman,1975).The main treatment approach arising from abehavioural conceptualization of OCD is exposure andresponse prevention. Exposure and response preventionconsists of exposure to the provoking stimuli, eitherinvivo or imaginal and secondly, not allowing thepatient to perform the ritual after the exposure. In1966, initial reports of the efficacy of this approachemerged with Meyer (1966), who successfully treated twoobsessive compulsive clients. Later studies showedthat the treatment was successful in 10 out of 15severely dysfunctional patients. The successesmaintained themselves with a 5 to 6 year follow-up(Meyer & Levy, 1973; Meyer, Levy & Schnurer, 1974).Several controlled studies have confirmed the promisingresults of the early uncontrolled clinical trials(Emmelkamp & Van Kraanen, 1977; Foa & Goldstein, 1978;Marks, Hallam, Connolly, & Philpott, 1976). Long-termmaintenance studies showed that improvement held overtime. For example, Emmelkamp and Rabble (1983)conducted a 4 year follow—up on 60 obsessive compulsive19disorder clients who had been treated with exposure andresponse prevention. Overall, they found improvementwas maintained on subjective self—ratings of obsessivecompulsive symptoms, anxiety and depression. Overall,global ratings showed 57% were much better, 23% wereimproved and 20% were unchanged. Other controlledstudies, such as Mawson, Marks and Ramm (1982), suggesteven greater long—term gains. Thirty—seven of 40obsessive compulsive disorder clients were followed for2 years. Results indicated that 75% of participantswere much improved, 5% improved, and 20% unchanged.Similar results suggesting stability of gains werereported by Robertson (1979), Foa and Goldstein (1978)and Marks, Hodgson and Rachman (1975) in long-termfollow- up studies varying from 2 to 4 years.What can be said to summarize the studies withhundreds of obsessive compulsive clients treated withexposure and response prevention? First, the outcomerates have been remarkably consistent. Approximately65—75% of clients treated with this procedure haveimproved and remained improved at long—term follow—up.Most patients in these studies have had chronic20problems with the mean duration of disorder ofapproximately 10 years and were at the top end ofscales in both subjective distress and avoidance.Given the consistent results with such severepopulation, we can be reasonably confident of theefficacy of this approach.Why then, if this approach to treatment was sosuccessful, was there a need to develop alternativeexplanations and treatment models? Two possiblesources of explanation may be considered. First,limitations of the behavioral theory and treatment ofOCD provided the opening to consider alternateexplanations. Some of these weaknesses will beexplored below. Second, the growing dissatisfactionwith behavioural theory across areas of emotionaldisorders, lead in part to the emergence of cognitiveapproaches to psychopathology.Upon turning to the limitations of behaviouraltherapy for the treatment of OCD, the issue of clinicalversus statistical significance becomes apparent(Meehl, 1966). We know that exposure plus response21prevention has been demonstrated as a statisticallysignificant form of treatment. That is, when comparedto wait—list, placebo, alternative forms of therapy,etc., exposure plus response prevention is a superiorform of treatment. In addition, we know that this typeof treatment produces a statistically significantreduction in symptoms. Yet, in the therapist’s officeare these differences meaningful? The answer may beboth yes and no. It may be argued that clients andtherapists alike, observe a marked decrease insymptomotology with this type of treatment. However, adecrease in obsessive thoughts, washing, checking etc.is not the same as long—term absence of symptoms.Exposure plus response prevention helps but does notremove the symptom cluster (Rachman, 1983). Therefore,it may be of use to describe this treatment asclinically helpful but certainly not a panacea.Second, the issue of relapse is apparent inworking with those who have gone through behaviouraltherapy (Foa, 1979). The clinical pattern of adecrease in symptoms at the end of treatment to befollowed by an insidious return of symptoms is noted by22many clinicians. In other instances, clients may besymptom free for a time and then experience a sudden,intense return of symptoms (Rachman, 1992). In manyinstances stressors that are seemingly unrelated to thecondition appear to trigger the relapse. Re— examiningthese triggers from the cognitive perspective that willbe outlined later may be of benefit and possibly shedsome light on to how these seemingly disparate sourcestrigger OCD symptoms.Third, in severe cases with individuals who have avery complex, elaborate system of obsessive thoughtsand compulsions such as in the case illustration, theremay be from the onset, little if any, response tobehavioural treatment (Foa, Stekette, Grayson &Doppelt, 1983). In these instances, many clinicianswould agree that we are working around the periphery ofthe problem and move from targeting one minute, ritualor obsession to another. The heart or core of thisdisorder, remains elusive.Fourth, behaviour therapy is shown to be lesseffective with overvalued ideation. Wernicke (1900)23first addressed the issue of overvalued ideation anddefined it as being neither an obsession or a delusionbut a single strongly held belief so dominant that itpreoccupies the individual’s life. The distinctionsbetween delusions and overvalued ideation are blurredand some sources such as the DSM—IIIR suggest a direct,albeit circular, approach to separating the concepts.According to these authors, beliefs amenable to changeshould be considered overvalued ideation and those thatcannot be changed, delusions.Foa (1979) provided a more precise definition bydefining overvalued ideation as client assignment of ahigh and unremitting probability to a feared negativeconsequence when not in the presence of the feared cue.Steketee and Foa (1985) suggest that this type ofovervalued ideation is a negative prognostic sign forbehavioural treatment. These authors suggest that theovervalued ideation must be assuaged before exposureand response prevention can be conducted. Again, reexamining this apparent treatment failure through acognitive lens may be fruitful in understanding why24overvalued ideation does not respond to behaviouraltreatment.The Wave of Cognitive AnalysesIn response to the introspection and subjectivismof psychoanalytical theory, post-war behaviorism placedlimits around what could be studied within psychology.Observable behaviour was of paramount importance.Mental processes were viewed as being outside of thepsychological realm and thus, inferences were notallowed. Precision of measurement and recording ofbehaviour were hallmarks of this era. Psychologicalproblems were translated into behavioural terms andtreatment consisted of changing behaviour.The late 1960’s marked the beginning of a periodof great change in North American Psychology in whichthe seeds of cognitive psychology were planted. Abalance between the subjectivism and psychoanalyticalanalyses of the pre—war era and the concrete, non—inference post—war behaviorism was greatly needed. Bythe mid-1960’s it was becoming apparent that the strictnonmediational, S—R approach was neither expansive nor25rich enough to account for all of human experience(Breger & McGaugh, 1965; Mahoney, 1974). Severalpioneering studies such as Bandura’s (1965, 1971)classic experiments demonstrating vicarious learningand Mischel’s (Mischel, Ebbesen, & Zeiss, 1972) work ondelay of gratification clearly were not explainableusing traditional S—R behavioural theory. In a reviewof historical and philosophical bases of cognitive-behavioural therapy Dobson and Block (1988) indicateduring this time there were several attempts to expandexisting behavioral models through classifyingcognition as a form of “covert” behaviour (e.g. Homme,1965). These models gave voice to the growingdissatisfaction within behavioural quarters of theexisting theoretical framework.Along side this dissatisfaction with behaviouraltheory was the parallel rejection of psychoanalyticalmodels of psychopathology and its reliance onunconscious, subjective processes and long—termtherapy. Reviews of the efficacy of treatmentapproaches stemming from this tradition were less thanpromising (Eysenck, 1969; Luborsky, Singer & Luborsky,261975; Rachman & Wilson, 1980). In fact, Rachman andWilson (1980) in their extensive review of theliterature conclude that “there still is no acceptableevidence to support the view that psychoanalysis is aneffective treatment” (p. 76)A third impetus for the rise of cognitivebehavioural theory was the recognition that behaviourtherapy targeted only the behaviour of many problemsthat were viewed as multifaceted. For example, thecomplexity of obsessive—compulsive symptomotology wasreduced to examining the compulsive behaviour withouttargeting accompanying obsessions. The compulsivebehaviour was the sole target of therapy. And so,treatment appeared to be successful for the targetproblem but left therapists and researchers aware thatentire problems or major parts of problems were notbeing addressed (Dobson & Block, 1988).The cognitive—behavioural approach asexemplified by the works of Beck (1964; 1967) began tobridge and address this dissatisfaction with existingtheoretical frameworks. At the heart of Beck’s (1964;271967) approach to depression, as well as to allcognitive—behavioural theories, lies three fundamentalprinciples: a) cognition affects behaviour b)cognitions may be monitored and changed c) behaviourchange may be produced through cognitive change (Dobson& Block, 1988). Beck proposed the following cognitiveframework for depressive symptomatology:the cognitive model postulates threespecific concepts to explain thepsychological substrate of depression:1) the cognitive triad, 2) schemas and3) cognitive errors (faulty informationprocessing)... The cognitive triadconsists of three major cognitivepatterns that induce the patient toregard himself, his future, and hisexperience in an idiosyncraticmanner... The concept of schemas is usedto explain why a depressed patientmaintains his pain-inducing and self-defeating attitudes despite objectiveevidence of positive factors in hislife. Relatively stable cognitivepatterns form the basis for theregularity of interpretations of aparticular set of situations. Faultyinformation processing consists ofsystematic errors in the thinking of thedepressed person which maintain thepatient’s belief in the validity of hisnegative concepts, despite the presenceof contradictory evidence (e.g.overgeneralization, magnification,personalization, etc.)(Beck, Rush, Shaw, & Emery, 1979,#pp. 10—li)28This cognitive theory of depression, the firstmajor theory of its kind, arose from clinicalobservation and empirical investigation (Beck, 1964;1967). Within this theoretical framework, depressionis regarded as an affective response to a series ofnegative beliefs about the self. Beck (1967) stressesthat these negative core beliefs are the result ofautomatic thoughts which may be below awareness.During the course of therapy the client is encouragedto monitor and attend to these rapid, discretethoughts. The client and therapist then examine theadaptiveness of these cognitions. Cognitive therapyattempts to modify thoughts that are seen asirrational. Central to this model is the technique of“collaborative empiricism” which involve directingclients to collect evidence and rationally test theirnegative beliefs pertaining to the self, future andenvironment. Support for the efficacy of cognitivetherapy for depression is well documented (e.g.,Blackburn, Bishop, Glen, Whalley, & Christie, 1981;Murphy, Silmons, Wetzel, & Lustman, 1974; Rush, Beck,29Kovacs & Hollon, 1977; Shaw, 1977; Teasdale, Fennell,Hibbert & Aides, 1984).These successes have promoted an extension ofcognitive analyses to a range of psychological problemsincluding anxiety disorders. Yet, a potential weaknessof this model may be that at its core it is untestable.That is, how can we disprove that there are underlyingcognitions (e.g. Coyne & Gotlib, 1983; Seligman, 1988)?At what point can it be acknowledged that a client isexperiencing non—cognitive depression or anxiety?While it is clear that depressed patients doselectively perceive negative information as morethreatening to themselves than do nondepressed patientsit remains unclear whether depressive cognitions alwaysprecede depression. Authors such as Seligman (1988)suggest automatic thoughts may be merely epiphenomena,and that both the thoughts and depression/anxiety havea cause that lies elsewhere (Beidel & Turner, 1985;Hallam, 1985).Until recently, learning theorists postulated thatagoraphobics were frightened of public places, driving,30etc. because these situations had become associatedwith anxiety. For example, Wolpe and Rachman (1960)argued that: “Any neutral stimulus, simple or complex,that happens to make an impact on an individual atabout the time that a fear reaction is evoked, acquiresthe ability to evoke fear subsequently.. .there will begeneralizations of fear reactions to stimuli resemblingthe conditioned stimulus and avoidance of theseplaces/situations will follow” (p.145). This straightconditioning model was readily translated intobehavioural treatment. If one could learn to befearful, through repeated pairings of an unconditionedstimulus with a conditioned stimulus, then one could“unlearn” the association by breaking the connection.Behavioural treatments attempted to do this throughgradually exposing agoraphobics to fearfulsituations/places and asking them to remain there untilthe anxiety lessened. This approach was found to beeffective but often left residual fear (e.g.,Emmelkamp, 1982; Kazdin & Wilson, 1978; Marks, 1987;Rachman & Wilson, 1980; Wilson, 1982).31As of late however, “conditioning theory hasundergone a major revision”, (Dickinson, 1987, P. 57).New approaches to conditioning as exemplified by theworks of Mackintosh (1983) and Rescorla (1980, 1988)suggest that “even the simplest forms ofconditioning. . . involve cognitive processes, thatlearning can occur without reinforcement. . . strictcontiguity between the response or stimulus and thereinforcer is neither necessary nor sufficient forconditioning” (Dickinson, 1987, p. 57—58). These neo—conditioning models see the organism as an active,“information seeker, using logical and perceptualrelations among events.. .to form sophisticatedrepresentation of its world” (Rescorla, 1988, p., 154).The flexibility and scope of conditioning has beenshown to be far greater than what was originallythought (Davey, 1988; Mackintosh, 1983; Rescorla, 1980,1988). Interestingly, Mackintosh (1983) suggests thatthe purpose of conditioning is to allow organisms todetermine the source or cause of events that aresignificant. These cognitive reconceptualizations of32conditioning were soon expanded upon and extended tothe anxiety disorders.By the mid 1980’s cognitive perspectives of theanxiety disorders began to emerge and have deepened ourunderstanding of fear processes. In 1986, David Clarkwhile working with his close collaborator, PaulSalkovskis, introduced a cognitive theory of panic thatsubsumed previous explanations of agoraphobia:panic attacks result from the catastrophicmisinterpretation of certain bodilysensations.The sensations which aremisinterpreted are mainly those which areinvolved in normal anxiety responses (e.g.,palpitations, breathlessness and dizziness)but also include some other sensations.The catastrophic misinterpretation involvesperceiving these sensations as much moredangerous than they really are, and inparticular, interpreting the sensations asindicative of an immediate, impendingdisaster. Examples of catastrophicmisinterpretations would be a healthyindividual perceiving heart palpitations asevidence of an impending heart attack;perceiving a shaky feeling as evidence ofimpending loss of control and insanity...(Clark, 1988, p.73)This fresh approach lead to the emergence of amodel suggesting that people with agoraphobia were not33simply frightened of situations associated with fear(i.e. classical conditioning) rather it was the beliefthat the physical sensations they were experiencingwere indicative of imminent catastrophic danger. Forexample, agoraphobics often report they are frightenedof “going crazy”, “having a heart attack”, or “losingcontrol” in situations ranging from being home alone tofreeways and supermarkets. Several studies have nowdemonstrated that these irrational beliefs are centralto panic and agoraphobia (Clark,l986; Barlow,l987). Itis presumed that by targeting these criticalcognitions, we are pinpointing the key of thisdisorder, a perceived threat to oneself. Irrationalbeliefs tied to imminent threat are thought to drivefear and the ensuing avoidance. Treatment studiestargeting key cognitions have shown this method to beextremely successful (see for example, Barlow, Cohen,Waddell, Vermilyea, Klosko, Blanchard & Di Nardo, 1984;Ost, 1988; Barlow, Craske, Cerny & Klosko, 1989; DeRuiter, Rijken, Garssen & Kraaimaat, 1989; Kiosko,Barlow, Tassinari & Cerny, 1990; for a review seeMargraf, Barlow, Clark & Telch, 1993).34Cognitive analyses of obsessions appeared inthe mid 1970’s shortly after their fruitfulintroduction into depression research. Analysis of thecontent of obsessions by researchers such as Akhtar,Wig, Verma, Pershad, and Verma (1975) and Jenike, Baer,and Minichiello (1986) found consistent themes of dirtand contamination, followed by aggression and violence,religion, and sex. Others have examined the risk—taking pattern of obsessionals noting the reluctance ofthese individuals to engage in any behaviour associatedwith risk (Steiner, 1972). However, the results ofthis investigation have not been replicated. Carr(1974) has suggested that the characteristic sense ofdoubt and indecision associated with this disorder maybe the result of a heightened sense of the probabilityof an unfavorable outcome. Laboratory tasks such ascard—sorting have confirmed that obsessives requiremore information than non—obsessionals before making adecision associated with risk (Beech, 1974; Milner,Beech, & Walker, 1971).35Cognitive-Behavioural Analysis of OCDRecently, Salkovskis (1985) presented aninnovative cognitive—behavioural formulation ofobsessions and compulsions. He begins his analysis bybridging the cognitive framework as proposed by Beck(e.g. Beck, Epstein & Harrison, 1983) and obsessions.He argues that this can best be achieved throughelucidating the relationship between intrusive thoughtsand negative automatic thoughts. Salkovskis draws onthe work of Rachman and Hodgson (1980) in comparingobsessional thoughts and automatic thoughts as proposedby Beck (1976):The major differences between thesenegative automatic thoughts and obsessionsseem to lie in the perceived intrusiveness,immediate accessibility to consciousness andthe extent to which they are seen as beingconsistent with the individual’s beliefsystem. This last differences isparticularly important, insofar as Beck’sview of cognitions producing affectivedisturbance rests on their perceivedrealistic and plausible nature, and theiracceptance by the individual experiencingthem. By contrast, obsessions areunacceptable, irrational and implausible.Obsessions are incongruent with theindividual’s belief system, unlike negativeautomatic thoughts which are an expression ofit (Salkovskis, 1985, p. 573).36He then argues that obsessions may serve to elicitautomatic thoughts and that persistent intrusivethoughts will occur only if there is some type ofnegative evaluation of the obsession. He states thatmood disturbances and discomfort over havingexperienced intrusive thoughts will only occur if theindividual attaches extreme, averse personalimplications to having the thought. He also speaks of apredisposing factor of depression which can increasethe range of triggering stimuli, and personalsignificance given to intrusive thoughts. He viewsobsessions as intrusive thoughts that the clientinterprets as being a sign that they will beresponsible for harm to themselves or others. In orderto prevent the harm from occurring, the client engagesin behaviour attempting to suppress or neutralize theintrusive thoughts. Thus, it is not simply theoccurrence of an intrusive thought or the negativeassociation with having the thought but rather, theresponsibility that arises from having the thought thatis the critical feature:That is, obsession—provoked automaticthoughts or images revolve around personal37responsibility, the possibility that ifthings go wrong it might well be the persons’own fault. Such responsibility may beindirect as well as direct, so that thepossibility of preventing harm caused byexternal agents is equally potent. Clearly,such ideas of responsibility would lead toself—condemnation in vulnerable individualsto the extent that such responsibility . . isabhorrent to them. Such ideas ofresponsibility can extend to having had thethought itself; that is, if the personbelieves that they are responsible for theirown thoughts.. .the content of which isabhorrent to them, then they presumablyregard themselves as being responsible forbeing a bad or evil person unless they takesteps to ensure their blamelessness. Theaffective disturbance usually described asarising from the obsession or intrusionactually arises from such automatic thoughtsabout the intrusion rather than from theintrusion itself (Salkovskis, 1985, p. 574).According to Salkovskis, once responsibility istaken for the potential occurrence of harm, then theensuing attempt to prevent harm is almost inevitable.The means of preventing harm is some form ofneutralization, either compulsive behaviour orcognitive rituals. In terms of treatment implications,Salkovskis (1985) proposes that therapy should“concentrate not on modifications of intrusions. . .buton automatic thoughts consequent on the intrusions, andon the beliefs which give rise to these” (p.