UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Interpersonal assessment of psychopathy Foreman, Michael Ernest 1988-12-31

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
831-UBC_1988_A1 F67.pdf [ 13.84MB ]
Metadata
JSON: 831-1.0107157.json
JSON-LD: 831-1.0107157-ld.json
RDF/XML (Pretty): 831-1.0107157-rdf.xml
RDF/JSON: 831-1.0107157-rdf.json
Turtle: 831-1.0107157-turtle.txt
N-Triples: 831-1.0107157-rdf-ntriples.txt
Original Record: 831-1.0107157-source.json
Full Text
831-1.0107157-fulltext.txt
Citation
831-1.0107157.ris

Full Text

INTERPERSONAL ASSESSMENT OF  PSYCHOPATHY  By MICHAEL ERNEST  FOREMAN  B.Sc., The University of British C o l u m b i a ,  1979  M.A., The University of British C o l u m b i a ,  1983  A THESIS SUBMITTED  IN PARTIAL FULFILLMENT  THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE (Department  of.Psychology)  We accept this thesis as to the required  August  conforming  standard  THE UNIVERSITY OF BRITISH  @  STUDIES  COLUMBIA  1988  Michael Ernest Foreman,  1988  OF  In  presenting  degree  at  the  freely available copying  of  department publication  this  of  in  partial  fulfilment  University  of  British  Columbia,  for reference  this or  thesis  thesis by  for  his  this thesis  and study. scholarly  or  her  Department of The University of British 1956 Main Mall Vancouver, Canada V6T 1Y3  \J  DE-6(3/81)  Columbia  the  requirements  I agree  that the  I further agree  purposes  may  representatives.  for financial  permission.  of  be  It  gain shall not  for  Library  an  advanced  shall  make  that permission  for  granted  head  is  by  understood be  the that  allowed without  it  extensive of  my  copying  or  my  written  ABSTRACT  This study was concerned with  the relations  representations of psychopathy and  interpersonal  perceptions.  From 147  inmates seen in a federal  security prison, 79 of the men provided comparisons.  between  medium-  complete data  for  Groups were defined under criteria from  (1)  the Psychopathy Checklist  (PC)  (Hare, 1985b), or  American Psychiatric Association  (2)  (1980, 1987) outlines  for Antisocial Personality Disorder  (APD).  Measures  were  derived  from the Interpersonal Adjective  (iAS-R)  (Wiggins, Trapnell, and Phillips, 1988 ) which  relate interpersonally  defined perceptions of  as locations within a circumplex Circle  (Wiggins, 1979, 1980).  as descriptive of  (1) self,  by a specific member of the  personality  space--Interpersonal  Self-ratings were  (2) ideal self,  thought seen by a friends, and  was also obtained  Scales-Revised  (3) self as  (4) self as thought  institutional staff.  from the specific staff members  descriptive of the particular  obtained  seen A  rating  as  inmates.  Comparisons were also made with respect to the specificity and sensitivity of MMPI profiles relevant to psychopathy.  Supplementary  considered  comparisons  used selected scales from the Adjective Checklist (Gough and Heilbrun, 1980) and Rosenberg's  (ACL)  (1965) Self-  esteem Scale.  These comparisons provided  manipulation  checks of the consistency of the data and contributed the interpretive generalizability  of the  to  results.  The primary hypotheses were that a group of individuals defined as psychopathic would show in representations obtained  from self-rated  differences  and  rated descriptions, with respect to circumplex  otherlocation  and derived difference scores from the IAS-R, in comparison to groups considered  non-psychopathic.  Results indicated differential  perceptions,  particularly by staff members, which provided  good  discriminations of groups based on the PC but not for groups defined by APD.  Circumplex  locations of  psychopaths defined by the PC were consistent expectations  for the Interpersonal Circle.  with  The  discriminative utility of group differences was much higher for the PC-defined groups than for APD to the base rates for these different The results are discussed contribution  categorizations.  in terms of  to the nomological  (1) their  network for the  of psychopathy as represented by the PC, limitations of the study, and  relative  concept  (2) specific  (3) the evident  confusion  which can result from the use of measures assumed to relate to the behavioural  to  'psychopath,' but that rely on primarily  descriptions.  iv TABLE OF CONTENTS Abstract  ...  **  List of Tables  vi  List of Figures  1X  Acknowledgement  x  Introduction  .  1  Assessment techniques relevant to the present research .... The Minnesota Multiphasic Personality Inventory (MMPI) . Antisocial Personality Disorder The Psychopathy Checklist (PC) . . . ... The Interpersonal Adjective Scales (IAS and IAS-Revised) Supplementary Assessments  10 10 18 21 25 37  Purpose Method Subjects . Setting Personnel Materials Procedure Scoring  54  54 60  61 62  . .  62 67  ...»  Results Diagnostic criteria Psychopathy Checklist (PC) PC Factor I scores Antisocial Personality Disorder (APD and APD-R) Minnesota Multiphasic Personality Inventory (MMPI)  •  .....  Diagnostic congruence Dependent measures Interpersonal Adjective Scales-Revised (IAS-R) Adjective Checklist (ACL) and Rosenberg Scale Comparisons of IAS-R descriptions within diagnostic groups Octant locations Coordinate distributions and group analyses Distance measures Comparisons using the selected ACL scales Discriminant analyses with the ACL scales  (MANOVAs)  70 70 70 71 71 72 75 80 80 89  ..  94 94 100 107 Il5 121  mmmmmmdc  TABLE OF CONTENTS - cont'd. Supplementary analyses Polar coordinates Self-esteem and "self" versus "ideal"  121 121 126  Discussion Overview and basic methodological issues Diagnostic considerations Relations of the dependent measures to the PC and APD Practical implications and theoretical considerations Summary and Conclusions Speculations  and suggestions  127 127 134 139 ..... 147 150  for further  research  151  References  1.5 3  Appendix A Appendix B  173  vi LIST OF Table I  TABLES  Cleckley's characteristies of the psychopath (Cleckley, 1982)  8  Table II  MMPI group definitions  Table III  Items of the Psychopathy (Hare, 1985b)  Table IV  Adjectives of scales BC/DE from the IAS-R (Wiggins, Trapnell, & Phillips, in press)  Table V  Hypotheses and possible alternative  Table VI  Distribution of subjects refusing or omitted from participation grouped with respect to the PC  Table VII Table VIII Table IX Table X  Table XI  Table XII  Table XIII  Table XIV  . ..  17  Checklist  22 ....  relations  Means and standard deviations of age and • of education across groups  30  ..  51  57  years .  58  Overall frequency distributions of MMPI group categories for the parent sample (N = 113) ...  73  Overall frequency distributions of MMPI group categories for the sub-sample (N = 7 9 ) .......  74  Lambda ( X ) coefficients of predictive association and Kappa (K) coefficients of diagnostic agreement across the different criteria based on the parent sample  76  Lambda ( X ) coefficients of predictive association and Kappa (K) coefficients of diagnostic agreement across the different criteria for the sub-sample (N = 79)  79  Summary statistics for the overall distributions of point coordinates and modal octant locations for the sample N = 79  82  Summary statistics for selected scales of the Adjective Checklist and the Rosenberg scale obtained from self-description (N = 79)  90  Correlations of IAS-R scales from set 1 with the selected ACL scales (and the Rosenberg scale) (N = 79) ...  91  vii LIST OF TABLES  Table XV  Table XVI  Table XVII  Table XVIII  Table XIX  Table XX  Table XXI Table XXII  Table XXIII  Table XXIV  Table XXV Table XXVI  Table XXVII  - cont'd.  C o r r e l a t i o n s of IAS-R s c a l e s from set 1 w i t h the selected ACL scales (and the R o s e n b e r g scale) (N = 79) O v e r a l l frequency d i s t r i b u t i o n s of octant a s s i g n m e n t s across the IAS-R rating sets (N = 79)  92  •  93  P r o p o r t i o n s of PC groups o c c u p y i n g adjacent paired octants assigned from s e l f - d e s c r i p t i o n and staff description (N = 79)  95  P r o p o r t i o n s of Factor I g r o u p s occupying a d j a c e n t paired octants assigned from selfd e s c r i p t i o n and staff d e s c r i p t i o n (N = 79) ...  97  P r o p o r t i o n s of APD groups o c c u p y i n g adjacent paired octants assigned from s e l f - d e s c r i p t i o n and staff description (N = 79)  98  P r o p o r t i o n s of APD-R g r o u p s o c c u p y i n g adjacent paired octants assigned from .self-description and staff description (N = 79)  99  M e a n s (and SDs) of Dom and Lov c o o r d i n a t e s across groups defined by the PC  102  M e a n s (and standard d e v i a t i o n s ) of Dom and Lov c o o r d i n a t e s across g r o u p s defined by Factor I  103  M e a n s (and standard d e v i a t i o n s ) of Dom and Lov c o o r d i n a t e s across g r o u p s defined by APD (DSM-111 ; APA, 1980)  105  M e a n s (and standard d e v i a t i o n s ) of Dom and Lov c o o r d i n a t e s across g r o u p s defined by A P D - R (DSM-III-R; APA, 1987)  106  S u m m a r y statistics of d i s t a n c e m e a s u r e s the overall sample (N = 79)  109  for  M e a n s and standard d e v i a t i o n s of d i s t a n c e m e a s u r e s and within cell c o r r e l a t i o n s for the PC groups  110  M e a n s and standard d e v i a t i o n s of d i s t a n c e m e a s u r e s and within cell c o r r e l a t i o n s for the Factor I groups  Ill  vii.i.  LIST OF TABLES - cont'd. Table XXVIII  Means and standard deviations of distance measures and within cell correlations for APD vs. not APD ..... 112  Table XXIX  Means and standard deviations of distance measures and within cell correlations for APD-R vs. not APD-R  113  Means and standard deviations of the selected ACL scales and Rosenberg scale for the PC groups  116  Means and standard deviations of the selected ACL scales and Rosenberg scale for the Factor I groups  117  Means and standard deviations of the selected ACL scales and Rosenberg scale for the APD groups  118  Means and standard deviations of the selected ACL scales and scale for the APD-R groups  119  Table XXX  Table XXXI  Table XXXII  Table XXXIII  Table XXXIV  Rosenberg  Means and (standard deviations) of vector length across Psychopathy Checklist groups (N = 79)  122  Table XXXV  Means and (standard deviations) of vector length across Factor I groups (N = 79) ...... 123  Table XXXVI  Means and (standard deviations) of vector length across APD vs. APD (N = 79)  124  Means and (standard deviations) of vector length across APD-R vs. not APD-R (N = 79)  125  Table XXXVII  ix LIST OF Figure 1  FIGURES  Circumplex structure of the IAS-R (Wiggins, Trapnell, & Phillips,  in press)  Figure 2  Obtained 2-factor solution based on N = 113  circumplex . ...  Figure 3  Mean Dom and Lov coordinate locations of IAS-R response sets for the overall sample (N = 79)  84  Figure 4  Means Dom and Lov coordinate locations of IAS-R response sets for the PC groups  85  Figure 5  Mean Dom and Lov coordinate locations of IAS-R response sets for the Factor I groups  86  Mean Dom and Lov coordinate locations of IAS-R response sets for the APD (DSM-111) groups  87  Mean Dom and Lov coordinate locations IAS-R response sets for the APD-R (DSM-III-R) groups  88  Figure 6  Figure 7  for the  ...  28 81  of  X  ACKNOWLEDGEMENTS  I would  like to take this opportunity to thank my wife,  Georgie, for her years of loving support, and to apologize her and our children, Amanda and Greg, for the number of I couldn't spend at home.  Many thanks are also due to Bob Hare for his in completing  school.  assistance  the project, and again to him and  Jerry Wiggins and Dimitri Papageorgis encouraging  evenings  Thanks, too, to my mother for her  considerable help over the too many years of graduate  and support  to  to  for their constructive  input to the completion of the thesis presented  I would also like to thank Adelle Forth for her  and here.  able  organization of the research team and data collection, and  to  Karen Harlos, Tim Harpur, Jim Hemphill, and Sherrie Green for their contribution  to the data collection and  scoring.  I must also thank Sharon Greye and the staff of the Matsqui  Institution Health Care Centre for their patience and  help, providing us space and  incorporating us into their  Many thanks also to the inmates who contributed and to the institutional made this study  possible.  to this  routine. study,  staff who gave their time and effort  and  1'  INTRODUCTION This study provides an evaluation of a variety of strategies with respect  to the identification  characterization of the psychopath adult male offenders.  and  in a sample of  A particular focus  assessment  incarcerated  is the use of an  interpersonal model of personality assessment, designed  to, among  other things, evaluate dissimulation  in self-report obtained  such a sample.  is in the potential  The primary  interest  from  differences among self-rated and other-rated descriptions of groups formed by differing diagnostic criteria on measures derived from a personality used  assessment strategy not  in criminal populations.  The purpose  is to test  expectations based on the Interpersonal Circle--a model of personality  previously  circumplex  (Kiesler, 1985; Leary, 1957; Wiggins,  1979,  1980, '1982)--with respect to groups classified by the criteria of the Psychopathy Checklist  (Hare, 1980, 1985a, 1985b) or by the  diagnosis of antisocial personality disorder defined by the American Psychiatric Association  (APA, 1980,  Psychopathy represents a particularly personality disorder  problematic  in terms of nosological  theoretical understanding, and efforts for  1987).  reference,  intervention.  It is  perhaps the most baffling and poorly understood disorder, and  is seen to be particularly  personality  resistant  interventions, whether correctional, medical, or  to  psychological  (e*g., Cleckley, 1982; Hare, 1970, 1981; Hart, Kropp, & Hare, 1987; Weiss, 1986).  The development  form and nature of this disorder  of further  insights  into the  requires continued efforts to  systematically delineate the essential characteristics  of the  psychopath, and develop appropriate assessment strategies.  In  this way, steps may be made toward more meaningful management the psychopathic personality as he  (to date little research  involved females) may be encountered  in correctional or  of  has  treatment  settings. Descriptively, there psychopathy  is good consensus for the concept of  in terms of an apparent  relationships with others and devoid  McCord  impulsive antisocial  of conscience or remorse  Hare, 1970, 1979, 1982  incapacity to form meaningful behaviour  (Buss, 1966; Cleckley,  1982;  1986; Hare & Cox, 1978; Maher,  1966;  & McCord, 1964; Millon, 1969, 1981; Weiss,  1986).  However, there has been considerable debate over the  relative  importance of the psychological aspects of the disorder lack of empathy) versus the behavioural manifestations antisocial acts or delinquency.  Pichot  debate.  He concludes that the rift derived  nosologies of German and Anglo-French  of  (1978) provides  interesting historical perspective of the development from  origins.  (e.g.,  an  of this  independent Historically,  3' "psychopathies"  represented  a class of disorders akin to current  conceptions of the personality disorders, with reference to the broader usage of "psychopathy" as any disease of the mind  (e.g.,  Dor land ' s Illustrated Medical Pi ct ionary, 2 6 th Ed., 1981) . Gradually, the reference to the psychopathic personality came to focus on the expression of antisocial behaviour any obvious mental defect. attributed  to Pritchard  In England, the early  conception  in 1835 as "moral insanity" emphasized  "congenital deficiency of moral sense" was echoed by the French  in the absence of  (Pichot, 1978, p. 56) and  in 1866 as "reasoning  insanity."  focussed on behaviour as the manifestation of the disorder related  "perverse  instincts" or "depraved  of its expression.  and  conception  retained the emphasis on a "personality disorder" focussing aspects believed  Both  feelings" as the basis  As the usage evolved, the German  the characterological  a  to underlie the  while the English and French conceptions have favoured  upon  behaviour, "moral  insanity" as defined by "abnormally aggressive or seriously irresponsible conduct  " (Mental Health Act for England and  Wales,  1959) . Current manifestations  of this debate may be seen in the  reliance of some authors on behavioural patterns as selective criteria  (e.g., American Psychiatric Association  (APA), 1980,  1987; Robins, 1966) or in an emphasis on personality characteristics an integrated  (e.g., McCord  & McCord, 1964).  Others  provide  approach, giving relatively equal weight to both  •  characterological  '  •  '  •  dispositions and behavioural expressions  Buss, 1966; Hare, 1980; Maher, 1966).  (e.g., Blackburn & Maybury, 1985).  however, there is agreement on the fundamental  is the typical expression of that core  1982; Hare, 1986; McCord  & McCord, 1964).  for  Generally,  role of  personality as the core of psychopathy and that behaviour  (e.g.,  Arguments still recur as  to which are of relative primacy or if both are necessary classification  4  antisocial (e.g., Cleckley,  Similarly,  it should  be agreed that conceptual discontinuity will result where behaviour  is emphasized  to the exclusion of consideration of the  personality or interpersonal style.  Behaviour may constitute  most tangible aspect of the psychopath, but it is not for the diagnosis;  different motivations a "psychopathic"  and, moreover,  debate.  quite  individuals who may  possess  relationship with their environment do not  necessarily exhibit the pattern of poorly behaviour common to incarcerated 1981).  sufficient  its necessity may also be a point of  Clearly, similar patterns of behaviour may occur with  the  samples  integrated  antisocial  (cf. Hare, 1986;  Millon,  A basic complaint, however, has been the difficulty  of  reliably and validly assessing the "personality" of the psychopath. Consequently, characterization  of this disorder has proven  problematic due to the application of different emphases and apparent  loss of focus.  the  This point has been one of concern and  confusion for research with the psychopath  (Hare, 1979, 1980,  '•5 1986; Hare & Cox, 1978; Millon, 1981), where assessments  with  adult males have relied upon various definitions such as single or composite scale profiles Personality  Inventory  from the Minnesota  Multiphasic  (MMPI), behavioural checklists,  scales, or the criteria for Antisocial Personality (APD) , (APA, 1980, DSM-111) .  self-report  Disorder  As outlined by Hare and Cox  (1978),  this variety of assessment contributes to the confusion of antisocial behaviour or criminality per se in place of the more clinically meaningful of conscience or Problems  concept of psychopathy emphasizing a lack  empathy.  in reaching some clear consensus for  purposes are exemplified  in  research  (but by no means unique to) a recent  article by Blackburn and Maybury  (1985) concerning  the  homogeneity of samples defined by various criteria. recognizing the apparent  sources of confusion  in  Although  identifying  "the" psychopath, Blackburn and Maybury argue against the primacy of affective characteristics  and assert a need to apply  involving both affective and  impulsive-aggressive  characteristics  criteria  behaviour  in order to obtain a more homogeneous sample.  is likely that such a sample may be more homogeneous  in terms of  their behaviour, but also likely that they do not adequately uniquely  represent the psychopathic character.  point of confusion  Another  or  frequent  in the literature, and which also arises  in  Blackburn and Maybury's article, relates to the experience of anxiety and the suggestion  of the primary vs.  It  secondary  psychopath as being free of anxiety or subject to anxiety, spectively. Harpur stated:  It has been responded  re-  to most succinctly by Hare and  (1986) who, in reference to the "secondary  psychopath,"  "They may be neurotic, anxious, socially withdrawn  psychotic criminals, but they are not psychopaths" The utility of the concept  (p. 150).  in applied settings  is similarly  confused, as indicated by a recent newspaper article  (Still,  1987, May 21) in which a psychiatrist with some expertise forensic consultations apparently dismissed  or  in  the prognostic  usefulness of a diagnosis of "psychopath" with a comment to the effect that three out of four Federal psychopathic.  This estimate  (Canadian) prisoners  are  is clearly out of keeping with  rigorous definitions of the psychopath  more  (Hare., 1983, 1985a), and  with the potential predictive utility with respect to such behaviour as parole violations  (Hart, Kropp, and Hare,  1988).  These problems typically derive from the emphasis placed upon an individual's current antisocial behaviour and history of delinquency, which are indeed common  in incarcerated  The problem thus becomes, from the point of view of the ability to make valid and  samples. assessment,  reliable discriminations with  such  populations based upon some coherent theoretical frame of reference. Although perhaps not widely acknowledged, and misrepresented  (e.g., Blackburn  & Maybury, 1985), Hervey  (1941/1982. 6th ed.) provides a most coherent the psychopathic personality.  occasionally Cleckley  representation  Cleckley acknowledges the  of  likely  7  tendency of the psychopath  to come to public attention as a  result of his behaviour, but argues there  is a more  affective deficit at the core of the psychopathic  remarkable  personality.  Using the analogy of a "semantic aphasia" in accounting  for  lack of emotional  suggests  relationship with the world, Cleckley  that the psychopath can intellectually  relate to others but  the  lacks  understanding of the emotional, connotative meaning of communications and, hence, lacks the capacity to value  others.  It should be noted that this affective deficit appears  distinct  from the blunted schizophrenia; expression.  or flattened affect associated  the psychopath  with  is capable of a range of  affective  Rather, this deficit refers more to a restricted  shallow affective capacity  or  in which self-interests are almost  always placed above the feelings of others  (cf. Raine,  1986).  Cleckley refers to the psychopath's ability to vocalize appropriate affective  relationships but with an apparent  shallowness or absence of feeling, likening  "a reflex machine  that can mimic the human personality perfectly" p. 228).  (Cleckley,  1982,  He notes that the basic demonstration of the  psychopath's  inability  to relate emotionally  behave by any appropriate standards.  is the failure to  This lack is thus  described  in terms of interpersonal attachment or sincerity, related as a failure to appreciate the emotional experience of others. Cleckley's criteria are listed The principal  in Table 1.  interest of the present  research  is to further  TABLE I  Cleckley's characteristics of the psychopath ( C l e c k l e y , 1 982)  1.  superficial charm and good "intelligence";  2.  absence of delusions and other signs of irrational t h i n k i n g ;  3.  absence of "nervousness" or psychoneurotic manifestations;  4.  unreliability;  5.  u n t r u t h f u l n e s s or i n s i n c e r i t y ;  6.  lack of remorse or shame;  7.  inadequately motivated antisocial b e h a v i o u r ;  8.  poor judgement and failure to learn from e x p e r i e n c e ;  9.  pathologic egocentricity and incapacity for love;  10.  general p o v e r t y in major a f f e c t i v e relations;  11.  specific loss of i n s i g h t ;  12.  unresponsiveness in general interpersonal relations;  13.  fantastic and u n i n v i t i n g behaviour with d r i n k and sometimes without;  14.  suicide r a r e l y carried o u t ;  15.  sex life impersonal, t r i v i a l , and poorly i n t e g r a t e d ;  16.  failure to follow any life p l a n .  9  flesh out this characterological/interpersonal  aspect of  psychopathy with an assessment strategy based upon a model of the interpersonal domain:  the Interpersonal Circle  (IPC)  (Freedman,  Leary, Coffey, & Ossario, 1951; Kiesler, 1985; Leary, 1957; Wiggins, 1980, 1982). interest for several  The application of this model reasons.  development, psychopathy  is of  In terms of theory and  is clearly an interpersonal  the IPC may allow a means of mapping perceptions of this interpersonal  the self- and  style.  conceptual disorder;  other-rated  The demonstration  of  congruence between  ratings of psychopathy and expectations  its representation  on the IPC will thus contribute to construct  validity.  Moreover, the IPC has been recommended  for  as a structural  foundation for a possible revision of the diagnosis of the Personality Disorders as presented 1980)  in DSM-111, Axis II  (APA,  (Kiesler, 1985; McLemore & Benjamin, 1979; Widiger &  Frances, 1985) and population  its performance with respect to a criminal  is also of interest.  On the practical or applied  side, the IPC may provide a means of reliably assessing personality aspects of psychopathy considered who assert that only behaviour  the  so elusive by  is appropriate for assessment.  may also allow a better means to approach populations other the incarcerated criminal.  Further discussion of these  shortly.  A previous attempt to assess psychopathy  in terms of  It  than  issues  and a review of the psychometric qualities of one format of IPC will be presented  those  the  10 interpersonal style was made by Blackburn and Maybury Although their method was rather the circumplex may be useful psychopathic personality. of mentally disordered observational  idiosyncratic,  (1985).  it suggests  that  in the assessment of the  Despite procedural p r o b l e m s — a  inmates of the Broadmoor  rating scales with  interrater  sample  Hospital,  reliabilities  varying  from .32 to .66, and factor analyses of 36 variables based on 57 subjects--Blackburn representation circumplex  and Maybury  of interpersonal  of interest here.  (1985) obtained a 2-dimensional styles approximating  The present  the  research will  attempt  to expand on this approach using assessment techniques with demonstrated  psychometric  some  stability.  Assessment Techniques Relevant to the Present The Minnesota Multiphasic Personality  Research  Inventory  (MMPI)  The MMPI has a long history of use and contributes to a vast literature; the use of the MMPI as a criterion or measure  discriminative  in research with delinquent youth and adult  spans some forty years.  offenders  An early summary of research with  delinquents was provided by Hathaway and Monachesi recent uses with adults are exemplified  in 1953.  by Brown and  More  Gutsch  (1985) who use scale 4 elevations to identify the psychopath, or by Megargee typology  (1977, 1984) who uses the MMPI as the basis for a  in criminal  populations.  As outlined by Dahlstrom, Welsh, and Dahlstrom  (1972),  scale  11 4  (Pd - psychopathic deviate)  is intended to address the  and asocial subgroup of persons with psychopathic disorders"  (p. 195).  Thus, scale profiles  high point are typically  personality  involving scale 4 as a  regarded as representing  However, some secondary scale elevations  "amoral  psychopathy.  (e.g., 1 -  Hypochondiasis, 2 - Depression, 3 -Hysteria, 5 - MasculinityFemininity, 7 - Psychasthenia) are considered  to diminish  the  likely expression of antisocial behaviour or acting out, whereas the involvement of scales 6 (Paranoid), 8 (Schizophrenia), and 9 (Hypomania) likelihood code "49"  in particular, are considered to increase the of antisocial behaviour  is considered  Welsh, and Dahlstrom  (Lachar, 1978).  Lachar  two-point  the "classic psychopath" by Dahlstrom,  (1972) and has shown some  ability with psychopathic/non-psychopathic 1970) .  The  discriminative  criminals  (Hare,  (1978) refers to the 49 profile as a likely  sociopathic personality, and suggests that the 48 profile representative of "classical psychopaths" Major descriptive  features attributed  is  (p. 85). to scale 4  elevations are consistent with current definitions of psychopathy,  including  flagrant disregard  for social  inability to profit from punishing experiences, and shallowness  in relations with others.  is considered  to increase activity  values, emotional  The involvement of scale 9  level and the likelihood  acting out; scale 8 involvement may reflect the  of  interpersonal  withdrawal associated with schizoid tendencies, which may be  12 considered  similar to the emotional detachment of the psychopath.  It is apparent that the MMPI has some utility; but its broad application  in various populations,  and, more particularly, psychopath may be  its validity  its reliability over in identifying  time  the  questioned.  The development of scale 4 was organized with respect to its ability to discriminate a criterion group of delinquent from normative groups of adults used  in the MMPI  sample and from a sample of college students.  youth  standardization  The relevance of  the criterion group, "fairly young people, with more girls group than boys"  in the  (Dahlstrom, Welsh, & Dahlstrom, 1972, p. 196),  with a history of delinquency  ("amoral and asocial"  behaviour--  primarily minor crime and status offences), to the psychopath currently defined interpretation  is clearly debatable.  Moreover,  the  of this scale varies widely depending upon  population to which  it is applied, and various factor  have been proposed based on intercorrelations among scale  (e.g., Astin, 1959, Comrey, 1958).  the  patterns  items on the  Among normal males  scoring high on scale 4, such positive attributes as  adventurous  and courageous, sociable, enthusiastic, good-tempered,  generous,  and fair-minded have been applied; other normal samples have described as aggressive, unemotional  immature, irritable, leisurely, and  (Dahlstrom, Welsh, & Dahlstrom, 1972).  college males, scale 4 may be considered "rebelliousness,"  as  In normal  an index of  reflecting difficulty with  authority.  been  13  In correctional populations, scale 4 elevations would to be the most common aspect of MMPI profiles. in Dahlstrom, Welsh, and Dahlstrom  (1972:  Tables  populations.  provided  Appendix M, pp.  448) outline the proportions of two-point codes found  (N=183), 48%  of the profiles obtained had scale 4 as the high point  1970, cited  involved  scale 4 as the secondary peak  in Dahlstrom et al., 1972).  male military prisoners  peak  in another 21.6%  1972).  scale 4 as the peak secondary  (Brodskey, 1967, cited in Dahlstrom et 49/94 two-point codes  provide scale 4 as a primary or secondary peak  college freshmen secondary peak  males.  (N=258)  in 21.4% of  obtained profiles, with 49/94 profiles seen in 6.2% Meehl, 1951, cited  al.,  accounted  in the youth sample and 22.6% of the military  Comparative samples of normal adult Minnesota males  (Hathaway &  in Dahlstrom et al., 1972); a sample of  (N=l,537) provide scale 4 as a primary  or  in 29.1% of the profiles, with 49/94 profiles  in 9.5%  (Dahlstrom & Reifler, 1970, cited  1972).  Dahlstrom et al.  prominent  (McMahon,  involved scale 4 as the  The more discriminative  for 24.6%  and  Similarly, a sample of  (N=2,126) exhibited  in 41;4% of the profiles and  438-  in various  In a sample of male youthful offenders  another 20.1%  appear  in Dahlstrom et al.,  (1972) also note that scale 4 peaks  in profiles from alcoholic samples, homeless  delinquent subgroups, disciplinary  seen  are  vagabonds,  and sexual offenders within a  prison system, drivers with high frequencies of violations accidents, and various drug abuse groups  (p. 270).  and  Megargee and his associates  (Megargee, 1977, 1984;  Megargee  & Bohn, 1977; Megargee & Dorhout, 1977; Meyer & Megargee, have developed an MMPI typology within prison samples  1977)  (males, age  19-27 years) which provides data suggesting the relationship the previously mentioned scale profiles characteristics of the psychopath.  of  (i.e., 4, 49, 48) with  Megargee empirically  derived  groups based on a cluster analysis of profiles; ten profiles, obtained with relatively specific selection criteria, for 87% of the sample.  accounted  However, seven of the ten profiles  involved elevations on scale 4 with differentiation provided secondary peaks and relative elevations. accounted  As might be expected, correlated  descriptive data which would relate to characteristics were distributed being described superficial  instability.  self-report  and  other  "psychopathic"  among these seven groups,  some  as cold and aggressive, others as glib and  in relationships, and others as pleasant but  manipulative.  Problems with this typology may also derive Simmons, Johnson, Gouvier, and Muzyczka  Considering that  reliabilities of MMPI scale scores range from intervals on the order of eight months  on the  interval of ten  They found that only 14 of the 50 retained  original group classification.  from  (1981)  retesting of 50 inmates previously classified  Megargee MMPI typology at an average follow-up months.  profiles  for 62% of the sample or 71% of the classified  profiles.  provided  These seven  by  their  retest  .44 to .73 with  (Dahlstrom, Welsh, &  Dahlstrom, 1975) , such a result "clinical"  As a  inventory, MMPI profiles should be expected to change  over time. likely  is not entirely unexpected.  Further, the more specific rules become, the less  it is that profiles will prove  reliable.  An additional difficulty with the MMPI and "cookbook" for profile  identification  Marks & Seeman, 1963) types generally populations  (e.g., Gilberstadt  & Duker, 1954, or  is that the more specific rules for code  fail to classify a significant portion of  (Butcher & Tellegen, 1978; Payne & Wiggins,  Classification  strategies used by Megargee  classification  rates varying from 63% to 96%  Megargee, 1977).  Hare  as named above, found  rules  resulted  target  1968).  in  (Dorhout &  (1985a), using standard  "cookbook"  few profiles meeting the criteria.  rules Higher  rates of inclusion can be obtained with some relaxation of code criteria  (Hare, 1985a; Payne & Wiggins, 1958), and may be more  readily obtained where only the one or two high point peaks are used to identify the profile Dahlstrom, 1972) . addressed  (e.g., Dahlstrom, Welsh, &  Another difficulty not often  is the rate of occurrence of invalid  particularly among correctional populations are "required" to complete the form. identification of different  specifically profiles,  in which  individuals  Although the MMPI  allows  approaches to dissimulation on the  test, such profiles are not usually The validity of MMPI codes  included  in research.  in identifying the psychopath  defined by Cleckley criteria has been questioned primarily by  as  Hare and his associates. profiles related obtained  Relatively poor overlap of MMPI  to the "psychopathic personality" has been  in research comparing definitional criteria  1980, 1985a).  Further confusion  is likely to result  researchers using elevations on one or two scales  (e.g., Hare, from  (e.g., 4, or 4  and 9) in isolation from the overall scale profiles to the psychopath  (e.g., Ray & Ray, 1982).  scale 4 appears as a high point  "define"  It is apparent  in various populations  various descriptive associations.  Perhaps a more  that with  basic  correspondence of antisocial behaviour, drug or alcohol use, or conflict with authority or parental figures, explains the common occurrence of scale 4 elevations  in criminal  samples.  In view of the above problems, a relatively  straightforward  approach to classifying obtained MMPI profiles has been for the purposes of this research. under this system: identify the typical  Eleven profiles are proposed  three incorporate  invalid profiles,  three  "psychopathic" profiles, four are other  readily discriminable profiles, and one is "normal." for these profiles are outlined group 10  applied  ("other") may  in Table II.  Profiles  within  include elevations involving scale 4 with  secondary elevations on scales 1, 2, 3, 5, or 7. are not included among  Criteria  These profiles  "psychopathic" profiles as the  implications of guilt, remorse, anxiety, or contrition  whether  situational or otherwise are further confounds to an already tenuous representation  of the psychopath using the MMPI.  This  17  T A B L E II MMPI Croup Definitions  Category Croup  Criterion Rules raw scores  Invalid  1  - ? > 30  2  - F > 23 or  Implication test avoidance  F - K> 15  'fake bad 1  - L S 9 and K - F > 15  Sociopath  'fake good 1  Scale Elevations - 1 is only scale ^ 70T  'spike 4 - antisocial'  - 4 + 9 only scales S 70T  'sociopath'  - 4 + 8 ^ 70T; 9 & /or 6 may also be ^ 70T  'psychopath'  Other - involving 1 or more of scales 1, 2 , 3 , 7 as only scales ^ 70T  'neurotic'  - scales 6 & /or 8 are only ^ 70T or are highest elevations by 10T  'psychotic'  - scale 9 only elevation^ 75T 'hypomanic'  and scale 2 S 50T  10  - fails to meet any of above, but has scale elevations ^ 70T  'other'  Normal 11  - no scale elevation^  70T  'normal'  strategy  is intended  to simplify the information available  the MMPI profiles obtained, and to focus more specific on the "psychopathic" profiles and their overlap with groups otherwise  from  interest criterion  defined.  Antisocial Personality Disorder  (APD)  Diagnoses of APD using the criteria of the Diagnostic and Statistical Manual of Mental Disorders, 3rd ed. 1980) and  its recent revision  (DSM-111: APA,  (DSM-III-R; A P A , 1987), were based  on data available from interview, case history, and information.  Criteria for APD have focussed  behaviour as the "essential  vandalism,  physical cruelty"  parental  (APA, 1987, p. 342).  and would  grounds for arrest;  truancy,  The pattern must  irritability  and  of acts which would  be  aggressiveness;  impulsivity, and/or promiscuity.  lasting, close, warm, and responsible  Although  alluding  impaired  capacity  relationships"  (APA, 1980, p. 318; APA, 1987, p. 343), the application diagnosis may require little or no consideration As previously  persist  include financial, vocational, or  to interpersonal characteristics as a "markedly to sustain  behaviour.  fights, running away from home, and  irresponsibility; commission  recklessness,  and adult  signs are listed as "lying, stealing,  initiating  into adulthood  antisocial  feature" of this disorder, and  basically provide a checklist of childhood Common childhood  upon  file  of this  of  the  aspect.  suggested, the DSM-III criteria have been  faulted  for being too liberal  (Frances, 1980; Hare, 1980, 1981,  1983, 1985a; Millon, 1981).  The reliance on  behavioural  descriptions of a delinquent history prior to age 15, and recurrent antisocial behaviour since age 18 does not provide a distinction between  "persistent criminality"  (Trasler, 1978) and  the remorselessness  or lack of empathy considered central to the  psychopath• Hare  (1981) has outlined  the potential for confusion  APD and psychopathy, and demonstrated with various criteria.  Hare  the extent of  (1983) reported  between  agreement  "generally  good  agreement" between extreme group assignments of psychopathic APD diagnoses with a Kappa coefficient of  .83; however,  tendency for overinclusiveness with APD is reflected relative rates of diagnosis: checklist assessment. of diagnosis  Hare  and  the  in the  39% APD vs. 22% psychopathic by a (1985a) obtained more similar  in a sample of 229 federal  inmates, 38% APD vs. 33%  psychopathic, and also showed a high congruence of the however, only with the extreme groups  rates  diagnoses,  (Kappa = .79).  Some modest changes have been made to the criteria for APD under the new DSM-III-R  (APA, 1987); however, there is little  reason to suspect the changes will significantly affect frequency or distribution of the diagnosis samples.  The annotated comparative  III to DSM-111-R  in  the  incarcerated  listing of changes from DSM-  (Appendix D, DSM-III-R) only notes the addit ion  of an item to express the absence of guilt or remorse, provided  20 in response to "frequent criticism" of the DSM-III criteria APD.  However, the added  item--"lacks  remorse  (feels justified  having hurt, mistreated, or stolen from another)" 346)--is only  for in  (APA, 1987, p.  included as the tenth of a set of ten aspects of  adult behaviour  from which any four satisfy  the criterion.  This  addition obviously does little to realign the APD diagnosis respect to psychopathy  as there remain many  (actually 126) ways  an individual might meet the criteria within that section inclusion of a remorseless attitude.  with  The emphasis for  without  diagnosis  continues to be irresponsible and antisocial behaviour as the "essential  feature" for both DSM-III and DSM-III-R.  changes within the criteria under DSM-III-R changes  Other  are basically  only  in wording or emphasis; however, academic difficulties or  problems at school have been dropped  from the adolescent  section, which may serve to tighten the criteria for  sub-  adolescence,  and the criteria for a pattern of adult behaviour has had the age criterion  reduced to 15 years.  possible c o n s e q u e n c e — a n d diagnosis--are  The only other changes of  this to increase the application of the  the omission of the continuity  criterion,  i.e.,  that there has been no period of five years or more since age 15 without evidence of antisocial behaviour, and the relaxation of the overlap of other mental disorders,  i.e., that behaviour  now "not exclusive" to a course of Schizophrenic or Manic disorder  rather than  "not due to" such a disorder.  The overlap of APD with the other criteria used  in this  is  21 study was assessed, as was the agreement of the APD under DSM-111  (APA, 1980) and the DSM-III-R  The Psychopathy Checklist  diagnosis  (APA, 1987).  (PC)  The PC has been developed by Hare and his colleagues as a means of standardizing psychopath within  research criteria to identify  incarcerated male samples.  Cleckley had provided  the most  the  Considering  that  influential conception of the  psychopath, early efforts were given to globally rating  inmates  based on interview and file information as to how well they met Cleckley's criteria that this provided  (Hare & Cox, 1978).  With the perception  a very useful and conceptually-based  to assessing the psychopath, attention was given to a checklist  approach  standardizing  format which could be more readily utilized by other  researchers.  A 22-item checklist was developed which  good correspondence with previous global ratings and outlines were provided Frazelle, 1980) .  for its application  provided  (Hare,  1979),  (Hare &  Two items were subsequently dropped  (previous  diagnosis as a psychopath, and drug and alcohol abuse not  direct  cause of antisocial behaviour) with evidence that they contributed  little to the discriminative power of the checklist  (Hare, 1986) . III.  The current 20-item checklist  is shown in Table  The PC incorporates case history data allowing  evaluation of problem behaviour and characterological from interview which  longitudinal inferences  in total provide a more complete profile  in  22  TABLE III Items of the Psychopathy Checklist ( H a r e , 1 985b)  1.  glibness/superficial charm;  2.  grandiose sense of s e l f - w o r t h ;  3.  need for stimulation/proneness to boredom; pathological l y i n g ;  5.  conning/manipulative;  6.  lack of remorse or g u i l t ;  7.  shallow a f f e c t ;  8.  callous/lack of empathy;  9.  parasitic lifestyle;  10.  poor behavioural controls;  11.  promiscuous sexual b e h a v i o u r ;  12.  early behaviour problems;  13.  lack of realistic, long-term goals;  14.  impulsivity;  15.  irresponsibility;  16.  failure to accept responsibility for own actions;  17.  many s h o r t - t e r m marital relationships;  18.  juvenile delinquency;  19.  revocation of conditional release;  20.  criminal v e r s a t i l i t y .  23  keeping with the clinical concept of psychopathy. Administration of the checklist requires about half hours for review of available approximately  institutional files plus  two hours for a structured  rated on a three-point scale  one-and-one-  interview.  Items are  (0, 1, 2 where 0 indicates that  item does not apply, 1 indicates that its application questionable, and 2 indicates the item definitely  the  is  applies);  scores for the 20-item checklist thus may range from 0 to 40. information  is lacking for completion of a particular  item may be omitted and the score prorated.  If  item, the  Rather than  assuming  the measure to represent a continuum, i.e., a range of psychopathic tendencies, for research purposes the scale with cut-off scores to provide diagnostic  discontinuity.  Individuals with scores falling approximately deviation above the range psychopaths. considered  1 standard  (e.g., > 30) are considered  Those in the lower range  is used  to be  (e.g., < 20) are  non-psychopathic, and those in the middle range may be  of questionable  status.  In use with adult male inmates, the checklist has been demonstrated within  to yield high  (.85 to .93) raters, good test-retest  and high  internal consistency  Hare, 1983).  Consideration  more problematic what?".  reliability across  (.90)  (.84 to .91) and reliability  (.89),  (Schroeder, Schroeder, &  of the validity of the PC is somewhat  in that there is the question of "Compared  Other assessments  related to psychopathy may be  to  faulted  24  for being too broad or based on questionable criterion  groups.  However, the evidence  instances  is at least consistent.  In most  adult male prison inmates identified as psychopathic by the PC are also considered psychopathic by other assessments  (or as APD  by DSM-III), although the converse is not necessarily  true,  those identified as "psychopathic" by other devices may not be identified  as psychopathic by the PC  1985a, 1986; Schroeder et al., 1983).  (or as APD)  (cf. Hare,  of the checklist may be taken from its apparent utility providing clearly discriminable groups for  validity in  psychophysiological  research  (Hare, 1979; Hare & Cox, 1978), and the predictive  validity  in the contexts of parole performance  (Hart et al.,  (Wong, 1985).  The factor structure of the PC  (Harpur, Hakstian, & Hare,  its psychometric stability across samples and  importance of evaluating personality characteristics identifying  the  Apart from the  convergence with other assessments, demonstration of the  1988a) reflects  1983,  These results affirm  more stringent criteria afforded by the PC.  1987) or response to treatment  i.e.,  the psychopath.  Harpur et al.  samples providing a total subject pool of 982.  from  interrelationship of behavioural expression and  glibness/superficial  Factor  five  The obtained  solution was not orthogonal, but oblique, reflecting  psychological attributes.  in  (1988a) considered  2-factor solution to best represent data obtained  the  "core"  I, containing the  charm, egocentricity/grandiose  the  items sense of  a  self-worth, pathological  lying and deception, conning/lack  sincerity, lack of remorse or guilt, lack of affect and  of  emotional  depth, callous/lack of empathy, and failure to accept responsibility  for own actions, reflects the personality  characteristics considered central to psychopathy. isolated  the behavioural  items, reflecting a chronically  lifestyle and the expression of antisocial Differential  relationships  and external variables  Factor  II  unstable  behaviour.  are seen with Factors  (Harpur, et al., 1988b), again  I and  II  reflecting  the distinction of the personality and behavioural aspects of psychopathy  and, more broadly, criminality.  For example, APD  diagnosis correlated more highly with Factor with Factor impression  I (r = .42)  behaviour.  the  (Hare, 1985a) that the congruence obtained between  the  relates to the common aspects of  The relationship of Factor  I, as a more  measure of personality attributes, to the assessment strategy to be outlined of this  (r = .55) than  = 2.90, p < .005)., consistent with  PC and APD diagnoses  U  II  next  antisocial  specific  interpersonal  is a principal  interest  study.  The Interpersonal Adjective Scales  (IAS and  IAS-Revised)  A line of enquiry which has not been previously pursued this context and which may serve several to the characterization  interests with  of the psychopath  interpersonal assessment strategy evolving model of Leary  in  respect  involves an from the circumplex  (1957), as currently developed by Wiggins  (1979;  Wiggins, & Broughton, 1985; Wiggins, Trapnell, & Phillips, in press).  The major values of this orientation  of psychopathy are that:  (1) psychopathy  "interpersonal"  (2) defined  disorder;  to the assessment  is clearly an  scales of the  relate directly to the descriptive aspects of the profile;  (3) the IAS and  properties; and  psychopathic  IAS-R assessment have good  (4) with additional  circumplex  psychometric  research, this strategy may  be more generally applicable to assessment settings other correctional  institutions.  Psychopathy disorder:  than  is readily characterized  it is the apparent  as an  interpersonal  remorselessness and  superficiality  in relationships with others that constitute the basis of this personality disorder.  As such, it is the failure of the  psychopath to form affective  interpersonal  relationships or to  behave in keeping with social values that defines the These personality  aspects form a strong component  structure of Hare's checklist and  disorder.  in the  (Hare, 1979; Harpur et al.,  factor 1988a),  it is these aspects for which the circumplex model for the  assessment of interpersonal behaviour may be particularly  well-  suited. The development of the circumplex model of behaviour, and Wiggins' particular,  interpersonal  is described by Wiggins  Trapnell, & Phillips,  in press).  interpersonal  adjective scales in  (1979, 1980, 1982;  Wiggins,  The basis of the model as  developed by Wiggins provides eight adjectival scales which, by  27 the nature of their  intercorrelations, can be related as a  circumplex  in a two-dimensional  oriented  space  The dimensions are defined by two orthogonal represented  as Status  corresponding  participants  (affiliation),  (love) consequences  (self and other)"  combinations derived  clear-cut  for both  (Wiggins, 1979, p. 398).  underlying structure of this circumplex  status to oneself  interpersonal  interactions that have relatively  (status) and emotional  1).  components  to the theoretical definition of  events as "dyadic social  (dominance) and Love  (see Figure  The  is related to the eight  from the granting or denying of love and/or  and/or the respective other.  Thus, in terms of  the interpersonal circumplex, psychopathy can be conceived of as the granting of love and/or status to oneself while denying both love and status  to the other--an arrogant, cold,  calculating  personality. The adjectival scales expected to relate to psychopathy other measures correlated  to them can be seen to concur well with  previous definitions of the psychopathic character. arrogant-calculating  and  profile  The  (variable label BC), has been used  by Wiggins to characterize the narcissistic personality  disorder  of DSM-111  content  (Keisler, 1985; Wiggins, 1982); however, the  and correlates of this scale have obvious application psychopath. exaggerated  As discussed by Wiggins characterization  to the  (1982), narcissism  of the BC profile  is an  involving  grandiosity, lack of empathy for others, acting out, feelings of  90° PA  .8  A s s u r e d - Dominant  135°  BC  •  • ArrogantCalculating  *  —  .4  WarmAgreeable -.4  AloofIntroverted •  .8  -.4  UnassumingIngenuous  FG  225°  NO  GregariousExtraverted  ColdHearted  t  45°  JK Unassured -Submissive  315°  -.8  HI  270°  Figure 1 Circumplex s t r u c t u r e of the I A S - R (Wiggins, T r a p n e l l , & Phillips, in p r e s s )  29  special entitlement, and exploitative  relations with others.  adjacent variable  as cold, quarrelsome, or  (DE), characterized  The  cold-hearted behaviour, has been related to the paranoid personality disorder as representing excessive hypervigilance, hypersensitivity restricted  affectivity  to the behaviour of others, and  (cold and unemotional)  The expected placement of personality Leary and Coffey  (1955: cited  suspiciousness,  (Wiggins,  1982).  disorders within the IPC by  in Widiger  & Kelso, 1983)  located  the psychopathic and sadistic personalities at the DE pole. studies relating the IAS to contemporary Murray's  inventories  From  assessing  (1938) taxonomy of needs, correlates and analysis of  content commonalities serve to further define the scales  (Wiggins  & Broughton, 1985) . Correlates of BC aggressiveness,  (arrogant-calculating)  emphasize  impulsivity, rationalization of  conflict with authority, and exploitiveness.  behaviour,  Item  factor  loadings reflect arrogance, exploitiveness, ready expression anger  (verbal), and competitiveness.  Variables  of  related to DE  (cold-quarrelsome), depict manipulativeness and the absence of warmth, cooperation, or nurturant behaviour. comprising  these two scales  in Table IV.  The  adjectives  (BC and DE) of the IAS-R are  It can be seen that the content of these  relates well to the characterization  listed  scales  of the psychopath, and  it is  the bctant bounded by these two scales which can be hypothesized to capture the psychopathic profile as obtained from ratings by  3 0  T A B L E IV Adjectives of scales B C / D E from the I A S - R (Wiggins, T r a p n e l l , S Phillips, in p r e s s )  (BC)  Arrogant-calculating  (DE)  Cold-hearted  cocky  ruthless  crafty  ironhearted  cunning  hardhearted  boastful  uncharitable  wily  coldhearted  calculating  cruel  tricky  unsympathetic  sly  warmthless  others.  In that dissimulation of self-report may be expected  the norm rather than the exception for a psychopathic self-report profiles were expected to be of  as  group,  interest more for the  relative disparity of self versus others' ratings across  groups  rather than for veracity of self-depict ion.  were  expected  Psychopaths  to provide more positive self-evaluations  in greater  disparity to the perceptions of others when compared with ratings for other  inmate groups.  similar  This effect m a y , of course, be  moderated by the manipulative facility and superficial charm of the individual, with the result that the most psychopath may escape  identification  successful  in this assessment  (or the  checklist) by having convinced others of his basic sincerity good  intentions.  An assessment strategy  in which the  and  IAS-R  provides several derived measures of disparities which may counter such results will be outlined In discussing Wiggins  presently.  the psychometric properties of the  (1979) presents data which support general  construct  validity, acknowledges a confound, and suggests high consistency  for the scales.  