#581).38Empirical evidence for this model comes from foursources which primarily pertain to obsessions andcognitive neutralizing rather than compulsivebehaviour. In addition, the work conducted comesalmost exclusively from non—clinical samples.Dickerson and colleagues examined the relationshipbetween the salience of intrusive thoughts andaversiveness in a normal population. England andDickerson (1988) found that the magnitude ofuncontrollability of an intrusive thought was linked tothe salience of the thought rather than the content ofthe thought. Additional evidence comes from Edwardsand Dickerson (l987a) who found that intrusive thoughtsare more salient than neutral thoughts as measured bythe length of time required to replace the thought.These authors suggest that these data are consistentwith the notion that the reaction or response to theintrusive thought is more important than the content ofthe thought.The second area of research provides a start forexamination of the potential links betweenresponsibility and compulsive behaviour. Salkovskis39(1989) reports a study by Salkovskis and Dent (1989) asindicating evidence for a relationship betweenresponsibility and compulsive behaviour. In thissurvey study of 243 nonclinical participants,individuals were asked to rate their beliefs concerningresponsibility for harm, threat and loss. Theseauthors found that individuals who reported compulsivebehaviour scored higher on the Maudsley Obsessional—Compulsive Inventory (MOC-I)(Hodgson & Rachman, 1978)than those who did not perform compulsive behaviour.Salkovskis (1989) indicates that this result providessupport for a link between responsibility andcompulsive behaviour. The finding that those whoreported compulsive behaviour scored higher on the MOCIis not surprising given that two of the four sub—scalesmeasure precisely this phenomena. However, the resultdoes little in elucidating the relationship betweenresponsibility and compulsive behaviour. Salkovskisand Dent’s (1989) data showing that subjects whoperformed compulsive behaviour had higher scores onbelief ratings of responsibility for harm but not forthreat or loss is consistent with responsibility being40a critical component for compulsive behaviour. Afurther analysis separating those who demonstratedcompulsive Checking from those who compulsively washedwould be of benefit in determining if responsibility isa critical component for all compulsive behaviour or islimited to a sub-group.A third line of evidence in support of Salkovskis(1985) theory examined the relationship betweenneutralizing activities and discomfort and persistenceof intrusive thoughts. Wegner, Schneider, Carter andWhite (1987) asked non-clinical subjects to think ofanything except for a white bear and then measured theoccurrence and contents of thoughts. They found thatthese instructions actually increased the frequency ofthoughts surrounding white bears in comparison to acontrol group. Using a similar line of thought,Salkovskis, Westbrook, Davis, Jeavons and Gledhill(1989) had non-clinical subjects who experienced bothintrusive thoughts and cognitive neutralizing behaviorrecord one intrusive thought on a loop tape. Half ofthe subjects were instructed to use their usualneutralizing thought and the other half were instructed41to perform a distraction task. Thoughts that wereneutralized resulted in greater discomfort on Time 2 incomparison to the distraction group. Salkovskis (1989)interprets this finding as support for the premisethat neutralizing behaviour increases discomfortassociated with intrusive thoughts. It may be moreprecise to say that this experiment provides evidencein support of cognitive neutralizing leading toincreased discomfort of intrusive thoughts. Furtherresearch is necessary to determine whether other typesof rituals (e.g.checking and washing) produce similarincreases in discomfort.The final evidence Salkovskis (1989) cites insupport of his model comes from treatment outcomestudies. Emmelkamp, Visser and Hoekstra (1988) andEmmelkamp (1988) have both shown that cognitiveinterventions attempting to alter the manner in whichpatients interpret intrusive thoughts were as effectiveas exposure plus response prevention.Overall, these studies provide early support forthe Salkovskis (1985) formulation of OCD. However, the42focus of the research to date and ensuing support forthe theory is primarily from studies examiningintrusive thoughts and cognitive rituals. Salkovskis(1985) asserts that the onset of all behaviouralrituals involve cognitive mediation but not necessarilythe maintenance. Little, and then only passing mentionhas been given to compulsive behaviour. Salkovskisargues a one to one relationship between heightenedresponsibility and neutralizing activity . It must beassumed then that he views all cognitive rituals asbeing cognitively mediated. He does not separatecognitive rituals from behavioural rituals, andtherefore one must also conclude that all behaviouralrituals are also cognitively mediated. Clinicalintuition suggests otherwise. That is, even in thedevelopment of the disorder, people describe instanceswhere they wash or check without being aware of anyobsessional thought. Salkovskis’ assertion thatrituals involve cognitive mediation is readilytestable. One could either change the intrusivecognition and then test to see if the ritual hasdisappeared or change the ritual and see if the43intrusive cognition is no longer troublesome. Futureresearch is necessary to examine if in the early stagesof the disorder behavioural rituals have the same oneto one relationship, between obsessions and rituals, asSalkovskis proposes.Second, a comment pertaining to the theory as awhole is warranted. Salkovskis’ model of obsessionsand compulsions lacks specificity. Fleshing out theskeletal draft is greatly needed. For example, asdiscussed above, are all forms of compulsions equallyinfluenced by an exaggerated sense of responsibility?Salkovskis’ model does not account for any variationfrom the direct link between obsessions, responsibilityand compulsions proposed in the currentconceptualization. Salkovskis (1985) posits thatneutralising is invariably present in OCD. Thishypothesis awaits empirical validation. The model mayneed to consider cases where obsessions withoutneutralizing occur, or instances where a less thanperfect relationship between obsessions and compulsionsis present, as well as determining the potentiallyunique sequela of the various types of rituals.44Third, it may be of benefit to clarify the roleof triggering stimuli. At this point it appears to besomewhat of a wildcard that does not allow for specificpredictions. That is, what factors determine whetherthe stimuli will trigger the intrusive thought? In thecurrent model if a specific stimulus does not producean intrusion and subsequent negative evaluation it isdefined as non-salient. Similarly, if the stimuli doesproduce the intrusive thought it is considered salient.This type of analysis is circular and adds littleclarity. It may be of benefit to examine therelationship between the triggering stimuli andresponsibility and derive specific predictions basedupon this relationship. For example, how do varyinglevels of responsibility influence the saliency oftriggeiing of stimuli? What is the relationshipbetween varying levels of responsibility triggeringstimuli and depressive affect? More generally, what isthe relationship between responsibility and depression?Exploring the effects of variations in responsibilityon the critical components in the model such asintrusive thoughts, triggering stimuli, and specific45compulsions may begin to provide the necessary detailto bring a sharper focus to this model.Fourth, the role of depression in this modelremains unclear. Salkovskis (1985) discussesdepression as serving two roles and being of importanceat various stages in the model. First, he argues thatdepression can be conceptualized as a predisposingfactor which may influence processing at a variety oflevels. Second, he views depression arising as aresult of increased responsibility and the negativeevaluation stemming from it. Mood acting as amodulating influence is certainly intriguing but givesus little predictive power unless we can begin todetermine what mechanisms are responsible fortriggering the presence or absence of depressed affect.Reynolds and Salkovskis (1992) began to address thisissue through comparing positive and negative intrusiveand their effects on mood. Induced happy and sad moodsappeared to differentially affect intrusive thoughts.Future research may need to examine differences betweennegative automatic thoughts and intrusive thoughts andmechanisms responsible for each type of thought. In46addition, it may be interesting to examine what factorscontribute to depression and associated negativeautomatic thoughts being present at various stages andwhat factors influence it.Despite these potential limitations, Salkovskis’model is a highly testable and intriguing one whichbegins to consider the complexity of this disorder.Examining how variations in responsibility affectobsessions and compulsions is greatly needed to givethis conceptualization greater specificity andtestability. In addition, further discussion of theinfluence of mood and its relationship to levels ofresponsibility would be of benefit.As is evident in accounts such as Salkovskis’(1985), research on obsessions has been primarilycognitive in focus. In contrast, the research oncompulsive checking has been almost exclusivelybehavioural.Investigators have attempted to delineatetypes of compulsions associated with ObsessiveCompulsive Disorder. For example, Hodgson and Rachman47(1977) factor analysed questionnaire data and found twomain forms of compulsions: checking and cleaning.Overall, 52% of their patients had checking rituals and48% had cleaning rituals. Others such as Stern and Cobb(1978) found similar percentages of cleaners andcheckers in their sample and also report that 40% oftheir patients had repeating rituals (doing things bynumber). Jenike, Baer, and Minichiello (1986) foundthat 79% of their patients had checking rituals, 21%had counting rituals and 58% had cleaning rituals.Overall, these investigations point to two main typesof compulsions: washing and checking.Rachman (1974) identified a major differencebetween these two types of rituals. He noted thatcompulsive washing occurs after the person is exposedto some type of perceived contaminant, germ or dirt.The cleaning is restorative and serves to reestablish asense of safety and control. On the other hand,checking functions to ward off some future harm orimperfection from occurring. Checking rituals arepreventative. Steketee, Grayson and Foa (1985) failedto find any further differences between washers and48checkers in a detailed examination. However, as is thecase with any negative result, it is unclear whetherthere are no true differences or if the correct probewas not used.Responsibility as a Major Determinant of OCDRachman and Hodgson (1980) noted the difficultiesinvolved in attempting to provoke the urge to check ina group of subjects diagnosed as having compulsivechecking problems. Unlike cleaning compulsions, whichcan be readily provoked through exposure to acontaminant in the experimenter’s laboratory/departmentstore/office etc., checking rituals were difficult orimpossible to initiate and produced mild or no anxietyin the laboratory. They suggested that if the study isconducted in the experimenter’s laboratory, compulsivecheckers feel that the experimenter is responsible forany negative outcome that may result from not checking.With this lessening of responsibility, the checkerfeels little discomfort when asked to exposehim/herself to situations in which she/he wouldnormally check. Rachman and Hodgson (1980) observed49that “if the obsessional subject is divested wholly orpartly of responsibility for the act, he or sheexperiences little discomfort”. These authors alsonoted that subjects felt most anxious and worried whenthey experienced the urge to check in the absence ofthe experimenter. It was suggested that “the presenceof another person may serve to reduce the obsessional’ssense of responsibility for the act and hence, allowthem to experience less discomfort when someone else ispresent.” (Rachman & Hodgson, p.181) Clinicalimpression as well as documented reports (Rachman &Hodgson, p. 1979) note the increase in anxiety, urge tocheck, and worry when the checker is asked to resistthe urge to check when the therapist is absent.Obsessionals reported that it is particularly difficultto resist the urge to check in their own homes whilealone. This finding together with earlier experimentalanalyses of compulsive checking confirming that a majortheme in compulsive checking is a fear of ha (RachmanHodgson, 1980) has led to the following cognitiveanalysis.50The central thesis is that compulsive checkingstems from a combination of an inflated fear of harmplus an exaggerated sense of personal responsibility.Related to this position is Salkovskis’ (1985)view that obsessions involve an “inflated belief in theprobability of being the cause of serious harm toothers or self or failing to avert harm”. In thisanalysis, Salkovskis distinguishes between an increasedbelief in the probability of harm and an increasedprobability of being the cause of harm. He emphasizesthe role of an inflated probability of being the causeof harm, or an increased sense of personalresponsibility over harm in obsessional problems.Attribution ResearchMost of the research on the maladaptive nature ofresponsibility comes from attributional researchexamining the role of self—blame in adjustment afterserious, unanticipated negative events (e.g. Abramson,Seligman, & Teasdale, 1978; Janoff—Bulman, 1979; Lerner& Miller, 1978; Shaver, 1985; Taylor, 1983; Walster,1966; Wortman, 1976) These events are thought to call51into question the notion that people are motivated tobelieve that the world is controllable and predictable(Heider, 1958; Kelley, 1971). When a traumatic eventoccurs it challenges this belief and leads to thequestion of “why?” and leads to attributingresponsibility for such events to oneself or others(Kelley, 1971; Weiner, 1985).Typically, self-blame has been assumed to bemaladaptive (e.g. Becker, Skinner, Abel, Howell, &Bruce, 1982; Af fleck, Allen, McGrade, & McQueeney,1982; M.A. Graham, Thompson, Estrada, & Yonekura, 1987:Kiecolt—Glaser & Williams, 1987; Meyer & Taylor, 1986;Moulton, Sweet, Temoshok, & Mandel, 1987). Thetheoretical position driving these studies and others,argues that self-blame is maladaptive because itundermines self—esteem, triggers feelings ofhelplessness and thus increases the likelihood ofdepression (Abramson, Seligman, & Teasdale, 1978;Peterson & Seligman, 1984).However, others have found that self—blame isadaptive. This theoretical position argues that self-52blame is adaptive because it defends against randomnessand uncontrollability and contributes to a sense ofsafety through viewing the world as being predictableand controllable (e.g. Bulman & Wortman, 1977; Lerner &Miller, 1978; Shaver, 1985; Tennen, Affleck, Allen—McGrade, & Ratzan, 1984; Tennen, Affleck, & Gershman,1986; Timko & Janoff—Bulman, 19857; Wortman, 1976).Yet a third group of investigators have found norelation between self—blame and adjustment (e.g. Miller& Porter, 1983; Silver, 1982; Taylor, Lichtman & Wood,1984; Witenberg, Blanchard, Suls, Tennen, Mccoy, &McGoldrick, 1983). Janoff-Bulman (1979) attempted tobridge these apparent inconsistencies by proposing thattwo types of self—blame are present, behavioural andcharacterological. Behavioural self-blame is thoughtto be adaptive if the victim directs the blame atcontrollable, specific behaviours. This type of blameis thought to be related to better adjustment becauseit may allow the individual to hang on to the beliefthat the world is a predictable, controllable, safeplace and thus, protect oneself from future negativeevents. On the other hand, characterological self—53blame is said to be maladaptive because it involvesattributions to the self which are stable anduncontrollable and thus, does not provide the samesense of control. Unfortunately, empirical support forthis differentiation is equivocal (see Turnquist,Harvey & Andersen, 1988). Overall, the mixed set ofresults has led several authors to conclude thatdespite the voluminous number of studies conducted inthis area over the past 15 years, little has beenfirmly established about the role of self—blame inadjustment (see Michela & Wood, 1986; Turnquist et al.,1988).After reviewing the work in this area, it appearedlimited in use for the focus of this research forseveral reasons: a) attributional theory generallyfocuses on how individuals assign responsibility afterthe occurrence of an event, b) the work focuses onunexpected events, and c) the negative event is a realoccurrence as opposed to obsessional’s characteristicthoughts pertaining to imagined future harm/danger.When examining the dynamics (specifics) of compulsivebehaviour, it is important to consider we are examining54how an individual assigns responsibility prior to anevent. Second, compulsive checkers overpredict theprobability of the occurrence of negative events. Theyare constantly vigilant, and by checking attempt toprevent the negative event. This vigilance and threatof danger is very different from the phenomena studiedin attribution research examining how people cope aftera single, unexpected event. Third, compulsive checkersare trying to prevent future, imagined danger or harm.The participants in the attributional research areindividuals who have experienced a serious, negativeevent. Fourth, this study proposes that compulsivecheckers have a stable, exaggerated sense ofresponsibility. This heightened sense ofresponsibility differs from participants inattributional research who may feel responsible for anoutcome in a limited instance but do not maintain aconsistent, stable state of heightened responsibilityfor all negative events. Although attributionalresearch is not central to the development of specifichypotheses and predictions of this study for thereasons discussed above, it may be of benefit when55exploring the origins of a heightened sense ofresponsibility and will be discussed at that point.Fear of Criticism and OCDPsychodynamic and learning theorists alike havesuggested that rituals are a means of preventingcriticism (e.g. Cameron, 1947; Dowson, 1977; Rosen,1975). Early behaviourists such as Dollard and Miller(1950) postulated that if a child is frequentlycriticized or punished for being unclean or having notdone a task correctly, the act of washing or checkingmay produce a reassuring effect. That is, thewashing/checking is associated with prevention ofcriticism and is therefore anxiety reducing.Other researchers, such as McFall and Wollersheim(1979), have attempted to delineate some of the beliefsassociated with OCD. They indicate that two of theprimary faulty cognitions are:1. One should be perfectly competent, adequateand high-achieving in all possible respects in order to56be worthwhile and to avoid criticism or disapproval byothers/oneself;2. Making mistakes or failing to live up to one’sperfectionistic ideals should result in punishment orcondemnation.Turner, Steketee, and Foa (1979) attempted toexamine this hypothesis through administering a fear ofcriticism questionnaire to a group of obsessives and acontrol group of simple phobics. They found thatobsessives were more fearful of criticism than thecontrol group of simple phobics. Thyer, Curtis, andFechner (1984) attempted to replicate these findingswith social phobics and agoraphobics serving as controlgroups, but failed to confirm the Turner et al. (1979)findings, suggesting sensitivity to criticism may notbe specific to OCD.However, another possible interpretation of thenegative results may be that the measures of criticismor perfectionism were not tailored to the specificbeliefs of harm associated with this disorder.57Beliefs of HarmAlthough not systematically evaluated, severalauthors (e.g. Rachman & Hodgson, 1980; Steketee & Foa,1980) have noted that compulsives fear negativeconsequences will occur if the urge to check/clean isnot carried out. Most washers have strong expectationsof disease, death or physical impairment of themselvesor their spouse/children. Compulsive checkers oftenfear that they will be responsible for either: 1)something bad happening (e.g., fire, injury, as aresult of leaving the iron on) or 2) something notbeing perfect; not okay (e.g., something is not rightif the picture is crooked and I may be criticized).Foa (1979) also made reference to the vagueness of somewashers’ and checkers’ anticipated harm. Herimpression is that some fear that the“anxiety/discomfort will last forever” and will lead to“going crazy” or “losing control”. If Foa’s impressioncan be confirmed experimentally, these types ofirrational thoughts can be readily translated into the“collaborative hypothesis testing” framework ofcognitive therapy. At this point, however, the58specific beliefs of harm held by people with OCD havenot been studied empirically and have received onlypassing mention.Responsibility, Controllability and OCDThere is an interesting paradox between the extentof responsibility assumed by compulsive checkers andthe level of perceived controllability over anticipateddanger or harm. Compulsive checkers may accept or takeon high levels of responsibility for anticipated harmeven in the absence of control. It is difficult toconceive of being responsible for any event if one isunable to exert any control over the outcome. Yet,compulsive checkers appear to do exactly this. Forexample, H.G. who repeatedly checks window locks in herhome, will indicate she is highly responsible ifsomeone breaks into her home, but that she has nocontrol over whether or not she will be burglarized.Another example is of C.R. who checks electricalappliances repeatedly fearing there may be anelectrical short—circuit. He indicates he would feelhighly responsible if the short was to occur, yet also59says he has little control over whether the circuitwill short.These examples point to the double bind of aperceived loss of control over an external threatcoupled with high sense of responsibility in compulsivecheckers. Given the frequency with which those withanxiety disorders, other than OCD, complain of a lossof control or insufficient control over anticipateddanger or harm, it may be timely to investigate itsimportance in OCD.Prominent anxiety investigators such as Mineka andKihlstrom (1978) and Seligman (1975) point to thecentrality of a loss of control in the development andmaintenance of anxiety disorders. Human and animalstudies alike, suggest control is intimately linkedwith anxiety through perceptions of lack of controlover negative events (Geer, Davison & Gatchel, 1970;Neale & Katahn, 1968; Pervin 1963; Staub, Tursky, &Schwartz, 1971: see also Miller, 1979, for a review).In a classic study on the role of control andresponsibility, Rodin and Langer (1977) had one group60of nursing home residents take responsibility andcontrol for the arrangements of items in their room, aswell as for their time. This group evidenced lowermortality rates, 18 months later, than a control group.Langer (1979) as well as Sanderson, Rapee and Barlow(1989) suggest it is not control in itself that iscritical but rather the illusion of control.The influential work of Seligman and colleagues(1968; 1975) on helplessness and uncontrollabilitydemonstrated that animals who have a history ofunpredictable, uncontrollable events are more apt toexperience emotional disturbances. The typical patternfor animals with this history is to appear extremelyanxious and then later to become severely depressed.The parallels between this sequence in animalsand the high rates of co-morbidity of OCD anddepression (Foa, 1979) readily tie in with the presentanalysis of OCD.High levels of responsibility and an exaggeratedsense of harm, combined with little controllabilityover external threat may result in the characteristic61pattern of compulsive checking. It is predicted thatfor people with OCD, the sense of responsibility foranticipated harm is independent of a sense of controlover the pertinent threat.Immediacy of Threat in OCDPanic is not limited to those suffering from PanicDisorder. Rather, panic is experienced across theanxiety disorders and is commonly reported in thegeneral public (Barlow, 1988; Norton, Harrison, Hauch,& Rhodes, 1985). In the current investigation we areinterested in determining whether and when perceivedpanic occurs. Central to perceived panic is theimmediacy of the presumed threat or danger. In panicdisorder, imminent threat is produced by catastrophicmisinterpretations of bodily sensations (Clark, 1986).It may be that any stimulus which produces imminentthreat will lead to panic. In contrast, if no imminentthreat is present, there should be no panic. It ishypothesized that in OCD, the immediacy of threat isvariable. For compulsive checkers, some future dangeror harm is prevented through ritualized checking;62rituals are preventative (Rachman & Hodgson, 1980).Therefore, for those who check repeatedly threat isfuturistic. Differential predictions can be made withrespect to frequency of perceived panic based uponimmediacy of perceived danger.It is predicted that the more immediate the threatthe greater the likelihood of perceived panic.BaseratesWe have little information about many of thecognitive processes associated with OCD. For example,we do not know the baserates for sensitivity tocriticism, probability for negative events, or thebaseline level of control they experience when dealingwith fear related events. This study attempts togather such information using a within subject designso that each subject may serve as her/his control. Thedetails of the design will be described under the“method” section.63SummaryCognitive analyses of panic have proven to befruitful and have improved our understanding of thenature and treatment of this disorder. The extensionof cognitive analyses to other anxiety disorders hasbegun, but to date there is an absence of informationabout the role of cognition in compulsive checking.This gap in the cognitive analysis of OCD, and theexperimental observations of the role of responsibilityin this disorder, are the impetus for the proposedresearch. The central argument presented is thatcompulsive checking stems from a combination of aninflated fear of harm plus a heightened sense ofresponsibility. The primary purpose of the proposedresearch is to test the hypothesis that responsibilityis a major determinant of OCD. Examination of thecombination of inflated fear and heightenedresponsibility will await future research.64Hypotheses and PredictionsHypothesis 1There is a causal relationship between perceivedresponsibility and compulsive checking.Prediction la. An increase in perceived responsibilitywill be followed by an increase in perceiveddiscomfort.Prediction lb. An increase in perceived responsibilitywill be followed by an increase in the urge to completea check.Prediction ic. An increase in perceived responsibilitywill be followed by an increase in the estimation oflength of time needed to complete a check.Prediction ld. A decrease in perceived responsibilitywill be followed by a decrease in perceived discomfort.65Prediction le. A decrease in perceived responsibilitywill be followed by a decrease in the urge to check.Prediction if. A decrease in perceived responsibilitywill be followed by a decrease in the length of timeneeded to check.Hypothesis 2Estimations of the timing of anticipated threatinfluence the probability of perceived panic.Prediction 2a. The more immediate the anticipatedthreat, the greater the probability of perceived panic.Prediction 2b. The less immediate the anticipatedthreat, the lower the probability of perceived panic.Hypothesis 3Variations in perceived responsibility do notalter estimations of the probability of anticipatedharm.66Prediction 3a. An increase in perceived responsibilitywill be followed by no change in the probability ofanticipated harm.Prediction 3b. A decrease in perceived responsibilitywill be followed by no change in the probability ofanticipated harm.Hypothesis 4Variations in perceived responsibility do notalter estimations of the seriousness of anticipatedharm.Prediction 4a. An increase in perceived responsibilitywill be followed by no change in the estimatedseriousness of anticipated harm.Prediction 4b. A decrease in perceived responsibilitywill be followed by no change in the estimatedseriousness of anticipated harm.67Hypothesis 5Variations in perceived responsibility do notalter estimations of when the anticipated harm willoccur.Prediction 5a. An increase in perceived responsibilitywill be followed by no change in when anticipated harmwill occur.Prediction 5b. A decrease in perceived responsibilitywill be followed by no change in when anticipated harmwill occur.Hypothesis 6Perceived responsibility for anticipated harm isindependent of a sense of control over the pertinentanticipated threat.Prediction 6a. An increase in perceived responsibilitywill not alter estimations of amount ofcontrollability.68Prediction 6b. A decrease in perceived responsibilitywill not influence estimations of amount ofcontrollability.Hypothesis 7Variations in perceived responsibility will alterestimations of the likelihood of anticipated criticism.Prediction 7a. An increase in perceived responsibilitywill be followed by an increase in the estimatedlikelihood of anticipated criticism.Prediction 7b. A decrease in perceived responsibilitywill be followed by a decrease in the estimatedlikelihood of anticipated criticism.Hypothesis 8Variations in perceived responsibility will alterestimations of the severity of anticipated criticism.Prediction 8a. An increase in perceived responsibilitywill be followed by an increase in the estimatedseverity of anticipated criticism.69Prediction 8b. A decrease in perceived responsibilitywill be followed by a decrease in the estimatedseverity of anticipated criticism.Hypothesis 9Variations in perceived responsibility will alterestimations of the timing of anticipated criticism.Prediction 9a. An increase in perceived responsibilitywill be followed by a decrease in the estimated lengthof time before anticipated criticism.Prediction 9b. A decrease in perceived responsibilitywill be followed by an increase in the estimated lengthof time before anticipated criticism.Hypothesis 10Perceived responsibility for anticipated harm isindependent of a sense of control over anticipatedcriticism.70Prediction ba. An increase in perceivedresponsibility will not alter estimations of amount ofcontrollability over anticipated criticism.Prediction lob. A decrease in perceived responsibilitywill not influence estimations of amount ofcontrollability over anticipated criticism.71METHODExperimental Design and OverviewIn order to test the central thesis of the currentstudy, that compulsive checking stems from anexaggerated sense of personal responsibility, it wasnecessary to examine the influence of levels ofperceived responsibility on critical dependentmeasures. Hence, the current study is a within—subjectrepeated measures design with levels of perceivedresponsibility (high, control, low) as the independentvariable. Given the difficulties in collecting thisinaccessible sample, a within—subject design was chosento maximize power. One of the major drawbacksassociated with this type of design is carryovereffects. After preliminary testing of the manipulationit seemed unlikely that such effects would be presentand thus, the within subject design was chosen.The effects of variations in perceivedresponsibility were examined on several dependentmeasures. In all cases dependent measures are based on72the client’s perception. For ease of expression, inthe remainder of this thesis I will refer to thedependent measures without adding the term “perceived”.However, in all cases the dependent measures are basedupon the client’s perception. Thus, the shorthand of“responsibility” refers to perceived responsibility,“criticism” refers to perceived criticism and so on.The dependent measures for the current study are: urgeto check, perceived discomfort, responsibility(manipulation check), likelihood of anticipated threat,severity of anticipated threat, control overanticipated threat, timing of anticipated threat,estimation of the time needed to check, likelihood ofanticipated criticism, severity of anticipatedcriticism, control over anticipated criticism, and thetiming of anticipated criticism.Subjects received each of the followingconditions: 1) high responsibility 2) lowresponsibility 3) control—check 4) control-clean. Thecontrol—clean condition was designed to address aseparate series of hypotheses than those discussedherein. It is described in this section, and will be73analysed later, only in terms of its impact (if any) onthe three remaining conditions. It will be referred toas a secondary study.All participants were first assigned to one ofthe two control conditions (control—check, control—clean). The order of the control conditions wascounterbalanced. In both control conditions, there wasno manipulation of the sense of perceivedresponsibility. The control conditions differ only inthat in the control-check condition, the urge to checkis provoked whereas in the control—clean condition theurge to clean is provoked.After completion of the control conditionssubjects were assigned to the two variations inresponsibility conditions. Order of the perceivedresponsibility conditions was counterbalanced. In thehigh responsibility condition, the subjects’ perceivedresponsibility for a personally relevant negative eventwas deliberately increased. In the low responsibilitycondition, the subjects’ perceived responsibility for a74personally relevant negative event was transferred tothe experimenter.Subjects were assessed before and after eachexperimental manipulation on all dependent measures.The pre—assessment consisted of a behavioural approachtest where the urge to check/clean was provoked andseveral self—report measures of anxiety. The post—assessment was identical to the pre—assessment.Sublects. The subjects were 30 volunteers who checkedcompulsively. The secondary study was comprised of 10volunteers who cleaned compulsively. Participation inthe study was restricted to individuals who: 1) wereover 18 years old, 2) met DSM-IIIR criteria forobsessive compulsive disorder, and 3) had no history ofpsychosis.Further inclusion criteria required thatparticipants report a minimum score of 70/100 (on a 0to 100 verbal analogue scale where “0 is no discomfortand “100” is extreme discomfort) when the urge to checkwas provoked.75In order to be considered appropriate for thesecondary study additional participants were requiredto report a minimum score of 70/100 on the subjectiveunits of discomfort scale when the urge to clean wasprovoked.ProcedureSublect Recruitment. Subjects were recruited asfollows:1. Therapists at the Psychology Clinic, U.B.C.and Health Psychology Clinic, U.B.C asked clients ifthey were interested in participating in the study.Those clients who expressed a willingness toparticipate were asked to complete a consent form whichallowed the investigators to contact them by telephone.2. The investigator advertised in newspapers, andparticipated in interviews on radio and television.3. Notices were posted on bulletin boards inhospitals, mental health clinics and “self-help”bookstores.76Initial Screening Many of the potential subjectsfirst made contact by telephone. During this contact,a brief interview was conducted to screen outindividuals who were inappropriate for the study.Screening criteria were as follows:1. Callers were required to answer affirmativelyto the first question on the obsessive compulsivesubsection of the Anxiety Disorders Interview Schedule- Revised (ADIS-R) (DiNardo, Barlow, Cerny, Vermilyea,Vermilyea, Himadi, & Waddell,l985) The specificquestion was: “Have you had to repeat some act over andover again which doesn’t seem to make sense and thatyou don’t want to do? For example, washing somethingover and over again or counting things, or checkingsomething repeatedly like locked doors, importantpapers or retracing driving routes?”2. In addition to an affirmative answer on theabove question, callers were required to meet either ofthe two following criteria:77a) wash or check more than 60 minutes per day, orb) indicate that their checking/cleaningdifficulties significantly interfere with their life.The following information was given to potentialsubjects over the telephone:We certainly appreciate your calling today, butbefore we give you the details of our study, we wouldlike to ask you a few questions. Is that Okay withyou?(Administer telephone interview)It sounds like you have been having some problemswith checking/cleaning, and because of this, we wouldbe interested in finding out more about yourexperience. If you agree to be in our study, we wouldask you to take part in a short interview. At thattime we would ask you more about your pattern ofchecking/cleaning. In addition, we would ask you to gointo four situations where you would usuallycheck/clean. After you go into the situation, we wouldask you some questions about your thoughts and feelingswhile you were checking/cleaning. Finally, we wouldask you a few more questions about your experience.It is our experience that people’s reactions tochecking/cleaning are different at home than when theyare away, so we would conduct the study in your home,if you agree. The study takes roughly four hours andwe would be giving you a information package onchecking and cleaning problems, which includes some78treatment alternatives. We are very interested in yourthoughts and feelings about checking/cleaning and wouldreally appreciate your participation. Our long termgoals of this research are to help others, likeyourself, by developing an effective treatment program.The more we can find out about your experience withchecking/cleaning, the better we can be at puttingtogether a treatment program that works.Initial Evaluation The purpose of the initialevaluation was to determine if potential subjects meetDSM-IIIR criteria for OCD. The evaluation andassessments were conducted in the subject’s home. Inorder to achieve the above objective subjects wereevaluated using the Obsessive Compulsive Disordersection of a structured interview designed to assessanxiety disorders (ADIS-R; DiNardo, et al. 1983). Inaddition, they were asked to complete the MaudsleyObsessional Compulsive Inventory (MOCI), a scale whichmeasures obsessive compulsive symptoms (Hodgson &Rachman, 1977).Experimental Procedure If subjects met DSM-IIIRcriteria and were within clinical range on the MOCIthey then received the four conditions. Each conditionconsisted of a pre—assessment, experimental79manipulation and a post—assessment. The pre—assessmentand post—assessment were identical and involved abehavioural approach test (BAT) and completion of allself—report measures of anxiety. The BAT was anindividually tailored approach test which involvedexposing the subject to an object/situation where s/heusually checks/cleans. Subjects were assessed afterthis exposure using self—report measures. It isimportant to note that there were no invivo basedbehavioural observations but rather that theparticipants were asked to complete several self—reportmeasures after the urge to check was provoked in theBAT. The self—report measures were as follows: urgeto check, perceived discomfort, perceivedresponsibility (manipulation check), likelihood ofanticipated threat, severity of anticipated threat,control over anticipated threat, the timing ofanticipated threat, estimation of the time needed tocheck, likelihood of anticipated criticism, severity ofanticipated criticism, control over anticipatedcriticism, the timing of anticipated criticism,80negative cogriitions and bodily sensationsquestionnaires.A research assistant “blind” to the order of theexperimental conditions and major hypotheses conductedthe pre and post assessments.Instructions for BAT Task Determination Theexperimenter determined which tasks were used for theBAT’s by asking the following questions:“You mentioned, that you check/cleanover and over again, if I was to ask you to___________at this moment, and notcheck/clean at all, how much discomfort doyou think you would experience? I’d like youto use a scale where “0” is “no discomfort”and “100” is “extreme discomfort”.“If I was to ask you to___________atthis moment, and not check at all, andthen something bad happened or something was not perfect, how responsible wouldyou feel? I’d like you to use a scalewhere “0” is “not at all responsible” and“100” is “completely responsible”.The interviewer used trial and error until thesubject reported a score of “70” on the perceived81discomfort scale and “50” on perceived responsibilitywith three separate tasks.All the tasks chosen for the BAT’s were randomlyassigned to condition.Instructions Prior to First BAT “In a moment we willask you to go into four situations where you previouslyhave checked/cleaned. We will ask you to go into eachsituation twice. Each time you go into the situation,we will ask you not to check/clean and then we will askyou some questions about your experience. There are noright or wrong answers. Each question is designed tobe answered quickly and is not meant to be thoughtabout too much.”After completion of the pre-test and all dependentmeasures, the experimental manipulation was introduced.Depending on their assignment to condition, subjectswere encouraged to increase, decrease or not changetheir perceived responsibility for an anticipatednegative event. The specific instructions given foreach condition was as follows.Low Responsibility Condition “You mentioned that youcheck_____over and over again. In a moment we willask you to go into a situation where you usually wouldcheck . People who check sometimes worry thatsomething bad will happen or what they do won’t be82perfect. They sometimes worry that damage of some kindmay occur, that someone will be hurt, or that they willhave to undo any damage or harm. People who have theseworries sometimes feel that something bad will happenor is not perfect when they are in situations like theone you are about to be in. Except this time, I wantyou to know that I will take complete responsibility ifanything bad happens or anything is not perfect. Youare i2 responsible for anything that happens or is notperfect.I will take on complete responsibility. I willbe to blame if anything bad happens or is not perfect.Your name will not be mentioned. I will also takeresponsibility for any damage that may occur and willpay back or undo any damage or harm that may occur.I would like to put this agreement in writing.There is one further point, I would like to mention andthat is, I am asking you to give up your responsibilityin this one limited instance. I would like you to takethis position seriously. It may be that outside ofthis situation, you are a highly responsible person andwill remain so. I am solely asking you to give up yourresponsibility in this one situation. Do you have anyquestions? Do you need any further information orevidence before you feel confident that I will in facttake on the full responsibility? (ADD THESE ITEMS TOCONTRACT)Could we go over the contract, just to make surewe both understand the agreement? First, as you cansee here in writing, I will take completeresponsibility for anything that may happen or is notperfect as a result of not checking. Second, I will beto blame if anything bad happens or is not perfect andyour name will not be mentioned. Third, myresponsibility applies to this situation only. Fourth,I will undo any damage or harm that may occur as aresult of not checking. I would like you to sign hereas a witness, that I must assume completeresponsibility. I would also like you to initial thesechanges/additions/deletions that we have made to make83this contract tailored to your specific pattern ofchecking. (Have participant sign form) You know youcan count on me.”High Responsibility Condition “You mentioned that youcheck_____over and over again. In a moment we willask you to go into a situation where you usually wouldcheck . People who check sometimes worry thatsomething bad will happen or what they do won’t beperfect. They sometimes worry that damage of some kindmay occur, that someone will be hurt, or that they willhave to undo any damage or harm. People who have theseworries sometimes feel that something bad will happenor is not perfect when they are in situations like theone you are about to be in. Except this time, I wantyou to know that you will have to take completeresponsibility if anything bad happens or anything isnot perfect.You are responsible for anything that happens oranything that is not perfect as a result of notchecking. You will be asked to take on completeresponsibility. You will be to blame if anything badhappens or if anything is not perfect. Your name willbe mentioned. You will also take responsibility for anydamage that may occur and will be asked to pay back orundo any damage or harm that may occur.I would like to put this agreement in writing.There is one further point, I would like to mention andthat is, I am asking to you to assume responsibility inthis one limited instance and we hope you will takethis position seriously. It may be that outside ofthis situation, you are a highly responsible person andthat will remain so. I am asking you to be completelyresponsibility in only this one situation. Do youhave any questions? Do you need any additionalinformation or evidence to make sure that you really dotake on the full responsibility? (ADD THESE ITEMS TOCONTRACT)84Could we go over the agreement, just to make sureof our understanding? First, as you can see here inwriting, you will take complete responsibility foranything that may happen or anything that is notperfect as a result of you not checking. Second, youwill be to blame if anything bad happens or anything isnot perfect and your name will be mentioned. Third,your responsibility applies to this situation only.Fourth, you will be asked to undo any damage or harmthat may occur as a result of not checking. I wouldlike you to sign here to guarantee that you will assumecomplete responsibility. I would also like you toinitial these changes/additions/deletions that we havemade to make this contract tailored to your specificpattern of checking. I will sign as a witness to yourtaking on the responsibility. I know I can count onyou.”Control-Check Condition “You mentioned that you check______over and over again. In a moment we will askyou to go into a situation where you usually checkPeople who check sometimes worry that somethingbad will happen or what they do won’t be perfect. Theysometimes worry that damage of some kind may occur,that someone will be hurt, or that they will have toundo any damage or harm. People who have these worriessometimes feel that something bad will happen or is notperfect when they are in situations like the one youare about to be in. We would like you to try to treatthis test like a realistic situation that you encounterregularly.We would like to get an understanding of yourtypical pattern of checking, so please try and treatthis situation as you usually do. (As before) we willask you to go into____situation and then we will askyou some questions about your experience. Do you haveany questions?85I would like to put this agreement in writing. Asyou can see here in writing., we would like you to 1)treat this situation as you usually do and 2) we wouldlike you to try to treat this behavioral test like arealistic situation - the kind of situation youencounter regularly.Is there anything that we can add/change in thiscontract, that will help you treat this situation asyou usually do? I would like you to sign here toacknowledge our agreement. I will sign as a witness.”Control—Clean Condition “You mentioned that you clean______over and over again. In a moment we will askyou to go into a situation where you usually clean_____•People who clean sometimes worry thatsomething bad will happen or what they do won’t beperfect. They sometimes worry that damage of some kindmay occur, that someone will be hurt, or that they willhave to undo any damage or harm. People who have theseworries sometimes feel that something bad will happenor is not perfect when they are in situations like theone you are about to be in. We would like you to tryto treat this test like a realistic situation that youencounter regularly.We would like to get an understanding of yourtypical pattern of cleaning, so please try and treatthis situation as you usually do. (As before) we willask you to go into____situation and then we will askyou some questions about your experience. Do you haveany questions?I would like to put this agreement in writing. Asyou can see here in writing, we would like you to 1)treat this situation as you usually do and 2) we wouldlike you to try to treat this behavioral test like arealistic situation - the kind of situation youencounter regularly.86Is there anything that we can add/change in thiscontract, that will help you treat this situation asyou usually do? I would like you to sign here toacknowledge our agreement. I will sign as a witness.”Post—assessment The post—assessment followed eachexperimental manipulation and consisted of the samestandard behavioral approach test and completion of theself—report measures of anxiety.Structured Interview At the very end of thesession, clients took part in a structured interviewwhich focused on history of trauma, responsibility andexplored possible sources of a heightened sense ofperceived responsibility.MeasuresAnxiety Disorders Interview Schedule — RevisedThe Anxiety Disorders Interview Schedule - Revised(ADIS—R) (DiNardo et al., 1985) is a detailedstructured interview designed to assess anxietydisorders. For purposes of this investigation, onlythe Obsessive Compulsive Disorder Subsection was used.This interview has high internal reliability for87diagnosis of OCD with Kappa Coefficients ranging from.825 to .857 (Barlow, 1987; DiNardo, O’Brien, Barlow,Waddell, & Blanchard, 1983)Maudsley Obsessional Compulsive InventoryThe Maudsley Obsessional—Compulsive Inventory(MOCI) (Hodgson & Rachman, 1977) is a 30 item true-false questionnaire designed to measure ObsessiveCompulsive symptomatology. Principal-componentanalysis has revealed four separate subscales, checking(9 items), cleaning (11 items), slowness (7 items) anddoubting (7 items). This measure has good internalconsistency and test—retest stability, as well as goodconvergent and divergent validity with the ninesubscales of the Symptom Checklist-90-Revised (Sanavio& Vidotto, 1985; Sternberger & Burns, 1990). It hasbeen used widely with both clinical and non-clinicalpopulations (e.g. Perse, Greist, Jefferson, Rosenfield,& Dar, 1987 Sher, Mann & Frost, 1984;).88Verbal Analogue ScalesVerbal analogue scales (0 to 100) were used toderive scores for the 12 dependent variables: urge tocheck, perceived discomfort, perceived responsibility(manipulation check), likelihood of anticipated threat,severity of anticipated threat, control overanticipated threat, the timing of anticipated threat,time needed to check, likelihood of anticipatedcriticism, severity of anticipated criticism, controlover anticipated criticism, the timing of anticipatedcriticism. These scales are widely used inbehavioural research on anxiety disorders, and areknown to correlate with several indices of autonomicarousal (Thyer, Papsdorf, Davis, & Vallecorsa, 1984).Behavioural Approach TestsBehavioural Approach Tests (BAT’s) have commonlybeen used to assess compulsive checking in bothresearch and clinical settings (e.g. Roper, Rachman &Hodgson, 1973; Steketee & Foa, 1985). IndividualizedBAT’s may be used to assess personally relevant89behaviours in naturalistic situations. In the currentstudy the BAT was tailored to the participant’sindividual pattern of checking. For example, if thesubject reported checking door locks repetitively, thenthe subject was given the following instructions in theBAT:“We would like you to lock the door and then walkaway from it without checking it at all.”Other examples of typical tasks chosen and instructionsfor BAT’s:Example 2 “We would like you to turn the stove on andthen off and then walk into the next room withoutchecking it at all.”Example 3 “We would like you to follow the route younormally take home without retracing your path.”Example 4 “We would like you to adjust the thermostatto 20 C and then walk into the next room withoutchecking it at all.”90RESULTSSublect CharacteristicsThirty people participated in the one-session, 4hour study. Sixty percent (18/30) of participants werefemale and 40 percent (12/30) of participants weremale. The mean age was 36.76 (S.D. 15.37) years (SeeTable 1).MedicationFifty percent (15/30) of participants were onmedication. Five of the 15 participants were onclomipramine (Anafranil), 3 of the 15 were onfluoxetine (Prozac), and the remaining 7 out of 15 wereon some other form of anxiolytic or anti—depressantmedication. None of the participants were on anti—psychotic or major tranquilizers.All participants were instructed not to altertheir regualar dosage of medication nor take any otherform of medication for twenty—four hours prior to thetesting.91TABLE 1SUBJECT CHARACTERISTICS(n=3 0)FEMALE (n=18) 60%MALE (n=12) 40%AGE (n=30) 36.76(15.37)92Sublect ScreeningWe received 423 phone calls for this study.Despite our attempts to be specific in our publicity,many