statistically  significant  Sex difference data  IAS,  internal  reveal  (p < .03 - .0001, N > 600)  differences  in a pattern that suggests an even split of the circumplex with a line rotated slightly clockwise from the principal dimension status.  Interestingly, the dimension that  of  is orthogonal to this  line passes through the octant of interest here,  BC/DE,  postulated  The mean  to contain the psychopathic profile.  32  differences across scales suggest small sex  differences  (approximately  (identifying)  .25sJ  with respect to seeing  as either self-assured relatively  (dominant), or unassured  large differences  (.5s^ - .75s) where  oneself  (submissive), but ascribing  characteristics to self as arrogant, calculating, cold, and quarrelsome ingenuous  (males > females) or warm, agreeable,  (females > males)  (Wiggins, 1979).  unassuming,  These  differences  are slightly attenuated but retain the same pattern when the revised scales  (IAS-R)  using  (Wiggins, Trapnell, & Phillips,  in  press). These differences may be interpreted apparent confound with social desirability to self-report with evaluative words the desirability  in relation to the in responding  (Wiggins, 1979) .  inherent  In that  in this instance seems to involve portrayal  of  social sexual stereotypes, high scores on the BC/DE scales may also result from a tendency of males to endorse items masculine. research.  This confound has implications for the present In that scales BC/DE demonstrate associated  characteristics with  (see Wiggins & Broughton, 1985) as  "negative masculinity"  1979) and Machiavellianism is expected  considered  negative  correlations  (M-; Spence, Helmreich, & Holahan, (MACH; Christie and Geis, 1970), it  that psychopaths, as more sophisticated  would minimize these endorsements more affable and easy-going.  respondents,  in favour of self-portrayals  This supposition  expectation that psychopaths would score lower  led to the (by self-  as  33  description) on scales BC/DE than other  inmate groups, and would  score higher on scales JK/LM which depict the individual as more modest and  agreeable.  Internal consistency estimates for the eight scales of the IAS over a combined  sample of university students  "meet a reasonably stringent (©<•> .80)"  requirement of  (Wiggins, 1979, p. 408).  the scales of  (N = 610) all  internal  consistency  The alpha coefficients  for  interest here, obtained over four subsamples of  student population  (N's range 100-152), ranged from .845 to  These results support the internal cohesiveness of these The revised scales demonstrate similar  levels of  the  .889.  scales.  reliability,  with alphas ranging from .749 to .856 in a total sample of  1161  (Wiggins, Trapnell, & Phillips, in press). External validation of the scales  is well demonstrated  recent research by Buss, Gomes, Higgins, and Lauterbach In an effort to demonstrate differences among use of interpersonal  of situational aspects and personality (1987) found good congruence between  (1987) .  individuals  "tactics of manipulation" and the  in  in the  relation  variables, Buss et al.  interpersonal styles and  different scales from the IAS circumplex.  That is, different  tactics were found to correlate with scales from the IAS in a way that  is consistent with  particular  interpretation of the scales.  Of  interest was the manipulative capacity demonstrated  those identified within octant BC  (Arrogant-Calculating).  individuals were more likely to use all of the  identified  by  These  34  manipulative tactics, rather than rely on one or two as was more typical or other groups defined by the IAS scales. is certainly consistent with the characterization "calculating"  interpersonal style.  have demonstrated  Such a result of a  Gifford and O'Connor  similar patterns of behaviour  (1987)  consistency  in  relation to scales from the IAS, with a more subtle measure of interpersonal  distance.  The potential utility of this assessment of evolves from  personality  its development and structural characteristics  its ease of administration. respondents are instructed scale, ranging from  The IAS-R  is a 64 adjective  (1) extremely  target).  list;  to rate each adjective on an 8-point inaccurate to  (8) extremely  accurate, as to how well the word describes them designated  Completion  (or some other  time for the list  quite brief, on the order of ten to fifteen  is generally  minutes.  A potential problem, however, for the use of adjective within the population too difficult. difficulty usage  of interest here  (e.g., "kind" vs. "unauthoritative")  research, a glossary was appended  (e.g., "cunning" - skillful synonyms  level of  or familiarity  oriented  in manipulating  list  providing  tendencies  others) or simple  (e.g., "jovial" - happy; good sense of humour).  glossary was provided  of  For the purposes of this  to the adjective  in terms of interpersonally  lists  is that some words may be  Not all words may represent a similar  (e.g., "outgoing" vs. "perky").  definitions  and  The  in order to better standardize the word  presentation and level of understanding  for individuals who may  otherwise have been uncertain of a word's meaning. serves to orient the respondents  It also  to the interpersonal  intent of  the words, suggesting styles of relating to others rather intrapersonal qualities or  characteristics.  Once completed, the list can be scored to yield independent  to provide two coordinates corresponding  (Phillips, 1983).  Either  coordinates  computation  in summarizing the scale results as a point  dimensional  space which  in a 2-  is located within a particular  the circle and at a certain distance from the centre. octant  identifies the individual's  and the distance from the centre  likely  combined  to the dimensions of  (e.g., Laforge, 1977) or polar  within the circumplex results  eight  (no items overlap) scales, and these scales  Love and Dominance  than  octant of The  interpersonal  style,  (vector length) may be taken  as the strength of association within that octant as a function of response variability across the eight scales.  Vector  is also looked upon as a potential measure of relative or rigidity of interpersonal style Kiesler, 1985; Widiger Phillips, & Trapnell,  (e.g., Chartier,  length  pathology  1984;  & Frances, 1985; Wiggins, 1982;  Wiggins,  in press), particularly within a given  octant. The scoring and circumplex disorders.  interpretation  of point  locations within  the  offers a model for representations of personality Hypothetical placements of the various  personality  disorders have been suggested  (e.g., Kiesler, 1985; Leary &  Coffey, 1955; Wiggins, 1982), and some research has ratings of personality disorders to the circumplex (e.g., Plutchik  & Platman, 1977).  related structure  Widiger and Kelso  (1983)  point to the advantages of the Interpersonal Circle in providing structure to the organization of the personality disorders, and in affording diagnostic flexibility with respect to dimensional measures and prototypic representations  rather than the  assumptions of discreet, classical categories.  These  characteristics of the scales have appeal here in providing a means to organize the conceptual coherence of  different  approaches to the psychopath with respect to perceptions  (by self  and others) of the psychopath's  obvious  limitation  interpersonal style.  lies in the actual representations of  perceptions as obtained by  An  these  self-report.  In addition to obtaining a basic profile of  self-reported  self-perception using the IAS-R, it was felt to be of some tactical value and potential slightly more complicated  theoretical  interest to pursue a  set of self-report profiles.  self-rating profiles were obtained with the IAS-R: yourself; 2) describe your  Four  1)  describe  ideal self--the person you would  like to be; 3) describe yourself  as you think your friends  describe you; 4) describe yourself  as you think  "  •-"  specific member) of the institutional staff would describe The tactical value is in the aim of being able to  best would  (a you.  counter  the efforts of evasive respondents attempting to distort self-report.  Because  it was expected  the  individuals would elect to  bias response toward more favourable representations, the utility of different  instructional  sets was to provide contrasts of  relative discrepancies or difference scores among the different profiles across groups.  This strategy was hoped to provide a  pattern of measures that could discriminate across  groups  although the content of self-representation may be less than veridical.  Theoretically  interesting  results would also  derive  from discrepancies among representations of self as they may relate to self-perception and  insight  in the psychopath.  An additional source for further comparisons of discrepant ratings was also obtained from others' ratings of the Institutional staff reasonably individuals provided  inmates.  familiar with the inmates as  ratings on the IAS-R which could be  contrasted with the self-report profiles.  The possible  interpretations and implications of these contrasts will be discussed  shortly.  Supplementary  Assessments  In order to provide comparative analyses with the to the IAS-R, the Adjective Checklist  (ACL; Gough & Heilbrun,  1980) and Rosenberg's Self Esteem Scale also obtained  as  responses  (Rosenberg, 1965) were  self-report.  The ACL has an extensive research background, as is  reflected  in its having attained  8th Mental Measurements Yearbook most often referenced  26th position  (1978)  list of the 100 most used and  assessment devices  in psychology.  is an alphabetical list of 300 adjectives.  The current ACL Manual  The ACL  Respondents  check off those adjectives which they consider descriptive.  in Buros'  simply  to be self-  (Gough & Heilbrun,  1980)  outlines 37 scales which may be scored with reference to normative data. demonstrated  Analyses by Wiggins and Broughton  (1985) have  that several of the ACL subscales have high  commonalities within the circumplex  space of the IAS and  numerous zero-order correlates with the IAS scales of here.  provide  interest  In the interest of providing adequate representation,  keeping the overall number of scale comparisons within  yet  reason, a  specific subset of ACL scales was selected with reference to their demonstrated  relations to the IAS.  To compare  self-  descriptions, ACL scales were selected which best represented IAS octants of interest here: (arrogant-calculating); DE submissive); JK  PA  (assured-dominant);  (cold-hearted); HI  (unassuming-ingenuous);  Nine ACL scales were selected; three scales:  and LM  Unfavourable; and six scales considered Murray's needs: Abasement, and  BC  (unassured(warm-agreeable).  'modus operandi 1  Total Checked, Number Favourable, and  (MO)  Number  representative of  Achievement, Dominance, Autonomy,  Aggression,  Deference.  The first of these scales  th  (Total Checked) provides both a  measure of "loquacity" and a criterion against which the other scales are standardized  for scoring  (Gough & Heilbrun,  Standard score transforms are provided  1980).  in the ACL Manual  based  upon differing tendencies  to endorse few or many  broken  The remaining scales are scored  into five groups.  reference to these standardized Megargee  group  adjectives, with  scores.  (1984) has found the ACL to provide  discriminative  differences of adjective endorsements and scale scores when contrasted across typology.  inmate groups formed on the basis of his MMPI  Overall, responses were notable for the rates of  positive self-descriptions. commented  upon similar  Sutker, De Santo, and Allain  response styles with the ACL seen in a  sample of antisocial men and women participating program.  Sutker et al.  in a drug  abuse  (1985), however, did not provide the more  typical scale score results. selected here, Megargee  (1985)  With respect to the ACL  scales  (1984) found that, overall, his  inmate  groups scored essentially at the mean of the normative range; but some pronounced differences were apparent when the inmates' ratings were compared to ratings of the inmates provided institutional staff psychologists.  Large discrepancies  evident with the scales Number Favourable, Number  self-  by were  Unfavourable,  Aggression, Achievement, and Dominance, wherein the  inmates'  self-ratings all reflected more positive descriptions than  those  made by staff. These results supplement the pattern of results  expected  40  with the IAS-R, and comparative analyses and correlations of scale representations will serve to enhance the generalizabi1ity however,  of results obtained.  in the functional vocabulary  complete the ACL.  interpretive  A problem  remains,  required to  adequately  Given a list of 300 adjectives,  it was not  practical to provide a glossary as was done with the IAS. consistency of results may be affected usage in an inmate sample as compared  The  by differences of word to student  samples.  The Rosenberg self-esteem scale was chosen as a simple (10 items) providing an additional measure of positive regard which can be compared with the ACL and IAS-R included  in contrasts of criterion groups.  reasonable reliability and validity  1965; Silber  & Tippett, 1965) as a measure of  is attractive  results, and  (Rosenberg,  "self-acceptance,"  in its brevity and ease of administration.  is useful here as a supplementary consistency  self-  The scale has  demonstrated  and  scale  It  "manipulation check" or test of  in responding across the self-report profiles.  If  the profiles are consistent and meaningful, one would expect  to  see a reasonable correlation between the Rosenberg scale and a measure of the discrepancy between ideal  representations of self and an  (e.g., low self-esteem should correspond  discrepancy).  Campos  to a greater  (1986) has presented data suggesting a  relation of self-esteem with congruence of depictions of self an  ideal-self.  and  PURPOSE  The purpose of this research was to characteristics of  interpersonal  style  investigate  in incarcerated  psychopaths using the 64-adjective format of the Adjective Scales-Revised  (IAS-R)  criminal  Interpersonal  (Wiggins, Phillips & Trapnell,  in press; W i g g i n s , Trapnell, & Phillips, in press).  Additional  self-report profiles were obtained with the 300-word  Adjective  Check List  (ACL)  esteem Scale  (Gough & Heilbrun, 1980) and the Rosenberg  (Rosenberg, 1965).  20-item Psychopathy Checklist  Psychopathy was defined by the  (PC)  (Hare, 1985b), and  made with the Minnesota Multiphasic Personality  comparisons  Inventory  (Hathaway & McKinley, 1951) and the diagnostic criteria Antisocial Personality Disorder of the DSM-III the DSM-III-R  Self-  (MMPI)  for  (APA, 1980) and  (APA, 1987).  Although the IAS-R has not been used specifically  with  clinical samples, the theoretical structure of the scales for specific expectations  allows  regarding the location of the  psychopath within the circumplex space.  It is also possible to  compare the relative locations of groups defined under the various diagnostic criteria. assessed  In this way differences can be  in the construct validity of the various criteria  as  42  providing groups which can be meaningfully distinguished another.  Do men  identified  differ from men considered  as psychopathic by the PC criteria non-psychopathic, and is the  representation of the psychopath consistent with  expectations?  Are men who meet the criteria for APD or APD-R different those not meeting the criteria? this  from one  from  These questions are central to  research. The concurrent validity of the diagnostic  assessment  strategies can be organized with respect to groups formed the range of PC scores. PC constituted  over  The group defined as psychopathic by the  the criterion against which the congruence of APD  diagnosis and MMPI profiles was assessed.  Although  this  represents an essentially nominal choice, it is not entirely arbitrary.  The focus of this research  interpersonal the PC.  is to investigate an  representation of the psychopath as identified  It is apparent from the review of the literature  there continues to be considerable debate over what adequately  that  criteria  represents the psychopath; this research also  the opportunity  affords  to further assess the congruence of different  diagnostic criteria and to compare the  interpersonal  representations associated with these criteria.  Of interest  are comparisons of diagnostic congruence when the criterion replaced by high scores character"  by  in Factor  I of the P C — t h e  (Hare, 1979; Harpur, Hare, & Hakstian,  also is  "remorseless 1988a).  Each of these diagnostic assessments can be taken as  43  independent definitions of criterion groups under which  the  IAS-R, the ACL, and the Rosenberg scales can be evaluated respect to the discrimination  of the resultant groups.  profiles for groups defined by the PC and PC Factor of primary  2-dimensional  circumplex yielding point  I scores  it is possible to compute and compare different  Euclidean profiles  However, simple Euclidean distance  information which  is basic to the circumplex model,  location, and relative distance from the origin. information could also be obtained  loses i.e., octant  Thus, more  from the contrast of polar  coordinates, yielding a difference between angles and lengths.  This coordinate system  analysis program devised by Phillips (1983), in reviewing Leary's  relative  is the basis of an  (1983) .  Madison and  Paddock  (1957) approach to the analysis of  in circumplex models, recommend adopting a system  similar to the use of polar coordinates  relating  relative  distance from the origin but using the arc segment between rather than the relative angular calculation  locations.  orientation, the use of angles was  the more reasonable  points  Since the  of the arc is rather more complicated and would  provide a directional considered  are  locations as profile  (Wiggins, 1982).  variability  IAS-R  structure of the IAS-R as a  distances between points representing  vector  The  interest.  Given the previously described  summaries,  with  choice.  These two approaches, Euclidean distance and polar  not  44  coordinates, provide two ways to analyse the contrasts  between  points.  summarize  The Euclidean distance measure can be used to  the simple  linear distances between points and forms the basis of  the contrasts following, while polar coordinates can be used to summarize octant  locations and vector  lengths which are more  relevant to the description of interpersonal styles the circumplex.  The principal  interest was  in the  across groups of the relative distances between corresponding  to five  identified  by  comparison  points  conditions:  1)  self versus  ideal self;  2)  self versus self as think friends  3)  self versus self as th ink inst itut ional staff  4)  self versus self as seen by others  see; see;  (inst itut ional  staff) ; and 5)  self as th ink staff see versus self as seen by staff.  Based on expectations  for the psychopathic character, patterns of  differences among these distance measures were expected groups.  A central question with  respect to assessment with  psychopath concerns dissimulation with self-report. expectations may effect different across these distance  for the the  Various  interpretations of patterns  measures.  With respect to the first measure, self versus ideal self, the psychopath was expected  to show lesser distance between  points than other respondents for at least  two reasons:  these  (1) as  45  facile and manipulative, sensitive to impression management, was expected himself  that he  (all subjects are male) would  portray  in a favourable way both as self-descriptive  descriptive; and the psychopath  it  and  ideal  (2) it is consistent with the egocentric view of  that he may see self as ideal.  The second  measure, self versus self as think friends see, was also expected to show lesser distance than other groups  in the belief  that the  psychopath, more so than others, would wish to portray himself  as  well liked by others with whom he should be expected to have good relat i ons. C o n v e r s e l y , with the third measure, self versus self as think  inst itut ional staff see, the psychopathic group was  expected  to show greater distance than other groups as a result  of a performance pressure to consider the perceptions of  others,  i.e., in the knowledge that certain staff would make ratings of them and may not think well of them.  The fourth measure, self  versus self as seen by staff, if consistent with the pattern proposed, should also be a maximum for the psychopathic  group.  The last distance measure, self as think staff see versus self as seen by staff, has interesting  implications for the self-  insight or social awareness of the psychopath; psychopath  aware of how he comes across to others?  expected that the psychopathic group would presented  i.e., is the  as a challenge--could  respond  It was to a task  they accurately predict how they  are seen by a particular member of the institutional staff?  To  46  the extent that the psychopathic group can provide congruent with the perceptions of others,  profiles  it may be argued  that  they can at least articulate the effect their behaviour has on the perceptions of others. empathy,  Although not representative  it is relevant to role-taking--an hypothesized  in the psychopath  of deficit  (Gough, 1948).  The capacity of psychopaths to merely understand or grasp the perceptions of others or the effects their behaviour may have on the feelings of others  is not clear.  Does the psychopath  blithely pursue his own ends in the absence of any for others  understanding  [as I believe Cleckley conceived] or does he act in  self-interest despite his understanding of others'  reactions?  That the psychopath can verbally present an understanding of social roles and expectations  is generally agreed  1966; Cleckley, 1982; Hare, 1970, 1978; Trasler,  (i.e., Buss, 1978).  However, the foregoing hypotheses need also be with  respect to other  interpretations of the possible  for the distortion of self-report. may be considered  considered  relevant:  At least four  motivations  interpretations  first, in keeping with  Cleckley's  formulation, an inability to appreciate the perceptions of others (a specific desirability  loss of insight); second, the obvious effects of on self-report; third, egocentric distortion or, in  effect, the denial of negative characteristics; or,  fourth,  egocentric manipulation or the desire to "con" others. probably the case that we cannot know which of these  It is  conditions  hold, as the interplay of these possible motivations may alternative explanations to all the foregoing  provide  expectations.  However, the test of a postulated pattern of disparities  among  perspectives was expected to reflect better for one interpretation  than  another.  It would be of considerable approach Cleckley's  interest to have a means to  "semantic aphasia," the defect postulated  account for the inability of the psychopath to appreciate meaning of principles of behaviour  (although able to  apparent understanding) and the resulting the effects of his behaviour on others.  the  verbalize  loss of insight as to This concept, in  relation to the foregoing hypotheses, suggests that provides  to  psychopath  representations of himself based on his egocentric  view rather than the interpreted perceptions of others.  self-  The  expectation here would have the psychopath providing a consistent appraisal of himself without regard to the perspective;  instructed  thus, the psychopath's discrepancy scores should  the least across groups for all contrasts of self-rated Recent  research has pursued psychophysiological  of lexical analysis  in male offenders and psychopaths  attempt to identify differences be related to psychopathy 1988).  be  profiles.  correlates in an  in language processing which may  (see Hare, Williamson, and  Considerable data have accumulated  Harpur,  across studies using  various techniques and modalities demonstrating differences the language processing of psychopaths which may relate to  in  48  differences 1988).  in cerebral organization or utilization  In that differences have been shown  (Hare et al.,  in the responses to  affective words and connotative meanings, one may expect this to affect the psychopath's capacity to organize his responses to self-description with evaluative In this  instance, similar to that just  inconsistencies reflected  outlined,  in the psychopath's approach to the task would be  in the absence of any identifiable pattern of  responses. expected  words.  That  is, with the IAS profiles psychopaths would be  to fall toward the centre of the circumplex  discriminations  of description) and show minimum distances  the sets of self-descriptions. particular  (no across  Similarly, one would expect  no  differences with the ACL "MO" scales of favourable and  unfavourable adjectives  if the psychopaths were, in fact, to  ascribe no significance to the words. pointed out by Hare et al.  However, it appears, as is  (1983) and others, that the psychopath  does make use of the usual literal meaning of words and has the capacity to use them in a systematic way others).  (e.g., to manipulate  The difference with the psychopath  one, likened  seems a more  to "knowing the words but not the music,"  subtle  in which  he generally makes appropriate use of words but does not show consistency between his words and his behaviour  (cf. Hare et al.,  1988). Another  interpretation of the psychopath's  tendency to  distort self-report may be related to the manipulation  of  49  desirability which could operate in three ways. social desirability, the psychopath's  First, as simple  responses can be seen as  reactive to their perceptions of the values of others fellow  inmates).  show relatively  (i.e.,  The pattern of results under this condition may lesser distances between self and idea 1, and self  and self as think friends see than the other groups; but no difference from others with respect to the contrast of self and self as think staff see in the absence of any vested the staff perceptions.  in  A second desirability mechanism may be  seen as egocentric denial of negative traits. have more  interest  This process may  impact on the portrayal of self and  ideal than on the  perceptions of others and as a consequence would yield a pattern of discrepancies wherein the distance between self and would be minimized  relative to the other groups, but the  contrasts of self and self as seen by others  (friends and  would both show no difference across groups.  desirability  This  here  the  that others may not think well of them.  The third  distortion can be seen as an egocentric,  pro-active  response style (i.e., the  staff)  The suggestion  is that psychopaths think well of themselves despite understanding  ideal  in an effort to manipulate or "con" the system  researcher). last desirability manipulation  is interesting  in that  it implies the psychopath sees himself as one who can be seen by others as affable and sincere  (i.e., to expect  and by corollary, to possess some "insight"  the "con" to work)  into behaviour and  50  his effect on others, seen as a "nice guy." psychopath  i . e . , k n o w how to behave in order to be This obviously suggests, then, the  to be aware of and responsive to principles of  appropriate behaviour but to reject them in favour of more egocentric pursuits of immediate goals.  The consideration of the  psychopath as one who rejects social values and authority mentioned by Buss This implication  (1966, p. 433) and by Sarason is also considered  (1978, p. 304).  tentatively by Cleckley  (1982, pp. 229, 238-239) with respect to the issue of for behaviour.  is also  It is obviously a central  culpability  issue as to whether  the  psychopath elects to act in the knowledge of potential consequences  or acts without the capacity to  appropriate behaviour.  Hare  "understand"  (1970) points out that  the  psychopath may be untroubled by discrepancies between  his  behaviour and social expectations, but that does not mean he is unaware of such  discrepancies.  A summary of the hypothesized  relationships among the IAS  response sets under the expectation of dissimulation and three alternative  interpretations  is provided  in Table V.  As can be seen from Table V, discrimination among alternative sets of hypotheses  regarding distance measures  also derive from the corresponding coordinates.  the  relations of polar  With the exception of differential  locations, the polar coordinates should provide  octant differences  across groups even if the distance measures fail to do so.  should  51  TABLE V 1  Hypotheses and possible alternative relations. I  Basic expectations - dissimulated self-report  ! I I ; s  Self vs. ideal Self vs. as friend Self vs. as staff Self vs. by staff As staff vs. by staff II  I  | i J  P P P P  < < > >  NP,M NP,M NP,M NP,M  P < NP ,M  Self Ideal As friend As staff  P P P P  > > > <  Vector Length  270 270 270 180  HI/JK HI/JK HI/JK PA/BC  P P P P  By staff P < 180  BC/DE  P > NP,M  > NP ,M > NP ,M > NP ,M > NP ,M  Alternative relations A.  ! i  POLAR COORDINATES Angle Octants  DISTANCES  Egocentric self-view ("semantic aphasia") POLAR COORDINATES Angle Octant  DISTANCES Self vs. ideal Self vs. as friend Self vs. as staff Self vs. by staff As staff vs. by staff B.  fi i  P P P P  < < < >  NP ,M NP,M NP,M NP,M  P > NP,M  270 270 270 270  HI/JK HI/JK HI/JK HI/JK  P> P> P> P>  By staff P < 270  HI/JK  P > NP,M  Self Ideal As friend As staff  P P P P  > > > >  NP,M NP,M NP,M NP,M  Failure to discriminate ("lexical indifference") POLAR COORDINATES Angle Octant  DISTANCES Self vs. ideal Self vs. as friend Self vs. as staff Self vs. by staff As staff vs. by staff C.  Vector Length  NP,M NP,M NP,M NP ,M  Self — Ideal As friend — As staff ---  P > NPyM  By staff ---  P P P P  < < < >  — ---  Vector Length P< P< P< P<  NP,M NP,M NP,M NP,M  P> NP,M  Social manipulation ("the effective 'con'") POLAR COORDINATES Angle Octant  DISTANCES Self vs. ideal Self vs. as friend Self vs. as staff Self vs. by staff As staff vs. by staff  P P P P  < < < <  NP,M NP,M NP,M NP,M  P < NP,M  Self Ideal As friend As staff  P P P P  > > > >  Vector Length  270 270 270 270  JK/LM JK/LM JK/LM JK/LM  P> P> P> P>  NP,M NP,M NP ,M NP,M  By staff P > 270  JK/LM  P > NP,M  psychopaths, particularly as characterized by staff ratings, are expected  to be represented  in the upper  left quadrant of the  circumplex, the psychopathic group having most and the nonpsychopathic group least.  Vector  length may  itself provide  some  basis for discriminations.  As outlined by Wiggins, Phillips, and  Trapnell  length may be expected to apply best  (in press), vector  as a measure of extremity or "rigidity" within a given however, there vector  octant;  is as yet little information on the relation of  length alone to outside measures of deviance or general  psychopathology. The IAS-R may, thus, prove useful to further development the assessment  and characterization of the psychopath.  extent that group profiles can be obtained which discriminative there  in  To the  demonstrate  utility among the groups to be assessed  here,  is the potential for future research to pursue the  assessment  of psychopathy  in populations other than  criminal or mentally disturbed offender groups.  incarcerated  The emphasis of  Cleckley's portrayal of the psychopath has focussed more on the callous superficiality  of the character than on unlawful  behaviour and, given the abilities of Cleckley's psychopath  to avoid prolonged contact with  "true"  legal or psychiatric  intervention, the opportunities to develop assessment are rare remain  (cf. Widom, 1978).  rare  instruments  profiles  Such opportunities may, of  course,  in the absence of more methods to flag a profile with that may be used  in more general assessment  settings.  5 3  In summary, the principal objectives of this research are to evaluate the concurrent validity of assessments provided by Hare's  (1985b) checklist, DSM-III, and DSM-III-R, and the MMPI  for the identification  of the psychopath within an  incarcerated  adult male population and to evaluate the utility of the Interpersonal Adjective Scales - Revised Phillips, classified  in press) groups.  (Wiggins, Trapnell, and  in discriminating among the obtained  METHOD  Subjects Subjects were obtained from among the inmates of Matsqui Correctional  Institution—a  medium security Canadian  federal  facility for men serving sentences ranging from two years to life.  With  "cascade" through the federal correctional system  moving men down the security men may have been convicted  levels of various of crimes ranging  —  institutions--the from break and  enter or theft to murder. Participants were solicited by word-of-mouth, posters, and advertisements  in the inmate newsletter.  general research  Explanation of the  interests was made to the Inmate Committee  in an  effort to facilitate understanding of the independence of the research from any affiliation with correctional authorities and to provide assurances of confidentiality.  The men were  $5.00 per session for their time as an additional  incentive.  Formal consent was obtained when a subject was seen first session.  A brief explanation was given of the  interest as emanating  from the University of British  in the  research Columbia  Department of Psychology and that all information gained remain completely confidential.  offered  would  A consent form was provided  55  prior to beginning the interview which again outlined the basic research history.  interest  in obtaining systematic data concerning  Inmates acknowledged  participate  life  consent by signature to  in the interview; have the interview  videotaped;  provide access to institutional case management and  psychological  files; and complete the questionnaires for this research.  It was  made clear assurances of confidentiality were limited by the communication  of likely harm to self or others.  The men were  informed that they could withdraw from the research at any time without consequence, and that neither  their participation  nor  their withdrawal would have any effect on their status within the institution. For the purposes here, inmates completing  the questionnaires  had to have a minimum grade 8 education with English as their primary  language.  After 13 months of data collection, 147 men had in the initial diagnostic interview.  From these 147,  IAS-R protocols were obtained for 79 individuals.  participated complete  Data  collection was stopped at this point as, based on expectations for the potential size of mean differences seen in the Wiggins, Trapnell, and Phillips  (in press) normative data, power  calculations suggested a group size of n 1 25 to be good. sample of 79 could also be considered  The  adequate for multivariate  analyses based on a rule-of-thumb as a ratio of subjects to variables exceeding 5.  Although estimates of that ratio for an  optimal test might be considered  30 to 1 or more  (cf. Anderson,  1958), practical considerations must also hold some weight.  An  additional  18 provided complete IAS-R self-ratings, but  identified  staff members failed to complete ratings for them.  Another 16 men completed either  refused  (9), or were rating  IAS-R ratings for sets 1, 2, and 3 but  (6), failed to appear for subsequent  "unable"  appointments  (1) to complete set 4 which asked for a  from the perspective of a specific staff member.  additional  individuals with incomplete IAS-R protocols were  retained for analyses  involving the data which they did provide,  yielding a total sample of 113 for some  comparisons.  Thirty-four men provided no data for the present 11 refused; 22 were omitted due to education (11), inadequate English or an inability (2); and one study.  less than Grade 8 (3),  to complete a significant portion of the material  individual asked to have his data removed from the  by the Psychopathy Checklist  tests of proportions not differ  purposes:  (6), transfer from the institution  See Table VI for a breakdown of  categorized  These  non-participants (PC).  Based on simple  (Glass and Stanley, 1970), the groups did  in terms of their rates of  non-participation.  The 113 men providing usable data were an average age of 29.94 years  (SD = 7.57), with a range of 19 to 53 years.  average education was 10.60 years 16 years.  Their  (SD = 1.67), ranging from 8 to  The sample of 79 men who provided  complete  IAS-R  protocols were 30.13 years . (SD = 7.46) of age on the average,  TABLE  VI  Distribution of Subjects Refusing or Omitted From Participation Grouped with Respect to the PC PC Group (Psychopathic)  (Mid-Range)  (Non-Psychopathic) Test of Proportions  Refused IAS-R: ALL PART  4 2  5 10  2 4  N.S. N.S. N.S.  MMPI Omitted IAS-R  N.S. N.S. N.S. N.S.  6 2 3 2  Education Language Transferred Incomplete MMPI Language Transferred Total Group Number Based on N = 146  4  7 8  4 5  38  65  43  1  N.S. N.S.  C --n J  TABLE VII  Means arid Standard Deviations of Age and Years of Education Across Groups GROUP CLASSIFICATIONS (n)  Mean  AGE SD  EDUCATION Mean SD  1 2 3  (31) (47) (35)  30.19 29.19 30.71  7.52 7.04 8.37  10.23 10.55 11.00  1.41=" 1.69 1.80  N.S.  N.S.  1  (31) (29) (.53)  29.81 31.03 29.41  7.73 7.82 7.42  10.42 10.79  1.50 1.86 1.67  N.S.  N.S.  (75)  31.76 29.01  8.25 7.07  10.35  (38)  11.10  1.43 1.98  N.S.  p < .025  (66) (47)  28.65 31.74  7.02 8.00  TO. 33 10.98  1.56 1.75  p < .05  p < .05  (27) (27) (25)  29.74 29.30 31.44  6.79 6.98 8.69  10.18 10.85 11.08  1.24 1.85 1.58  N.S.  N.S.  (25)  6.91 7.19 7.73  11.10  10.44  N.S.  10.65  1.-42 1.94 1.52  N.S.  (34)  29.04 33.15 29.15  APD  (52) (27)  31.78 29.27  8.45 6.83  10.36 11.33  1.88  N.S.  p < .01  APD-R  (47) (32) -  29.15 31.56  6.82  10.30  8.21  11.28  N = 113  PC GROUP  FACTOR I GROUP APD APD-R N = 79 PC GROUP  FACTOR I GROUP  (20)  NOTE:  10.60  1.34 1.35 1.78  Age  DIFFERENCES Education  N.S.  p < .01  Within each set, Group 1 refers to those meeting the relevant criteria.  U1 00  59  with an age range of 19 to 53 years. 10.70 years  (SD^ = 1.60), ranging  Their average education  from 8 to 16 years.  groups formed by the different classification (Psychopathy Checklist  (PC), PC Factor  III, or APD under DSM-III-R reference to APD and APD-R education  (See Table VII).  same criteria.  =  t  (lll)  =  to be younger  in age and/or  (M = 28.65, SD = 7.02  = 2.18, p < .025) and to have  (M = 10.33, SD = 1.56 versus M = 10.98,  Under APD by DSM-I11, the 75 men meeting in age but tended  the  the  to have somewhat  less  (M = 10.35, SD. = 1.43 versus M = 11.10, £D = 1.98; 2.33, p. < .025).  From the subsample of 79, the 47 men  meeting the APD-R criteria did not differ have less education 1.78; t  with  2.06, p < .025) than the men not meeting  criteria did not differ education  DSM-  In the sample of 113, the 66 men  versus M = 31.74, SID = 8.00; '•t^-^  SD = 1.75; ^ ^ )  criteria  (APD-R), only the groups formed  meeting APD-R criteria tended  slightly less education  Of the  I scores, APD under  showed differences  was  Under APD-R by DSM-III, the 52  the criteria similarly did not differ in age but  tended to have less education  (M = 10.36, SD = 1.34 versus  M = 11.33, SD = 1.88; _t ( 7 7 ) =  2.64, p < .01) than those not  meeting the  to  (M = 10.30, SD = 1.35 versus M = 11.28, S £ =  ( 7 7 ) ; = 2.79, p < .005).  men meeting  in age but tended  criteria.  Overall, the men participating cooperative and  interested.  in this research were quite  A good rapport was generally  established with the men seen.  Once accepted as representing a  60  research  interest  authorities, detailed  independent of the Corrections Service or other  the men were typically quite willing  to provide  life histories and complete the questionnaires  Assurances of confidentiality  served to remove any threat of  personal consequence and many men appeared opportunities  provided.  for conversation.  to enjoy  the  As such, the assessment  situation may have compromised generalization  to an  assessment and may not reflect the responses given  institutional in an applied  context.  Setting Matsqui  is a relatively open institution.  Inmates within  the general population have free movement between the  living  unit, grounds, vocational training centre, academic centre for upgrading or university coursework, and work settings during most daytime hours with the exception of counts before lunch and dinner when all  inmates must be accounted  for.  After the dinner  hour, inmates may only have access to a more limited portion of the facilities,  including the gymnasium, hobby shops, library, or  games room unless specifically provided with a pass. within the segregated  unit have little or no access to the rest  of the institution, and times for their appointments had to be restricted  generally  to the evening hours when the general  population was under more restricted Security staff  Inmates  access.  (men and women) are highly visible within  the  institution, but are not armed.  Major control points, such as  the living unit tiers or main entrance, are run from secure with electronic gate controls.  The perimeter of the  is fenced by a double series of 12-foot chain  rooms  institution  links topped by  barbed wire and watched by video surveillance and  perimeter  guards. Interviews and testing for the current conducted  in a room provided  were  for this purpose within  institution's Health Care Centre. advance and passes provided  research  the  Appointment times were set in  on the day prior to the  The room was quite comfortable, approximately  appointment.  15 feet  with carpeting, blinds, plants, and wall posters.  Office  furniture consisted of a desk, three chairs, a filing and a table with a computer and monitor.  square,  cabinet,  Interviews were  videotaped with a camera and recorder placed beside the desk.  Personnel The research personnel consisted graduate students ranging  of one female and two male  in age from 25 to 35 years, and two  female research assistants aged 27 and 32.  As projects  than this one were being conducted concurrently, participating  could see up to four different  inmates  individuals  connection with various aspects of the research  other  in  program.  Interviews were conducted by all research personnel; the selfreport questionnaire specific to this research were all  62  administered  by the author.  Materials Rating The PC scale each  forms were completed  (0-2) as to how well the inmate met the description item.  1987).  The APD diagnoses were completed with respect  press)  in DSM-III  (APA, 1980) and DSM-III-R  The IAS-R  adjectives completed by simply  adjective  indicating which adjectives  the statement presented.  The MMPI  are  The Rosenberg  (Rosenberg, 1965) is a ten-item questionnaire  item is endorsed on a four-point scale  The  list of 300  appropriate for the descriptive task.  in which  each  (1-4) as to agreement  with  (Hathaway and McKinley,  was the 556-item booklet format, each statement being to as true or  Rosenberg  (1-8) as to its accuracy of description.  (Gough and Heilbrun, 1980) is an alphabetic  considered  to  (Wiggins, Trapnell, and Phillips, in  is a 64-adjective list completed by rating each  on an 8-point scale  for  (APA,  Self reports were obtained with the IAS-R, ACL,  scale, and MMPI.  Scale  criteria.  (Hare, 1985b) consisted of 20 items rated on a three-point  criteria provided  ACL  for the PC and APD  1947)  responded  false.  Procedure Men participating with a semi-structured  in the research were first  interviewed  protocol outlining educational and work  histories, psychological or health problems, family and  other  63  relationships, drug use, juvenile and adult criminal history, and general questions self and others for subsequent interviewer  relating to attitudes and perceptions  (see Appendix A). review.  noted  videotaped  At the conclusion of the interview, the  impressions regarding the inmate with  to verbal style, behaviour, and Institutional  The interview was  files were reviewed for  respect  attitudes.  case management, medical, and  criminal record, staff  regarding  psychological  information relating to past history, impressions, psychiatric contacts, and  psychological testing or  reports.  The interviewer completed PC ratings and APD criteria on the basis of interview and available file information. independent  A second,  rating of the PC and APD criteria was made on the  basis of the videotape of the interview and from file information. At the completion of the interview, the inmate was booked for a subsequent appointment for the completion of the IAS-R, ACL and Rosenberg  scale.  fixed, and a standard  Presentation of these questionnaires introduction  was  provided:  What I would like you to do today is complete a few questionnaires, word lists actually, to describe yourself in a few different ways. All together these questionnaires take about an hour to complete. The first thing I need is a bit of backgrounds information, . . . . The inmate's age, birthdate, education and upgrading any history of or current  reading difficulties, and  employment history were requested.  achieved, recent  The IAS-R was then presented  four times  in a fixed order with both a written and  instructional set.  The inmate was provided with a glossary  reference if he was unsure of the meaning of a word Appendix  verbal for  (see  A).  The IAS-R sequence was presented  as  follows:  Here is the first one, please read the instructions and then I will explain it a bit . . . . What I would like you to do here is describe yourself as an individual, on average, not just here in the institution but you as a person, by rating each of these words on the list below as to how accurately it describes you. Use the numbers from the scale at the top: 1 means the word is 'extremely inaccurate'--it doesn't describe you at all, or some number over to 8 which means the word is 'extremely accurate'--it fits you to a 'T.' So the idea is to rate how accurate the word is in describing you as a person. If you are unsure of the meaning of a word, you can look it up on this other list, which gives you an explanation of how the word is meant. Some of the words are a bit odd so if you're not sure please look it up. If you're still not sure, ask and I'll try to explain it. Any questions about it? O.K., go ahead. At the completion of the first set, it was taken away and the second set provided with a written and verbal  introduction:  This is the same list of words again, but this time I would like you to describe an 'ideal self,' a perfect character for you, sort of the person you would most like to be. So this time rate the words as to how accurately they would describe a perfect character for you. At the completion of the second set, it was taken away and the third set provided with verbal and written  instructions:  O.K., this is the same list again, but this time I'd like you to describe yourself as you think a friend of yours would describe you. Think of someone who knows you pretty well and I'd like you to rate the words as you think they would describe you.  At the completion of the third set, it was taken away and fourth set provided with verbal and written  the  instructions:  O.K., last time for this one, same list again, but this time sort of like the last one but a bit more specific. This time I'd like you to describe yourself as you think some member of the institution staff would describe you. I'd like you to think of someone in particular who has some knowledge of you; I know you don't have a lot of opportunity or interest to talk to staff, or that they necessarily know you, but perhaps your casemanager or work supervisor or someone else you can think of is likely to have an impression of you based on what they see of you. I'd like you to think of someone in particular and note their name, so that for the flip side of this, I can send the same form to the person you name to have them describe you. O.K.? Once the inmate had  indicated someone, the instruction  continued  O.K., good, so now describe yourself as you think he'll/she'll describe you, and then I'll send one to them and ask them to describe you. You won't see theirs and they won't see yours; just try to describe yourself as you think they will, as accurately as you can. At the completion of the fourth set, it was taken away and the ACL was provided with verbal and written  instructions:  O.K., thanks, this next one is a different list of words and this time you don't have to rate the words, just indicate with a check or an 'X' which words describe you. this is to describe yourself again, as you think you are, as an individual, on average, not just in here. Go through the list and check off those words which describe you and leave them blank if they don't. Some of these words are a bit odd, so if you're not sure of a word, please ask and I'll try to explain it. Upon completion of the ACL, it was taken away and Rosenberg scale presented with the following verbal  the  instructions  O.K., here's the last one; just indicate how you'd agree with each of these statements."  With the completion of the Rosenberg scale, the inmate was thanked  for his participation and queried as to whether he would  be available at another time for another questionnaire  if  necessary. Staff particular  ratings of the inmate were obtained by sending staff member, named by the inmate, forms  returnable envelope using the institution mail.  him/her aware of  letter  indicating  in a  The staff  thus received a copy of the IAS-R and glossary with instructions and a cover  written  the interest  in having  it being sent and had given his consent, and that all remain  confidential.  At the conclusion of the research, a representative  sample  respondents was selected and these individuals were  asked to provide a hypothetical descriptive "average  member  rate the indicated person, that the individual named was  responses would  of staff  the  rating of the  inmate" seen at this institution, based on their own  experience.  Responses were received  from eight staff  six men and two women, representing case management vocational  instructors, and security staff.  The MMPI was administered approximately  "stereotype." required  As the MMPI was also  as part of psychological  induction to the institution,  were  staff  at a separate time as it  1 1/2 hours to complete.  routinely administered  personnel,  These responses  pooled to provide a descriptive reference point for perceptions which may be related as an inmate  members:  testing  with  it was given again only to those  inmates who did not have one on file  (had refused testing) or, if  it was on file, the previous administration was more than six months prior to the current assessment.  Inmates completing  MMPI were provided with the 566-item format, hand True/False answer sheets, and written and verbal  the  scorable instructions:  This questionnaire is a bit longer, but looks worse than it is. It takes about 1 1/2 hours to finish. You may have seen it before. This one is a list of statements with which you might agree or not as being true for you. The statements range from simple preferences like 'I like mechanics magazines' to other statements of beliefs, problems, or concerns. So the idea is to read the statement and consider it, on average, as being true or false for you. They don't require much thought, so you can go through it quite quickly. If you have any question as you go through, please ask. O.K.? Go ahead. Throughout the administration of these questionnaires, administrator  remained  completed  forms.  the  in the room and  read while the  the  inmate  Scoring The PC totals for two raters were averaged and the averaged totals greater than or equal to 30 were classified as the psychopathic criterion group.  The averaged  totals which  were  less than or equal to 20 were classified as non-psychopathic, those greater  than 20 and less than 30 constituted  or mixed group. retained  and  the mid-range  The PC totals averaged for two raters, were also  for correlational analyses.  provide assessment  of interrater  The independent  reliability.  ratings  Where ratings were  10 7 discrepant by more than five points, a third was scored and paired  independent  to the closest of the discrepant  Totals for the PC items comprising Factor 1988a) were summed over the two raters. to a maximum of 36.  two.  I (Harpur, et al.,  Totals could thus range  Total scores ranging from 28 to 36  constitute the high group  rating  inclusive  (psychopathic), from 20 to 28 the mid-  range, and less than 20 the low group  (non-psychopathic).  total scores were also retained for correlation  These  analyses.  APD ratings were completed as 3, 2, 1 where 3 indicates not APD, 2 indicates possible APD, and 1 indicates definite APD the criteria for each of DSM-III and DSM-III-R.  under  The use of  "possible" was used here to indicate those men who may meet all but one of the criteria the APA  in either set A or set B as defined by  (1980, 1987) for the diagnoses.  These three point  ratings were used to assess the congruence of APD ratings across both criteria, as well as collapsed criterion ratings of 3 (3, 2) o f l'(l) indicating not or definite APD.  Interrater  reliabilities were taken from a subset of the total The IAS-R  results were entered  sample.  into a computer for  scoring  point profile summaries, polar coordinates, and calculation of distances between points across the different The point coordinates  instructional  sets.  (Dominance, Love) and polar coordinates  are  based on the raw scale scores standardized with reference to the means and standard deviations obtained  from the cumulative  sample  of 1,162 college students reported by Wiggins, Trapnell, and  69 Phillips  (in press).  derived as weighted  The Dominance and Love coordinates  are  linear combinations of the standardized  scale  scores: Dom =  (. 3PA + .212 BC + .212 NO - .212 FG - .212 JK - .3HI)  Lov =  (.296LM + .209 NO + .209 JK - .209 FG - .209 BC- .296 DE)  and the polar coordinates calculated with respect to the point locations  (Phillips, 1983).  obtained  The ACL was handscored  scales Total Checked, Number Favourable, Number  for the  Unfavourable,  Achievement, Dominance, Aggression, Autonomy, Abasement, and Deference.  These scale scores  (with the exception of Total  Checked) were then transformed  to standard scores according  the norms provided  in Gough and Heilbrun  Total Checked and the standardized comparative and correlational  (1980).  to  The raw score  scale scores were entered  for  analyses.  The Rosenberg Scale was totalled with respect to low selfesteem, i.e., appropriate  item scores were reflected so that high  score totals would correspond to low self-esteem. range from 10 to 40.  