of those who called requested information forother problems (ie. relationship problems, depressionetc.). If the individual indicated they washed orchecked repeatedly we conducted the telephone screen(see methods section). Hence, of these 423 calls weconducted telephone screens with 221 individuals. Ofthese 221 individuals, 38 appeared suitable for ourstudy and we arranged appointments to go to theirhomes. This low acceptance rate for the study, had todo with our stringent recruitment guidelines. Many ofthe callers who reported checking repeatedly, indicatedon the telephone that either it did not significantlyinterfere with their lives, or mentioned theyoccasionally checked repeatedly but that it wasn’t on aregular or consistent basis. Thus, these individualswould not meet criteria for OCD and were deemedinappropriate for this study.93Of the 38 homes we went to, 8 people did notexperience an urge to check in the presence of theexperimenter and thus, for these individuals we wereunable to proceed with the study. We noted that 6/8people who did not experience an urge to check,appeared to be less severe in overall OCDsymptomotology than those who participated. That is,in comparison to the remainder of the sample thesepeople scored lower on the ADIS (x=4.2 (1.4) vs. (x=6.l(1.79) and MOCI x=2.2 (1.34) vs x=l9 3 (6.34). Nofurther information was collected from theseindividuals so they could not be compared on any othercharacteristics. Of the remaining 2/8 who met criteriafor OCD these individuals reported signficantinterference but failed to report an urge to check inour presence. In both these cases, the participantsindicated that they only checked late at night beforeretiring for bed and that our presence made them feelmore safe.94Missing DataIn limited instances participants reported thatthey were unable to answer specific questions. In mostcases, it appeared that participants would perseverateand would fail to make a decision. After prompting fora reply if the participant was still unable to answer,or simply indicated that they “didn’t know” these datawere coded as missing.Missing data comprised less than 1% of the totaldata set and appeared to be randomly distributed acrossthe data set. That is, no apparent pattern relating tomissing data was evident.These missing data were recoded with values takenfrom randomly sampling the distribution with the meanand standard deviation of each condition. Othermethods such as simply replacing the missing data withthe mean and standard deviation of each condition wouldartifically reduce the variance of each condition.Thus, this approach of obtaining scores throughsampling the distribution to replace missing datamaintains the actual variance of each condition.95Preliminary Analyses Means and standard deviationswere calculated for each dependent measure, pre andpost manipulation and are reported in Tables 2 and 3.Major AnalysesAlpha level Given that this is exploratory researchand therefore conclusions will be tentative and will befully tested again, the risk of Type I errors need nothave the overriding importance it must have in othercircumstances. Indeed if power is overly reduced,there is a risk that findings might not be noted andduly followed up. Therefore, an alpha level of .05will be used for multivariate tests.96TABLE 2MEANS AND STANDARD DEVIATIONSPRE SCORES (n=30)ResponsibilityUrge to CheckDiscomfortLikely/BadSev/BadTime/BadFinish CheckControl/BadPanicLik/CriticismSev/CriticismControl/CritTime/CritCONTROL54.67 ( 9.00)74.67 (18.64)65.83 (15.60)50.70 (32.64)61.57 (32.13)2.76 ( 1.28)6.80 (11.74)42.83 (32.73)1.67 ( .80)57.87 (38.74)59.63 (34.70)35.30 (30.94)3.00 ( 1.93)HI RESP55.83( 8.62)70.33(19.16)65.29 (17.57)56.67 (20.48)62.21 (24.37)2.60 ( 1.33)5.00 ( 4.73)43.83 (29.50)1.57 ( .68)52.60 (26.89)55.43 (31.19)37.67 (30.92)2.57 ( 1.38)LO RESP53.70( 6.25)74.77(21.69)65.47 (21.23)55.83 (24.73)62.97 (23.62)2.37 ( 1.13)5.43 ( 7.63)40.83 (30.43)1.73 ( .74)55.33 (30.48)56.53 (28.25)38.73 (30.78)3.03 ( 1.61)97Urge to CheckDiscomfortLikely/BadSev/BadTime/BadFinish CheckControl/BadPanicLik/Criticism5ev/CriticismControl/CritTime/Crit**p<.001* p<.0lTABLE 3MEANS AND STANDARD DEVIATIONSPOST SCORES (n=30)HI RESP(20.06) 90.17(24.55)(fl!.T9’1flTResponsibility 58.73LO RESP F16.47(23.14) 65.09**68.7360.2353.8366 .772 .567.0343.831.6056. 3062.8734.703.00(25.70)(22.57)(24.77)(25.32)( 1.01)(12.02)(30.73)( .72)(37.62)(29.92)(33.02)( 2.04)79.50(25.78)72.17(27.50)59.57(30.87)73.50(26.98)2.48( 1.25)9.00(12.48)33.43(39.41)2.03( .93)73.77(31.60)74.30(31.19)37.13(39.03)2.54( 1.30)35.00(35.04)26.13(31.40)25.90(24.18)43.13(32.59)2.63( 1.57)2.40(3.40)28.13(25.93)1.20( .48)16.87(25.79)28.73(32.44)34.40(38.28)4.27( 2.29)17. 97**22. 06**15. 69**13. 91**.154 .58*1.6911. 52**26. 81**19. 29**.086.04*98Were th?re baseline differences?In order to ensurethere were no differences between conditions prior tomanipulation a mulitivariate analysis of varianceforrepeated measures (MANOVAR) was conducted on all premeasures.The within—subject factor was condition(control-check, highresponsibility,lowresponsibility). AMauchley sphericity test was notsignificant. No significant differences betweenconditions werefound using either multivariate(E(4,26)=.70, =n.s.) or univariate results(F(26,92)=.58, =n.s.). These resultsconfirm nodifferences in baseline measuresprior to manipulationof perceived responsibility.Order EffectsIn order to test for order effects,the order in which subjects receivedthe conditions(control-check, control—clean, hi responsibility, lowresponsibility) wasexamined.The two controlconditions (control—check,control—clean) werealways given first and the orderofthese two conditions was counterbalanceth Forthepurpose of analysis, control—check was assigneda99value of one andcontrol—clean avalue of two. Thepossible order could be either one, two or two, one.The two experimental conditions (highresponsibility,low responsibility) were always assigned aftercompletion of thecontrol conditions. The order ofthese two conditions was counterbalanced. The highresponsibility condition was assigned a numericequivalent of three, and low responsibility a numericequivalent of four. The possibleorder was either3, 4or 4, 3. Overall then, there were four possible orders: 1,2,3,4; 1,2,4,3; 2,1,3,4; 2,1,4,3. These fourpossible combinations were used inthe analysesdescribed below.In order to examine if order in whichconditions were received influencedthe dependentmeasures, amulitivariate analysis of variance forrepeated measures (NANOVAR) was conducted on all premeasures.The between—subjects factor was order(combination 1, combination 2, combination 3,combination 4). Thewithin-subject factorwascondition (control-check, high responsibility, lowresponsibility). Inorder to minimizethe probability100of a type II error, univariate results,with strongerpower to detect differences, were used. No main effectof order (E(3,26)=l.05, =n.s.), condition((2,52)=l.O6, p=n.s.), nor Order X Conditioninteraction ((3 ,26)=l.12,=n.s.) was present.Theseresults suggest that the order in whichsubjectsreceived the conditions had no effecton the baselinedependent measures.Given that there were no differencesin thedependent measures prior to the experimentalmanipulation, nor were order effectspresent, allsubsequent major analyses were performed usingpostscore results.In order to test the series of predictions forthis investigation, a MANOVAR was conducted on all postmeasures. This approach of including all post measuresin one MANOVAR was used for severalreasons. First, itprovides protection against Type 1 error. Second, itbest controls for type one error rates through takinginto consideration that several dependent variables areassumed to be correlated and it mayreveal differences101not shown in separate ANOVAs. According to Tabachnickand Fidell (1988), “.. .when responses to two dependentvariables are considered in combination, differencesmay become apparent. Thus, MANOVA which considersdependent variables in combination, may sometimes bemore powerful than separate ANOVAS”.For purposes of this analysis, the between—groupfactor was Order (combination 1, combination 2,combination 3, combination 4). The within-subjectfactor was Condition (control-check, highresponsibility, low responsibility). A main effect forCondition ((2,52)=4.23, <.03 emerged. A Mauchleysphericity test was not significant. The significantmain effect suggests that there were differencesbetween the three conditions in dependent measures.Univariate results will be described below. There wasno main effect for Order ((3,26)=.98,=n.s.), norOrder x Condition interaction ((3,26)=l.32,=n.s.).Therefore, the order in which subjects received theconditions did not appear to influence the dependentmeasures. In addition, there was no interaction102between the order of conditions and the conditionsthemselves.Did the manipulation work?Repeated measures analyses of variance (ANOVAR)were then computed for each post measure comparing thecontrol condition to each of the experimentalconditions (high responsibility, low responsibility).In order to control for the number of comparisons beingmade alpha level was set at p=.025 for all ANOVARresults. Significant results were then followed upusing Dunnett planned comparisons.The first question to be put to the test is whetherthe manipulation of perceived responsibility waseffective. A univariate ANOVAR revealed that there wasa difference between the three conditions on perceivedresponsibility ( (2,58)=65.02, p<.001). Dunnettplanned comparisons indicated that the measure ofperceived responsibility was significantly higher inthe high responsibility condition than in the controlcondition (t=4.85, p<.0l)(See Table 3). Similarly,Dunnett planned contrasts showed that perceived103responsibility scores were significantly lower in thelow responsibility than in the control condition(t=6.52, p<.Ol). These analyses confirmed that themanipulation was successful; perceived responsibilityscores in the high responsibility condition were higherthan the control condition and perceived responsibilityscores in the low responsibility condition weresignificantly lower than in the control condition.Is there a causal relationship between perceivedresponsibility and compulsive checking?Hypothesis 1 stated that there is a causalrelationship between perceived responsibility andcompulsive checking. In order to address thispostulate, predictions 1 through if were tested.Prediction la stated that an increase in perceivedresponsibility would be followed by an increase inperceived discomfort. In contrast, Prediction idindicated that a decrease in perceived responsibilitywould be followed by a decrease in perceiveddiscomfort. An ANOVAR indicated there was a differencebetween conditions in perceived discomfort104((2 , 58)=22 .06, p_<.001). Dunnett planned comparisonsrevealed that there was a trend for perceiveddiscomfort levels to be greater in the highresponsibility condition than in the control condition(t=l.66, p<.l0). Dunnett tests found that perceiveddiscomfort scores were significantly lower in the lowcondition than in the control condition (t=4.74,p<.01). Therefore, Prediction la and ld were partiallyconfirmed; there was a trend for an increase inperceived responsibility to be followed by an increasein perceived discomfort and a decrease in perceivedresponsibility.105Table 4Dunnett Planned ComparisonsPost Scores (n30)CONTROL HI RESP LO RESP tResponsibility 58.73 (20.06) 90.17(24.55) 4.85**58.73 (20.06) 16.47(23.14) 6.52**Urge to Check 68.73 (25.70) 79.50(25.78) 1.3968.73 (25.70) 35.00(35.04) 4•35**Discomfort 60.23 (22.57) 72.17 (27.50) 1.6660.23 (22.57) 26.13 (31.40) 4.74**Likely/Bad 53.83 (24.77) 59.57 (30.87) .8953.83 (24.77) 25.90 (24.18) 434**Sev/Bad 66.77 (25.32) 73.50 (26.98) 1.1166.77 (25.32) 43.13 (32.59) 3.91*** p<.05.01106Table 4(Cont’d)Dunnett Planned ComparisonsPost Scores (n=30)CONTROL HI PESP2.56 ( 1.01)2.56 ( 1.01’)Panic 1.60 ( .72) 2.03 ( .93) 2.50*1.60 ( .72) 1.20 (.48) 2.31** p<.05.01Time/Bad1.0 RFP2.48 ( 1.25)tn/a2 . 63 ( 1 .57 ) n/aFinish Check 7.03 (12.02) 9.00 (12.48) .887.03 (12.02’) 2.40 (3.40) 2.07*Control/Bad 43.83 (30.73) 33.43 (39.41) n/a43.83 (30.73) 28.13 (25.93) n/aLik/Criticism 56.30 (37.62) 73.77 (31.60) 2.19*56.30 (37.62) 16.87(25.79) 4.95**Sev/Criticism 62.87 (29.92) 74.30 (31.19) 1.5062.87 (29.92) 28.73 (32.44) 4.47**107Table 4(Cont’d)Dunnett Planned ComparisonsPost Scores (n=30)CONTROL HI RESP LO RESP tControl/Crit 34.70 (33.02) 37.13 (39.03) n/a34.70 (33.02) 34.40 (38.28) n/aTime/Crit 3.00 ( 2.04) 2.54 ( 1.30) .723.00 ( 2.04) 4.27 ( 2.29) 3.01*** p<.05**p<.01108was followed by a decrease in perceived discomfort.Prediction lb stated that an increase in perceivedresponsibility would be followed by an increase in theurge to check and Prediction le put forward thatdecreases in perceived responsibility would result indecreases in the urge to check. The three conditionswere shown to be significantly different in the urge tocheck (E(2,58)=17.97, .<.O01). Dunnett plannedcomparisons showed that there was a trend for the urgeto check to be greater in the high responsibilitycondition than in the control condition (t=l.39,p<.l0). This same test revealed that the urge tocheck was significantly lesser in the lowresponsibility condition than in the control condition(t=4.35, p<.0l). To summarize, prediction lb and lewere partially verified: there was a trend for anincrease in perceived responsibility to be followed byan increase in the urge to check and a decrease inperceived responsibility was followed by a decrease inthe urge to check.109Prediction lc stated an increase in perceivedresponsibility would be followed by an increase in theestimated length of time needed to complete a check andPrediction le proposes that a decrease in perceivedresponsibility would be followed by a decrease in theestimated time needed to complete a check. An ANOVARshowed that the three conditions were significantlydifferent in the estimated length of time needed tocomplete a check ((2,58)=4.58, p<.Ol). Dunnettplanned contrasts found no significant difference inestimated length of time needed to complete a checkwhen comparing the high responsibility condition to thecontrol condition (t=.88, p=n.s.). However, asignificant difference emerged when examining theestimated length of time needed to finish checking inthe low responsibility versus control conditions(t=2.07, p<.05). Therefore, prediction lc was notconfirmed. There was no significant difference in theestimated length of time needed to complete a check inthe high responsibility versus control conditions. Incontrast prediction le was confirmed; decreases in110perceived responsibility were followed by decreases inthe estimation of the time needed to complete a check.Do estimations of the timing of anticipated threatinfluence the probability of perceived panic?Hypothesis 2 stated that predictions of the timingof anticipated threat would influence the probabilityof perceived panic. Prediction 2a proposed that themore immediate the anticipated threat, the greater theprobability of perceived panic. Similarly, Prediction2b stated that the less immediate the anticipatedthreat, the lower the probability of perceived panic.Given that there were no differences between conditionsin the timing of anticipated threat (see below),Pearson Correlations were computed, collapsed oncondition, between the timing of anticipated threat andperceived panic scores. No significant correlation wasfound (r=.12,p=n.s.). Therefore, prediction 2a and 2bwere not confirmed: the timing of anticipated threatdid not appear to influence the probability ofperceived panic.111Do variations in perceived responsibility alterestimations of the probability of anticipated harm ?Hypothesis 3 postulated that variations inperceived responsibility would not alter estimations ofthe probability of anticipated harm. In order to testthis, Predictions 3a and 3b were put to the test.These predictions proposed that increases in perceivedresponsibility (Prediction 3a) or decreases inperceived responsibility (Prediction 3b) would both befollowed by no change in the probability of anticipatedharm. An ANOVAR revealed that the three conditionswere significantly different (L(2,58)=15.69, p<.OO1).Dunnett planned comparisons found that decreases inperceived responsibility were followed by decreasedestimated probability of anticipated harm (t=4.34,p<.Ol). No differences were found between increases inperceived responsibility and estimated probability ofanticipated harm (t=.89, p=n.s.). Therefore,prediction 3a was confirmed; increases in perceivedresponsibility did not produce changes in theprobability of anticipated harm. In contrast,prediction 3b was not confirmed; decreases in perceived112responsibility lead to decreases in the probability ofanticipated harm.Do variations in perceived responsibility alterestimations of the seriousness of anticipated harm?Hypothesis 4 postulated that variations inperceived responsibility would not alter estimations ofthe seriousness of anticipated harm. Predictionsarising from this premise, proposed that increases inperceived responsibility (prediction 4a) and decreasesin perceived responsibility (prediction 4b) would bothbe followed by no change in the seriousness ofanticipated harm. An ANOVAR showed that the threeconditions were significantly different ((2,58)=l3 .91,p<.OO1). Dunnett planned comparisons revealed thatthe seriousness of anticipated harm was significantlyless in the low responsibility than in the controlcondition (t=3.91, p<.Ol). No difference was found inseverity of anticipated harm in the high responsibilitycondition as compared to the control condition (t=l.ll,p=n.s.). Therefore, Prediction 4a was confirmed: therewas no difference in seriousness of anticipated harm in113the high responsibility condition as compared to thecontrol condition. Prediction 4b was not confirmed;decreases in perceived responsibility were followed bydecreases in the severity of anticipated harm.Do variations in perceived responsibility alterestimations of when perceived harm will occur?Hypothesis 5 stated that changes in perceivedresponsibility would not influence estimations of whenperceived harm would occur. Two predictions weretested to examine this hypothesis. Predictions 5aproposed that increases in perceived responsibilitywould not change estimations of when the perceived harmwould occur. Prediction 5b postulated that decreasesin perceived responsibility would be followed by nochange in when perceived harm would occur. An ANOVARfound no difference between conditions in whenperceived harm would occur ((2,58)=.l5, p=n.s.).Therefore, Predictions 5a and 5b were confirmed.Variations in perceived responsibility did not alterestimations of when perceived harm would occur.114Do variations in perceived responsibility foranticipated harm influence amount of control over thepertinent anticipated threat?Hypothesis 6 stipulated that changes in perceivedresponsibility were independent of a sense of controlover the pertinent anticipated threat. Two predictionswere put forward to test this postulate. Prediction 6astated that an increase in perceived responsibilitywould not change degree of control and prediction 6bproposed that a decrease in perceived responsibilitywould not alter subjects’ sense of control. An ANOVARshowed no difference between the conditions((2,58)=1.69, p=n.s.). To summarize, Prediction 6aand 6b were confirmed: variations in perceivedresponsibility were independent of a sense of control.Do variations in perceived responsibility influence theprobability of anticipated criticism?Hypothesis 7 postulated that changes in perceivedresponsibility would result in changes in theprobability of anticipated criticism. Prediction 7astated that increases in perceived responsibility would115lead to increases in the probability of anticipatedcriticism and Prediction 7b indicated that decreases inperceived responsibility would lead to decreases in theprobability of anticipated criticism. The threeconditions were found to be significantly different inthe likelihood of anticipated criticism (E(2,58)=26.81,p<.OOl). Dunnett planned comparisons revealed thatincreases in perceived responsibility were followed byincreases in the likelihood of anticipated criticism(t2.19, p <.05) and decreases in perceivedresponsibility were followed by decreases in thelikelihood of anticipated criticism (t=4.95, p <.01).In sum, prediction 7a and 7b were confirmed: increasesin perceived responsibility were followed by increasesin the likelihood of anticipated criticism anddecreases in perceived responsibility were followed bydecreases in the likelihood of anticipated criticism.Do variations in Derceived responsibility alterestimations of the severity of anticipated criticism?Hypothesis 8 stated that perceivedresponsibility would influence estimations of the116severity of anticipated criticism. Prediction 8asuggested an increase in perceived responsibility wouldbe followed by an increase in the severity ofanticipated criticism and Prediction 8b proposed that adecrease in perceived responsibility would be followedby a decrease in the severity of anticipated criticism.An ANOVAR found a significant difference betweenconditions ((2 ,58)=l9.29, p<.OOl). Dunnett plannedcomparisons revealed that there was a trend forseverity of anticipated criticism to be greater in thehigh responsibility condition than in the controlcondition (t=1.50, p <.10). This same test found asignificant difference between the low responsibilitycondition and control condition in severity ofanticipated criticism (t=4.47, p <.01). Therefore,prediction Ba and 8b were partially confirmed. Therewas a trend for increases in perceived responsibilityto produce increases in the severity of anticipatedcriticism. Decreases in perceived responsibilityresulted in decreases in the severity of anticipatedcriticism.117Do variations in perceived responsibility alterestimations of the timing of anticipated criticism?Hypothesis 9 postulated that changes in perceivedresponsibility would influence predictions of thetiming of anticipated criticism. Prediction 9aproposed that increases in perceived responsibilitywould be followed by decreases in period of time beforeanticipated criticism. Similarly, Prediction 9bsuggested that decreases in perceived responsibilitywould be followed by increases in estimated length oftime before anticipated criticism. An ANOVAR found thethree conditions to be significantly different((2,58)=7.83, p<.OOl). Dunnett planned comparisonsshowed that the low responsibility condition wassignificantly different from the control condition(t=3..Ol, p<.Ol). This same test found no differencebetween the high responsibility and control conditions(t=.72, p=n.s.). Therefore, prediction 9a was notconfirmed: increases in perceived responsibility didnot lead to a decrease in the estimated length of timebefore anticipated criticism. In contrast, prediction9b was confirmed: decreases in perceived responsibility118lead to an increases in estimation of the estimatedlength of time before anticipated criticism.Do variations in perceived responsibility influenceestimations of the amount of control over anticipatedcriticism?Hypothesis 10 stated that variations in perceivedresponsibility would not alter estimations of theamount of control over anticipated criticism.Prediction ba postulated that increases in perceivedresponsibility would not lead to increases in controlover anticipated criticism and Prediction lOb statedthat decreases in perceived responsibility would notlead to decreases in control over anticipatedcriticism. A repeated measures ANOVA found nodifferences between conditions in control overanticipated criticism ((2,58)=.O8,p=n.s.). Predictionba and lOb were therefore verified: variations inperceived responsibility did not appear to influenceamount of control over anticipated criticism.119CorrelationsPearson correlations were computed between alldependent variables for each condition and are reportedin Tables 5—7. Given the numbers of correlations beingcomputed, alpha level was set at p<.Ol. Despite thisadjustment the probability of a type 1 error remainsexceedingly high and thus, interpretation is tenuous atbest.Correlations that are noteworthy in the controlcondition include the correlation between perceivedresponsibility and perceived discomfort (r=.48, p<.Ol).In the high responsibility condition, perceivedresponsibility was correlated with the urge to check(r=.77, p<.OOl), perceived discomfort (r=.61, p<.OOl),likelihood of anticipated threat (r=.44, p<.Ol), andseverity of anticipated threat (r=..67, p<.OOl,likelihood of anticipated criticism (r.58, p<.OOl),and severity of anticipated criticism (r=.59, p<.OOl).In the low responsibility condition, perceivedresponsibility was correlated with urge to check(r=.52, p<.Ol), perceived discomfort (r=.58, p<.OOl),120estimated length of time needed to finish checking(r=.44, p<.0l), and perceived panic (r=.56, p<.OOl,likelihood of anticipated criticism (r=.85, p<.00l),and severity of anticipated criticism (r=.65, p<.00l).Of interest are correlations that are presentin one of the experimental conditions but absent inanother. There are significant correlations betweenestimated length of time needed to finish checking andperceived responsibility, and perceived panic andperceived responsibility, in the low responsibilitycondition. These links disappear in the highresponsibility condition. In addition, correlationsbetween perceived responsibility and likelihood ofanticipated threat, and perceived responsibility andseverity of anticipated threat in the highresponsibility condition are not found in the lowresponsibility condition.Trauma HistoryDuring the course of the structured interviewparticipants were asked if they had experiencedphysical, sexual or emotional abuse before 18 years of121age. Definitions were given for each category.Physical abuse was defined as “excessive physicalpunishment (i.e., more than a spanking)”. Sexual abusewas defined as “any sexually inappropriate act (forexample, fondling, masturbation, oral, anal or vaginalintercourse) or acts with sexual overtones to meetsomeone else’s sexual or emotional needs”. Emotionalabuse was defined as “force to perform cruel ordegrading tasks, or repeated humiliation andanticipated criticism (see appendix for precisequestions).Twenty three out of 30 participants or 73%,reported yes to at least one of these categories. Fourof the 30 (13%) participants met criteria for physicalabuse, 19 out of 30 met criteria for sexual abuse (63%)and 24 out of 30 (80%) met criteria for emotionalabuse.Given that this interview was exploratory noformal baserates have been established and thuscomparison of these scores with other data could not bedone.122TABLE 5PEARSON CORRELATIONSCONTROL CONDITION(N= 30 ) *URGE DISC LIKBURGE 1.00**.51 .21 .26DISC .51* 1.00** .48* .65**RESP .21 .48* 1.00** .50*LIKB .26 .65** .50* 1.00**SEVB—.11 —.04 .18 .34FINISH .14 .10—.15 .26CONTROLBAD .19 .06 .06 .06PANIC .05 .20 .33 .38LIKCZ—.02 .25 .30SEVCZ —.06 .34 .40 •53*CONTROLCZ -.08 .25 .19-.01TIM/BAD— .19 —.41— .14TIMINGCZ —.28 —.07 .21—.26SEVB FINISH CONTROLBAD PANICURGE—.11 .14 .19 .05DISC —.03 .10 .06 .20RESP .18 —.15 .06 .32LIKB .34 .26 .06 .38SEVB 1.00** •45*—.11 .33FINISH •45* 1.00’—.08 .30CONTROLBAD -.11 -.08 1.00**-.15PANIC .33 .30 —.15 1.00**LIKCZ .19 .24 .04 .18SEVCZ •43* .32—.13 .22CONTROLCZ -.16 -.19 .36-.23TIM/BAD .07 .01 .00-.17TIMINGCZ —.26 —.18—.13 —.14* p<.O1.001123TABLE 5 (CONT’D)PEARSON CORRELATIONSCONTROL CONDITION(N=30)*LIKCZ SEVCZ CONTROLCZ TIM/BADURGE —.02 —.06 —.08 —.19DISC .25 .34 .25 —.41RESP .30 .40 .19 —.14LIKB .61** •53* —.01SEVB .19 •43* —.16 .07FINISH .24 .32 —.19 .01CONTROLBAD .04 -.13 .36 .00PANIC .18 .22 —.23 —.17LIKCZ 1.0** .69** .02 —.39SEVCZ .69** 1.0** 02 —.28CONTROLCZ .02 .02 1.0** .16TIM/BAD —.39 —.28 .16 1.0**TIMINGCZ —.19 —.12 •47* •34TIMCZURGE -.28______DISC -.07RESP .22LIKB - .26SEVB -.26FINISH - .18CONTROLBAD -.13PANIC -.14LIKCZ -.19SEVCZ - .12CONTROLCZ .47TIM/BAD - .34TIMINGCZ 1.0*URGEDISCRESPLIKBSEVBFINISHCONTROLBADPANICLIKCZSEVCZCONTROLC ZTIM/BADTIMINGCZ*LEGENDURGE TO CHECKDISCOMFORT EXPERIENCEDRESPONSIBILITYLIKELIHOOD OF THREATSEVERITY OF THREATTIME TO FINISH CHECKINGCONTROL OVER THREATREPORTED PANICLIKELIHOOD OF CRITICISMSEVERITY OF CRITICISMCONTROL OVER CRITICISMTIMING OF THREATTIMING OF CRITICISM124TABLE 6PEARSON CORRELATIONSHIGH RESPONSIBILITY CONDITION(N=30)*URGE DISC RESP LIKBURGE 1.O0** •77** •45*DISC 1.OO** .61** .65**RESP .61** 1.0O1 •44*LIKB •45* .65** •44* 1.OO**SEVB •45* .65** .67** •54*FINISH .34 .09 .18 .22CONTROLBAD .06 .12 .24 .01PANIC •57** .65** .30 .60**LIKCZ •57** .63** .58** .70**SEVCZ •57** .67** •59** .76**CONTROL/CZ - .50* -.46* —.30 -.20TIM/BAD .03 -.10 .11 -.29TIMING/CZ — .51* —.50* —.46* —.40SEVB FINISH CONTROLBAD PANICURGE •45* .34 .06DISC .65** .09 .12 .65**RESP .67** .18 .24 .30LIKB •54* .22 .01 .60**SEVB 1.00** .03 —.01 .33FINISH .03 1.00** .12 .23CONTROLBAD -.01 .12 1.00** .12PANIC .33 .23 .12 1.00**LIKCZ •53* .30 .11 •44*SEVCZ •59** .36 .08 •53*CONTROL/CRIT -.37 -.15 •53* -.14TIM/BAD .09 —.30 —.08 —.02TIMINGCZ —. 20 —. 26 —. 23 —. 27* P<.01.001125TABLE 6 (CONT’D)PEARSON CORRELATIONSHIGH RESPONSIBILITY CONDITION(N= 30) *LIKCZ CONTROLCZ TIM/BADURGE •57** —.50*. .03DISC .63** .67** —.46* —.10RESP .58** •59** —.30 .11LIKB .70** .76** —.20 —.29*SEVB 53* •59** —.37 .09FINISH .30 .36 —.15 —.30CONTROLBAD .11 .08 •53* -.08PANIC •44* •53* —.14 —.02LIKCZ 1.0** .91** —.29 —.23SEVCZ .91** 1.0’ —.31 —.12CONTROL/CRIT -.29 —.31 1.0** .08TIMING/BAD —.23 —.12 .08 1.0**TIMINGCZ —.36 —•47* .29 .18TIMCZURGE -.51*______DISC —.50RESP -.46*LIKB -.40SEVB -.20FINISH —. 26CONTROLBAD -.23PANIC -.27LIKCZ -.36SEVCZ -.47CONTROL/CRIT .29*TIM/BAD .18TIMINGCZ 1.0**URGEDISCRESPLIKBSEVBFINISHCONTROLBADPANICLIKCZS EVC ZCONTROLCZTIM/BADTIMINGCZ*LEGENDURGE TO CHECKDISCOMFORT EXPERIENCEDRESPONSIBILITYLIKELIHOOD OF THREATSEVERITY OF THREATTIME TO FINISH CHECKINGCONTROL OVER THREATREPORTED PANICLIKELIHOOD OF CRITICISMSEVERITY OF CRITICISMCONTROL OVER CRITICISMTIMING OF THREATTIMING OF CRITICISM126TABLE 7PEARSON CORRELATIONSLOW RESPONSIBILITY CONDITION(N=30)*URGE RESP LIKBURGE 1.00** .78** .52* •44*DISC .78** 1.00** .58** .51*RESP .52* .58** 1.00**.40LIKB •44* .51*.40 1.00**SEVB .32 .37 .41 .35FINISH .42 .37 •44* .17CONTROLBAD 37 .37 .39 .21PANIC .51* .61** .56**LIKCZ •43* •55** .85** •47*SEVCZ .41 •45* .65** •35CONTROL/CZ .24 .24 .11 —.01TIM/BAD -.11 -.13 .18 —.28TIMING/CZ —.16—.03 —.34 —.13* p<.o1**p<.001SEVB FINISH CONTROLBAD PANICURGE .32 .42 .37 .51*DISC .37 .37 .37RESP .41 •44* .39 .56**LIKB •35* .17 .21SEVB 1.00** .48* .00 .41FINISH .48* 1.00** .41 .10CONTROLBAD -.00 .41 1.00** .10PANIC .41 .10 .10 1.00**LIKCZ .38 .21 .21SEVCZ .38 .11 —.02 .61**CONTROL/CRIT .08 .16 .71** .06TIM/BAD -.16 .02 -.03 -.13TIMINGCZ —.25 .01 .21 —.20127TABLE 7 (CONT’D)PEARSON CORRELATIONSLOW RESPONSIBILITY CONDITION(N=30)*LIKCZ SEVCZ CONTROLCZ TIM/BADURGE •43* .41 .24DISC •55** •45* .24 —.13RESP .85** .65** .11 .18LIKB •47* .35 —.01 —.28SEVB .38 .38 .08 —.16FINISH .21 .11 .16 .02CONTROLBAD .21 -.02 .71** -.03PANIC .77** .61** .06 —.13LIKCZ 1.0** .76** .02 .01SEVCZ 76 1.0** —.18 .03CONTROL/CRIT .02 -.18 1.0** -.06TIM/BAD .01 .03 —.06 1.00**TIMINGCZ —.40 —.27 .25 .09TIMCZURGE -.16______DISC —.03RESP -.34LIKB -.13SEVB -.25FINISH .01CONTROLBAD .21PANIC -.20LIKCZ -.40SEVCZ —. 27*CONTROL/CRIT .25*TIM/BAD .09TIMINGCZ 1.0**URGEDISCRESPLIKBSEVBFINISHCONTROLBADPANICLIKCZSEVCZCONTROLC ZTIM/BADTIMINGCZ*LEGENDURGE TO CHECKDISCOMFORT EXPERIENCEDRESPONSIBILITYLIKELIHOOD OF THREATSEVERITY OF THREATTIME TO FINISH CHECKINGCONTROL OVER THREATREPORTED PANICLIKELIHOOD OF CRITICISMSEVERITY OF CRITICISMCONTROL OVER CRITICISMTIMING OF THREATTIMING OF CRITICISM128FIGURE 1SCHEMATIC REPRESENTATION OF RESUlTSURGE TO CHECKDISCOMFORTPANICINCREASES LIKELIHOOD OF CRITICISMSEVERITY OF CRITICISMRESPONSIBILITYURGE TO CHECKDISCOMFORTLIKELIHOOD OF THREATDECREASES SEVERITY OF THREATTIME TO FINISH CHECKINGPANICLIKELIHOOD OF CRITICISMSEVERITY OF CRITICISMTIMING OF CRITICISM129Summary of ResultsHypotheses and PredictionsHypothesis 1There is a causal relationship between perceivedresponsibility and intended compulsive checking.Prediction la. An increase in perceived responsibilitywill be followed by an increase in perceiveddiscomfort.This prediction was partially confirmed. There was atrend for an increase in perceived responsibility tolead to an increase in perceived discomfort (t=l.66, p<.10).Prediction lb. An increase in perceived responsibilitywill be followed by an increase in the urge to completea check.This prediction was partially confirmed. There was atrend for an increase in perceived responsibility to130lead to an increase in the urge to check (t=l.39, p<10).Prediction ic. An increase in perceived responsibilitywill be followed by an increase in the estimated lengthof time needed to complete a check.This prediction was not confirmed (t=.88, p=n.s.).Prediction ld. A decrease in perceived responsibilitywill be followed by a decrease in perceived discomfort.This prediction was confirmed (t=4.74, p <.01).Prediction le. A decrease in perceived responsibilitywill be followed by a decrease in the urge to check.This prediction was confirmed (t=4.35, p <.01).Prediction lf. A decrease in perceived responsibilitywill be followed by a decrease in the estimated lengthof time needed to check.This prediction was confirmed (t=2.07, p <.05).131Hypothesis 2Estimations of the timing of anticipated threatinfluence the probability of perceived panic.Prediction 2a. The more immediate the anticipatedthreat, the greater the probability of perceived panic.This prediction was not confirmed.Prediction 2b. The less immediate the anticipatedthreat, the lower the probability of perceived panic.This prediction was not confirmed.Hypothesis 3Variations in perceived responsibility do notalter estimations of the probability of anticipatedharm.Prediction 3a. An increase in perceived responsibilitywill be followed by no change in the probability ofanticipated harm.This prediction was confirmed (t=.89, p=n.s.).132Prediction 3b. A decrease in perceived responsibilitywill be followed by no change in the estimatedprobability of anticipated harm.This prediction was not confirmed (t=4.34, p<.Ol).Hypothesis 4Variations in perceived responsibility do notalter estimations of the seriousness of anticipatedharm.Prediction 4a. An increase in perceived responsibilitywill be followed by no change in the estimatedseriousness of anticipated harm.This prediction was confirmed (t=l.ll, p=n.s.).Prediction 4b. A decrease in perceived responsibilitywill be followed by no change in the estimatedseriousness of anticipated harm.This prediction was not confirmed (t=3.91, p<.Ol).133Hypothesis 5Variations in perceived responsibility do notalter estimations of when perceived harm will occur.Prediction 5a. An increase in perceivedresponsibility will be followed by no change in whenperceived harm will occur.This prediction was confirmed.Prediction 5b. A decrease in perceived responsibilitywill be followed by no change in when perceived harmwill occur.This prediction was confirmed.Hypothesis 6Perceived responsibility for anticipated harm isindependent of a sense of control over the pertinentanticipated threat.134Prediction 6a. An increase in perceived responsibilitywill not alter estimations of amount ofcontrollability.This prediction was confirmed.Prediction 6b. A decrease in perceived responsibilitywill not influence estimations of amount ofcontrollability.This prediction was confirmed.Hypothesis 7Variations in perceived responsibility will alterestimations of the likelihood of anticipated criticism.Prediction 7a. An increase in perceived responsibilitywill be followed by an increase in the likelihood ofanticipated criticism.This prediction was confirmed (t=2.19,p<.05).Prediction 7b. A decrease in perceived responsibilitywill be followed by a decrease in the likelihood ofanticipated criticism.135This prediction was confirmed (t4.95, p<.Ol).Hypothesis 8Variations in perceived responsibility will alterestimations of the severity of anticipated criticism.Prediction 8a. An increase in perceived responsibilitywill be followed by an increase in the severity ofanticipated criticism.This prediction was partially confirmed (t=l.50,p<.lO). There was a trend for increases in perceivedresponsibility to be followed by an increase in theseverity of anticipated criticism.Prediction 8b. A decrease in perceived responsibilitywill be followed by a decrease in the severity ofanticipated criticism.This prediction was confirmed (t=4.47, p<.Ol).Hypothesis 9Variations in perceived responsibility will alterestimations of the timing of anticipated criticism.136Prediction 9a. An increase in perceived responsibilitywill be followed by a decrease in the estimated lengthof time before anticipated criticism.This prediction was not confirmed (t=.72, p=n.s.).Prediction 9b. A decrease in perceived responsibilitywill be followed by an increase in the estimated lengthof time before anticipated criticism.This prediction was confirmed (t=3.Ol,p<.0l) .Hypothesis 10Perceived responsibility for anticipated harm isindependent of a sense of control over anticipatedcriticism.Prediction ba. An increase in perceivedresponsibility will not alter estimations of amount ofcontrollability over anticipated criticism.This prediction was confirmed.Prediction lob. A decrease in perceived responsibilitywill not influence estimations of amount ofcontrollability over anticipated criticism.This prediction was confirmed.137138DISCUSSIONResults confirm there was a very powerfulmanipulation of responsibility in this study. Therewas approximately a 74/100 point difference in the meanperceived responsibility score reported in the highresponsibility condition versus the low responsibilitycondition.Decreases in perceived responsibility produceddecreases in perceived discomfort, urge to check,probability of anticipated harm, severity ofanticipated harm, estimated length of time needed tofinish checking, perceived panic, likelihood ofanticipated criticism, timing of anticipated criticismand severity of anticipated criticism. The onlyvariables not significantly affected by decreases inperceived responsibility had to do with control andtiming of anticipated harm: control over the bad thinghappening, control over anticipated criticism andtiming of anticipated harm.139On the other hand, increases in perceivedresponsibility lead to increases in perceived panic andlikelihood of anticipated criticism. Several trendswere present. There were trends for increases inperceived responsibility to lead to increases inperceived discomfort experienced, urge to check, andseverity of anticipated criticism. Increases inperceived responsibility did not effect the estimatedprobability of anticipated harm occurring, nor severityof anticipated harm, estimated length of time needed tocomplete a check, timing of anticipated harm, timing ofanticipated criticism, control over anticipated harmnor control over anticipated criticism (See Figure 1).To summarize, the four main conclusions from thisanalysis were:1) there is a connection between decreases inperceived responsibility and compulsive checking2) decreases in perceived responsibility affect boththe estimated probability of anticipated harm and itsconsequences1403) there is no relationship between variations inperceived responsibility and extent of controllability4) there is a relationship between perceivedresponsibility and anticipated criticism.Unexpected FindingsThere were several unexpected findings in thisstudy. First, the effects of variations in perceivedresponsibility appeared to be stronger in the lowresponsibility condition than in the highresponsibility condition. Upon examination of themeans for perceived responsibility, the weaker effectspresent in the high responsibility condition may inpart be due to a ceiling effect. Perceivedresponsibility scores in the control condition wereroughly 60/100 and therefore participants may beconsidered moderately responsible from the start.Attempting to increase perceived responsibility fromthis already high starting point was possible butquickly reached maximum scores (the mean was 90.17).Twenty-six of 30 participants scored 90/100 or above inthis condition suggesting overall a high mode but there141was little room to move beyond a 90/100 point mean.Examination of the means on urge to check, discomfortexperienced likelihood of bad thing happening, severityof anticipated harm, and estimated length of time tofinish checking were all in the predicted direction.Due to the small sample size it was not possible to doa formal analysis on this group of 26 participants whoresponded more vs. the 4 participants who respondedless to the manipulation. A rudimentary analysis foundno differences between those participants who respondedmore and those who responded less to the manipulationon the dependent measures.Means were computed for all variables fromparticipants who scored 90 or above in the highresponsibility condition and compared to the sample of30 (see Table 8). Although not formally analysed, incomparison to the sample of 30, the means appeared tobe even further in the predicted direction. Again,this rudimentary look at the data suggests ceilingeffects may have been present.142One possible solution to this apparent ceilingeffect would be to start with lower perceivedresponsibility scores in the control condition.However, this type of design may lead to floor effectsin the low responsibility condition. A secondpossible solution would be to increase sample size andsubsequent power. The feasibility of collecting asample of greater than 30 with this population would besomewhat daunting and impractical.Hence, the present design, despite its limitationsappears to be the best possible option. Futureresearch may consider examining solely increases inperceived responsibility or decreases in perceivedresponsibility to clarify whether the smaller effectsfound in this study in the high responsibilitycondition were mitigated by ceiling effects.A second unexpected finding appeared uponanalyzing the data pertaining to Hypothesis 2 whichstated that estimations of the timing of anticipatedthreat would influence the probability of perceivedpanic. This hypothesis was not confirmed. In fact,143Table 8Deviations(15.58)(22.72)(4.06)(29.39)(20.25)( 1.36)(13 .26)(40.81)( .91)(27.01)(26.40)(33.02)(1.07)Hi Responsibility ConditionPerceived responsibility >90/100Means and Standard(n=26)Urge to Check 86.15Discomfort 76.92Responsibility 97.69Likely/Bad 63.54Sev/Bad 78.27Time/Bad 2.58Finish Check 9.65Control/Bad 37.15Panic 2.12Lik/Criticism 78.19Sev/Criticism 78.50Control /Criticism 34.70Time/Criticism 2.46144Pearson Correlations between timing of anticipatedthreat and all other variables collapsed on condition,as well as broken down by condition, failed to turn upany significant association. Examination of thedistribution of scores on timing of anticipated threatproduced no clear indication as to the negativefinding. Scores on this variable appeared to be randomwith no clear pattern emerging.It may be that this question was rather abstractfor those with compulsive checking problems. Theprecise wording on this question was “When would the“bad thing” happen?” For example, if the “bad thing”was that their house would burn down, asking exactlywhen that would occur might seem too removed orabstract for individuals who compulsively check.Compulsive checkers may be too involved in preventingthe negative event from occurring to consider exactlywhen it would occur. Future research may need toconsider finding a different way of probing theimmediacy of anticipated threat. Alternatively, it maybe that the timing of anticipated threat has little todo with perceived panic. Perceived panic scores in145this investigation were more highly correlated withdiscomfort experienced, severity of anticipated harm,likelihood and severity of anticipated criticism.Perhaps the timing of when anticipated harm would occuris of little consequence for those with this disorder.Again, future research is needed to determine whatassociation, if any, there is between immediacy ofanticipated threat and perceived panic.Theoretical ImplicationsThe results of the current investigation providesupport for Salkovskis’ (1985) cognitive model ofobsessions and compulsions. As Salkovskis’ (1985)proposes, the results of this study suggest perceivedresponsibility does appear to play a central role incompulsive checking. The effects appear to be moststriking when perceived responsibility is reduced.Salkovskis’ (1985) position of influencing perceivedresponsibility in order to decrease compulsive ritualsrather than attempting to modify beliefs of harmreceived support from this study. On a more clinicalnote, and as someone who has painstakingly implemented146behaviour therapy, the effects of a reduction inperceived responsibility using cognitive procedureswere dramatic. With several participants, I observed amarked change in overt signs of anxiety such asfidgeting, eye contact and hypervigilance. Some spokeof the freedom they felt in transferring theresponsibility, others were themselves surprised at howrelaxed they felt, still others spoke of having a heavyburden removed from their shoulders albeit temporarily.Future research may need to examine these apparentnoted changes more closely, but as a first attempt Iwas truly amazed at the verbal as well as non—verbalchanges when perceived responsibility was reduced.Results from this exploratory study help provide abasis for making more precise, clear hypotheses andpredictions about how variations in perceivedresponsibility influence compulsive checking. Futureresearch needs to examine whether an asymmetry existsin varying levels of perceived responsibility orwhether the results obtained are due to a ceilingeffect.147The results begin to provide explanation for someof the unexplained findings of earlier work. Forexample, Rachman and Hodgson (1980) noted that theywere unable to provoke the urge to check in theirlaboratory. They correctly concluded that perceivedresponsibility may have been a central feature of thisfailure to provoke checking. The central role ofresponsibility also helps explain the lack ofgeneralization in exposure plus response prevention.It may be that the effects of exposure were limitedbecause we were working around the edges of the fearnetwork rather than getting to the heart of thedisorder, as influencing responsibility appears to do.The results of this study begin to specify howdepression may be linked to compulsive checking.Salkovskis’ (1985) model suggests depression isconnected to compulsive checking but does not specifythe parameters of this relationship. This studysuggests that the links between responsibility andcontrollability, and responsibility and probabilityand severity of harm occurring may play pivotal rolesin modulating mood. I will elaborate on these148relationships later under the heading of cognitivebiases and depression and compulsive checking.MaintenanceCompulsive checking fueled by a heightenedsense of responsibility serves to foster safety throughpreventing danger from occurring. The results of thisstudy suggest compulsive checking may also be a meansof preventing criticism. Seligman (1975) viewssuccessful extinction as a function of how likely theindividual is to sample disconfirmation. For theindividual who compulsively checks, the only way toexpose oneself to disconfirmation is to not perform theavoidance behaviour, or in other words, to refrain fromchecking. Unfortunately, for individuals with thisproblem, disconfirmatory experiences of not anticipatedcriticism, and the absence of danger and harm cannotoccur until the person risks not checking. If one hasa heightened sense of responsibility and a propensityfor compulsive checking the results of this studysuggest several cognitive biases may contribute to the149individual not taking the necessary risk of notchecking leading to discomfirmatory evidence.The Role of Cognitive Biases in MaintenanceCompulsive checking, however futile, isultimately a method of searching for faults orimperfections to prevent future danger, harm orcriticism. Miller and Porter’s (1983) interpretationof increased sense of control with self—blame infersthat logically, there is a direct link betweenresponsibility and controllability. Compulsivecheckers may fail to experience this connection. Thecurrent study provides evidence to support the notionthat compulsive checking may be tied to the faultyassumption that if one is responsible and checksrepeatedly, control and safety will follow. Thisfaulty belief may be open to modification throughcognitive therapy and will be discussed below.Cognitive biases may be central in maintenance ofcompulsive checking through decreasing safety andincreasing a sense of failure and helplessness. Threemain sources of bias have been uncovered in this150investigation. First, that the probability of a badevent occurring changes with the transfer ofresponsibility. For the compulsive checker, decreasesin perceived responsibility result in decreases in theprobability of the bad event. Second, that theseriousness of harm changes with the transfer ofperceived responsibility. A similar pattern as withthe probability of the bad event was found: decreasesin perceived responsibility lead to decreases in theseriousness of harm. Third, that amount of controlexperienced does not change with variations inperceived responsibility. Therefore, regardless ofextent of perceived responsibility, amount of controlexperienced was not influenced.Taken together, these findings suggest a ratherhopeless scenario for the person with compulsiveüchecking: if one takes responsibility, then the beliefis that something bad will be more likely to happenand the bad event will be more severe than if someoneelse had the responsibility. This heightened sense ofsomething bad occurring and increased severity of harmwhen responsibility is taken is coupled with no changes151in perceived controllability. Notably, the sense ofheightened responsibility is not open to question.Overall, it can be stated that the compulsive checkerhas an increased sense of responsibility, likelihood ofharm, and severity of harm but no enhanced sense ofcontrol.Compulsive Checking and DepressionThe biases noted above readily bring to mind theliterature on uncontrollable events and its effects onemotional processing. In particular, Seligman’s (1975)work suggesting that the