Scores could  The MMPI profiles obtained were  classified  to groups on the basis of previously defined criteria  (Table II,  p. 17).  Thus, group membership was used as the basis for  analysis using the MMPI as a criterion. of categorical assignment was assessed sample.  Interrater  reliability  for a subset of the  70  RESULTS  DIAGNOSTIC  CRITERIA  This section will outline the obtained distributions of the diagnostic criterion groups; the following section will diagnostic agreement across Psychopathy Checklist  examine  criteria. (PC)  Group assignments under the PC were made with respect to the average of checklist scores obtained scores ranged  from 3.5 to 37  from two raters.  (maximum = 40) with a mean of  (SD = 7.26) in the sample of 113, and a mean of 23.96 in the sub-sample of 79. independent points)  The "adjusted"  23.84  (_SD = 7.85)  (use of a third  rater when the other two differed by more than 5  interrater  reliability was  .896 for the sample of 113 and  .903 for the sub-sample of 79.  The overall unadjusted  reliability  .780.  for an N of 142 was  the intraclass correlation with an N of  These  interrater  (A subsequent analysis of  over 5 pairs of 4 raters was also  .780  174).  In keeping with classification guidelines provided by Hare (1985b), men with averaged scores equal to or greater than 30 were categorized  as psychopaths, men with averaged scores  than or equal to 20 were categorized  less  as non-psychopaths, and  the  71 balance constituted  the mid-range.  into groups as follows: non-psychopathic;  The sample of 113 men  divided  31 psychopaths, 47 mid-range, and 35  the sub-sample of 79 provided groups of 27, 27,  and 2 5.  PC Factor  l_ Scores  The men were also scored from the PC with respect to Factor I item score totals raters.  (Harpur, et al., 1988a) summed for the two  Scores could  range from 0 to 36.  to or greater than 28 were classified  Men with scores  as psychopaths by  this  criterion; those with scores less than or equal to 19 were psychopathic, and the rest constituted  the mid-range.  scores ranged from 5 to 34 with a mean of 21.62 the sample of 113, and a mean of 21.91 sample of 79.  non-  Obtained  (SD = 6.89) in  (SD = 7 . 2 4 )  The sample of 113 provided  equal  in the sub-  groups as follows:  psychopaths, 29 mid-range, and 53 non-psychopathic;  31  the  subsample of 79 yielded groups of 25, 20 and 34 as defined by the Factor  I score  criterion.  Antisocial Personality Disorder  (APD and APD-R)  Diagnostic assignments of APD were made using the specified 1987) with  in DSM-III  (APA, 1980) and APD-R by DSM-III-R  information provided  institutional  files.  criteria (APA,  from the interview and. available  Ratings were made which resulted  (not APD, possible APD, definite APD) or 2-group  in 3-group  (not APD,  72 definite APD) assignments.  In the latter case, "not APD"  corresponds to the collapsing of the "not" and categories of the 3-group set.  Interrater agreements for the 2-  group assignments were calculated independent coefficients  "possible"  on a subsample for whom two  ratings were available; results were fair with Kappa of  .62 for APD and  .55 for APD-R.  The following  results are based on diagnoses provided by one rater  only.  In the sample of 113, 75 men met the criteria for  definite  APD and 66 met the criteria for definite APD-R, 10 were considered possible APD and 20 possible APD-R, 28 men were categorized  as not APD and 27 as not APD-R.  When only 2-group  assignment possibilities were used, 38 of 113 men did not meet the criteria  for APD and 47 did not meet the criteria for APD-R.  In the subsample of 79, 52 men met the criteria for definite APD and 47 for definite APD-R, 8 were considered possible APD and 13 possible A P D - R , 19 were categorized APD-R.  as not APD and 19 as not  Thus, as 2-group assignments, 27 of 79 men did not meet  criteria for APD and 32 did not meet the criteria for APD-R.  Minnesota Multiphasic Personality  Inventory  (MMPI)  MMPI profiles were obtained from 89 of the sample of 113 men and 64 of the sub-sample of 79  (see Tables VIII and IX for the  distribution of MMPI groups previously defined).  For the  purposes of this research the profiles commonly considered related to the  psychopathic  personality are of primary  to be  interest.  TABLE VIII Overall Frequency Distributions of MMPI Group Categories For the Parent Sample (N = 113) -9 MMPI 3 GROUP N = 113  1  "missing" "omits"  2  3  4  "Fake "Fake bad" good" "Spike 4"  8  7  5  6  4/9  4/8 "Neurotic" "Psychotic"  10  9 "Hypomanic"  "Other"  11 "Norm  24  4  4  1  9  6  5  0  2  4  48  6  (31) 2 (47) 3 (35)  8 11 5  1 3 0  1 2 1  0 0 1  4 3 2  2 2 2  2 3 0  :0 0 0  0 0 2  2 1 1  10 19 19  ' 1 3 2  b 1 (31) 2 (29) 3 (53)  8 6 10  2 0 2  1 1 2  0 0 1  4 2 3  2 2 2  2 3 0  0 0 0  0 0 2  3 1 0  8 12 28  1 2 3  APDL  (75) (10) (28)  19 3 2  3 0 1  4 0 0  1 0 0  5 2 2  4 0 2  4 0 1  0 0 0  0 0 2  2 0 2  28 5 15  5 0 1  APD-R1"  (66) (20) (27)  13 6 5  3 0 1  3 1 0  1 0 0  8 0 1  3 1 2  4 0 1  0 0 0  0 0 2  2 0 2  24 12 12  5 0 1  NOTES:  (a) (b)  Overall PC GROUP  h  FACTOR I GROUP  1  (c)  see Table II for definitions of the MMPI Group c a t e g o r i e s . group classifications are defined as 1 = P s y c h o p a t h i c , 2 = m i d r a n g e , 3 = non-psychopathic. group classifications are defined as 1 = Definite A P D , 2 = Possible A P D , 3 = Not A P D .  CJ  TABLE IX Overall Frequency Distributions of MMPI Group Categories For the Sub-Sample (N = 79) 2 "Fake "missing" "omits" bad" -9  MMPI GROUP N = 79 (n) Overall  1  4 3 "Fake good" "spike 4  4/9  10 9 "hypo4/8 "neurotic" "psychotic" manic" "other"  5  6  7  8  11 "normal"  15  2  2  1  8  3  4  0  2  4  32  6  PC GROUP"  (27) (27) (25  5 5 5  0 2 0  1 1 0  0 0 1  4 2 2  2 0 1  2 2 0  0 0 0  0 0 2  2 1 1  10 11 11  1 3 2  FACTOR GROUP"  (25) (20) (34)  5 3 7  0 0 2  1 0 1  0 0 1  4 2 2  2 0 1  2 2 0  0 0 0  0 0 2  3 1 0  7 10 15  1 2 3  APD  (52) (8) (19)  10 3 2  2 0 0  2 0 0  1 0 0  5 0 1  2 0 1  3 0 1  0 0 0  0 0 2  2 0 2  20 3 9  5 0 1  APD-R1"  (47) (13) (19)  9 3 3  2 0 0  2 0 0  1 0 0  8 0 0  1 1 1  3 0 1  0 0 0  0 0 2  2 0 2  14 9 9  5 0 1  NOTES:  (a) (b) (c)  see Table II for definitions of the MMPI group c a t e g o r i e s . group classifications are defined as 1 = P s y c h o p a t h i c , 2 = M i d r a n g e , 3 = Non-Psychopathic. group classifications are defined as 1 = Definite A P D , 2 = Possible A P D , 3 = Not A P D .  Thus, groups 4  ("spike 4"), 5 (49/94), and 6 (48/84 with 9 and/or  6) are the focus here.  In the sample of 113  profiles fit one of these groups  group 6).  20  (9 in group 4, 6 in group 5, and  5 in group 6); the sub-sample of 79 fitting one of these groups  (89 profiles)  (64 profiles) yielded  15  (8 in group 4, 3 in group 5, and 4 in  As can be seen from Table VIII, group 10  profiles failing to meet more straightforward  ("other")—  criteria—was  the  most common, constituting 53.9% of the larger sample and 50% of the sub-sample.  Interestingly, group 7 (a "neurotic"  was not represented.  Interrater agreement  for group  profile) assignments  was 100% for a sub-sample of 55 profiles.  DIAGNOSTIC  CONGRUENCE  Comparisons of diagnostic congruence have been made using lambda  (A) coefficients  when contrasting  (Hays, 1973) of predictive  association  3 group criteria and Kappa coefficients  diagnostic agreement  for collapsed  two group sets.  for  Comparisons  involving the MMPI are based on tests of proportions of the relevant MMPI groups across each contrasted  diagnostic set, as  well as lambda coefficients based on the overall distributions MMPI profiles.  It is also possible to consider  and sensitivity  of the relevant MMPI profiles  the  specificity  in comparison  assignments made by other criteria, particularly  to  the PC.  The ten possible pairwise contrasts of the five diagnostic criteria will be outlined  in the order  listed  in Table X, first  of  TABLE XXXIII Lambda (X ) Coefficients of Predictive Association and Kappa (K) Coefficients of Diagnostic Agreement Across the Different Criteria Based on the Parent Sample (N = 113) COOEFICIENT AB  A  Simple "Agree"  BEST  PC v s . Factor I v s . APD v s . APD-R  .524 .183 .115  ' ( - • P C ) . .545 (—•APD) .211 ( — PC) .121  .778 .183 .114  91.2% 59.3% 56.6%  Factor I v s . APD v s . APD-R  .012 0.0  ( — FI) .017 0.0  .076 .079  54.0% 54.9%  APD v s . APD-R  .471  (—•APD) .500  .529  77.9%  MMPI  L CONTRASTS0  MMPI v s . vs. vs. vs.  NOTES:  PC Factor I APD APD-R  (a)  (b) (c)  (d) (e) (f)  .038 .056 .019 .018  ( - > PC) (—* FI) (—»APD) ( _ > APD-R)  .076 .117 .053 .043  Specificity0  Test o f j Proportions (Approx. Z)  40% 40% 65% 75%  1.29 0.99 4.22 7.98  Sensitivity 25.9% 25.9% 17.8% 22.7%  Lambda coefficients are based on 3x3 group c o m p a r i s o n s , Kappa and simple agreements are based on 2x2 c o m p a r i s o n s ; X ^ g refers to the symmetric avera g e , AD^^J refers to the best predictive relation with the associated dependent variable in p a r e n t h e s i s . MMPI contrasts providing Lambda coefficients are based on the overall MMPI d i s r i b u t i o n . Specificity refers to the proportion of relevant MMPI profile groups (i.e., 4 , 5 , and 6 ) accounted for by the target criterion group (i.e., psychopathic or definite A P D ) . Test of proportions of extreme groups within the diagnostic c a t e g o r y . Sensitivity refers to the proportion of the target criterion group accounted for by the relevant MMPI profile g r o u p s . Test of proportions of extreme groups within the diagnostic c a t e g o r y .  Test of ^ Proportions (Approx. Z) 1.54  2.01  0.86  0.86  CTl  77 comparing  the PC to Factor I, APD, APD-R, and so forth, for  comparisons based on the parent sample of 113 men.  As can be  seen from Table X, the PC and the derivative Factor I scores provide the highest  level of agreement  predictive association  (X^  = .524).  however, that the correspondence Factor  (Kappa = .778) and It is interesting to note,  is less than perfect and that  I scores are a better predictor  totals than the  (Xbest  = .545) of the PC  converse.  The correspondence between the PC classification diagnoses of APD and APD-R  is clearly poor  (Kappas = .183 and  .114 respectively), and much lower than has been previously by Hare  (1981, 1983, 1985a).  and  reported  The differences  here  likely derive from having used the complete data set in a collapsed  2 x 2 comparison of those meeting or not meeting  the  criteria, rather than the association of extreme groups without inclusion of the mid-range as was reported by Hare The relations of the Factor somewhat worse still  I categories to APD and APD-R  (Kappas = .076 and  and,  that knowledge of  status with respect to APD-R diagnosis tells you  nothing about his possible Factor I group membership and versa.  are  .079 respectively)  within this data set at least, it is apparent an individual's  (e.g., 1985a).  vice  The level of agreement evident between APD and APD-R  surprisingly tightening  low  (Kappa = .529) and may reflect the effect of  the adolescent criteria for the APD-R diagnosis as  well as potential  rater  variability.  is  The MMPI profile group identified previously  contribute  little to the prediction of the other diagnostic group (see Table X). 5, and  The more circumscribed  categories  set of profile groups 4,  6 (total 20) do, however, show a modest level of  correspondence to the specific diagnostic criterion particularly APD-R. specificity  groups,  Comparisons here are based on the  of these profiles--defined  for the purposes here as  the proportion accounted  for by the specific criterion group, and  the s e n s i t i v i t y — d e f i n e d  here as the proportion of a specific  criterion group having  the relevant profiles.  The proportions  within the various diagnostic groups were tested as an approximate Z -statistic  (Glass and Stanley, 1970) comparing  extreme groups of each diagnostic set.  It is apparent  profiles are quite specific to the APD-R diagnosis obtained profiles), but are evident (22.7%) of those receiving  proportion  The association of  "psychopathic" MMPI profiles to the PC or Factor group  these  (75% of the  in only a small  the diagnosis.  the  I psychopathic  is rather modest, 40% of the profiles being specific to  these criterion groups and 25.9% of the criterion exhibiting  the relevant  groups  profiles.  Table XI provides the same diagnostic comparisons for subsample of 79.  The relationships among the various  maintain the same pattern as seen in the parent sample  the  criteria although  the absolute magnitude appears marginally greater for all comparative measures.  This suggests that the smaller sample is a  good representation of the parent sample and that the data  remain  TABLE XXXIII Lambda (A ) Coefficients of Predictive Association and Kappa (K) Coefficients of Diagnostic Agreement Across the Different Criteria for the Sub-Sample (N = 79) COEFFICIENT DIAGNOSTIC CONTRASTS  A  PC v s . Factor I vs. APD v s . APD-R  .608 .304 .250  FACTOR I v s . APD v s . APD-R  .097 .078  APD vs. APD-R  .559  MMPI  Simple "Agree"  ( ( (  PC) .635 PC) .327 PC) .269  .771 .317 .238  89.9% 65.8% 62.0%  73.3 28.7 24.9  4 4 4  ( (  FACI) .156 F A C I ) .133  .215 .183  60.8% 59.5%  13.6 10.1  4 4  (  A P D - R ) .563  .648  83.5  71.3  4  BEST  U  CONTRASTS0  MMPI v s . vs. vs. vs.  K  X  AB  PC Factor I APD APD-R NOTES:  X  AB  X  BEST  .081 ( PC) .087 ( FACI) .027 ( APD) .025 ( APD-R) (a)  (b) (c)  (d) (e) (f)  3  .154 .178 .074 .063  Specificity0 53.3% 53.3% 67.7% 80.0%  Test of d Proportions (Approx. Z) 1.59 1.59 2.64 3.72  X  V > .05 < .01 < .01  2  d.;f.  Sensitivity6 29,.6% 32,.0% 19..2% 25,.5%  P  < <  .00005 .00005 .0001 .0088 .0388  <  .00005  Test of ^ Proportions (Approx. Z) 1.55 2.26 0.318 1.34  Lambda coefficients are based on 3x3 group c o m p a r i s o n s , Kappa and simple agreements are based on 2x2 c o m p a r i s o n s ; A ^ refers to the symmetric avera g e , AgEST refers to the best predictive relation with the associated dependent variable in p a r e n t h e s i s . MMPI contrasts providing Lambda coefficients are based on the overall MMPI d i s r i b u t i o n . Specificity refers to the proportion of relevant MMPI profile groups (i.e., 4 , 5 , and 6 ) accounted for by the target criterion group (i.e., psychopathic or definite A P D ) . Test of proportions of extreme groups within the diagnostic c a t e g o r y . Sensitivity refers to the proportion of the target criterion group accounted for by the relevant MMPI profile g r o u p s . Test of proportions of extreme groups within the diagnostic c a t e g o r y .  P^  < .02  80 quite robust despite the loss of 30% of the sample.  DEPENDENT  MEASURES  Interpersonal Adjective Scales - Revised As a preliminary  evaluation of the adequacy of the  data, a principal components analysis of the (set 1) was conducted  (IAS-R)  self-descriptions  for the sample of 113 men.  consistent with a 2-component solution  IAS-R  The data  were  (eigenvalues of 3.26 and  2.57, all others < .7; accounting for 72.8% of the variance) which provided  a good distribution of the scales suggesting a  circumplex space  (see Figure 2).  The same analysis for the  subsample of 79 was similarly encouraging, with the 2-component solution accounting  for 69.7% of the variance.  Since complete IAS-R protocols  (sets 1 through 5) were only  obtained for the subsample of 79, the following outline of descriptive aspects of the IAS-R measures subsample.  is based on that  Table XII provides the summary statistics  from standardized  scores  the  derived  (see p. 69) yielding the DOM and LOV  coordinates, polar coordinates expressed as the angle corresponding associated  to the mean Dom and Lov coordinates with  vector  length, and the modal octant  sample of 79 across the five IAS-R  the  locations for  representations.  A  manipulation  check, or test of the adoption of different  perspectives  as instructed  representations,  for the different  the  IAS-R  is afforded by the assessment of the main  effect  Figure 2 Obtained 2-factor solution for the circumplex based on N = 113. Note: The obtained scale locations provide a clockwise rotation, convention reflects to counter-clockwise.  T A B L E XXXIII  Summary Statistics for the Overall Distributions of Point Coordinates and Modal Octant Locations for the Sample N = 79 Standard Deviation  TA C  D  1.064 1.336  1.  Dom .304 Love -.511 Self Angle 3 149.25 Vector Length 1.615 (BC) 2(21.5%) Modal Octant Dom Love Ideal Angle Vector Length Modal Octant  .982 .420 66.84 1.812 (NO) 8(26.61)  .837 1.499  Dom : Love 3. As Friend Angle Vector Length Modal Octant  .478 -.461 133.96 1.889 (DE) 3(24.1%)  1.133 1.653  Dom Love 4. As Staff Angle Vector Length Modal Octant  .432 -1.180 159.89 2.003 (DE) 3(38.0%)  .906 1.831  Dom Love 5. By Staff Angle Vector Length Modal Octant  .125 -1.377 174.81 2.027 (DE) 3(35.4%)  1.103 1.582  IMo SET  2.  Coordinates  Mean  —  .796 —  Minimum  Maximum  -3.020 -3.230 3.000 .090  3.250 2.800 342.000 4.35 —  —  —  .881  -.770 —3.060 1.000 .080  3.190 3.700 360.000 4.420 —  —  —  .920 —  -2.890 -3.900 4.000 .240  3.270 3.430 359.000 4.320 —  —  —  1.307  -2.140 -7.060 7.000 .080  2.850 2.96 347.000 7.080  —  —  —  1.143  -1.640 -5.310 45.000 .370  —  2.540 2.090 350.000 ; 5.440.  —  Staff Rating of "Average Inmate" (N = 8) SD  Mean Dom Love Angle Vector Length Modal Length NOTE: (a)  .344 -4.00 175.08 4.04 (DE) 3 (100%)  .480 1.399 —  1.403 —  Min  Max  -.63 -6.66 168.00 1,99 •  1.00 -1.98 188.00 6.67  ANGLE is derived as ARCTAN DOM/LOV using the mean DOM, LOV coordinates obtained for each set.  for the obtained Dora and Lov coordinates within a repeated measures MANOVA; that test is highly significant Figure 3 provides the point representations profiles obtained  (p. < .0005).  of the different  from the sample of 79, and Figures 4 through 7  show the profiles obtained  for the various group  criteria.  Overall, the average self-description provided by these men is surprisingly negative in that they tend to endorse descriptions placing them clearly within the calculating/cold  hearted" octants  is considerable variability  review of the scatter-plots  "arrogant-  (BC/DE) of the IAS-R.  as indicated by the ranges of obtained there  coordinates  in the group as a whole.  for these profiles  range of self-descriptions.  positive shift for the representation coming  However,  (Table XII),  to  outliers  There is an apparent  of an ideal-self, with most  to occupy octant NO and ascribing to positive  characteristics as warm and outgoing. perceptions  The depictions of  as a friend or as a staff member, as well as the  ratings provided by staff are relatively consistent and emphasize the negative aspects of a cold and aloof  re-  interpersonal  style, perceived both by staff and the inmates themselves. description staff  From  it is difficult  say that the variability derives from a few distinct versus a broad  adjectival  The  of an "average inmate" provided by a sample of the  respondents yields some suggestion of a staff-perceived  inmate stereotype. interesting  The relative location of this  'stereotype 1  for its descriptive associations and in making  it  is  90°  S I AF AS BS  = self = Ideal = as friend = as staff = by staff  PA  —14  135°  45*  BC  NO  as  'average inmate' bs  180° I D 6 -4-  •  af  S ;  -z  -2  FG  JK  225°  315° -J-4  HI 270°  Figure 3 Mean Dom and Lov coordinate locations of IAS-R response sets for the overall sample (N = 79)  85 S I AF AS BS  90°  PA  = self = Ideal = as friend = as staff = by staff  • Psychopaths ^Midrange • Non-psychopaths  135°  45° NO  BC  i i BS  S  AS  •  AF • AF  A  AS  AS  180° L  AF  0V LM  r  DE -2  BS  BS  -i  FG  JK 315°  225°  -2  HI 270° Figure H Mean Dom and Lov coodinate locations of IAS-R response sets for the PC groups.  86 90° PA 2-i  S I AF AS BS  = self = Ideal = as friend = as staff = by staff  • Psychopaths A Midrange • Non-psychopaths  135° BC  45° NO I  BS AS  •  S  AF  A •  AS  S  A A  BS  180° L DE -z  AF A  AS  AF  0°  -1  LM  BS  -i  FG 225°  JK 315° 1-2  HI 270° Figure 5 Mean Dom and Lov coordinate locations of IAS-R response sets for the Factor I groups.  87 S I AF AS BS  90°  PA  2 —,  = self = Ideal = as friend = as staff = by staff • •  135°  APD not APD  45°  BC  NO  i  i-  AF AF  AS  180°  IS AS  BS  2  J 0° LM  DE  BS  -i  FG  JK  225°  315° -2  HI  270°  Figure 6 Mean Dom and Love coordinate locations of IAS-R response sets for the APD (DSM-III) groups.  88 S I AF AS BS  90° PA  2—  = self = Ideal = as friend = as staff = by staff  • APD-R • Not APD-R  135°  45°  BC  NO  i-  BS  AS  AF AS  i  AF  180° L  DE -z  J 0° LM  -i BS  -1  FG  JK  225°  315e -J-2  HI  270°  Figure 7 Mean Dom and Lov coordinate locations of IAS-R response sets for the APD-R (DSM-lll-R) groups  89  evident that staff are providing reasonable discriminations the inmates on an individual  basis.  An indication of the potential relation of vector relative psychopathy  is provided by the correlations  between the PC scores and vector vector  length  length and Factor  in the sample of 79.  .248  (p = .014) with Factor  .235  length to  obtained  I scores and  When taken from  description, vector length correlates scores and  self-  (p = .019) with PC  I scores.  When taken  ratings of the inmates provided by staff members, vector correlates  .461  (jo < .0001) with the PC scores and  (p < .0001) with Factor  Adjective Checkl ist  among  from  length  .437  I scores.  (ACL) and Rosenberg' Sea le  The overall distributions of the selected ACL scales are quite congruent with the standardizing population of males reported  in the ACL Manual  summarized  (Gough and Heilbrun, 1980), and are  in Table XIII for the subsample of 79.  Although  evidently somewhat skewed, an adequate distribution was also obtained  for the Rosenberg scale.  scales obtained  Correlations  of the  from self-description with the selected  scales and the Rosenberg scale are reported  IAS-R ACL  in Table XIV for the  total sample of 113 and Table XV for the subsample of 79. be seen  in these tables, substantial correlations consistent  the expected a good  As can  relations for these scales were obtained and  level of consistency  in self-report.  with  reflect  90  TABLE XIII Summary Statistics for Selected Scales of the Adjective Checklist and the Rosenberg Scale Obtained from Self-Description (N = 79)  Mean ACLTOT  3  Standard Deviation  Minimum  Maximum  131.747  37.210  45  213  FAV  50.152  7.873  29  65  UNFAV  51.266  10.629  36  78  ACH  49.810  8.662  31  66  AUT  50.595  10.736  19  77  DOM  50.519  10.229  21  70  AGG  50.671  10.123  29  77  ABA  49.025  11.492  21  82  DEF  49.367  10.855  .16  71  ROSEN*5  18.278  5.000  10  30  NOTES:  (a)  ACLTOT refers to the simple total number of adjectives endorsed, the remaining ACL scale scores are T-scores taken from male norms provided in the ACL Manual (Gough & Heilbrun, 1980).  (b)  The Rosenberg scale is scored so that high scores reflect low self-esteem (maximum score = 40).  TABLE XIV CORRELATIONS OF IAS-R SCALES FROM SET 1 WITH THE SELECTED ACL SCALES (AND THE ROSENBERG SCALE) (N = 113)  ACLTOT  FAV  UNFAV  ACH  DOM  AUT  AGG  DEF  ABA  ROSEN  PA  .193 (.02)  .274 (.002)  -.165 (.04)  .570* (.0000)  .646* (.0000)  (.'oooo)  .413*  .375* (.0000)  -.644* (.0000)  -.434* (0000)  -.530* (.0000)  BC  .113 (.118)  -.119 (.105)  .174 (.033)  .111 (.122)  .193 (.020)  .367* (.0000)  .284 (.0012)  -.327* (/0000)  -.383* (.0000)  - . 109 (.1255)  .054  -.475*  (.0010)  .472 (.0000)  -.153 (.053)  . 559*  (!oooo)  .288  (.286)  (.381)  (.0000)  .462* (.0000)  .001 (.496)  - . 587* (.0000)  (.0000)  .493*  -.403* (.0000)  -.463* (.0000)  .213 (.012)  (.237)  -.032 (.368)  -.359*  .202  (.0000)  .102  .068  -.498*  (.0000  (.018)  .185 (.025)  -.123 (.097)  .516* (.0000)  .435*  - . 547*  -.395*  (!oooo)  -.720* i (.0000)  -.358*  (.016)  (.0000)  (!oooo)  .667* (.0000)  (!oooo)  .594* (.0000)  -.034 (.360)  (.141)  - . 104 (.136)  - . 150 (.056)  -.229 (.0074)  -.284 (.0012)  -.299 (.0007)  .323* (.0002)  .351* (.0001)  .107 (.130)  .222 (.009)  .473* (.0000)  -.463* (.0000)  .156 (.050)  -.020 (.415)  -.484* (.0000)  -.427* (.0000)  .301* (.0006)  .495* (.0000)  -.226 (.0080)  .012 (.142)  .559* (.0000)  -.526* (.0000)  .411* (.0000)  .458*' (.0000)  -.190  -.147  -.220  (.060)  (.0095)  .104 (-137)  -.623* (.0000)  NOTE:  (.022)  For a table of 80 c o r r e l a t i o n s , a conservative level of significance may be estimated as .05/80 = .0006; asterisks indicate correlations significant at that level or less ; numbers in parentheses are the associated probabilities  TABLE XV CORRELATIONS OF IAS-R SCALES FROM SET 1 WITH THE SELECTED ACL SCALES (AND THE ROSENBERG SCALE) (N = 79)  ACLTOT  FAV  UNFAV  AGG  ABA  DEF  ROSEN  .425* (.0000)  .432* (.0000)  - . 563* (.0000)  -.407* (.0001)  -.406* (.0000)  .128 (.130)  .206 (.034)  .393* (.0002)  .374* (.0003)  -.320 (.0020)  -.409* (.0001)  -.097 (.199)  .428* (.0000)  -.059 (.304)  .119 (.147)  .564* (.0000)  .523* (.0000)  -.364* (.0005)  - . 507* (.0000)  .146 (.099)  -.511* (.0000)  .386* (.0002)  -.284 (.006)  -.468* (.0000)  .187 (.049)  -.008 (.471)  .171 (.065)  - . 100 (.191)  .430* (.0000)  .007 (.476)  -.261 (.010)  .093 (.208)  -.457* (.0000)  -.695* (.0000)  -.366* (.0005)  -.477* (.0000)  .631* (.0000)  .416* (.0001)  .528* (.0000)  .008 (.472)  .203 (.037)  - . 198 (.041)  -.164 (.075)  -.257 (.011)  -.312 (.003)  -.429* (.0000)  .336 (.001)  . 384* (.0002)  .089 (.217)  .248 (.014)  .398* (.0001)  -.341 (.001)  -.002 (.494)  -.151 (.092)  -.495* (.0000)  -.444* (.0000)  .443* (.0000)  .517* (.0000)  -.074 (.259)  .088 (.222)  .432* (.0000)  -.376* (.0003)  .258 (.011)  .441* (.0000)  -.165 (.073)  -.064 (.286)  -.191 (.046)  .079 (.244)  567* (.0000)  .150 (.093)  .143 (.104)  .015 (.447)  .454* (.0000)  BC  .103 (.182)  -.212 (.031)  .276 (.007)  DE  .053 (.322)  -.451* (.0000)  FG  .037 (.373)  HI JK  NO  AUT  . 581 * (.0000)  PA  LM  DOM  ACH  NOTE:  •  Asterisks indicate correlations significant at P .05/80 = .0006; numbers in parentheses are the associated probabilities.  M  93  TABLE  XVI  Overall Frequency Distributions of Octant Assignments Across the IAS-R Rating Sets (N = 79)  IAS-R RATING  Octant 2 (BC) 3(DE) 4(FG) 5(HI) 6(JK) 7(LM) 8(NO) j_(PA)  1.  Self  21 .5  20.3  13.9  6..3  5.1  12.7  8 .9  11.4  2.  Ideal  25 .3  6.3  3.8  1,.3  0.0  21.5  26 .6  15.2  3.  As Friend  21 .5  24.1  7.6  5,.1  6.3  12.7  8 .9  13.9  4.  As Staff  29 .1  38.0  3.8  2 .5  6.3  5.1  8 .9  6.3  5.  By Staff  20 .3  35.4  16.5  7 .6  6.3  2.5  3 .8  7.6  NOTE:  Frequencies are expressed as Percentages of.the sample.  94  COMPARISONS OF IAS-R DESCRIPTIONS WITHIN DIAGNOSTIC Octant  Locations  The most basic summary descriptive the IAS-R  GROUPS  is given by the Octant  information derived  locations of the profile points  obtained; the overall frequencies obtained are outlined XVI.  from  in Table  For the purposes here, comparisons were made of the  frequency distributions of profile points falling of the circumplex  into quadr.ants  formed by adjacent octants as 2/3  arrogant-calculating, cold-hearted; 4/5  (BC/DE) -  (FG/HI) - aloof-  introverted, unassured-submissive;  6/7  ingenuous, warm-agreeable; and 8/1  (NO/PA) - gregarious-  extroverted, assured-dominant. be expected  that octants 2  As previously outlined, it was to  (BC) and 3 (DE) would capture  psychopathic profile, particularly obtained  (JK/LM) - unassuming-  the  from others' description.  