effects of a history ofuncontrollable events coupled with perceived dangerwill often lead to depression appears to be ofsignificance in this analysis. If one takes intoconsideration the added feature of being highlyresponsible in concert with having little control, thesituation is doubly stacked for experience of failureand helplessness. It seems reasonable that ifcompulsive checkers process negative events with thebiases delineated above, depression may be one of thenatural outcomes. That is, if one feels highly152responsible for danger or harm to others but has littlecontrol over preventing the catastrophic event, a senseof failure and overriding feeling of helplessness willfollow.The high co-morbidity rates between OCD anddepression gives further credence to this formulation(Rachman & Hodgson, 1980). In fact, moderate levels ofdepression are a negative prognostic indicator forbehavioural treatment of OCD (Foa, 1979). Theenveloping sense of hopelessness and helplessness thatis often linked with this problem may undermine themotivational factors and risk taking necessary forsuccessful repeated trials of exposure and responseprevention. Overall, it may be that the cognitivebiases intimately linked to an exaggerated sense ofresponsibility are the critical factors in maintainingcompulsive checking and associated depression.Compulsive Checking and CriticismThis study provides support for the notion thatcompulsive checking may serve as means of preventingdanger, or its emotional counterpart, criticism. More153specifically, we found that the likelihood and severityof anticipated criticism varied with levels ofperceived responsibility for compulsive checkers.Increases in perceived responsibility lead to increasesin the probability of anticipated criticism and in theseverity of anticipated criticism. Decreases inperceived responsibility resulted in decreases in thelikelihood of anticipated criticism and in the severityof anticipated criticism. According to the Just WorldHypotheses (Lerner, 1977) if one were highlyresponsible, had maximum control and something badhappened, the chance that one would be blamed and toextrapolate, criticized by others may increase as wellas the severity of anticipated criticism. The reverseholds true. However, the results of thisinvestigation, point to the paradox of being highlyresponsible and yet experiencing little control. Thispotentially helpless scenario may be thwarted throughcompulsively checking. The result is that thecompulsive checking may protect one from feelinghighly vulnerable to criticism.154Rachman (1975) points to an intriguingrelationship between inflated responsibility andcriticism. He proposed that only acts for which one isdeemed responsible is there likely to be feelings ofguilt or criticism. That is, if the individual whocompulsively checks is not held responsible for the actor the results of the act, then he/or she cannot becriticized for the action. He goes on to say that thereason most checking often occurs at home is that thissituation is one where we would be most likely to havethe highest level of responsibility. In addition itfollows that the fear of criticism is strongest inrelation to close friends and family who are often inour homes. Periods of depression serve to intensifysensitivity to criticism and feelings ofresponsibility.According to Rachman (1975) a potential weaknessof this analysis is that many family members ultimatelycriticize checking behaviour. He provides twoexplanations for the maintenance of checking despitethe presence of criticism. First, punishment of anyactive avoidance behaviour such as compulsive checking155may in some instances led to a increase in themagnitude and resiliency of the avoidance behaviour.Therefore, being criticized may serve to increaserather than decrease the checking. Second, thechecking behaviour may turn into stereotyped, activeavoidance which is relatively independent from itsoriginal source (Rachman, 1975).To summarize, compulsive checking may serve toprovide safety through preventing future disaster orcriticism, Cognitive biases surroundingcontrollability, probability and consequences of anegative event may undermine safety and keepcompulsive checkers in a powerless, depressive state.Limitations of Within Sublect DesignOne of the limitations associated with the choiceof a within subject design is the possibility thatparticipants responses may have changed systematicallyduring the course of repeated testings. This type ofchange, commonly referred to as a practice effect couldmanifest itself through producing either positive (i.e.general improvement) or negative effects (e.g.156fatigue). Only the condition given first is notaffected by practice effects. In order to minimizethis problem in the current study the order ofconditions was counterbalanced. In addition, analyseswere conducted to examine the possible impact of suchpractice effects. Although, results suggest nosignificant influence, the possibility of some errorbeing attributed to this effect cannot be ruled out.A second potential source of systematic varianceis the possibility of differential carry—over effects,which counterbalancing cannot control. In comparisonto practice effects which affect all treatmentconditions equally, carry—over effects produce specificeffects. For example, administering one treatment mayaffect a participants responses on a later conditionone way and on a different condition in another way.In order to reduce the effects of differential carryover effects, time was given between conditions untilapproximate baseline conditions were once againestablished. No indications of carry—over effects werepresent in this investigation but one cannot completelyrule out the potential influence.157Advantages of With-In Sublects DesignOne of the main advantages of a within—subjectsdesign is that each participant serves as her/his owncontrol. Through reducing the error associated witheach condition a direct increase in economy and poweris noted with this type of design. Given the practicaldifficulties inherent in recruiting a sufficient numberof participants who have clinically significantobsessive compulsive problems, a study which couldexamine the main questions of this study but stilleconomize on number of participants was necessary. Abetween—groups design which would require a minimum ofthree groups with roughly twenty participants per groupseemed infeasible with my level of resources. Thecurrent design appeared to provide the means to answersome key questions pertaining to the effects ofresponsibility on checking and also maximize theamount of data that could be obtained with a smallsample size and thus, a within—subject design waschosen.158Strengths and Weaknesses of Self—Report MeasuresThe current investigation used self—reportmeasures almost exclusively. The advantages of suchmeasures are several fold. First, given that anxietyis often defined by how the client feels, a measureasking clients precisely this, is most fitting.However, according to Rachman and Hodgson (1974)anxiety may be best construed as a set of looselycoupled components involving behavioural, cognitive andphysiological elements. In this study only thecognitive component was examined using self—reportmeasures. One of the primary purposes of thisexploratory investigation was to provide a cognitiveanalysis of compulsive checking. Thus, the use ofself—report measures examining the cognitions ofparticipants was warranted. Given the promisingresults of this investigation, it may be fruitful tolook at other facets of responsibility (behavioural,physiological) and the correlations between theseelements in future research.159Another reason that self—report measures werechosen for this investigation had to do with the levelof secrecy inherent in compulsive checking. Behaviouralmonitoring through an observer being present mayincrease self—consciousness to the point that it maysignificantly alter the behaviour (Steketee & Foa,1985). The primary purpose of this exploratoryinvestigation was to manipulate responsibility andmeasure critical dependent measures in the most simpleand straight forward manner. Now that some data hasbeen gathered, future research may want to considerusing a behavioural measure despite the potentialconfounds.Third, self—report measures were chosen becausemany of our questions involved asking clients abouttheir perceptions of future harm and danger — a centralfeature of compulsive checking. Compulsive checkinginvolves preventing some future harm or danger ratherthan compulsive washing which is primarily conducted torestore safety. Behaviourally examining future dangeris logistically difficult. However, assessing thiscomponent is quite simple using self—report measures.160Despite the strengths of self—report measuresseveral limitations are worth mentioning. First, arange of biases on the part of the participant may comeinto play using these measures. According to Kadzin(1985) there is a tendency to agree with items, toendorse values on the far ends of a rating scale, togive qualified responses, and to be inconsistent acrossitems when using self—report measures. In the presentinvestigation, these biases were thought to be equalacross conditions and there was no reason to believethere would be systematic differences betweenconditions. Thus, although this weakness is inherentwith the use of self—report measures, these biasesshould not affect the main hypotheses of this studyinvolving differences between conditions.A second limitation of self—report measures isthat they may result in socially desirable responding.That is, it may be that participants who are aware ofbeing tested would respond differently than they wouldwithout this knowledge. In order to address thispotential source of bias the current study ensuredconfidentiality, and indicated in the instructions that161there was no right or wrong answer to any testquestion. In addition, demand characteristics wereheld constant across conditions. Potential influenceof demand characteristics will be discussed later underthe heading entitled “limitations”.Upon examination of the costs and benefits ofusing self—report measures for this exploratory study,the strengths appeared to outweigh the weaknesses. Thefirst necessary step was to attempt to manipulateresponsibility and assess it as simply and clearly aspossible. With this groundwork being laid it may nowbe timely to add complexity to the assessmentprocedures. Future research may consider using a hostof measures to augment the use of self—report measuresin this investigation including behavioural andphysiological assessmentPossible Sources of a Heightened Sense ofResponsibilityGiven the intriguing findings coming from thisstudy it may be timely to begin to ask some of themajor questions associated with this disorder. How162does compulsive checking develop? What factorsmaintain it? Why is compulsive checking oftenassociated with depression? I will first speculateabout the development of the disorder. As Rachman(1975) noted, discussion of origins of compulsivechecking is purely speculative given that noprospective studies have been conducted. Difficultiesassociated with retrospective data and theimpossibility of verifying most information aredefinite limitations. With this cautionary note inmind, I will briefly discuss some possible sources of aheightened sense of responsibility.Cognitive theories of psychopathology concur thatthe formation of core maladaptive beliefs orassumptions often occur after significant life eventsin childhood. These core assumptions may lay dormantuntil a critical life event or stressor triggers thissemantic network in late adolescence or early adulthood(Salkovskis & Warwick, 1990). I will argue thatcognitive biases uncovered in this study serve tomaintain the high levels of responsibility andassociated depression. Treatment implications arising163out of this position will be then be explored.Finally, limitations of the present study will beconsidered.As Tallis (in press) noted, with very fewexceptions (Rachman, 1976; Leonard, Goldberger,Rapoport, Cheslow, & Swedo, 1990) little considerationhas been given to how specific early experiences mayinfluence later development of maladaptive coreassumptions and compulsive behaviour. He cites caseexamples which suggests a very precise link betweenblurring between thought and action, early life lossand later obsessive compulsive behaviour. Rachman(1993) refers to such phenomena as psychologicalfusion. That is, the patient appears to regard theobsessional thought and the forbidden action as beingmorally equivalent. In the following case examplesTallis (in press) describes how critical early eventsmay lead to later manifestation of the psychologicalfusion often associated with OCD:The S was a 49 year old married,woman, employed as a caterer in an oldpeople’s home. Obsessional symptomsincluded intrusive thoughts, numberrituals, and repeated checking of164household appliances. Typically,checking and ritualizing occurred inresponse to thoughts of beingresponsible for harm coming to others.the S had enormous difficultydistinguishing thought from action. Forexample, she would become extremelydistressed after preparing meals for theelderly, thinking that she may haveunwittingly mixed a poisonous substanceinto their food. These thoughts wereparticularly troublesome when the S wasgiven more responsibility than usual,for example, when taking her wards on anouting. Exploration of the origins ofthis characteristic revealed a criticallearning incident. As a child she hadbeen continually sexually abused by herfather. When she was 15 years of age,she recalled forming an intense wishthat he would “go” or be “taken away”,and prayed to this effect. Within aweek, her father was involved in a rail“accident”. He was instantly killed.There was some suspicion of suicide.This guilt was so unbearable that shereported wanting to die.In later life, she experiencedseveral occasions when she thought about“disasters” which were subsequentlyreported in the media. Although she didnot feel that she was the cause of thesedisasters, they served to reinforce herbelief that mental events and events inthe real world bore a close relationshipto each other.Case BThe S was a 43 year old woman,employed as a training officer in alarge company, married with threechildren. Obsessional symptoms included165intrusive thoughts, repeated checking ofhousehold appliances, and “retracing”journeys in order to investigate whethershe had been the cause of harm toothers. She described having low moodsince childhood. Obsessional checkingbehaviour emerged when she was in herearly 20’s. She was admitted tohospital (aged 43 years) afterattempting suicide. The suicide attemptwas precipitated by an event at work.She was given responsibility for atrainee, who was dismissed because ofmisconduct. The S began to think thatshe had telephoned the trainee’sparents, and informed them of aparticularly embarrassing incident whichthe young man had been party to. Shebecame convinced that her imaginarytelephone call would lead to the youngman being severely punished, andconsidered that he might take his ownlife as a direct result. Her guilt wasso extreme, that she attempted to escapethe aversive mood state by taking anoverdose of drugs with suicidalintention.While exploring the nature ofthought-action fusion during therapy,she volunteered the followinginformation: as a child ofapproximately six years, she prayed toGod that her grandfather would die. Theday following this “death prayer”, hergrandfather suffered a fatal heartattack. The apparent relationshipbetween thought and action was furthersecured at the age of 37, when she had aviolent argument with her mother. The Srarely expressed anger, and attributesher behaviour at the time to increasedemotional sensitivity (it being shortlyafter the birth of her third child).Two weeks after this argument the S’s166mother was diagnosed as having a fatalcancer. Thereafter, obsessionalsymptoms worsened, ultimately leading toher attempted suicide and hospitaladmission. (Tallis, in press)Talus (in press) cogently argues thatspecific early life connections between wishing harm toothers and it actually occurring shortly thereafter,lead in these case illustrations to an exaggeratedsense of responsibility and guilt. He proposes thatthe blurring between thought and action which waspresent in these examples, contributed to thedevelopment of compulsive behaviour. However, theprecise link between responsibility for thought andultimate negative consequence as detailed above may besolely one way to arrive at a stable, high level ofresponsibility. There may be various pathways oravenues to arrive at a heightened sense ofresponsibility. The uncovering of these parallelsystems awaits future research. Only longtitudinalresearch can begin to address these interesting albeitdifficult questions pertaining to etiology. I willbriefly review two models of the socialization ofmoral responsibility in hopes that future research will167begin to integrate these models and perspectives whenconsidering the etiology of a heightened sense ofresponsibility.Bandura’s (1971) social—learning model drawsheavily on instrumental conditioning and imitationlearning in hypothesizing that response patterns arelearned and maintained by conditioning history . Insome cases the learning involves some type of“behavioural deficit” which the individual fails toacquire responses that are adaptive. Bandura sees thistype of failure to learn necessary skills arising frominadequate modelling and reinforcements. He alsoindicates that response patterns may also be as aresult of an inability “to respond discriminately toimportant stimuli” (Bandura, 1969, p.299). Forexample, Bandura suggests that children mayconsistently receive approval (reinforcement) foreating at meal times and consistently receivedisapproval (punishment) for snacking at other times.Through using this type of discriminable stimuli andthe different schedules of reinforcement, the behaviourof snacking is brought under stimulus control.168Further, observing a model who is similar to the childin some way, or important in stature (such as a parent)being rewarded for not snacking may hold tremendouslearning power. Through faulty learning or a break ofpreviously acquired discriminative responses,behaviours that are inappropriate to the situation mayoccur.In the case of someone with a heightened sense ofresponsibility, Bandura’s model would suggest that achild may learn to be overly responsible through beingrewarded for taking on increasing responsibility andthrough punishing instances when the child is notresponsible. Further, a child may learn vicariouslythrough observing a parent who may be responsible orthrough watching friends receive praise or approval forbeing responsible. The development of a non—adaptivesense of responsibility may be related to the child’sinability to discriminate between instances when it isadaptive to be less responsible and instances when itis adaptive to be more responsible. It may be thatfaulty training (inconsistencies) in rewarding thechild when s/he is responsible that is pivotal.169Another pathway may be that the child is punished forbeing irresponsible which may also lead to an enhancedsense of responsibility.A cognitive developmental model of thesocialization of moral responsibility may beextrapolated from the area of moral reasoning.Kohlberg’s (1967;1969) model of the development ofmoral judgment is largely an extention of the earlierwork of Piaget (Piaget and Inhelder, 1969) who heldthat morality could best be construed as a system ofrules for conduct that a child develops from theinfluence of a child’s caregivers, other significantadults and the child’s own experiences.According to Piaget, the interaction ofindividual growth factors and socialexperiences instrumental in the developmentof moral judgment may be distinguished byreference to four successive and invariantstages of emergence. These include a periodof motor development, followed by anegocentric stage, leading to a period ofcooperative effort, and, finally, terminatingin the child’s recognition of moral principlein the establishment of social order”(Cohen, 1976, pp. 177—178).Kohlberg extends upon Piaget’s (Piaget andInhelder (1969) work through providing a more detailed170and complete account of moral development. Overall,Kohlberg proposes three levels and six stages of moraljudgment. Each individual is thought to go through thesame successive stages of development (Kohlberg, 1969).In Level one morality is external to the individual andthere is primarily an obedience and punishmentorientation. Individual standards are not important andobiective responsibility is of importance. In StageTwo morally right action parallels satisfying theperson’s needs and occasionally another’s needs. InLevel Two moral value is viewed as maintainingconventional order and the expectancies of others.Stage 3 is exemplified by the “good girl/boy”orientation with approval seeking and pleasing othersas being most critical. Stage 4 is construed as anauthority and social order maintaining orientation.There is an emphasis on “doing one’s duty” and showingrespect for authority. In Level Three moral valuerests in the self and shared standards, rights andduties. Stage 5 is viewed as a legal orientation.There is some awareness of an arbitrary element inrules. Duty is related to what is specified in the171contract and there is a general avoidance of violatingthe rights of others. In stage 6 individual conscienceand principal are of utmost importance. There is anemphasis on social rules and also to principles ofchoice ultimately residing in a sense of logicaluniversality. There is also a sense of using one’sconscience as a driving force and to mutual respect andtrust.Turning to the issue of a heightened sense ofresponsibility, Kohlberg’s model would suggest thatobjective responsibility or a direct relationshipbetween control and responsibility as suggested in thisthesis as the equivalent of a “healthy” sense ofresponsibility would be construed as the lowest stageand level of moral judgment. Further, this model wouldsuggest that the type of strict sense of conscience,often a defining feature of OCD, would be classified atlevel 3, stage 6, the highest stage and level of moraldevelopment. Thus, treatment focusing on influencingresponsibility through appealing to a logical orobjective sense of responsibility would be theequivalent of bringing the individual back to a lower172stage of moral development—such perils await theunsuspecting clinician.What is evident is that this model does notaccount for the non—adaptiveness of an exaggeratedsense of responsibility and strict conscience as notedwith those with OCD. Kohlberg’s model may be moresuccessful when applied to normal personalitydevelopment rather than the development of emergence ofunhealthy states such as compulsive checking. Inaddition, many researchers have raised seriousquestions about the validity of the stages (eg.,Bandura, 1969; Hoffman, 1970). These learningtheorists suggest that development is not a set ofdiscrete and unchanging stages but rather, thatdevelopment is construed as a dynamic interactionbetween maturational growth and learning.Why then do not all individuals with a heightened senseof responsibility go on to develop compulsive checking?Buffers and Alternate Ways of Coping.First, many adults find more adaptive ways ofcoping. It may be the absence of/or poor alternative173methods of coping which increases the likelihood ofdeveloping compulsive checking. Factors such as havingalternate sources of support and safety to turn to suchas grandparents, teachers, neighbors etc. as a childmay buffer the adult from developing compulsivechecking. Notably, if the child has these resources inher/his environment, he/she may develop healthier earlyrelationships, increasing the likelihood of an enhancedsocial support system as an adult.Second, compulsive checking is seen as one of anumber of non—adaptive ways an adult may cope in anenvironment. Various negative coping strategies suchas extreme aggressiveness, drugs, alcohol, extremepassivity, or overworking may all be used. Forexample, in the case of the family with highexpectations the child may become a overfunctioninghigh achiever who may be prone to burnout, stressillnesses etc.Third, the presence of alternate healthy adultchild relationships where the child is told directlythat s/he is fl responsible for a negative event, and174then followed up through action, may go a long way increating a more equitable sense of responsibility.Fourth, some individuals may find alternate waysof coping such as choosing occupations where aninflated sense of responsibility is beneficial. Mentalhealth workers, teachers, and supervisory positions allimplicitly hold added responsibility. Ability tohandle high levels of responsibility is rewarded inthese occupations.Implications for TreatmentTransferring of ResponsibilityLimitations to the present study are addressedbelow and serve as precautionary reminder of the needto replicate and determine extent of generalizability,before treatment approaches can be employed. Withthese precautions in mind, the present study suggestsa cognitive approach to treatment may be of benefit.As a start, examination of the extent of responsibilityassumed in situations associated with compulsivechecking is necessary. Once a baseline measure of175responsibility is obtained, exploring various pathwaysto sharing/transferring of responsibility for a givensituation may be of value.In the current study, responsibility wastransferred through establishing an agreement betweenthe client and the experimenter. The agreement fortransferring responsibility included acceptingresponsibility for the catastrophic event, for damagesincurred, through limiting the assignment ofresponsibility to one situation, and through adding anyadditional items to the contract that the individualrequested. The only limit was that the experimenterwould not check. Some participants requested that theexperimenter agreed to prevent the catastrophic eventat the first sign of danger. In addition, discussionof fairness in a job share analogy was used attemptingto appeal to their strict moral conscience. Forexample, the behavioural test was discussed as a jobshare asking the client to imagine that it was theexperimenter’s day at the job versus the client’s dayat the job. In addition, clients were asked toimagine changing or switching roles with a less176responsible family member and to assume that role inthe behavioural test. For others, the analogue of a“responsibility holiday” was used. Again imaginaltechniques were employed with clients to envisionpacking their suitcases, going on vacation and leavingbehind all their responsibilities for the given task.The experimenter was to take on the “job” while theywere away.It appeared that the experimenters’ assumingresponsibility for the consequences of the perceivednegative event was an important first step in shiftingresponsibility. That is, many indicated that they wereconcerned that the experimenter take responsibility forthe damages incurred, criticism received etc. Ratherthan working specifically on transferringresponsibility for the negative event itself, whichappeared to a belief more rigidly held on to, most ofthe clinical work was first conducted on assumingresponsibility for the consequences of the event. Oncethe experimenter assumed responsibility for theconsequences of the event, the client’s responsibilityfor the actual negative significantly diminished.177These clinical impressions require further examinationbut overall, the method in this study was to firstshift responsibility for the consequences of the event.The previous rigid hold on responsibility for theactual event then seemed to fall apart.In addition, it appeared that the contract used inthis study was of greater use for those individuals whowere concerned with litigation following the negativeevent. For individuals who were not concerned with thelegalities and more concerned with personalconsequences of the negative event appealing to theirhigh sense of morality regarding fairness was mostuseful. That is, using a starting point of sharingresponsibility where the experimenter was responsiblefor 50% and the participant was responsible for 50% andthen alternating into high or low responsibility basedupon fairness proved to be useful. Generally, thisgroup places high value on being fair and using thisinformation to shift responsibility seemed to be ofbenefit.178Hawton, Salkovskis, Kirk, and Clark (1989)have suggested challenging the extent of responsibilityassumed through using a pie-chart technique. In thismethod, estimates of personal responsibility areobtained, then additional factors contributing to thecatastrophic event are listed. Finally, a pie chart isdrawn and the client is asked to delineate portions ofpie associated with each factor other than personalresponsibility. According to these authors, personalresponsibility is factored in last because the clientwill otherwise resort to the bias of increasedresponsibility for self. This method of reducing theamount of responsibility experienced is currently beingempirically evaluated.In addition, developing a hierarchy ofsituations where clients rate the extent ofresponsibility over the occurrence of a negative eventfrom high levels to low levels may be of benefit.Working at transferring or sharing of responsibilitywith the therapist with low levels items leading tohigher level items of responsibility may be a useful179method. This type of hierarchy ranging from low tohigh awaits empirical testing.The methods described above could be readilyincorporated into a therapeutic context. The firststep may be to transfer responsibility for a givencatastrophic event and its consequences to thetherapist within session. The second goal could be totransfer the responsibility to a trusted companion orspouse, leading to transferring the responsibility toan imagined third party. Empirical validation ofthis approach is necessary.In addition, given that responsibility implicitlyshifts from location to location, home vs. awayexercises could be conducted. Varying the venue may beone way to practically work at shifting levels ofresponsibility. It may be incorporated into treatmentby noting that the probability of a negative event,severity, criticism etc. do not change through shiftingthe venue. Working at changing responsibility byvarying the location may be of benefit.180Cognitive BiasesThe second set of results coming from this study,involve the uncovering of cognitive biases that maycontribute to the maintenance of compulsive checking.Seminal workers such as Beck, Shaw and Emery (1985)have stressed the importance and pervasiveness of aheightened perception of danger across the anxietydisorders. In the present study, this perceptionappears to shift with a transferring of responsibility.Notably, the probability of the occurrence of anegative catastrophic event decreases with decreasesin perceived responsibility. The severity of thecatastrophic event holds a similar pattern. Finally,controllability does not change with shifts inperceived responsibility.As a result of these findings treatment may needto focus not solely on perceptions of dangerousness andperceptions of responsibility but how they work inconcert with controllability. Treatment may need toexplicitly focus on the connection betweenresponsibility and controllability. A therapist could181begin by expressing the impossible task the cliententertains- to be completely responsible with littlecontrol. Then working in collaboration with theclient, situations where the client assumesresponsibility with little control can be examined.Cognitive techniques to diminish or transferresponsibility in areas of low control, as describedbelow, may then be appropriate. Second, examining thefallacy of the probability of a negative event and itsconsequences changing with who is held responsible, maybe of use. It may be effective to employ similarstrategies as used in the treatment of panic (e.g.,Clark, 1989; Barlow, 1989) such as probabilityestimations of the bad event occurring with the addedstep of examining self vs. other probabilities, vanOppen et al. (1992) also discuss the usefulness ofexamining the formation of this double standard.According to these authors the double standard shouldbe made explicit and then suggest exploring avenuessuch as “What makes you different from others so thatdifferent criteria should be employed?”, as well asexamining early experiences which lead to this standard182being developed. Finally, behavioural experiments asemployed in the treatment of panic may also be ofbenefit.LimitationsSeveral limitations of the present study warrantdiscussion. I will discuss the weaknesses of thisstudy from what I consider to be the most important toleast important. First, this study was comprised of asample of volunteers from the community. It may bethat the sample was not representative of allcompulsive checkers. Characteristics other than thoserequired for inclusion in this study (i.e., clinicalseverity, moderate urge to check, significantinterference) may have biased this sample. Forexample, participants in this investigation may not belike other OCD clients in that they were not explicitlyseeking treatment. Most of what we know about thosewith OCD comes from individuals who are requestingtreatment. The individuals in the current study weregiven information about OCD but were not recruited withthe enticement of treatment. They were also given183referral sources if requested. It may be that thoseseeking information but not treatment are different insome way from the general population of thoseexperiencing compulsive problems.Second, the participants in this investigationallowed the experimenter and assistant to come intohis/her home. This ability to let outsiders into one’shome may not be typical of the general population ofthose with OCD. OCD often includes a high level ofsecrecy and in some instances an unwillingness to letothers into their private space for fear that theindividual(s) may intrude upon their highly controlledenvironment. Some unintentioned action such as sittingon a particular chair, touching an object etc. which isassociated with a ritual may result in continuouschecking by the individual. Some reported during thecourse of this study that if particular objects weretouched it may trigger them into hours if not days ofchecking. In addition, replication of this study isnecessary to rule out chance findings.184Third, the participants in this study wereindividuals who had at least a moderate urge to checkin the presence of the experimenter and her assistant.Results of the present study may not generalize tocompulsive checkers who do not experience urges tocheck in the presence of others. Future research isnecessary to determine generalizability.However, it is also important to note that thereare several reasons to believe that the samplerecruited for this investigation was like others withOCD. First, all participants indicated that compulsivechecking significantly interfered with their life.Second, they indicated that they experienced a level ofdistress congruent with a clinical range as a result ofcompulsive checking. Third, scores obtained on theADIS-R and MOC-I were within the range of scoresreported by those seeking treatment at anxiety centers,hospitals etc.In addition, demand characteristics may havecontributed to the findings. The instructions given toparticipants were designed to hold demand constant as185well as maximize the urge to check regardless of thecondition. The segment of the instructions givenacross conditions in which we attempted to do this isas follows:“people who check sometimes worry thatsomething bad will happen or what they dowon’t be perfect. They sometimes worry thatdamage of some kind may occur, that someonewill be hurt, or that they will have to undoany damage or harm. People who have theseworries sometimes feel that something badwill happen or is not perfect when they arein situations like the one you are about tobe in.”However, despite attempting to standardize demandthrough giving these instructions across conditionsparticipants may have picked up on subtle cues and feltthe need to be the “good participant”. Thus, similarto all psychological experiments demand cannot becompletely ruled out in this investigation. However,the very nature of the persistence and intensity ofcompulsive checking and the inability of our treatmentsto produce significant reductions in checking suggeststhat the results obtained are not likely to be theresult of demand.186If demand is in fact responsible for the reportedfindings, we need to capitalize upon these effects in atherapeutic setting. That is, if demand produced thesignificant reduction in compulsive checking as notedin this investigation, we need to investigate itspotential in a systematic fashion.A further limitation of this study is that littleis known about methods of transferring responsibilityand its critical factors. We used a conglomerate oftechniques to shift responsibility in the currentinvestigation. Future research may begin to examinewhich factors are of importance in shiftingresponsibility.Another potential source of bias was that we useda conglomerate of techniques to influenceresponsibility. With some individuals some techniquesappeared to have an immediate effect and with othersthe same technique appeared to be less successful. Inorder to ensure that there were no differences betweengroups in use of the techniques all individualsreceived all of the techniques. Future research will187need to address which techniques were essential andperhaps which techniques were superfluous.Last, this study looked at single trials of shiftsin responsibility and short term consequences. We donot know whether the manipulation of responsibilityemployed herein would be successful on repeated trials,over a longer period of time or for long—termconsequences. Again, these elements await empiricalexamination.ContributionsThis study is the first empirical based study toexamine the influence of perceived responsibility onmeasures critical to compulsive checking. Itsstrengths include the use of a community based samplewith participants whose checking is of clinicalseverity and use of an experimental design to examinethe effects of perceived responsibility on checking.In addition, it may be argued that it has high externalvalidity given the study was conducted in the homes ofparticipants and focused on situations which weretailored to individual patterns of checking.188ConclusionsTo summarize, enhanced sense of responsibilityappears to be causally connected to compulsivechecking. Various pathways to arriving at a heightenedsense of responsibility have been discussed. Cognitivebiases which may serve to reduce safety and maintain adepressive state have been explored. Finally, clinicalimplications of the current study as well itslimitations have been considered.Prior to this study, authors such as Rachman andHodgson (1980) and Salkovskis (1985) discussed thepossibility of responsibility being linked to OCD.However, little empirical work had been done tosubstantiate these theoretical speculations. With thisinvestigation we are now much clearer thatresponsibility plays an important role in compulsivechecking. This study will hopefully open the door toan area of research which should be pursued vigorously.189REFERENCESAbramson, L., Seligman, M., & Teasdale, J. (1978).Learned helplessness in humans: critique andreformulation. Journal of Abnormal Psychology,Z, 49—74.Affleck, G., Allen, D., McGrade, B., & McQueeney, M.(1982). Maternal attributions at hospitaldischarge of high risk infants. American Journalof Mental Deficiency, , 575-580.Akhtar, S., Wig, N., Verma, V., Pershad, D., & Verma,S. (1975). A phenomenological analysis ofsymptoms in obsessive—compulsive neurosis.British Journal of Psychiatry, ].1Z, 342-348.Alexander, G., DeLong, M., & Strick P. (1988).Parallel organization of functionally segregatedcircuits linking basal ganglia and cortex. AnnualReview of Neurosciences, 9, 357—381.American Psychiatric Association (1980). Diagnosticand statistical manual of mental disorders (3rded.). Washington, DC: Author.American Psychiatric Association (1987). Diagnosticand statistical manual of mental disorders (3rded., rev). Washington, DC: APA.Bandura, A. (1968). A social learning interpretationof psychological dysfunctions. In P. London andD. Rosenhan (Eds.), Foundations of abnormalpsychology. New York: Holt, Rinehart, andWinston.190Bandura, A. (1969). Social learning of moraljudgments. Journal of Personality and SocialPsycholov, j], 275-279.Bandura, A. (1971). Social learning theory.Morristown, N.J.: General Learning Press.Barlow, D. (1987). The classification of anxietydisorders. In G.L. Tischler (Ed.), Diagnosis andclassification in psychiatry: A criticalappraisal of DSM—III. Cambridge, England:Cambridge University Press.Barlow, D. (1988). Anxiety and its disorders. NewYork: Guilford Press.Barlow, D., Cohen, A., Waddell, M., Vermilyea, B.,Klosko, J., Blanchard, E. & Di Nardo, P. (1984).Panic and generalized anxiety disorders; Natureand treatment. Behavior TheraDy, j., 431-449.Barlow, D., Craske, M., Cerny, J. & Kiosko, J. (1989).Behavioral treatment of panic disorder. BehaviorTherapy, 2, 261—282.Beck, A. (1964). Thinking and depression: Theoryand therapy. Archives of General Psychiatry, jQ,561—571.Beck, A. (1967). Depression: Clinical, experimental,and theoretical aspects. New York: Hoeber.Beck, A., Epstein, N., & Harrison, R. (1983).Cognitions, attitudes and personality dimensionsin depression. British Journal of CognitivePsychotherapy.,1-16.191Beck, A., Rush, J., Shaw, F., & Emery, G. (1979).Cognitive therapy of depression. New York:Guilford.Becker, J., Skinner, L., Abel, G., Howell, J., & Bruce,L. (1982). The effects of sexual assault on rapeand attempted rape victims. Victimologv: AnInternational Journal, 2, 106-113.Beech, H. (1974). Obsessional states. London: Methuen.Black, A. (1974). The natural history of obsessionalneurosis. In H. R. Beech (Ed.), Obsessionalstates. London: Methuen.Blackburn, I., Bishop, S., Glen, A., Whalley, L. &Christie, J. (1981). The efficacy of cognitivetherapy in depression: A treatment trial usingcognitive therapy and pharmacotherapy, each aloneand in combination. British Journal ofPsychiatry, j, 181—189.Boulougouris, J. (1977). Variables affecting thebehaviour modification of obsessive—compulsivepatients treated by flooding. In J. Boulougouris,& A. Rabavilas (Eds.), The treatment of phobic andobsessive—compulsive disorders. Oxford: PergamonPress.Boulougouris, J., Rabavilas, A., & Stefanis, C. (1977).Psychophysiological responses in obsessive—compulsive patients. Behaviour Research andTherapy, j, 221—230.Breger, L. & McGaugh, J. (1965). Critique andreformulation of “learning—theory” approaches topsychotherapy and neurosis. PsychologicalBulletin, 63, 338—358.192Briere, J. (1992). Child abuse trauma: Theory andtreatment of lastinc effects. Newbury, CA: Sage.Breitner, C. (1960). Drug therapy in obsessionalstates and other psychiatric problems. Diseasesof the Nervous System, L, 31—35.Bulman, R., & Wortman, C. (1977). Attributions ofblame and coping in the “real world”: Severeaccident victims react to their lot. Journal ofPersonality and Social Psychology, 3, 351-363.Burgess, A. & Holmstrom, L. (1974a). Rape traumasyndrome. American Journal of Psychiatry, 11,981—985.Burgess, A. & Holmstrom, L. (1974b). Rape: Victims ofcrisis. Bowie, Md.: Robert J. Brady.Cameron, N. (1947). The psychology of behaviordisorders. Cambridge, Mass: Houghton-Mifflin.Carr, A. (1974). Compulsive neurosis: a review of theliterature. Psychological Bulletin, 311-318.Chambless, D., Caputo, G., Bright, P., & Gallagher, R.(1984). Assessment of fear in agoraphobics: Thebody sensations questionnaire and the agoraphobiccognitions questionnaire. Journal of Consultingand Clinical Psychology, , 1090-1097.Chambless, D. and Goldstein, A. (1982). Aaoraphobia:Multiple perspectives on theory and treatment.New York: Wiley.193Clark, D. (1986). A cognitive approach to panic.Behaviour Research and Therapy, 24, 461-470.Cohen, S. (1976). Social and personality developmentin children. New York: Macmillan.Courtois, A. (1988). Healing the incest wound. Adultsurvivors in therapy. New York: W.W. Norton &Company.Coyne, J. & Gotlib, I. (1983). The role of cognitionin depression: a critical appraisal.Psychological Bulletin, , 472-505.De Young, M. (1982). Sexual victimization of children.Jefferson, NC: McFarland.De Ruiter, C., Rijken, H., Garssen, B., & Kraaimaat, F.(1989). Breathing retraining, exposure and acombination of both, in the treatment of panicdisorder with agoraphobia. Behaviour Research andTherapy, 21, 647—655.Dickinson, A. (1987). Animal conditioning and learningtheory. In H.J. Eysenck, & I. Martin (Eds.),Theoretical foundations of behaviour therapy (pp.45-61). New York: Plenum Press.Di Nardo, P., Barlow, D., Cerny, J., Vermilyea, B.,Vermilyea, J, Himadi, W., & Waddell, M. (1985).Anxiety disorders interview schedule— revised(ADIS-R). Albany, NY: Phobia andAnxiety Disorders Clinic, State University of NewYork at Albany.Di Nardo, P., O’Brien, G., Barlow, D., Waddell, M. &Blanchard, E. (1983). Reliability of DSM-III194anxiety disorder categories using a new structuredinterview. Archives of General Psychiatry, 4Q,1070—1074.Dollard, J. & Miller, N. (1950). Personality andpsychotherapy: An analysis in terms of learning.thinking and culture. New York: McGraw-Hill.Dowson, J. (1977). The phenomenology of severeobsessive—compulsive neurosis. British Journal ofPsychiatry, 75-77.Edwards, S., & Dickerson, M. (1987) Intrusiveunwanted thoughts: a two stage model of control.British Journal of Medical Psychology, , 317-328.Emmelkaiup, P. (1982). Phobic and obsessive-compulsivedisorders. New York: Plenum Press.Emmelkamp, P. (1988, September). Cognitive andrelational factors in the treatment of anxietydisorders. Paper presented at the World Congressof Behaviour Therapy, Edinburgh.Emmelkamp, P. & Rabbie, 5. (1983, July). Four yearfollow—up of obsessive compulsive disorder afterpsychological treatment. Paper presented to WPA,Vienna, Austria.Emmelkamp, P., Visser, S., and Hoekstra, R. (1988)Cognitive therapy vs. exposure in vivo in thetreatment of obsessive—compulsives. CognitiveTherapy and Research, j., 103-114.England, S., and Dickerson, M. (1988). Intrusivethoughts: unpleasantness not the major cause of195uncontrollability. Behaviour Research andTherapy, , 277—279.Eysenck, H. (1979). The conditioning model ofneurosis. The Behavioral and Brain Sciences,155—199.Finkeihor, D. & Browne, A. (1986). The impact of childsexual abuse: A review of the research.Psychological Bulletin, , 66-77.Foa, E. (1979). Failure in treating obsessive—compulsives. Behaviour Research and Therapy, jZ,169—176.Foa, E. & Goldstein, A. (1978). Continuous exposureand complete response prevention in obsessivecompulsive neurosis. Behavior Therapy, 2, 821—829.Foa, E., Stekett, G., Grayson, J. & Doppelt, H. (1983).Treatment of obsessive—compulsives: when do wefail? In E. Foa & P. Emmelkamp (Eds.) Failures inBehaviour Therapy. New York: Wiley.Gorman, J., Liebowitz, N., Fyer, A., Dillon, D.,Davies, S., Stein, J., & Klein, D. (1988).Lactate infusion in obsessive compulsive disorder.American Journal of Psychiatry, 142, 864-866.Graham, M., Thompson, S., Estrada, N., & Yonekura, N.(1987). Factors affecting psychologicaladjustment to a fetal demise. Journal ofObstetrics and Gynecolocy. 157, 254 - 257.Heider, F. (1958). The psychology of interpersonalrelations. New York: Wiley.196Herman, J. (1992). Trauma and recovery. New York:Harper Collins.Hewitt, P. & Flett, G. (1990). Perfectionism in theself and social contexts: conceptualization,assessment, and association with psychopathology.Journal of Personality and Social Psychology. Inpress.Hodgson, R. & Rachman, S. (1972). The effects ofcontamination and washing in obsessional patients.Behaviour Research & Therapy, jQ, 111-117.Hodgson, R., & Rachman, S. (1977). Obsessional-compulsive complaints. Behaviour Research andTherapy, j, 389-395.Hoffman, N. (1971). Father absence and consciencedevelopment. Developmental Psychology, 4, 400-406.Hollander, E., DeCaria, C., Liebowitz, M. (1989).Biological aspects of obsessive compulsivedisorder. Psychiatry Annals, , 80-87.Hornsveld, R., Kraaimaat, F. & Van dam—baggen, R.(1979). Anxiety/discomfort and handwashing inobsessive—compulsive and psychiatric controlpatients. Behaviour Research & Therapy, 12., 223-228.Hursch, C. (1977). The trouble with rape. Chicago:Nelson Hall.197Janoff—Bulman, (1979). Characterological versusbehavioral self—blame: Inquiries into depressionand rape. Journal of Personality and SocialPsycholocw, j, 1798—1809.Jenike, M., Baer, L., & Minichiello, W. (Eds.). (1986).Obsessive—compulsive disorders: Theory andmanagement. Littleton, MA: PSG.Jenike, M., Baer, L., & Minichiello, W. (Eds.). (1990).Obsessive—compulsive disorders: Theory andmanagement, 2nd edition. Littleton, MA: PSG.Kazdin, A. and Wilson, G. (1978). Evaluation ofbehaviour therapy: Issues, evidence, and researchstrategies. Cambridge, Mass.: Ballinger.Kelley, H. (1971). Attribution in social interaction.Morristown, N.J.: General Learning Press.Kiecolt—Glaser, J. & Williams, D. (1987). Self—blame,compliance, and distress among burn patients.Journal of Personality and Social Psycholoav, ,187—193.Klosko, J., Barlow, D., Tassinari, R. & Cerny, J.