frequencies were tested by Chi-square for each of the  definitional criteria using the inmates' self-description and staff  ratings.  the  Given two ratings and four sets of  classifications, each Chi-square was assessed level of  The  at the  (.05/8)  .006  significance.  The frequencies of quadrant  location for  self-descriptions  and staff descriptions of the men categorized by the PC are outlined  in Table XVII.  for dissimulation  It is surprising, given the expectation  in this population, that the  self-descriptions  of these men show a modest tendency for differential  assignment  T A B L E XXXIII  Proportions of PC Groups Occupying IAS-R Adjacent Paired Octants Assigned from Self-Description and Staff Description (N = 79)  IAS-R ADJACENT OCTANT PAIRS PC GROUP  Self Description (Set 1)  Staff Description (Set 5)  2/3 (BC/DE)  4/5 (FG/HI)  6/7 (OK/LM)  8/1 (NO/PA)  7.4%  11.1%  2/3 (BC/DE)  4/5 (FG/HI)  6/7 (JK/LM)  18.5%  85.1%  7.4%  0.0%  : 8/1 (NO/PA)  Psychopathic  27  62.9%  Midrange  27  37.0  29.6  11.1  22.2  51.8  29.6  3.7  14.8  Non-Psychopathic  25  24.0  24.0  32.0  20.0  28.0  36.0  24.0  12.0  NOTES:  (a) (b)  2 * X  (6)  ( 6 )  = 1 2 . 7 8 , p < .05; . A f t B = . 1 4 3 , A  -(^PC)  = 23.01, p < .001; ^ g  = .195,  p c )  7.4%  = .231 = .288  KD Ul  10 7 in which  those  identified as psychopaths by the checklist tend to  identify with octants 2/3  (overall X 2 f r ,  = 12.78, p < .05).  — (b)  As  —  expected, staff descriptions show a strong tendency to place psychopaths  in octants 2/3  (~X2  (6)  = 23.01, p < .001), with 85% —  (23 of 27) of the psychopathic group so placed, compared 51.8%  (14 of 27) of the mid-range and 28% of the  with  non-psychopathic  group. When grouped with respect to scores on Factor  I of the PC,  (see Table XVIII) there is no evident differential assortment by self report  (X ^  = 10.36, p > .10), but staff  descriptions  again show a strong tendency to locate psychopaths 2  (X ,.,  — (6)  50%  = 21.57, p < .005) placing 88%  —  (22 of 25) compared  (10 of 20) of the mid-range and 35.8%  psychopathic  in octants 2/3 to  (12 of 34) of the non-  group.  Classification by the DSM-III  (APA, 1980) criteria for APD  (Table XIX) showed no difference between the groups meeting or not meeting the criteria when compared on self (X 2  ^ )  = 0.83, p > .50), but a strong trend —  (X  (3)  as APD in octants 2/3 identified  (3)  in staff  o  descriptions  1987),  report  as APD  =  15.63, P < .002) to place those  (69.3%: 36 of 52) versus those not  (29.6%: 8 of 27).  Under APD by DSM-III-R  (Table XX) groups did not differ by self =  2  '74' P > -40)  by staff descriptions  identified  and  showed a somewhat =  report  less robust trend  10.35, p < .02), with  of 47) of those identified as APD(-R) falling  (APA,  66.1%  in octants  2/3  (32  T A B L E XXXIII  Proportions of Factor I Groups Occupying Adjacent Paired Octants Assigned from Self-Description and Staff Description (N = 79) IAS-R ADJACENT OCTANT PAIRS a Staff Description (Set 5) Self Description (Set 1)  FACTOR I GROUP n  2/3 (BC/C (BC/DE)  Psychopathic  25  52.0%  Midrange  20  55.0  Non-Psychopathic  3 4  2 g  NOTES:  5  (a) (b)  4/5 (FG/HI) 8.0%  6/7 (JK/LM) 1.2.0%  8/1 (NO/PA) 28.0%  2/3 (BC/DE) 88.0%  .4/5 (FG/HI) 0.0%  6/7 (JK/LM) 4.0%  20.0  10.0  15.0  50.0  25.0  5.0  29.4  26.5  17.6  35.3  41.2  14.7  x  (6  ) = 10-36, £ > .10; X ^  ( 6 )  = 21.56, i < .005,  = .066,  A a b = .163,  n )  pl  =  .j  8/1 N0/P (NO/PA) 8.0% 20.0 8.8  .111  .244  U3  T A B L E XXXIII  Proportions of APD Groups Occupying Adjacent Paired Octants Assigned them from Self-Description and Staff Description (N = 79)  IAS-R ADJACENT OCTANT PAIRS Self-Description (Set 1)c  Staff Description (Set 5) 2/3 (BC/DE)  4/5 (FG/HI)  6/7 (JK/LM)  8/1 (NO/PA)  19.2%  69.3%  21.2%  1.9%  7.7%  22.2  29.6  29.6  APD GROUP  n  2/3 (BC/DE)  4/5 (FG/HI)  6/7 (JK/LM)  8/1 (NO/PA)  Meet APD  52  44,.3%'  21.2%  15.3%  Not Meet APD  27  37.0  18.5  22.2  NOTES:  (a) (b)  2  * (3) 2 x  ( 3 )  0.83  £ > .50;  = 15.63, p <  A a b = 0.0,  X  APD)  22.2  18.5  = 0.0  .002; X A B = .097, A ( _ > A p D ) = .222  U3 00  TABLE  XX  Proportions of APD-R Groups Occupying Adjacent Paired Octants Assigned from Self-Description and Staff Description (N = 79) IAS-R ADJACENT OCTANT PAIRS Self-Description (Set l) a  Staff Description (Set 5) b  APD-R GROUP  n  2/3 (BC/DE)  4/5 (FG/HI)  6/7 (JK/LM)  8/1 (NO/PA)  2/3 (BC/DE)  4/5 (FG/HI)  6/7 (JK/LM)  8/1 (NO/PA)  Meet APD-R  47  42.5%  25.5%  14.9%  17.1%  68.1%  21,3%  2.1%  8.3%  N O t  APD-R  3 2  4 0 , 6  NOTES:  1 2 , 5  (a)  x  (b)  x (3)  ( 3 )  2  2 1 - 9  = 2.74, p > =  10  2 5 : 0  3 7 , 5  .40',  -35> £ < -02'  A A B = 0.0, X { ^ A  AB  =  -090'  2 8 - 1  A p D  _R)  APD-R) =  1 8 , 8  1 5 , 6  0.0 -188-  V£>  versus 37.5%  (12 of 32) of those not identified  In terms of the more utilitarian  as APD(-R).  idea of predictive  association, knowledge of an individual's quadrant contribute to a reduction  location  in the probability of error  does  in  predicting group memberships within each of the diagnostic categories, particularly when using the staff  descriptions.  Lambda coefficients under each of the above conditions are noted in Tables XVII-XX.  The predictive utility of self  reports  ranges  from zero with both APD groups, to .231 or a 23.1% reduction the probability of error when predicting PC group given an individual's octant better, ranging from  .188  location.  membership  Staff reports do somewhat  (an 18.8% reduction) for predicting  APD-R to ^288 for the prediction  of PC groups.  COORDINATE DISTRIBUTIONS AND GROUP ANALYSES  (MANOVAs)  Comparisons were made of the Dom and Lov coordinate over all of the IAS-R sets across groups defined by the criteria using repeated measures MANOVA designs. MANOVA  results  in  values various  Significant  (p < .025) were followed by univariate ANOVAs with  significance set at p < .005  (.05/10), and these were followed  post-hoc comparisons using Tukey's HSD at p < .05 and  by  Scheffe's  comparisons at p < .01. Based on the averaged multivariate tests of significance  for  group differences within a repeated measures design, only the PC groups  (Wilk's F,. ,, n l  = 4.75, p < .001) and the Factor  I groups  101 (Wilk's F  —(4,150)  = 4.96, p < .001) yielded  significant  —  differences across the Dom and Lov coordinates.  Univariate  comparisons for the PC groups indicated only the Dom and Lov coordinates obtained (Dom5: • F  from staff  ratings to differ at p < .005  = 10.6, p— = .0001; Lov5: F— ,(z,/ _ _b) = 9.4,  p = .0002).  Post-hoc contrasts by Tukey's HSD at p < .05  indicated the psychopathic group was scored higher on Dom  than  either the mid-range or non-psychopathic groups and lower  than  both other groups on Lov. indicated  Scheffe's contrasts at p < .01  the differences to be retained only for  psychopathic versus non-psychopathic groups. PC groups are summarized  the  The results for  the  in Table XXI.  Table XXII outlines the results from the Factor I groups. Univariate comparisons were most significant for the obtained  from staff ratings  Lov5: F ^ 2  (Dom5: F  = 7.63, p = .0001).  ^  75)  =  coordinates  14.59, p <  .00005;  Post-hoc contrasts by  HSD at p < .05 indicated that both the mid-range and  Tukey's  psychopathic  groups were scored higher by staff on Dom and that the psychopathic group was scored  lower than both the mid-range  non-psychopathic groups on Lov. retained  Scheffe's contrasts at £ < .01  these differences only between the psychopathic and  psychopathic groups.  A difference was also evident across  for self ratings of ideal on the dimension (Lov2: F  and  76)  =  6  - H ' E. < .004).  were significant only at the .05 level and  groups  Lov  Post-hoc comparisons  here  indicated both  non-  the  TABLE XXI  MEANS (AND SDS) OF DOM AND LOV COORDINATES ACROSS GROUPS DEFINED BY THE PC GROUP COORDINATES  3  P (n=27)  M (n=27) .172 ( 1 . 2 8 5 ) - . 6 5 7 (1.271)  NP (n=25) .080 ( .800) .053 (1 .171)  cotLr n c  DOM 1 LOV 1  .644 ( .984) - . 886 (1.410)  T nc Al lUtML  DOM 2 LOV 2  1 . 0 7 2 ( .902) - . 0 1 7 (1.446)  AS FRIEND  DOM 3 LOV 3  .780 ( .929) - . 6 6 4 (1.741)  .411 ( 1 . 2 9 2 ) - . 6 1 8 (1.667)  .226 (1 .118) - . 0 7 1 (1 .533)  AS STAFF  DOM 4 LOV 4  .639 ( .767) - 1 . 5 5 0 (1.985)  .480 ( 1 . 0 0 0 ) - 1 . 3 5 5 (1.672)  .157 ( .903) - . 5 8 9 (1 .745)  BY STAFF  DOM 5 LOV 5  .740 (1.117) - 2 . 2 5 8 (1.439)  - . 0 0 8 ( .849) -1.271 (1.395)  - . 3 9 5 ( .688) - . 5 3 9 (1 .472)  .951 ( . 8 4 2 ) .238 ( 1 . 5 6 1 )  .918 ( .783) 1.090 (1 .295)  COMPARISONS rnnunTNATF^ COORDINATES  Univariate F(2,76)  Tukey HSD (p < .05)  SELF  DOM 1 LOV 1  2.21 3.69  P > NP  IDEAL  DOM 2 LOV 2  0.24 4.16  P < NP  AS FRIEND  DOM 3 LOV 3  1.65 1.02  AS STAFF  DOM 4 LOV 4  1.94 2.03  BY STAFF  DOM 5 LOV 5  10.61 9.43  NOTE:  (a)  p = .0001 p = .0002  P > M, NP P < M, NP  Scheffe (p < .01)  P > NP P < NP  G r o u p s are: P = P s y c h o p a t h i c , M = M i d r a n g e , NP = N o n - P s y c h o p a t h i c .  T A B L E XXII  MEANS (AND STANDARD DEVIATIONS) OF DOM AND LOV COORDINATES ACROSS GROUPS DEFINED BY FACTOR 1 GROUP 3 P ( n=25)  COORDINATES  M (n=20)  NP ( n=34)  SELF  DOM 1 LOV 1  .749 -.814  .968) 1.445)  .438 (1.096) .893 (1.086)  -.101 ( .987) -.063 (1.288)  IDEAL  DOM 2 LOV 2  1.038 .052  .901) 1.446)  .938 ( .766) .189 (1.424)  .966 ( .850) 1.050 (1.374)  AS FRIEND  DOM 3 LOV 3  .817 .398  .955) 1.632)  .656 (1.000) -1 .186 (1.401)  .125 (1.249) -.081 (1.708)  AS STAFF  DOM 4 LOV 4  .644 -1.468  .739) 2.089)  .447 (1.045) -1 .219 (1.347)  .268 .923 -.944 (1.893)  BY STAFF  DOM 5 LOV 5  .792 -2.311  .891) 1.511)  .248 (1.108) -1 .092 (1.275)  -.438 ( .684) -.857 (1.526)  COMPARISONS Univariate F (2,76)  COORDINATES SELF  DOM 1 LOV 1  5.33, p < .007 3.59, p < .04  IDEAL  DOM 2 LOV 2  0.09 6.11 , p < .004  AS FRIEND  DOM 3 LOV 3  3.18, p < .05 2.99  AS STAFF  DOM 4 LOV 4  1.25 0.59  BY STAFF  DOM 5 LOV 5  NOTE:  14.59, p < .00005 7.63, p = .001  (a)  Tukey HSD (p < .05)  Scheffe (p < .01)  P > NP  P > NP  —  P, M < NP — —  —  P, M > NP P < M , NP  - - -  P > NP P < NP  G r o u p s are: P = Psychopathic, M = Midrange, NP = N o n - P s y c h o p a t h i c .  104 psychopathic and mid-range groups to have scored non-psychopathic group. significance  (F  ^  75)  One other measure =  5.33, £ < .007) suggesting  significant reflecting  interaction  for the Factor (Wilk's Z(i6  the  non-psychopathic  (£ < .01) on the Dora dimension taken from The overall MANOVA  self-description.  I groups also revealed a  138)  =  2  -53,  < .002)  the alternate directions of group differences  the significant contrasts outlined, and the relationship of the mid-range  across  inconsistent  group.  The Dom and Lov coordinates and univariate Fs obtained the APD and APD-R groups are outlined no overall comparisons  reached  the  approached  psychopathic group to score higher than the group  lower than  in Tables XXIII and  for  XXIV;  significance.  Discriminant analyses based on the Dom and Lov  coordinates  provided an overall hit rate of 58.4% for the three group classifications under the PC, and performed best for the psychopathic group in correctly classifying  70.4%.  As  expected,  based on the post-hoc comparisons from the foregoing MANOVAS, coordinates obtained from the staff descriptions contributed to the discrimination. first discriminant and  Pooled within-group correlations  function were -.643 for the Dom5  the most  to the  coordinate  .606 for Lov5. The analysis for the Factor  I groups provided a hit rate of  65.8% overall, and correctly classified group.  72% of the psychopathic  Again the Dom and Lov coordinates obtained  from staff  105 TABLE XXIII M e a n s (and S t a n d a r d Deviations) of DOM and LOV Coordinates A c r o s s G r o u p s Defined by APD ( D S M - I I I ; A P A , 1980)  GROUP COORDINATES  0  APD-R  (n=52)  N o t APD  (n=27)  Univariate F ^  SELF  DOM 1 LOV 1  .357 .632  (1.083) (1.376)  .203 -.277  (1.040) (1.247)  IDEAL  DOM 2 LOV 2  .966 .363  .867 (1.544)  1.012 .531  ( .790) (1.432)  .054 .220  AS FRIEND  DOM 3 LOV 3  .519 -.598  (1.026) (1.718)  .401 .198  (1.332) (1.514)  .189 1.04  AS STAFF  DOM 4 LOV 4  .510 -1.408  ( .896) (1.894)  .283 .740  ( .923) (1.647)  1.12 2.41  BY STAFF  DOM 5 LOV 5  .252 1.897  (1.060) (1.440)  .120 .376  ( .882)  2.44 20.55  NOTES:  (1.363)  JJ)  .369  1.26  p > .10 i < .00005  (a)  G r o u p s are APD = meet DSM-III c r i t e r i a for •Antisocial'Personality D i s o r d e r , not APD = do not m e e t the c r i t e r i a .  (b)  U n i v a r i a t e F for 2 groups c o r r e s p o n d s to a T-test as  ~(v2)  =  V^(v1,v2)'  106 TABLE XXIV M e a n s (and S t a n d a r d D e v i a t i o n s ) of DOM and LOV C o o r d i n a t e s A c r o s s G r o u p s D e f i n e d by APD-R ( D S M - I I I - R ; A P A , 1987)  GROUP COORDINATES  APD-R  3  (n=47)  Not A P D - R ( n = 3 2 )  b  Univariate F ( i , 7 7 )  DOM 1 LOV 1  .271 -.606  (1.028) (1.368)  .353 -.371  (1.130) (1.297)  .111 .588  DOM 2 LOV 2  .871 .314  ( .890) (1.569)  1.144 .576  ( .736) (1.401 )  2.06 .577  AS FRIEND  DOM 3 LOV 3  .429 -.676  ( .977) (1.684)  .552 -.144  (1.343) (1.577)  .223 2.00  AS STAFF  DOM 4 LOV 4  .469 -1.544  ( .885) (2.034)  .379 -.644  ( .947) (1 .342)  BY STAFF  DOM 5 LOV 5  .350 -1 .875  (1 .096) (1.506)  -.205 -.645  ( .780) (1.415)  C F I P s t L r  T n C M  NOTES:  "  .187 4.83 £ <  6.07 13.33  p< £=  (a)  G r o u p s are A P D - R = m e e t criteria f o r antisocial personality d i s o r d e r ( D S M - 1 1 1 - R ) . n o t APD-R = do not m e e t the c r i t e r i a .  (b)  UnivariateT  for 2 groups c o r r e s p o n d s to a T - T e s t as =  *U(v1,v2)'  .04  .02 .0005  10 7 ratings contributed most, and correlated first discriminant  function.  The Lov3  .610 and -.418 to the  (as friend) and  (ideal) coordinates contributed best to the second function, correlating contributed  relatively  only for the remaining  .638 and  Lov2  discriminant  .575, although the second  little to group discrimination,  function  accounting  14% of the variance after the first  function. Interestingly, although not achieving significance foregoing MANOVA, the discriminant  in the  function based on the Dom and  Lov coordinates for the two group discrimination of Yes/No APD-R  2 did reach significance  (X  = 21.80, p < /02).  (staff rating) contributed most to the correlating .421).  .700 followed by Dom5  This provided  classified  discrimination,  (- .472) and Lov4  (as staff:  an overall hit rate of 68.4%, and  68.1% of those categorized as APD-R.  functions for the APD groups did not reach (X 2 (io)= I***  Here Lov5  The  correctly  discriminant  significance  p < .05) within this set of analyses.  Distance Measures The comparisons of primary to the hypothesized  interest to this research  relate  differences expected across the sets of IAS-R  responses for the various criterion groups, particularly those defined by the PC and Factor I scores.  for  Figure 3 (p. 84)  shows the location of the profiles for the total subsample of 79 men.  10 7 The pattern of distance measures taken among the five sets of responses were analyzed design.  in a oneway between groups MANOVA  Although demonstrating differences  in the obtained  coordinates of the 2-dimensional system for the PC and Factor I groups, no significant  results were obtained  distances between the point  for the pattern of  locations for any of the defined  groups. Figures 4 to 7 (pp. 85-88) show the point the mean Dom and Lov coordinates obtained  locations based on  for the sets of  IAS-R  profiles for the different diagnostic criterion groups, and the means and standard XXIX.  Although  deviations are shown  it might be suggested  relatively different profile pattern, contrasts suffer  in Tables XXV  through  that the means display a it is apparent that the  from the high variability obtained.  Indeed,  only the comparisons for the APD groups survived the multivariate BOX-M test for homogeneity of the dispersion  matrices.  However, results of some interest are apparent patterns of within cell correlations obtained measures across the various groups outlined  in the  for these distance  in the tables.  Inspection of the correlations significant at p < .001 suggests differential patterns  in the abilities of members of the various  groups to consider the perspectives of others and, in particular, to predict the perceptions of specific staff  members.  For example, Table XXVI contains the within correlations among the distance measures obtained  cell for the groups  109  TABLE XXV S u m m a r y Statistics of D i s t a n c e M e a s u r e s For the Overall Sample (N = 7 9 )  Distance  Mean  Standard Deviation  Minimum  Maximum  1.543  1.034  0.17  4.83  Self v s . as Friend (SVSF)  1.001  0.656  0.09  3.55  S e l f v s . as (SVSSO)  1.350  1.377  0.13  7.81  2.060  1.232  0.18  6.90  1.056  0.37  5.25  Self v s . Ideal  (SVSI)  Staff  Self v s . by S t a f f (SVSO) As Staff v s . by S t a f f (S0VS0)  « 1  7 Q o '798  TABLE XXVIII MEANS AND STANDARD DEVIATIONS OF DISTANCE MEASURES AND WITHIN CELL CORRELATIONS FOR APD VS. NOT APD  PC GROUP  DISTANCE SVSF  SVSI 1.  2.  3.  Mean SD  Psychopathic  Midrange  Non-Psychopathic  " (  1.500 .877)  SVSI SVSF  (  1 .147 .714)  1.344 (1.457)  2.451 (1.302)  2.064 (1.276)  (  Mean SD  1.623 (1.309)  (  .927 .628)  1.546 (1.488)  2.025 (1.410)  Mean SD  1.502 ( .881)  (  . 924 .618)  1.144 (1.172)  (  SVSF  SVSSO  SVSO  .687 *  -.207 -.238  .476  SVSO  -.024  - ,.079  .227  .070  - ..000  .215  NOTE:  CORRELATIONS 1 5  .719**  SVSF  SVSI  SVSSO  Non-Psychopathic (n==25) SVSO  SVSI  SVSF  SVSSO  SVSO  .449  *  .405  .479  .001  .280'  .605**  .154  .155  -.060  1.681 .825)  1.636 (1.001)  1.676 .786)  Midrange (n=27)  (n=27)  SVSSO  SOVSO  SOVSO  SVSO  SVSSO  WITHIN CELL Psychopathic  3  .248  .010  .422  -.071  --.100  -.034  -.048  --.198  .281  (a)  SVSI = Self v s . Ideal; SVSF = Self v s . Friend; SVSSO = Self v s . as S t a f f ; SVSO = Self vs. By Staff; SOVSO = As Staff vs. by Staff.  (b)  Asterisks indicate correlations significant at ^jd < .01, ** jd < .001.  .524*  TABLE  XXVII  MEANS AND STANDARD DEVIATIONS OF DISTANCE MEASURES AND W I T H I N C E L L C O R R E L A T I O N S FOR T H E F A C T O R 1 G R O U P S DISTANCE  FACTOR 1 GROUP  Psychopathic  Mean SD  1.495 ( .907)  2.  Midrange  Mean SD  (  Mean SD  1.768 (1.155)  3.  Non-Psychopathic  1.219 .911)  (  .768 .395)  (  1.046 .685)  W I T H I N CELL Psychopathic SVSI SVSF  SVSSO  SVSI  SVSF  -.260  .486  SVSO  -.040  -.040  .238  .056  .037  .237  NOTE:  .747**  (  1.819 .838)  1.938 (1.271)  (  1.599 .652)  (  1.765 .995)  CORRELATIONS Non-Psychopathic  (n=20) SVSSO  SVSO  SVSF  SVSI  SVSSO  (n=34) SVSO  .631**  .276  .221  SVSSO  SOVSO  SVSO  .937 .596)  1.555 (1.575)  Midrange  (n=25)  SVSF  (  2.001 (1.359)  2.419 (1.399)  1.401 (1.504)  1.127 ( .752)  SOVSO  SVSO  SVSSO  SVSF  SVSI 1.  3  .079  .162  -.073  .424  .214  .060  .509  .013  .577*  .283  .436  .003  .070  -.111  -.186  .447* .193  (a)  SVSI = S e l f v s . I d e a l ; S V S F = S e l f v s . F r i e n d ; S V S S O = S e l f v s . as S t a f f ; S V S O = S e l f v s . by S t a f f ; S O V S O = As S t a f f v s . by S t a f f .  (b)  Asterisks indicate correlations significant at  <  . 0 1 , ** £  <  .001.  .238  T A B L E XXVIII M E A N S AND S T A N D A R D D E V I A T I O N S OF D I S T A N C E M E A S U R E S AND W I T H I N CELL C O R R E L A T I O N S FOR APD V S . NOT A P D  DISTANCE  APD GROUP  1.  2.  D e f i n i t e APD  Not APD  Mean SD  1.626 (1.032)  (  1.053 .700)  1.447 (1.399)  2.209 (1.388)  Mean SD  1.382 (1.038)  (  .902 .561)  1.161 (1.337)  (  W I T H I N CELL D e f i n i t e APD SVSI SVSF  .414*  SVSSO  .012  Not APD  SVSSO  1.774 .804)  (  1.636 .906)  SVSI  SVSO  (n=27)  SVSF  SVSSO  SVSO  .512* .424*  SVSO  -.122  .016  .332  SOVSO  -.001  -.019  .145  NOTE:  1.881 (1.125)  CORRELATIONS  (n=52)  SVSF  SOVSO  SVSO  SVSSO  SVSF  SVSI  .566**  .  .314  .507*  .232  .323  .360  -.090  .065  .334  .314  (a)  SVSI = Self v s . I d e a l ; SVSF = S e l f v s . F r i e n d ; SVSSO = S e l f v s . as S t a f f ; SVSO = Self v s . by S t a f f ; SOVSO = As S t a f f v s . by S t a f f .  (b)  A s t e r i s k s indicate c o r r e l a t i o n s s i g n i f i c a n t at ^jd <  . 0 1 , ** jd <  .001.  T A B L E XXIX M E A N S AND S T A N D A R D D E V I A T I O N S OF D I S T A N C E M E A S U R E S AND WITHIN CELL C O R R E L A T I O N S FOR A P D - R v s . NOT A P D - R  DISTANCE  A P D - R GROUP SVSI 1.  2.  Definite APD-R  Not A P D - R  SVSF  SVSSO  Mean SD  1.521 (1.057)  (  1.022 .710)  1.523 (1.524)  2.224 (1.426)  Mean SD  1.574 (1.015)  (  .972 .577)  1.096 (1.099)  (  WITHIN  CELL C O R R E L A T I O N S  SVSF  SVSSO  1.819 .839)  (  1.554 .720)  Not A P D - R (n=32) SVSO  SVSI  SVSF  SVSSO  SVSO  .399  SVSF  .548**  SVSSO  .510**  .475**  SVSO  .269  .225  .570**  SOVSO  .088  .083  .085  NOTE:  1.964 (1.213)  6  Definite APD-R ( n=47) SVSI  SOVSO  SVSO  . 502*  -.051  .444  - . 125  .042  .256  -.045  -.018  .221  ,507*  (a)  SVSI = Self v s . Ideal; S V S F = S e l f v s . F r i e n d ; SVSSO = Self v s . as S t a f f SVSO = Self v s . by S t a f f ; SOVSO = As Staff v s . by S t a f f .  (b)  A s t e r i s k s indicate c o r r e l a t i o n s s i g n i f i c a n t at j^p < . 0 1 , **p  <  .001.  defined by the PC.  Given that staff show a tendency to rate  members of the psychopathic  group most extremely  (highest  the  Dom  score and lowest Lov score) as previously outlined, and that distance measures for self  ratings versus staff  relatively high magnitude,  it may be taken that the  correlation  (.227) between self versus as staff  versus by staff  (SVSO) seen  by staff  ratings have a low  (SVSSO) and  in the psychopathic group and  same group's notably high correlation  self  the  (.719) between self  (SVSO) and as staff versus by staff  the  versus  (SOVSO) reflect a  particularly poor ability for members of this group to predict the staff perceptions of them.  By the same line of  reasoning,  the mid-range group under the PC shows the best facility predicting staff p e r c e p t i o n s , and the non-psychopathic  for  group  takes a rather middling position suggestive of poor prediction the staff perceptions.  For the groups formed by Factor  XXVII), the psychopathic group  I  (Table  retains the pattern suggesting a  poor prediction of the staff perceptions of them; however, mid-range and non-psychopathic  the  groups do not retain the same  relations as seen with the PC groups.  The changes  latter groups likely reflect the changed group  in these  membership  effected by the different c r i t e r i a , and suggests the nonpsychopathic predicting  group to be relatively best  staff perceptions.  (although not good) at  The relative magnitudes, and  patterns of correlations are not nearly as striking groups defined by APD and APD-R  (Tables XXVIII and  for  the  XXIX).  of  COMPARISONS USING THE SELECTED ACL  SCALES  Means and standard deviations of the obtained scale scores based on self-description Factor  for the groups defined by the PC,  I, APD, and APD-R are shown in Tables XXX through  respectively.  Overall differences were noted only for the PC  groups; however, rather anomalous results were obtained Factor  I groups  in which evidently significant  differences were not reflected A oneway MANOVA (Wilk's F  XXXIII  (20 134)  =  for  the  univariate  in the overall test by MANOVA.  indicated significant  differences  1*86, E < .020) for seIf-descriptions  among  the selected ACL scales across the groups defined by the PC. Subsequent univariate analyses at £ < .005  (.05/10)  differences among the groups on the number of adjectives Autonomy (F ^ 2 76)  (UNFAV) endorsed  (F ^ =  76)  =  (F( 2 76)  =  unfavourable  8.15, £ .001), and  7.59), £ < .001), and  9.32, £ < .0005) scales.  indicated  the  Deference  Post-hoc contrasts  using  Tukey's HSD at £ < .05 and Scheffe's at £ < .01 indicated  the  psychopathic group to have scored higher than both the mid-range and non-psychopathic  groups on the Unfavourable Adjective  Scale  at £ < .05, but to differ only from the non-psychopathic group at £ < .01.  Differences on the Autonomy scale were evident  only  between the psychopathic and non-psychopathic groups in which former group scored higher  (£ < .01).  On the Deference  both the psychopathic and mid-range groups scored  the  scale,  lower than the  TABLE XXXIII  Means and Standard Deviations of the Selected ACL Scales and Rosenberg Scale for the PC Groups GROUP Midrange Mean SD (n=27)  ACL SCALES  Psychopathic Mean SD (n=27)  ACL Total Favourable Unfavourable Achievement Dominance Autonomy Aggression Abasement Deference  139,,15 49.,48 57.,04 49.,52 53..52 55.,78 55,.22 44,,67 43..89  41,.77 8,.62 11,.69 7,.43 8 .58 TO..91 10 .86 10 .76 10 .95  116,.85 49..85 50,.11 49,.41 49 .63 50 .56 49 .56 48 .37 49 .00  17 .70  4 .75  17 .70  Rosenberg  Non-Psychopathic Mean SD (n=25)  29,.94 8,.46 9,.94 • 10,.12 11,.31 10 .10 9 .32 12 .15 . 9 .82 5 .88  139 .84 51 .20 46 .28 50 .56 48 .24 45 .04 46 .96 54 .44 55 .68  35,.48 6,.46 6,.88 8,.50 10 .25 8 .52 8,.50 9 .58 8 .56  19 .52  4 .12  COMPARISONS15 Univariate  Tukey HSD  Scheffe  F (2,76)  (£ < .05)  <p < .01)  ACL Total Ffevourable Unfavourable Achievement Dominance Autonomy Aggression Abasement Deference  3.50, 0.33 8.15, 0.14 1.93 7.59, 5.05, 5.28, 9.32,  Rosenberg  1.13  NOTES: (a)  (b)  £<  .04  p < .001  P  p < £< p < <  P P P P  .001 ,009 .007 .0005  — > M , NP — — > NP >NP > NP , M < NP  — P > NP — —P > NP P < NP P < NP  —  —  ACL Total refers to the total number of adjectives endorsed (max = 300), the remaining ACL scales are T scores taken with respect to male norms obtained from the ACL MANUAL (Gough & Heilbrun, 1980). T h e overall M A N O V A w a s s i g n i f i c a n t a t p <  .02.  T A B L E XXXIII  Means and Standard Deviations of the Selected ACL Scales and Rosenberg Scale For the Factor 1 Groups GROUP -a ALL SCALES  Psychopathic Mean SD (n=25)  Midrange Mean SD (n=20)  Non-Psychopathic Mean SD '(n=34)  128,; 04 ACL Total 49..84 Favourable Unfavourabl e 55..68 50,.20 Achievement 53,.52 Dominance 54 .64 Autonomy 53 .80 Aggression 44 .48 Abasement Deference 44 .00  39 .35 8 .54 12 .10 7 .57 8 .98 10 .41 10 .85 10 .37 10 .76  134..55 52..00 49,.70 50,.85 52,.65 52 ,90 51. 30 47 .35 49 .50  40,,21 7..26 9.,92 8..89 11,.25 12,.22 11,.57 12,.22 11 .54  132,,82 49,.29 48 .94 48 .91 47 .06 46 .27 48 .00 53 .35 53 .24  34 .63 7 .76 9 .06 9 .41 9 .70 8 .53 8 .03 10 .56 9 .04  16 .60  4 .47  18 .20  6 .25  19 .56  4 .27  Rosenberg  comparisons'3  Univariate F (2,76) ACL Total Favourable Unfavourable Achievement Dominance Autonomy Aggression Abasement Deference  0.19 0.77 3.38, 0.35 3.69, 5.59, 2.51 5.05, 5.87,  Rosenberg  2.63  NOTES: (a)  (b)  Tukey HSD (P < .05)  Scheffe (p < .01) — —  p < .04  P  >  NP  - - -  —  P < .03 .005 E <  P P  p < .009  P < NP P < NP  P < .004  > >  NP NP - -  - -  —  P > NP — — —  P < NP —  ACL Total refers to the total number of adjectives endorsed (max = 3 0 0 ) , the remaining ACL scales are F scores taken with respect to male norms obtained from the ACL MANUAL (Gough & Heilbrun, 1980). The overall MANOVA did not reach significance _(£>  .05)  T A B L E XXXIII  Means and Standard Deviations of the Selected ACL Scales and Rosenberg Scale for the APD Groups  ACL SCALES 5  APD Mean SD (n=52)  ACL Total  Not APD Mean SD (n=27)  Univariate b F (1,77)  132.06  39.88  131.15  32.14  0.01  Favourable  49.69  8.55  51.04  6.44  0.52  Unfavourable  53.60  10.95  46.78  8.47  7.97, £ <  Achievement  49.44  8.56  50.52  8.98  0.27  Dominance  50.98  10.12  49.63  10.57  0.31  Autonomy  52.50  12.00  46.93  6.49  5.04  Aggression  52.15  10.91  47.81  7.81  3.36  Abasement  47.14  12.27  52.67  8.94  4.29  Deference  47.04  11.70  53.85  7.30  7.59  Rosenberg  17.98  4.95  18.85  5.13  0.54  NOTES:  (a)  ACL Total refers to the total number of adjectives endorsed (max = 300), the remaining ACL scales are F-scores taken with respect to male norms obtained from the ACL MANUAL (Gough & Heilbrun, 1980).  (b)  The overall MANOVA did not reach significance (p > .15)  TABLE  XXXIII  Means and Standard Deviations of the Selected ACL Scales and Rosenberg Scale for the APD-R Groups  GROUP ACL SCALE  3  APD-R Mean SD (n = 47)  Not APD-R Mean SD (n = 32)  Univariate 5 F (1,77)  ACL Total  129..79  38.,02  134..63  36,.39  0..32  FAVOURABLE  49.,81  8,.16  50.,66  7..54  0,.22  UNFAVOURABLE  53..11  10..54.  48.,56  10..32  3 .60  ACHIEVEMENT  48..81  8,.61  51..28  8,.67  1 .56  DOMINANCE  50,.85  10..00  50..03  10,.70  0 .12  AUTONOMY  52,.40  10,.45  47,.94  10,.76  3 .40  AGGRESSION  51 .72  10,.00  49,.13  10,.26  1 .26  ABASEMENT  46 .72  11,.51  52,.41  10 .76  4 .89  DEFERENCE  47 .13  10,.14  52..66  11,.19  5 .20,  ROSENBERG  17 .36  4 .70  19,.63  5 .19  NOTES:  (a)  (b)  4 .05  ACL Total refers to the total number of adjectives endorsed (max = 300), the remaining ACL scales are F-scores taken with respect to male norms obtained from the ACL MANUAL (Gough & Heilbrun, 1980). The overall MANOVA was not significant (p > .04).  non-psychopathic group at p < .05, but the difference was only retained for the psychopathic vs. non-psychopathic groups at p_ < .01.  Two other scales, Aggression and Abasement,  significance  (£ < .009 and p < .007, respectively)  approached  in the  univariate analyses reflecting a p < .05 difference on Aggression with the psychopathic group scoring higher than the nonpsychopathic group, and a p_ < .01 difference on Abasement with the psychopathic group scoring lower than the group.  Thus, compared  non-psychopathic  to the non-psychopathic group, the  psychopaths as defined by the PC described  themselves with more  unfavourable adjectives, and as more autonomous and  less  deferrent; modest trends also suggest them to describe as more aggressive and surprising quite  less abasing.  These results are again  in appearing to be rather realistic  inconsistent with the expectation  self-report. differences difference  themselves  self-appraisals,  for dissimulation  on  Moreover, these results do not evolve from simple in "loquacity" as there was no evident  in the total number of adjectives  group  endorsed  (F = 3.50, p < .035; no significant post-hoc — (2,76) —  contrast).  As mentioned, the multivariate analysis for the groups formed by the Factor (F. ~~ (20,-I-D^N 134)  =  I criterion did not reach  significance  1.57, £ < .05) , and only two of the ten  contrasts met p <  .005  (Autonomy and Deference) .  univariate  The  multivariate group comparisons for APD and APD-R were not significant and no univariate contrasts met p < .005.  Discriminant Analyses with the ACL Scales Discriminant  analysis for the PC groups using the ACL  scales  did not perform as well as the IAS-R coordinate system, yielding a 57% overall hit rate and best group members at 68%. discrimination,  identifying the  non-psychopathic  Scales contributing most to the  in order, were Deference, Number  Unfavourable,  Autonomy, Abasement, and Aggression. None of the functions for classification of the other criterion groups were  SUPPLEMENTARY Polar  significant.  ANALYSES  Coordinates  Use of polar coordinates  (vector  length and  angular  placement) as a bivariate system proved to be problematic  for  analysis, given that the full circular array of angles cannot be used to derive simple arithmetic means and standard Considered  deviations.  individually, vector length for the different  perspectives across the criterion group sets did not significance  IAS-R  reach  for any of the comparisons made, although, as  previously mentioned, vector  length did show an  appreciable  correlation with PC scores, particularly when taken from staff descriptions  (r = .461, n = 79).  See Tables XXXIV-XXXVII  summaries of the vector lengths obtained  for  for the different  groups. Similarly, tests of "angular preference" or a tendency to a  122 TABLE X X I I I Means and (Standard Deviations) of Vector Lenqth Across Psychopathy Checklist Groups (N = 79)  T/\C N CCT  SL1  Psychopathic Vector Length (n = 2 7 )  1. "Self"  4 . "As Staff"  5 . "By Staff"  (n =  Non-Psychopathic Vector Length (n = 2 5 )  Overall Vector Length  (  1.769 .977)  (  1.747 .760)  (  1.306 .505)  (  1.615 .796)  (  1.752 .954)  1.785 ( .914)  (  1.908 .785)  (  1.812 .881)  1.714 ( .773)  (  1.889 .920)  2 . "Ideal'  3 . "As Friend"  Midrange Vector Length  1.989 .955)  1.951 (1.017)  (  2.253 (1.475)  2.063 (1.234)  1! 668 (1.162)  2.003 (1.307)  2.763  1.777 (1.033)  1.502 ( .869)  2.027 (1.143)  (1.122)  T A B L E XXXVII  Means and (Standard Deviations) of Vector Length Across Factor I Groups (N = 7 9 )  IAS-R SET  Psychopathic Vector Length (n=25)  Midrange Vector Length (n=20)  Non-Psychopathic Vector Length (n=34)  1. "Self"  1.811 ( .941)  1.614 ( .824)  1.472 ( .643)  2. "Ideal"  1.734 ( .947)  1.615 ( .911)  1.986 ( .803)  3. "As Friend"  1.864 ( .903)  1.929 ( .990)  1.884 ( .918)  2.212 (1.579)  1.850 (1.038)  1.940 (1.245)  2.755 (1.179)  1.808 ( .856)  1.619 (1.025)  4. "As Staff" 5. "By Staff"  T A B L E XXXVII  Means and (Standard Deviations) of Vector Length Across APD versus NOT APD (N = 7 9 )  MEET APD Vector Length (n=52)  NOT MEET APD Vector Length (n=27)  Self  1.663 ( .884)  1.522 ( .595)  Ideal  1.802 ( .953)  1.832 ( .740)  As Friend  1.905 ( .970)  1.857 ( .832)  As Staff  2.180 (1.353)  1.663 (1.161)  By Staff  2.337 (1.167)  1.428 ( .825)  TABLE  XXXVII  Means and (Standard Deviations) of Vector Length Across APD-R versus NOT APD-R (N - 79)  MEET APD-R Vector Length (n=47)  NOT MEET APD-R Vector Length (n=32)  Self  1.664 ( .741)  1.543 ( .879)  Ideal  1.173 ( .957)  1.870 ( .767)  As Friend  1.903 ( .863)  1.868 (1.013)  As Staff  2.311 (1.456)  1.550 ( .891)  By Staff  2.369 (1.206)  • 1.523 ( .830)  126 particular angular orientation within samples based on the Rayleigh  test  (attributed  to Lord Rayleigh circa 1880;  1977) did not provide any positive results. be related  Norcliffe,  The problem here may  to the small sample sizes within groups relative to  the potential dispersion of specific  angles.  Another basic test of the angular distributions for has already been outlined correspond  in reference to octant locations  to the relative frequencies of angles within  segments of the circle and tested by Chi-square 1972; Norcliffe, 1977).  groups which  arc-  (cf. Mardia,  It was shown that tendencies  for  differential assortment across adjacent octant pairs was quite evident, particularly with reference to the staff groups formed by the Factor  ratings and  I criterion.  Self Esteem and "Self" versus As a simple supplementary  "Ideal"  test of the adequacy or  consistency of the obtained data and the derived distance  scores,  a correlation was computed between the Rosenberg scale as a measure of self-esteem and the distance measure "self ideal"  (SVSI).  This result  (.342 with N = 7 9 ;  versus  .469 with N = 113)  suggests an adequate level of correspondence consistent with idea that lower self-esteem  (a higher score) relates to a larger  disparity between one's view of self versus a self  ideal, and  reflects reasonably well on the consistency of the data from the  self-reports.  the  obtained  127  DISCUSSION  Overview and Basic Methodological  Issues  Although not wholly supporting hypotheses  considered  of central  interest to this study,  results were generally consistent characteristics IAS-R.  the specific expectations  conceptually  Checklist and the construct  as identified by the  The results demonstrated  differences  discriminative  utility  descriptions  identified  were  the  of the diagnostic  and  consideration  study.  Perhaps the most basic threat  similar  those  those  implications of these results following  of the limitations of this  of any study  identified  space defined by  I will return to further discussion  theoretical  in the  as p s y c h o p a t h s by the PC where they  to be expected within the Interpersonal IAS-R.  structure  in relation to groups formed by the PC versus  on APD c a t e g o r i e s , and showed placement of  individuals  conceptual  validity of the circumplex  of interpersonal  the  Psychopathy  of the IAS-R.  based  to  of the psychopath which are assessable with  validity of the psychopath  by the IAS-R  the  in pointing  As such, the results contribute to both the  and  to the interpretive  is its generalizabi1ity  validity  to populations more or  to the sample targeted by the study.  The present  less  study  12 8 is affected  to some extent by the characteristics of the men seen  and by the circumstances under which they were seen. confounds will have the greatest effect on any practical application  These  anticipated  of the assessment used here.  The foremost characteristic of the men seen for this study was their willingness to participate. seen, these men may be considered programs and other opportunities within the  Compared  to others not  to be relatively more active in for activities or  interests  institution, to be more open to participation  psychological  research oriented  in  to criminality, and/or to be less  engaged by the "con code" or institutional ethic proscribing manner of cooperation with "the system." solicit volunteers and negotiate with  Through efforts to  individuals and groups such  as the Inmate Committee, it became clear that such was not considered institution.  participation  trivial by many of the men within  the  Considerable effort was given to assuring  inmates of the confidentiality of any material obtained relation to any particular the research  interest.  in  institution  Also, offering the men money for  participation must be considered motivation and  the  individual, and of the independence of  interest from any association with the  or other authorities.  any  to have affected  their  The selection of subjects was  affected by the need for education and English in order to complete the adjective  their  language  also criteria  lists.  Thus, at basis, the sample of men seen for this research  is  not necessarily correctional  representative of the general population of  inmates, or even of the population  specific to this  particular  institution.  may differ  in the relative prevalence of diagnostic  used here.  It is not known, for example, how  categories  Based on research with the PC, the prevalence of  psychopathy within approximately distribution  inmate samples has been found to be  22-27%, but may vary as functions of the of scores obtained and selected cut-offs.  We do not  know whether psychopaths may be more or less likely than inmates to choose to participate offered.  others  in the research  other  programs  However, comparable distributions of scores have been  shown by Wong  (1985) with a random sample of inmates using  scoring based on file information A further confound  alone.  to the possible practical utility of the  assessment strategy used here derives from the setting or circumstances considered  in which the men were seen; this may also be  to have affected the obtained distance measures which  formed the primary hypotheses for this study.  As was mentioned,  considerable effort was given to assurances of  confidentiality  and  independence from correctional authorities; as such, the  obtained self-reports may not reflect the response styles to be obtained when the men believe their responses to be of some consequence within the system.  If so, this may have worked  against the expectations for dissimulated  self-reports.  In the  absence of "threat" or some perceived consequence resulting  from  130 how they chose to portray  themselves, these individuals may have  felt no particular pressure to depict themselves favourable way.  in a more  Therefore, one might still assume that  differential patterns of dissimulation would result particularly  the psychopaths, believed  some effect within the institutional Statistical  if the men,  the assessment  to have  setting.  influences on the results are seen to derive  from high and heterogeneous  variability across groups and  the dependent measures which may tend to obscure differences at the higher  levels of analysis  across  potential  (i.e., MANOVA),  particularly when contrasting the more specific derived  measures.  Several sources of variability may be seen to operate on the responses obtained with the word lists, not the least of which the basic motivation of the individual to respond reasonably meaningful way.  is  in some  The content of the adjective  scales  themselves and the level of understanding of the word usage which the men may possess was an initial concern which was moderated some extent for the IAS-R by the provision of a glossary. the glossary  itself did no particular  to  That  violence to the intent of  the IAS-R was supported by the results of the obtained from the self descriptive profile, and the resulting placements of the profiles being consistent with  circumplex  octant  theoretical  expectations. Approach to the task by individual subjects ranged from very fast rates of completion of the adjective lists with, one may  131 suspect, very superficial painstaking  levels of attention, to such  efforts that two subjects were considered  complete the task.  unable to  Personal expectations and the role of  stereotypes are also likely to influence the responses particularly when asked  to consider the perspective of a member  of the institutional staff. himself  offered,  One may expect an inmate to consider  to be seen by staff as a "bad ass" as, to many, this is  what the system demands. too, would  Similarly,  it was expected  suffer from stereotyped perceptions.  that  staff,  To assess  this,  a rating of "an average inmate" was requested at the conclusion of the study from a representative sample of the staff who had been asked to provide ratings of inmates participating study.  in the  This anchor point description was seen to fall at  essentially the DE pole of the circumplex and constituted  a far  more extreme rating than was obtained on average as staff  ratings  of individual  inmates.  This would suggest the staff to be  cognizant of a stereotyped  "average  made ratings of individual  inmates which were not unduly  influenced by the stereotype. that the staff  inmate" and to generally  Moreover, the results  ratings provided the basis of group  obtained with the  have  indicated  discrimination  IAS-R.  Two additional considerations might be raised with to the inmates' completion of the instructed their selection of staff:  respect  self-report sets and  differential effects of memory  abilities on the consistency of self-report, and  systematic  selection bias.  With respect to the first, memory  effects,  concurrent data were available providing measures of  intellectual  and memory function collected  for a different  research  interest.  Comparisons of these measures  (Hare and Forth, 1988) showed  no  group differences for the groups defined herein and support contention  that there were no differential abilities to  and manipulate responses across the instructed sets.  remember  Although  not specifically assessed, selection bias was not expected evident across groups as the inmates typically complained little opportunity or interest  in relating  to staff and  expected a negative perception  from them.  It was also  that the derived measures would  the  to be of  generally considered  reflect the ability to predict  staff perceptions, and, as such, did not rely on a particular consistency of negative or positive  expectation.  Given the above considerations of potential variability inaccuracies  in the data, the obtained  quite promising.  On the more global  distributions obtained  and  results may be seen as  level of the octant  for the criterion groups and the overall  consistent pattern of dependent measures, the data are quite good.  Initial consideration for the power of the tests was put  at approximately seen  in the data reported by Wiggins, Trapnell, and  (in press), using .025.  .80 for groups of 25, based on mean  simple  differences Phillips  contrasts and a Type I error rate of  That the multivariate comparisons of the Dom and  coordinates  for groups based on the PC resulted  Lov  in differences  133 significant at £ < .001 suggests that the IAS-R may prove fruitful for further  research.  A further consideration  in the interpretation of the results  is the number of comparisons made and the consequent overall I error rate.  As an exploratory study, a relatively  alpha level was used  liberal  (.01), taken in relation to each of the  dependent comparisons of interest. measures used  Thus, each of the derived  in the MANOVAs--Dom and Lov  coordinates,  interprofile distances, and vector length--were  considered  independently, and comparisons of the defined criterion (by PC, Factor (.10/4).  Type  groups  I, APD, and APD-R) were each evaluated at £ < .025  Although  this would suggest a nominal  Type I error rate of  experiment-wise  .30 for the MANOVAs alone, it may be seen  that the nominal alpha  levels are quite arbitrary and that the  same pattern of results would be retained even if the alpha levels were set lower.  Comparisons were not made for  which may be defined under Factor 2 in order to avoid inflation of the overall alpha level and  groups further  in the expectation  Factor 2 shows considerable overlap with APD  (see Appendix  Given the interests of the study in comparing profiles diagnostic sets,  that B).  across  it was felt heuristic value would be better  served by a systematic use of a relatively more statistical criterion.  liberal  The consistent pattern of results  from the more global non-parametric comparisons, through  ranging simple  correlations, to the more specific contrasts based on the MANOVAs  134 and subsequent post-hoc comparisons may contribute more to the inferential process and suggestions for further  research than a  simple yea or nay based on MANOVAs at a more stringent  level of  s ign i f icance. Given the above considerations, strengths of the present generalizabi1ity,  research.  it is possible to review the With respect to issues of  the results have inherent  limitations but are  not atypical of other studies in which consenting volunteers obtained  from a correctional  sample with defined  institution sample.  As such, a  characteristics was obtained which  provides  the basis for comparisons of diagnostic criteria within sample, and which demonstrates patterns of differences selected dependent measures that show reasonable statistical  the on  levels of  reliability and good conceptual congruity.  now return to a discussion of the implications of these  Diagnostic  I shall results.  Considerations  The central  issues for diagnostic usage with the psychopath  relate to the poor consensus for definitional  criteria and the  potential for misapplication where "psychopathic are assumed  are  characteristics"  to apply despite the use of inappropriate or  incomplete criteria.  Current  literature continues to provide  examples where specific reference  is made to research with  "psychopaths" when, in fact, the basis for inclusion of APD by DSM-111  (APA, 1980) or more or less  is diagnosis  significant  elevations on Scale 4 (Pd - psychopathic deviate) of the MMPI. Such continued misapplication contributes misinterpretation confusion  to ongoing  of the meaning of these terms and  in the interpretation of research with  continued  "the  psychopath." Comparisons of the base rates of individuals meeting criteria under the PC or APD are a basic potential for confusion.  illustration of the  Based on the total sample of 113  subjects, 27.4% were identified the Factor  the  as psychopaths by both the PC and  I criteria, whereas 66.4% met the criteria for APD by  DSM-III and 58.4% as APD by DSM-III-R. demonstrates  The subsample of 79  the more usual convention with  research  the PC in which approximate thirds of the obtained  involving  distribution  of scores are taken to facilitate group comparisons; here met the PC criterion and 31.7% met the Factor  I score  34.2%  cut-off.  The APD criteria for DSM-III were met by 65.8% and by 59.5% for DSM-111-R  criteria.  Although these rates result from diagnoses  made by one rater only and may be somewhat high compared the rates w h e r e two raters agree, they are generally with past research  to  consistent  (e.g., Hare, 1981, 1983, 1985a) and point  again to the fact that, although most,  if not all, of the  individuals meeting the PC criterion also meet the criterion APD, the converse  for  is far from true--APD cannot be considered  synonymous with psychopathy  as a more rigorously defined  The more specific measures of concordance rates and  concept.  136 contributions  to the prediction of group membership  across  criteria underscore the above base rate differences.  Poor  rates  of agreement were obtained between the PC and APD criteria either DSM-III or DSM-III-R respectively). previously  (Kappas of  .183 and  for  .114,  These rates are much lower than have been  reported  (e.g., Hare, 1983, 1985a) and, as mentioned,  may relate to the inclusion of all subjects than agreement  for the extreme groups only.  could appear much more favourable yielding Kappas of  (here, 113) rather Indeed, results here  if based only on the extremes,  .621 with APD by DSM-I11 or .527 by  The rates of agreement were somewhat worse still when the APD criteria to groups based on the Factor  DSM-III-R. comparing  I scores, and may  further emphasize the conceptual differences of APD as grounded in persistent antisocial behaviour versus psychopathy emphasizing  as  the lack of empathy.  Contributions  to the predictions of group membership  on knowledge of membership under another criterion again  based reflect  the poor levels of association seen between the PC criteria APD. Factor  Symmetric  lambda coefficients ranged  from 0 between  and  the  I groups and APD-R to .183 between the PC and APD,  indicating  little reduction  predictions.  in the probability of error for such  Although somewhat better relations were seen in the  subsample of 79  (compare Tables X and XI, pp. 76 and 79), the  overall results  reflect poor levels of association and  the misrepresentation  inherent  in assuming  emphasize  the diagnosis of APD  137 to be equivalent to psychopathy as identified by the PC. The utility of the MMPI to be particularly poor. specific criteria  in defining the psychopath  is seen  The low rates of profiles conforming  (e.g., Gilbertstad  & Duker, 1965; or Marks  & Seeman, 1963) precluded group comparisons.  Comparisons  were  instead made with respect to the specificity and sensitivity profiles considered  typical of the psychopath  other criterion groups. MMPI profiles of which association  to  of  in relation to the  The sample of 113 men provided only 89 20 were considered  of "psychopathy"  relevant to the  (i.e., elevations  typical  > 70 T on scale  4 - psychopathic deviate--alone, or with scales 9 - hypomania, . 8 - schizophrenia, or 6 - paranoia).  These 20 profiles were then  sorted by their specificity--the proportion being specific to the criterion group of interest  (i.e., psychopathic by the PC or  definite APD), and their sensitivity--how many of the criterion group actually had the profile.  relevant  It was found that  profiles were most specific to APD-R  (75% versus 40% for PC  psychopathy), and might be suggested  to reflect the  correspondence of common behavioural elements--family conflict with authority, poor school performance, abuse, and so forth. sensitivity and appear  Overall, the profiles show  problems,  substance little  in only 23.2% of those definite APDs who  provided profiles, up to 34.8% of those identified psychopathic by the PC or Factor profile.  the  as  I criteria and who provided a  Although the psychopathic group defined by Factor I  138 demonstrated  a significantly different  proportion  (Z^ = 2.26, jd < .02) with a relevant profile than among the nonpsychopathic group, the individuals  in the psychopathic group did not provide a  profile considered Table XI  it is still to be seen that more than 65% of  "diagnostic" of psychopathy.  (p. 79), show relatively  relevant profiles  (15) obtained  Values shown  larger proportions  in  relating  the  in the subsample of 79, but the  same pattern of results can be seen. It would appear that MMPI profiles associated "psychopathy" have a similar relationship to more  with stringent  criteria as seen between APD and the PC criteria.  That  is, the  majority of relevant profiles may be common to the other criterion  (here, APD), but the majority of the criterion group do  not exhibit the relevant profiles. particularly poor representation  The MMPI  is seen to be a  of psychopathy as defined by the  PC. From the foregoing  it should be clear that failure to  discriminate between APD, MMPI criteria, and psychopathy identified by the PC can only contribute to confusion  as  in efforts  for research with the psychopath as was noted ten years ago by Hare and Cox  (1978).  There are obvious problems  in assuming  APD  or MMPI scale 4 elevations to be equivalent to psychopathy, as it is quite unlikely that one is relating to a common  population.  Continued misuse of these terms contributes to a "halo" common to categorical assignments wherein one assumes  effect  all  139 associated  features are present, given the diagnosis, when in  fact only a minimum subset required to meet the criteria actually applicable. DSM-III-R  This is particularly  (APA, 1987) wherein the inclusion  are  evident with APD under of a feature  to a remorseless attitude is somehow considered  relating  to make the  diagnosis more consonant with the concept of psychopathy.  It is  quite apparent  that  in the structure of the diagnostic criteria  this feature need not be considered diagnosis. served present  Individuals receiving  in order to apply  such diagnoses may be better  if specific reference were made to the features for the application.  the  In correctional and  considered  forensic  treatment settings particular care should be given to avoid attributing  a "psychopathic"  label on the basis of  loose  criteria.  Relations of the Dependent Measures to the PC and APD Descriptive differences were obtained with the IAS-R in relation to the PC and Factor I groups which are consistent with the concept of psychopathy and which were not evident in comparisons with the APD criteria. The most basic level of analysis utilizing the  circumplex  structure of the IAS-R derives from the relative frequencies of octant  locations seen for the various groups.  introduction,  As outlined  in the  it was expected that octants 2 and 3 (BC/DE) would  capture the psychopathic profile, particularly  from staff  140 descriptions.  