(1990). A comparison of alprazolam and behaviortherapy in treatment of panic disorder. Journalof Consulting and Clinical Psychology, , 77-84.Kohlberg, L. (1967). Moral and religious education andthe public schools: a developmental view. In T.Sizer (Ed.), Religion and public education.Boston: Houghton Miffun.198Kohlberg, L. (1969). Stage and sequence: Thecognitive—developmental approach to socialization.In DA. Goslin (Ed.), Handbook of socializationtheory and research. Chicago: Rand McNally.Leonard, H., Goldberger,E., Rapoport, J., Cheslow,D., &Swedo,S. (1990). Childhood rituals: normaldevelopment or obsessive—compulsive symptoms?Journal of the American Academy of Child andAdolescent Psychiatry, , 17—23.Leonard, H., Swedo, S., Rapoport, J., Koby, E., Lenane,N., Cheslow, D., & Hamburger, S. (1989).Treatment of obsessive compulsive disorder withclomipramine and desipramine children andadolescent: A double blind crossover comparison.Archives of General Psychiatry, 4k., 1088-1092.Lerner, M. (1977). The justice motive: Somehypotheses as to its origins and forms. Journalof Personality, 4, 1—52.Lerner, N, & Miller, D. (1978). Just world researchand the attribution process: Looking back andahead. Psychological Bulletin, , 1030—1051.Long, 5. (1986). Guidelines for treating youngchildren. In McFarlane (Eds.)Mackintosh, N. (1983). Conditionina and associativelearning. New York: Oxford University Press.McFall, M., & Wollersheim, J. (1979). Obsessive—compulsive neurosis: a cognitive—behavioralformulation and approach to treatment. CognitiveTherapy and Research, 333-348.199MacLennon, R., & Jackson, D. (1985). Accuracy andconsistency in the development of socialperception. Developmental Psychology, 21, 30-36.Margraf, J., Barlow, D., Clark, D., & Telch, M. (1993).Psychological treatment of panic: Work inprogress on outcome, active ingredients, andfollow—up. Behaviour Research and Therapy, 31, 1—8.Margraf, J., Ehiers, A., & Roth, W. (1986). Biologicalmodels of panic disorder and agoraphobia: areview. Behaviour Research & Therapy, 24(51, 553-567.Marks, I. (1987). Fears. phobias. and rituals.Oxford: Oxford University Press.Marks, I., Hodgson, R., & Rachman, S. (1975).Treatment of chronic obsessive compulsive disordertwo years after invivo exposure. British Journalof Psychiatry, 122, 349-364.Mawson, D., Marks, I., & Ramm, L. (1982). Clomipramineand exposure for chronic obsessive—compulsiverituals: III. Two year follow-up and furtherreadings. British Journal of Psychiatry. 140, 11-18.Meyer, C. & Taylor, S. (1986). Adjustment to rape.Journal of Personality and Social Psychology, Q,1226—1234.Meyer, V. & Levy, R. (1973). Modification of behaviorin obsessive—compulsive disorder. In H.E. Adams &P. Unikel (Eds.), Issues and trends in behaviortherapy. Springfield, Il: Charles C. Thomas.200Meyer, V., Levy, R., & Schnurer, A. (1974). Thebehavioral treatment of obsessive—compulsivedisorder. In H. Beech (Ed.), Obsessional states.London: Methuen.Michela, J., & Wood, J. (1986). Causal attributions inhealth and illness. In P. C. Kendall (Ed.),Advances in cognitive behavioral research andtherapy, , 179-234.Miller, D. & Porter, C. (1983). Self—blame in victimsof violence. Journal of Social Issues, 39, 139—152.Milner, A., Beech, H., Walker, V. (1971). Decisionprocesses and obsessional behavior. BritishJournal of Social and Clinical Psychology, jQ, 88—89.Moulton, J., Sweet, D., Temoshok, L, & Mandel, J.(1987). Attributions of blame and responsibilityin relation to distress and health behavior changein people with AIDS and AIDs-related complex.Journal of Applied Social Psychology, 17, 493-506.Mowrer, 0. (1939). A stimulus—response analysis ofanxiety and its role as a reinforcing agent.Psychological Review, 4, 553-565.Murphy, G., Simons, A., Wetzel, R. & Lustman, P.(1984). Cognitive therapy and pharniacotherapy:Singly and together in the treatment ofdepression. Archives of General Psychiatry, ill33—41.201Myers, J., Weissman, M., Tischler, C., Hoizer, C., III,Orvaschel, H., Anthony, J., Boyd, J., Burke, J.,Jr., Kramer, M., & Stoltzman, R. (1984). Six-month prevalence of psychiatric disorders in threecommunities. Archives of General Psychiatry, ill959—967.Ney, P., Moore, C., McPhee, J., & Trought, P. (1985).Child abuse: a study of the child’s perspective.Child Abuse and Neglect,.1Q, 511-518.Norton, F., Harrison, B., Hauch, J., & Rhodes, L.(1985). Characteristics of people with infrequentpanic attacks. Journal of Abnormal Psychology,, 216—221.Ost, L. (1988). Applied relaxation vs. progressiverelaxation in the treatment of panic disorder.Behaviour Research and Therapy, , 13—22.Perry, H., & Perry, S. (1959). The schoolhousedisasters: Family and community as determinantsof the child’s response to disasters. Washington:National Academy of Sciences.Persons, J. and Foa, E. (1984) Processing of fearfuland neutral information by obsessive—compulsives.Behaviour Research and Therapy, , 259-265.Peterson, C., & Seligman, M. (1984). Causalexplanation as a risk factor for depression:theory and evidence. Psychological Review:347—374.Piaget, J. & Inhelder, B. (1969). The psychology ofthe child. New York: Basic Books.202Rachman, S. (1974). The meanings of fear.Middlesex: Penguin Books.Rachman, S. (1975). Obsessional-compulsive checking.Behaviour Research and Therapy, j4, 269-277.Rachman, S. (1983). Obstacles to the sucessfultreatment of obsessions. In In E. Foa & P.Emmelkamp (Eds.) Failures in Behaviour Therapy.New York: Wiley.Rachman, S. (1992) Personal communication.Rachman, S. (1992). Obsessions, responsibility andguilt. Behaviour Research and Therapy.Rachman, S. and Hodgson, R. (1980) Obsessions andcompulsions. Englewood Cliffs, N.J: Prentice-Hall.Rachman, S. and Maser, J. (1988). Panic:Psychological perspectives. Hilldale,N.J. :Lawrence Erlbaum.Rachman, S., and Wilson, G. (1980). The effects ofpsycholoaical therapy (2nd ed.). Oxford: PergamonPress.Rapoport, J. (1991). Recent advances in obsessive—compulsive disorder. Neuropsychopharmacology, ,1—10.Reed, G. (1969a). Under-inclusion-a characteristic ofobsessional personality disorder 1. BritishJournal of Psychiatry, 781-785.203Reed, G. (1969b). Under-inclusion-a characteristic ofobsessional personality disorder 2. BritishJournal of Psychiatry, 115, 787-790.Reed, G. (1977). Obsessional cognition: performanceon two numerical tasks. British Journal ofPsychiatry, jJ.Q, 184—185.Rescorla, R. (1980). Pavlovian second—orderconditioning. New Jersey: Erlbaum.Rescorla, R. (1988). Pavlovian conditioning. AmericanPsychologist, 43, 151—160.Reynolds, M & Salkovskis, P. (1992). Comparison ofpositive and negative intrusive thoughts andexperimental investigation of the differentialeffects of mood. Behaviour Research & Therapy,Q, 273—281.Robertson, R. (1979). A controlled investigation ofthe treatment of obsessive compulsive disorders.MD dissertation, Middlesex Hospital, London.Rholes, W., Jones, M., & Wade, C. (1988). Children’sunderstanding of personal dispositions and itsrelation to behavior. Journal of ExperimentalChild Psycholoay, 4, 1-17.Rholes, W., & Ruble, D. (1984). Children’s understand-ing of dispositional characteristics of others.Child Development, , 550—560.Roper, G., and Rachman, S. (1976). Obsessionalcompulsive checking: Replication and development.Behaviour Research and Therapy, , 25-32.204Roper, G., Rachman, S., and Hodgson, R. (1973). Anexperiment on obsessional checking. BehaviourResearch and Therapy, 271-277.Rosen, M. (1975). A dual model of obsessionalneurosis. Journal of Consulting and ClinicalPsychology, 4_3, 453-459.Rotenberg, K. (1982). Development of characteristicconstancy of self and others. Child Development,53, 505—515.Rush, A., Beck, A., Kovacs, M., & Hollon, S. (1977).Comparative efficacy of cognitive therapy andpharmacotherapy in the treatment of depressedoutpatients. Cognitive Therapy and Research,.,17—37.Salkovskis, P. (1989) Cognitive—behavioural factorsand the persistence of intrusive thoughts inobsessional problems. Behaviour Research andTherapy,.Z, 677-682.Salkovskis, p. (1985) Obsessional-compulsive problems:A cognitive—behavioural analysis. BehaviourResearch and Therapy. 571-583.Salkovskis, P. and Dent H. (1989). Intrusive thoughts,impulses and imagery: cognitive and behaviouralaspects.Salkovskis, P. and Warwick, H. (1988). Cognitivetherapy of obsessive—compulsive disorder. In C.Perris, I. Blackburn, & H. Perris (Eds). Theoryand practice of cognitive therapy. Heidelberg:Springer-Verlag.205Sanavio, E., & Vidotto, G. (1985). The components ofthe Maudsley obsessional—compulsive questionnaire.Behaviour Research and Therapy, , 659-662.Shaver, K. (1985). The attribution of blame:causality, responsibility, and blameworthiness.New York: Springer-Verlag.Shaw, B. (1977). comparison of cognitive therapy andbehavior therapy in the treatment of depression.Journal of Consulting and Clinical Psychology, 4,543—551.Sher, K., Frost, R. and Otto, R. (1983). Cognitivedeficits in compulsive checkers: an exploratorystudy. Behaviour Research and Therapy, 21, 357-363.Silver, R. (1982). Coping with an undesirable lifeevent: A study of early reactions to physicaldisability. Unpublished doctoral dissertation,Northwestern University, Evanston, Ii.Steele, B. & Pollock, D. (1968). A psychiatric studyof parents who abuse infants and small children.In C.H. Kempe & R.E. Helfer (Eds.), The batteredchild (2nd ed.). Chicago: University of ChicagoPress.Steketee, G., & Foa, E. (1985). Obsessive compulsivedisorder. In D.H. Barlow (Ed.), Clinical handbookof psycholoaical disorders. New York: GuilfordPress.Steketee, G., Grayson, J. & Foa, E. (1985). Obsessivecompulsive disorder: Differences between washersand checkers. Behaviour Research and Therapy, ,197—201.206Steiner, J. (1972) A questionnaire study of risk-taking psychiatric patients. British Journal ofMedical Psycholoy, 365—374.Stern, R., & Cobb, J. (1978). Phenomenology ofobsessive—compulsive neurosis. British Journal ofPsychiatry, JJ.2., 233-239.Summit, R. (1983). The child sexual abuseaccommodation syndrome. Child Abuse & Neglect, 2,177—193.Tabachnick, B. & Fidell, L.(1988). Using multivariatestatistics. New York: Harper & Row.Talus, F. (in press). Obsessions, responsibility andguilt: two case reports suggesting a common andspecific etiology. Behaviour Research andTherapy.Taylor, 5. (1983). Adjustment to threatening events:A theory of cognitive adaptation. AmericanPsychologist, 3, 1161—1173.Taylor, S., Lichtman, R., & Wood, J. (1984).Attributions, beliefs about control, andadjustment to breast cancer. Journal ofPersonality and Social Psychology, 4, 489-502.Teasdale, J., Fennell, M., Hibbert, G. & Amies, P.(1984). Cognitive therapy for major depressivedisorder in primary care. British Journal ofPsychiatry, 14.4, 400—406.Tennen, H., Affleck, G., Allen, D., McGrade, B., &Ratzan, 5. (1984). Causal attributions and coping207with insulin-dependent diabetes. Basic andApplied Social Psychology, 5, 131—142.Tennen, H., Affleck, G., & Gershman, K. (1986). Self—blame among parents of infants with perinatalcomplications: The role of self—protectivemotives. Journal of Personality and SocialPsychology, , 690—696.Turnquist, D., Harvey, J., & Andersen, B. (1988).Attributions and adjustment to life-threateningillness. British Journal of Clinical Psychology.az, 55—65.Thyer, B., Curtis, G., & Fechner, S. (1984). Fear of‘criticism in not specific to obsessive—compulsivedisorder. Behaviour Research and Therapy. 22, 77—80.Thyer, B., Papsdorf, J., Davis, R., & Vallecorsa, S.(1984). Autonomic correlates of the subjectiveanxiety scale. Journal of Behavior Therapy andExperimental Psychiatry, j, 73-85.Timko, C., & Janoff-Bulman, R. (1985). Attributions,vulnerability, and psychological adjustment: Thecase of breast cancer. Health Psychology, 4, 521—544.Tsai, M., & Wagner, N. (1978). Therapy groups forwomen sexually molested as children. Archives ofSexual Behavior, 2, 417-421.Turner, R., Steketee, G., and Foa, E. (1979) Fear ofcriticism in washers, checkers and phobics.Behaviour Research and Therapy, jZ, 79-81.208Walster, E. (1966). Assignment of responsibility foran accident. Journal of Personality and SocialPsychology, 73-79.Weiner, B. (1985). “Spontaneous” causal thinking.Psychological Bulletin, , 74-84.Wernicke, C. (1900). Grundriss der Psychiatrie.Leipzig: Verlag von Georg Thieme.Witenberg, S., Blanchard, E., Suls, J., Tennen, M.,McCoy, G., & McGoldrick, M. (1983). Perceptionsof control and causality as predictors ofcompliance and coping in hemodialysis. Basic andApplied Social Psychology, 4, 319-336.Wortman, C. (1976). Causal attributions and personalcontrol. In J. H. Harvey, W.J. Ickes, & R. Kidd(Eds.), New directions in attribution research(pp. 23—51), Hilisdale, N.J.: Erlbaum.Wegner, D., Schneider, D., Carter, S. and White, T.(1987). Paradoxical effects of thoughtsuppressing. Journal of Personality and SocialPsychology, , 5-13.Weiss, K., & Weiss, S. (1975). Victimology and thejustification of rape. In I. Drapicin & E. Viano(Eds.), Exploiters and exploited: Vol. 5.Victiiuologv: A new focus. Lexington, Mass:Lexington Books.Wilson, G. (1982). Adult disorders. In G. Wilson &C. Franks (Eds.), Contemporary Behaviour Therav.New York: Guilford Press.Wortman, C. (1976). Causal attributions and personalcontrol. In J. H. Harvey, W.J. Ickes, & R.F. Kidd (Eds.), New directions in attributionresearch, (pp. 23-51). Hilisdale, NJ: ErlbaumZohar, J. (1989, December). Specificity ofclomipramine response in obsessive compulsivedisorder. Poster presented at the Annual Meetingof the American College of Neuro—psychopharmacology, Maui, Hawaii.209APPENDICES210APPENDIX AVERBAL ANALOGUE SCALES211212Pre____ PostTask #____Task Description :_____________________________Con_ Hi_ Lo______ _______ _ __ _ __SS No.1. How strong is your urge to check?Using a scale ranging from 0% to 100% where 0% is not at allstrong and 100% is extremely strong.(post— keeping in mind the agreement you have made).2. How much discomfort are you currently experiencing?Using a scale ranging from 0% to 100% where 0% is nodiscomfort and 100% is extreme discomfort.(post— keeping in mind the agreement you have made).3. If something bad happened right now or something was notperfect, how responsible would you feel?Using a scale ranging from 0% to 100% where 0% is notat all responsible and 100% is completely responsible.(post— keeping in mind the agreement you have made).4. What is the likelihood that something bad would happen?Using a scale ranging from 0% to 100% where 0% is notat all likely and 100% is extremely likely.(post— keeping in mind the agreement you have made).5. If something bad happened, how bad would it be?Using a scale ranging from 0% to 100% where 0% is notat all bad and 100% is extremely bad.(post- keeping in mind the agreement you have made).2136. When would something bad happen?(post— keeping in mind the agreement you have made).— seconds— minutes hours days weeks _months7. How long will it take you to finish checking (insert task)?(before you could leave the house, area etc.)(post— keeping in mind the agreement you have made).______minutes7b. How much control do you have over something bad happening?Using a scale ranging from 0% to 100% where 0 is no control and100 is complete control. (post - keeping in mind the agreementyou have made).%We define a panic attack as an INTENSE RUSH OF FEAR/ANXIETY OR AFEELING OF IMPENDING DOOM.8. Are you panicking? (please check)(post— keeping in mind the agreement you have made).Yes No Close to Panic_____9. Given that you are not able to check, what is the worst thingthat could happen?(post— keeping in mind the agreement you have made).21410. If (insert task) was not perfect, what is the likelihoodthat you would be criticized?Using a scale ranging from 0% to 100% where 0% is not atall likely and 100% is extremely likely.(post— keeping in mind the agreement you have made)._______%11. If (insert task) was not perfect, when would you becriticized?(post— keeping in mind the agreement you have made)._seconds _minutes _hours _days _weeks _months12. If you were criticized, how bad would it be?Using a scale ranging from 0% to 100% where 0% is notat all bad and 100% is extremely bad.(post— keeping in mind the agreement you have made).l2b. How much contro do you have over being criticized if(insert task) is not perfect? Using a scale ranging from 0% to100% where 0 is no control and 100 is complete control. (post -keeping in mind the agreement you have made)._______%215Pre____ PostTask #_____Task description:_________________________Con_Hi_Lo_______ ________ ______SS No.1. How strong is your urge to clean?Using a scale ranging from 0% to 100% where 0% is not atall strong and 100% is extremely strong.(post— keeping in mind the agreement you have made)__ _ _ %2. How much discomfort are you currently experiencing?Using a scale ranging from 0% to 100% where 0% is nodiscomfort and 100% is extreme discomfort.(post— keeping in mind the agreement you have made)3. If something bad happened right now/or was not perfect, howresponsible would you feel?Using a scale ranging from 0% to 100% where 0% is not atall responsible and 100% is completely responsible.(post— keeping in mind the agreement you have made).4. What is the likelihood that something bad would happen?Using a scale from 0% to 100% where 0% is not at all likelyand 100% is extremely likely.(post— keeping in mind the agreement you have made)._ _ __ %5. If something bad happened, how bad would it be?Using a scale ranging from 0% to 100% where 0% is not at allbad and 100% is extremely bad.(post— keeping in mind the agreement you have made).2166. When would something bad happen?(post- keeping in mind the agreement that you have made).seconds — minutes hours _days weeks _months7. How long will it take you to finish cleaning (insert task)?(before you could leave the house, area etc.)(post— keeping in mind the agreement you have made).minutes7b. How much control do you have over something bad happening?Using a scale ranging from 0% to 100% where 0 is no control and100 is complete control. (post - keeping in mind the agreementyou have made)._______%We define a panic attack as an INTENSE RUSH OF FEAR/ANXIETY OR AFEELING OF IMPENDING DOOM.8. Are you panicking? (please check)(post— keeping in mind the agreement you have made).Yes___No Close to Panic_____9. Given that you are not able to clean, what is the worst thingthat could happen?(post— keeping in mind the agreement you have made).21710. If (insert task) was not perfect what is the likelihood thatyou would be criticized?Using a scale ranging from 0% to 100% where 0% is not at alllikely and 100% is extremely likely.(post— keeping in mind the agreement you have made)._______%11. If (insert task) was not perfect when would you becriticized?(post— keeping in mind the agreement you have made)._seconds _minutes hours _days _weeks _months12. If you were criticized, how bad would it be?Using a scale ranging from 0% to 100% where 0% is not at allbad and 100% is extremely bad.(post— keeping in mind the agreement you have made).12b. How much contro do you have over being criticized if(insert task) is not perfect? Using a scale ranging from 0% to100% where 0 is no control and 100 is complete control. (post —keeping in mind the agreement you have made).APPENDIX BCONSENT FORM218219VOLUNTEERS NEEDEDRESEARCH ON COMPULSIVE CHECKING AND CLEANINGWe are conducting research on compulsive checking andcleaning and would value your participation. You can assist usin finding out more about compulsive checking and cleaning bysimply agreeing to volunteer approximately three to four hours ofyour time. If you are willing to participate, we will ask you totake part in a short interview. During this interview we willfirst ask you more about your individual pattern of checking orcleaning. Secondly, we will ask you to go into four anxietyprovoking situations of your choice. These situations may resultin your experiencing moderate levels of anxiety. Finally, wewill ask you some questions about your experiences.Participation in the study is entirely voluntary and you arefree to change your mind and withdraw at any time if you so wish.The research project has no bearing on any treatment you mayreceive, but it may help others, like yourself, in the long run.All the information collected during this study is strictlyconfidential. Participation in this research or unwillingness toparticipate, has no bearing on any treatment you may receive.When the project is finished we would be pleased to send youa description of the results. If you have any questions aboutthis study, please feel free to call:C. Lopatka, M.A. or S. Rachman, Ph.D.Department of PsychologyUniversity of British Columbia(604) 822—9028Participants Only: I have read and understood this descriptionof the study and agree to volunteer. I have received a copy ofthis description.Name (please print) SignatureAPPENDIX CSTRUCTURED INTERVIEW220221STRUCTURED INTERVIEWTrauma HistoryBefore the age of 18, were you physically,emotionally and/or sexually abused by any person whowas five or more years older than yourself?I’ll give you some definitions of physical, sexualand emotional abuse.Physical AbuseDid anyone use excessive physical punishment inyour family toward you? (ie. more than a spanking)Sexual AbuseDid anyone impose any sexually inappropriate act(for example, fondling, masturbation, oral, anal orvaginal intercourse) or acts with sexual overtones tomeet their own sexual or emotional needs?222Emotional AbuseDid anyone force you into performing cruel ordegrading tasks or humiliate or criticize yourepeatedly?Responsibility1. If you didn’t check____at all, and something badhappened would you be responsible for______occurring?2. If you didn’t check at all, and something goodhappened would you be responsible for_occurring?3. If you didn’t check at all, what is thelikelihood that something good/bad would happen?223GoodBad4. If you did check____completely, what is thelikelihood that something good/bad would happen?GoodBad5. Are you responsible solely for checking or areyou also responsible for____occurring?6. If you check____completely how much control doyou have over something bad happening?2247. If you don’t check_____completely how muchcontrol do you have over something good happening?8. Imagine for a moment that no one would ever know ifyou did or did not check__,how responsible wouldyou feel if something bad happened?9. Describe two checks you can remember clearly andvividly.225Situation 1.Description:Whose fault was it?How angry were you?How guilty were you?Situation 2.Description:Whose fault was it?How angry were you?How guilty were you?2269. If you didn’t check_____at all and something badhappened/something was not perfect, who is to blame?10. If you didn’t check_at all and something badhappened/something was not perfect, how angry would yoube?11. If you didn’t check_at all and something badhappened/something was not perfect, how guilty wouldyou be?12. If you didn’t clean_at all and somethingbad happened/something was not perfect, who is toblame?APPENDIX DRESPONSIBILITY CONTRACTS227228AGREEMENTI,_____________________,ofhereby agree to thefollowing conditions:1. I will try to treat this situation as I usually do.2. I will try to treat this test like a realisticsituation— the kind of situation I encounter regularly.3. The behavioral test applies only to the situationnoted below:4.5.I hereby agree to all of the above conditions and solemnlyswear to fulfill my obligation as noted above, dated thisday in the month of___________________Nineteen Hundred and Ninety One (1991) in the Province ofBritish Columbia.(Witness) (Signature)229RESPONSIBILITY AGREEMENTI,_____________________,of____ __ _ __ _ __ _ __ __ _ __ _ _hereby agree to thefollowing conditions:1. I will take complete responsibility for anything thatmay happen as a result of not checking.2. I will be to blame for any negative consequence as aresult of not checking.3. My responsibility applies only to the situation notedbelow:4. I will undo any damage or harm that may occur as aresult of not checking.5.6.I hereby agree to all of the above conditions and solemnlyswear to fulfill my obligation as noted above, dated thisday in the month of__________________,Nineteen Hundred and Ninety One (1991) in the Province ofBritish Columbia.(Witness) (Signature)230RESPONSIBILITY AGREEMENTI,_____________________,ofhereby agree to thefollowing conditions:1. I will take complete responsibility for anything thatis not perfect as a result of not checking.2. I will be to blame for anything that is not perfect as aresult of not checking.3. My responsibility applies only to the situation notedbelow:4. I will undo any damage or harm that may occur as a resultof not checking.5.6.I hereby agree to all of the above conditions and solemnlyswear to fulfill my obligation as noted above, dated this____ __ _ __ _ _day in the month of__________________,Nineteen Hundred and Ninety One (1991) in the Province ofBritish Columbia.(Witness) (Signature)APPENDIX EDEBRIEFING FORM231232DEBRIEFING FORMThe study that you kindly agreed to participate inexamines the role of responsibility in compulsive checking.All participants were encouraged to increase, decrease and notchange their sense of responsibility over negative events.We expect that when participants were asked to increase theirsense of responsibility they would feel more anxious and spendmore time checking than in situations when they were asked todecrease their sense of responsibility over negative events.In addition, we were interested in finding out more about yourthoughts, feelings and bodily sensations during periods ofanxiety.Your participation is greatly valued and it may assistus in finding more effective treatment approaches forObsessive Compulsive Disorder. We would be happy to provideyou with our findings when the study is complete. If you haveany further questions about this study, please feel free tocontact:C. Lopatka, M.A or S. Rachman, Ph.D.Department of PsychologyThe University of British Columbia822—9028If you are interested in reading further about ObsessiveCompulsive Disorder we recommend the following book:Steketee, G. (1990) Once is not enough. HarbingerPublications.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items