This was shown to be the case for the PC groups,  with staff descriptions demonstrating  significant  differential  assortment across groups; placing 85.1% of the psychopathic group  in octants 2 and 3, versus 28.0% of the  group and 51.8% of the mid-range. modest trend  non-psychopathic  Interestingly, there was a  < .05) for the groups to be differentially  by self-description, with 62.9% of the psychopathic group  sorted falling  in octants 2 and 3 versus 24.0% of the non-psychopathic group and 37.0% of the mid-range. descriptions  Similar results were obtained with  for the Factor  group being placed  I groups with 88% of the psychopathic  in octants 2 and 3, versus 35.3% of the non-  psychopaths and 50% of the mid-range.  There were no evident  differences by self-description by the Factor Staff descriptions also demonstrated assignments  staff  I groups.  differential  for the individuals meeting DSM-III criteria for APD,  placing 69.3% of the definite APDs in octants 2 and 3 versus 29.6% of those not meeting the criteria. argued  that there is some element of confound here in that 26 of  the 52 men identified the PC.  However, it might be  Thus,  were placed  as APD were also seen as psychopathic by  in essence, only 10 of the remaining 26  to octants 2 and 3.  staff showed a tendency  Under the DSM-III-R  (38.5%)  criteria,  (p. < .02) to place more of the APDs in  octants 2 and 3 (68.1% versus 37.5%), but the same confound applies with 22 of the 47 APD-Rs also being psychopaths by the PC.  Thus, here 40%  identified  as  (10 of remaining 25) of  141 those meeting APD-R criteria but not the PC criteria were placed in octants 2 and 3. Thus, at this point there are evident differences perceptions of others relating to the characteristics  in the  interpersonal  of the men identified as psychopaths by the PC  versus those receiving a diagnosis of APD.  Psychopaths  as  identified by the PC are more likely to be associated with characteristics as grandiosity, lack of empathy with  such  others,  attitudes of special entitlement, restricted affect or an absence of warmth, and exploitative  relations with others;  aspects  considered central to the psychopathic personality. interesting  to note that these descriptive characteristics  also common to the narcissistic character as falling octant BC  It is are  within  (Kiesler, 1985; Wiggins, 1982) of the circumplex  space.  These results perhaps reflect the dichotomy of the clinical concept of psychopathy as a psychological construct versus behavioural aspects of antisocial adjustment more associated with the APA personality. been pursued  typically  (1970, 1987) definition of Antisocial  (A basis for the analysis of this dichotomy in recent  the  research by Harpur, et al.  has  (1988b)  investigating external correlates of Factors I and 2 of the Psychopathy Checklist  (see Appendix  B).)  The results are reiterated, and given a stamp of better statistical  relatively  reliability, in the repeated measures  of the Dom and Lov coordinates.  analyses  Post-hoc comparisons also allow  142 more specific measures.  identification of group differences on these  From these analyses differences were evident among  groups defined by the PC and Factor for the APD or APD-R groups. derived  the  I criteria, but were not seen  For the PC groups,  differences  from the Dom and Lov coordinates of the staff  ratings;  post-hoc contrasts at £ < .05 indicated the psychopathic group to be scored higher on the Dom dimension and lower on the Lov dimension  than either of the mid-range or  non-psychopathic  groups; however, these differences were retained only for the extreme groups at £ < .01. Groups formed by the Factor  I criterion also showed  differences on the Dom and Lov coordinates of the staff but here both the mid-range and psychopathic groups were  ratings, scored  higher (p < .05) than the non-psychopathic group on the Dom dimension and the psychopathic group was scored  lower  (p < .05) than both  the mid-range and non-psychopathic groups on the Lov  dimension.  Again, these differences were only retained for the extreme groups at p < .01. dimension  The Factor  for representation  I groups also differed on the Lov  of an ideal self, with both the mid-  range and psychopathic groups scoring non-psychopathic significant  trend  group.  (p < .05) than  the  One other measure suggested a  in which the psychopathic group scored  than the non-psychopathic description.  lower  group on the Dom dimension for  higher self-  143 Thus, the results based on the dimensional structure of the IAS-R  reflect the differential perceptions of others in which  psychopaths are quite distinct from non-psychopaths at least, as being more dominant and assured and less warm or affiliative; a pattern suggestive of an arrogant, cold and calculating personal style.  The additional differences seen in the Factor I  group suggest some refinement comes to differ affiliative  inter-  in which the non-psychopathic  group  in terms of the depiction of a more nurturant or  ideal relative to the others, and the extreme  reflect a difference  in self-perceived  groups  dominance or social  status. Comparisons based on the derived distance measures between the point  representations of the IAS-R protocols were  disappointing; no differential patterns were evident the criterion group comparisons.  As outlined  particularly  dissimulation  in the response  for self-description  defined by the PC and Factor XXIX providing  assumptions  profiles,  in the psychopathic  I criteria.  in any of  in the  Introduction, expectations here were predicated upon incorporating  taken  group  Review of Tables XXVI -  the group means and standard deviations of the  distance m e a s u r e s , and Figures 5 - 7  depicting the point  locations suggests that, although the groups appear to occupy relatively different and Factor  I groups  locations  in the space  (as seen for the PC  in the DOM/LOV analyses outlined above), the  pattern of distances between points  is quite  similar.  As already suggested, one implication here is that  the  144 psychopathic group was surprisingly candid.  It is quite apparent  that, on average, the men in this group endorse which  self-descriptions  incorporate the characteristics of octants BC/DE, contrary  to the expectations of them denying such attributes. interpreted  in either of two ways:  This may be  first, that given the  research situation with assurances of confidentiality, felt no particular pressure to portray themselves  in a more  favourable way; or, alternatively, that they responded with a sex role stereotype of exaggerated This latter aspect was outlined  these men  (negative)  in keeping  masculinity.  in the Introduction with  to sex differences and social desirability (1979) and Wiggins and Broughton  outlined  in Wiggins  (1985); however, the  expectation  was that psychopaths would tend to avoid this pattern. obtained  respect  The  result may be a combination of the above  interpretations, and the suggestion repeated  that this pattern  may not obtain  their responses to  if these individuals believed  have some personal  consequence.  Another problem evident  in the analyses of these measures  stems from the high and heterogeneous variability which tends to overwhelm the differences obtained.  This aspect might be  moderated with a more substantial sample  size.  The distance measures did, however, yield some results of interest with  inspection of the within cells correlation  for these measures across groups.  Patterns evident among  matrices these  matrices for the PC and Factor I groups appear consistent with  145 Gough's  (1948) hypothesis of a deficit of role-taking  psychopath.  It would appear that those identified as  here show a particular  Thus, even though appearing of self  psychopaths  inability to predict staff perceptions of  them, despite their own relatively negative  representations  in the  rather candid  self-descriptions.  in their  ("insight"?), psychopaths may still not  appreciate how poorly they may be seen by others.  Again, these  differences were not evident for those men identified  as APD.  Group comparisons with scales from the ACL taken as selfdescription were also interesting and also only  provided  demonstrable differences among the groups formed with the PC. Here, too, results suggested  rather realistic self-appraisals by  the psychopathic group, differing psychopathic group  (p < .01) from the non-  in endorsing more unfavourable  adjectives,  scoring higher on Autonomy, and lower on Deference and  Abasement;  an additional near difference suggested  (p < .05)  on Aggression. ACL Manual  a higher score  Based on descriptive associations provided  (Gough and Heilbrun, 1980), this profile of  high and low scores suggests characteristics  in the  relatively  such as pessimistic,  changeable mood, quick to take offense, feelings of bitterness and hostility  toward others, indifference for the feelings of  others, egotistic, headstrong, a view of others as rivals to be vanquished, risk-taking, and impulsive behaviour which leads to conflict with others.  frequently  Such characteristics are quite  consistent with perceptions of the psychopath; however,  it should  146 be noted that the differences obtained were only relative to the non-psychopathic group  in particular and do not reflect  scores on the scales described.  extreme  Overall, the scale score  were within the normative range suggested by Gough and  results  Heilbrun  (1980, p. 48) as lying between 40T and 60T; the maximum seen for the psychopathic group was 57T (DEF)  (UNFAV) and the minimum  43.9T  (See Table XXX, p. 116). As has been suggested, the obtained  results are notable  in  being quite specific to the psychopath as defined by the PC and Factor  I criteria.  Groups defined by the criteria for APD as  defined by either DSM-III  (APA, 1980) or DSM-III-R  not show similar descriptive differences.  This  specificity  contributes to the view that APD does not capture attributes whether self-described  (APA, 1987) do  "personality"  or perceived by others, but  appears more generally to reflect an individual's tendency criminal  to  behaviour.  The performance of discriminative IAS-R coordinates  in differentiating  or APD criteria further  functions based on the  the groups defined by the PC  illustrates these differences.  It was  seen that the IAS-R Dom/Lov coordinates provided an overall hit rate and correctly predicted  70.4% of the psychopathic  58.2% group.  While these results are not particularly high, it is a substantial  improvement over the base rate for the psychopathic  group of some 34%.  The false positive  resulted from the mid-range group  rate of 37.6% primarily  (29.6%); the rate for members  of the non-psychopathic group predicted quite low  (8% : 2 of 25).  to be psychopaths  was  The discrimination of the Factor I  groups was somewhat better with an overall hit rate of 65.8% and correct prediction of 72% of the psychopathic group. positive only 5.9%  rate was reduced to 25.9%, and  incorrectly  The  false  predicted  (2 of 34) of the non-psychopaths as psychopaths.  Although the numbers appear good for the discrimination the APD g r o u p s — a n  of  overall hit rate of 70.9% for APD and  68.4%  for APD-R--they do not represent much gain over the base  rates  for these diagnoses--65.8% APD and 59.5% APD-R.  Moreover, it  should be recalled that much of the discrimination may  derive  from the fact that 26 of the 52 APDs were psychopaths by the PC, as were 22 of the 47 APD-Rs. Overall,  it becomes quite clear that the data obtained  supportive of the conceptual validity of the psychopath assessed by the PC in having characteristics  not seen  Antisocial Personality.  identifiable  are  as  interpersonal  in those receiving a diagnosis of Similarly  (if circuitously),  these  results support the construct validity and potential utility of the IAS-R as an assessment strategy for use in this problematic  Practica1  typically  population.  Implications and Theoretical  As indicated, the obtained  Considerations  results suggest some potential  utility for the IAS-R as an assessment strategy for use with  criminal populations.  However, it is apparent that the  differences seen derive primarily from staff perceptions of the inmates and not from readily discriminable differences based on simple self-report.  Moreover, the differences are  relative  across groups and do not, at this point, lend themselves to ready or likely replicable cut-off scores for identification psychopath.  A possible application might be in the use of the  IAS-R for staff  ratings, helping to characterize and  interpersonal perceptions of individual  standardize  inmates in assessment  contexts where the desire to identify such attributes salient, e.g., treatment progress or pre-release In terms of theoretical  is quite  reviews.  issues, it is apparent the  does contribute to the characterization of the psychopath  of the  in identifying a discriminable  and that its location within the defined  IAS-R  "average"  interpersonal  style,  Interpersonal Circle is  consistent with expectations for this personality type.  There  has been debate regarding this location, however, with Leary and Coffey  (1955) expecting  the DE pole to represent the sadistic and  psychopathic personalities, and Kiesler  (1985) arguing  octant BC as representative of the psychopath. here, however, that BC/DE  (CD?) would provide a more  location as representing both cold-hearted characteristics.  It was  It is interesting  and  against considered  likely  manipulative  to note that the staff  description of the "average inmate" fell at the DE pole (coordinates Dom: 0.34, Lov: -4.0) suggesting a criminal  stereotype as "cold-hearted" and essentially without a relative interpersonal status.  The psychopath, however, was clearly  as one who is both cold-hearted  and domineering  reflecting characteristics  in common with the  personality defined by APD  (1980, 1987).  Another aspect of the circumplex interest  distance of a point representation applicability  of vector  that  is of  length or the  from the origin, and  its  (see Wiggins, Phillips, and Trapnell, in  To the extent that the scores on the PC may be  considered vector  narcissistic  as an index of pathology or rigidity of  interpersonal style press).  (manipulative),  representation  relates to the conceptualization  seen  a continuum of psychopathology,  length  this interpretation  of  is consistent here, demonstrating a correlation of  .461 between PC scores and the vector  length obtained  from staff  ratings of the inmates taken over the sample of 79 men.  150  SUMMARY AND  CONCLUSIONS  From 147 men seen over 13 months of data collection, 79 provided  complete data for the analyses of interest here.  Issues  of generalizability have been discussed, but the sample may be considered  adequate for the purposes of diagnostic  and the assessment of differences among derived  comparisons  the descriptive  aspects  from the dependent measures as they apply to the  different diagnostic groups.  These latter differences  reflected  staff perceptions of interpersonal characteristics as described by the IAS-R which provided groups based defined  reasonably good discrimination  on the PC and Factor  for APD.  of  I criteria but not for groups  The discriminative utility of these  differences  was seen to be much higher for the PC and Factor I groups than the APD and APD-R groups relative to the base rates for these different  categorizations.  Thus, these results contribute to the nomological for the construct  validity of the Psychopathy Checklist and  reflect well on potential applications of the IAS-R samples.  network  in clinical  Speculations and Suggest ions for Further With interest  respect to the population targeted here, and the in the representation of IAS-R  distances  inter-profile  it is suggested that such patterns may  in a different context for the assessment.  particularly  That  is,  and no consequence from their  participation may have resulted pattern of description  in a relatively more  that would  result from an  assessment with assumed consequences presentations.  forthright  institutional  resulting from their  This raises an issue for  informed consent  selffor  in such settings where the goal may be in developing an  assessment strategy with practical applications. there  still  for the psychopathic group, the present context with  assurances of confidentiality  research  point  relying on expectations of differential patterns of  dissimulation, obtain  Research  As  is an ethical onus to inform subjects as to the  "research" interests  of the research and their own freedom from any obligation or consequence there  in participating.  As an applied  interest, however,  is a desire to know how the results of assessment may be  expected general  to obtain  in the context of its application.  This is a  issue for assessment and also affects the use of the  Psychopathy Checklist  in contexts other than  its  research  applications. There concerning  is also  interest  in obtaining more  information  the differential application of the Factor I items of  the PC versus the total PC in the assessment of psychopathy. has been suggested, Factor aspects considered  I may assess the  Factor  I and IAS-R  Further  institutions.  Research  the relation of other measures of narcissism results within  populations would also be of  provide  samples, may have broader  in contexts beyond correctional  investigating  characterological  central to psychopathy and, as less dependent  on behaviours common to incarcerated utility  As  incarcerated  and  to  other  interest.  research using the IAS-R  in clinical samples would  information necessary to assess the potential  differential utility of IAS-R profiles  in various  clinical  populations, and whether such profiles may differ from those seen here.  It may be seen that ratings provided by  relatively  familiar but objective others could be useful  in the  characterization  related  personality  of interpersonal perceptions  to other  or psychiatric disorders, and contribute to  "prototype" development of the personality disorders to the Interpersonal  Circle.  in relation  REFERENCES American Psychiatric Association (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, D.C.: Author. American Psychiatric Association (1987). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, D.C.: Author. Anderson, T.W. (1958). An introduction to multivariate statistical analysis. New York: Wiley. Astin, A.W. (1959). A factor study of the MMPI psychopathic deviate scale. Journa1 of Consulting Psychology, 23, 550-554. " Blackburn, R. (1987). Two scales for the assessment of personality disorder an antisocial populations. Personality and Individual Differences, 8_(1), 81-93. Blackburn, R., & Maybury, C. (1985). Identifying the psychopath: The relation of Cleckley's criteria to the interpersonal domain. Personality and Individual Differences, 6_, 375-386. Brodsley, S.L. (1967 ) . Collected papers in prison psychology. Council for Research and Evaluation Project 6-66, U.S. Disciplinary Barracks, Fort Leavenworth, Kansas. Brown, H.J., & Gutsch, K. (1985). Cognitions associated with a delay of gratification task: A study with psychopaths and normal prisoners. Criminal Just ice and Behavior, . 12 (4) , 453-462. Buros, O.K. Yearbook Buss, A.  (Ed.). (1978). Eighth Mental Measurements (Vol. 1, p. xxxviii), New Jersey: Gryphon.  (1966). Psychopathology. New York: Wiley.  Buss, D.M., Gomes, M., Higgins, D.S. & Lauterback, K. (1987). Tactics of Manipulation. Journal of Personality and Social Psychology, 5j2 (6) , 1219-1229. Butcher, J.N., & Tellegen, A. (1978). Common methodological problems in MMPI research. Journal of Consulting and Clinical Psychology, 46(4), 420-428.  Campos, A. (1986). Self-ideal congruence as a measure of self-esteem. Psychological Reports, 58J3), 729-30. Chartier, B.M. (1984). The meaning and measurement of interpersonal flexibility. Unpublished doctoral dissertation, University of Saskatchewan. Christie, R., & Geis, F. (1970). Studies New York: Academic Press.  in Machiavellianism  Cleckley, H. (1982). The mask of sanity, Louis, MO.: Mosby.  (rev. ed.). St.  Comrey, A.L. (1958). A factor analysis of items on the MMPI psychopathic deviate scale. Educat ional and Psychological Measurement, 18, 91-98. D a h l s t r o m , W.G., & Reifler, C.B. (1970). MMPI Code patterns for entering freshmen. Unpublished materials, 1970. D a h l s t r o m , W.G., Welsh, G.S., & Dahlstrom, L.E. (1972). An MMPI Handbook (Vol. I, rev. ed.) . Minneapolis: U. of Minnesota. D a h l s t r o m , W.G., Welsh, G.S., & Dahlstrom, L.E. (1972). An MMPI Handbook, (Vol. II). Minneapolis: U. of Minnesota. Dor land's illustrated medical dictionary. sixth edition. Philadelphia: Saunders.  (1981). Twenty-  Frances, A. (1980). The DSM-III personality disorders section: A commentary. American Journal of Psychiatry, 137, 1050-1054. Freedman, M . , Leary, T., Ossario, A . , & Coffey, H. (1951). The interpersonal dimension of personality. Journa1 of Personality, 20, 143-161. Gifford, R., & O.Connor, B. (1987). The interpersonal circumplex as a behavioral map. Journal of Personality and Social Psychology, 52(5), 1019-1026. Gilberstadt, H., & Duker, J. (1965) . A handbook for clinical and actuarial MMPI interpretation. Philadelphia: Saunders Glass, G.V., & Stanley, J.C. (1970). Statistical methods education and psychology. Englewood Cliffs, NJ.: Prent ice-Hall.  in  Gough, H.G. (1948). A sociological theory of psychopathy. American Journal of Sociology, 53, 359-366. Gough, H.G., & Heilbrun, A.B. (1980). The Adjective Checklist Manual. Palo Alto: Consulting Psychologists Press. Hare, R.D. Wiley.  (1970). Psychopathy: Theory and research.  Toronto:  Hare, R.D. (1979). A research scale for the assessment of psychopathy in criminal populations. Personality and Individual Differences, 1, 111-119. Hare, R.D. (1980). Psychopathy. In H.M. Van Praag (Ed.). Handbook of biological psychiatry (Part II), (pp. 249261) . New York: Marcel Dekker. Hare, R.D. (1981). Psychopathy and violence. In J.R. Hays, T.K. Roberts, & K.S. Solway (Eds.), Violence and the violent individual (pp. 53-74). New York: Spectrum. Hare, R.D. (1982). Psychopathy and the personality dimensions of psychoticism, extraversion, and neuroticism. Personality and Individual D i f f e r e n c e s 3 , 35-42 . Hare, R.D. (1983). Diagnosis of antisocial personality disorder in two prison populations. American Journal of Psychiatry, 140 (7), 887-890. ~ ~ Hare, R.D. (1985a). Comparison of procedures for the assessment of psychopathy. Journal of Clinical and ConsuIting Psychology, 53 (1), 7-16. Hare, R.D. (1985b). The Psychopathy Checklist. Unpublished manuscript, Department of Psychology, University of British Columbia, Vancouver, Canada. Hare, R.d. (1986). Twenty years of experience with the Cleckley psychopath. In W.H. Reid, D. Dorr, J.I. Walker, & J.W. Bonner (Eds.), Unmasking the psychopath, pp. 3-27). New York: Norton. Hare, R.D., & Cox, D. (1978). Clinical and empirical conceptions of psychopathy, and the selection of subjects for research. In R. Hare & D. Schalling (Eds.), Psychopathic behaviour: Approaches to research (pp. 1-21) . Chichester: Wiley. Hare, R.D., & Frazelle, J. (1980). Some preliminary notes on the use of a research scale for the assessment of p s y c h o p a t h y - i n criminal populations. Unpublished manuscript: University of British Columbia.  Hare, R.D., & Harpur, T.J. (1986). Weak data, strong conclusions: Some comments on Howard, Bailey & Newman's use of the Psychopathy Checklist. Personality and Individual Differences, 7, 147-151. Hare, R.D., Williamson, S.E., & Harpur, T.J. (1988). Psychopathy and language. In T.E. Moffit & S.A. Medwick (Eds.), Biological contributions to cr ime causat i on. Dordrecht, Netherlands: Martinus Niijhoff. Harpur, T.J., Hakstian, A.R., & Hare, R.D. (1988a). The factor structure of the Psychopathy Checklist. Manuscript submitted for publication. Harpur, T.J., Hakstian, A.R., & Hare, R.D. (1988b) . Separable personality and behavioural traits underlie psychopathy Manuscript submitted for publication. Hart, S.D., Kropp, P.R., & Hare, R.D. (1988). Performance of male psychopaths following conditional release from prison. Journal of Consulting and Clinical Psychology, 56_(2), 227-232 . " ~ ~ Hathaway, S.R., & McKinley, J.C. (1951). The Minnesota Multiphasic Personality Inventory Manual (Rev.). New York: Psychological Corporation. Hathaway, S.R., & Monachesi, E.D. (1953). (Eds.). Analyzing and predicting juvenile delinquency with the MMPI. Minneapolis: University of Minnesota Press. Hays, W.L. (1973). Statistics for the social sciences ed.). New York: Holt, Rinehart & Winston.  (2nd  Kiesler, D. (1985). The 1982 Interpersonal Circle: An analysis of DSM-III Personality Disorders. In T. Millon & G. Klerman (Eds.) Contemporary Issues in psychopathy. New York: Guildford. Lachar, D. (1978) . The MMPI: CIinical assessment and automated interpretation. Los Angeles: Western Psychological Services. Laforge, R. (1977). Using the ICL. Unpublished Mill Valley, California.  manuscript:  Leary, T. (1957). Interpersonal diagnosis of personality. New York: Ronald. Leary, T., & Coffey, H.S. (1955). Interpersonal diagnosis: Some problems of methodology and validation. Journal of Abnormal and Social Psychology, 50, 110-124.  McCord, W. , & McCord, J. (1964 ). The psychopath: An essay on the criminal mind. Princeton: Van Nostrand. McLeraore, C.W., & Benjamin, L.S. (1979). Whatever happened to interpersonal diagnosis? A psychosocial alternative to DSM-III. American Psychologist, 34_(1), 17-34. McMahon, R. (1970). MMPI prof iles from a sample of North Carolina youth offenders. Unpublished materials. M a d i s o n , J., & Paddock, J.R. (1983). Assessing variability in circumplex models of personality. Journal of Personality A s s e s s m e n t , 47J4) , 390-395. M a h e r , B. (1966). Principles of Psychopathology. New York: McGraw-Hill. Mardia, K.V. Academic.  (1972). Stat ist i cs of d i rect ional data . London:  Marks, P. , & Seeman, W. (1963) . The actuarial description of personality: An at las for use wi th the MMPI. Baltimore: W i l l i a m s and Wilkins. Megargee, I. (1977). I: The need for a new classification system. Cr iminal Just ice and Behavi or, 4_(2), 107-114. Megargee, I. (1984). A new classification system for criminal offenders, VI: Differences among the types of the A d j e c t i v e Checklist. Criminal Just ice and Behavior, 11(3), 349-376. Megargee, I., & Bohn, M.J. (1977). IV: Empirically determined characteristics of the ten types. Criminal Justice and Behavior, 4(2), 149-210. M e g a r g e e , I., & Dorhout, B. (1977). Ill: Revision and refinement of the classificatory rules. Criminal Just ice and Behavior, 4(2), 125-148. M e y e r , J., & Megargee, I. (1977). II: Initial development of the system. Criminal Justice and Behavior, 4_(2), 114-124. Mi 1Ion, T. (1969). Modern psychopathology. W.B. Saunders. Mi lion, T.  Philadelphia:  (1981). Disorders of personality. New York: Wiley.  Murray, H.A. Oxford.  (1938). Explorations  in personality. New York:  158 Norcliffe, G.B. (1977). Inferential statistics geographers. London: Hutchinson.  for  Payne, F.D., & Wiggins, J.S. (1968). Effects of rule relaxation and system combination on classification rates in two MMPI "cookbook" systems. Journal of Consulting and Clinical Psychology, 32_(6), 734-736 . Phillips, N. (1983) . Selection of items with circumplex properties. Unpublished manuscript, University of British C o l u m b i a , Vancouver, Canada. Pichot, P. (1978). Psychopathic behaviour: A historical overview. In R. Hare & D. Schalling (Eds.), Psychopathic behaviour: Approaches to research (pp. 55-70). Chichester: Wiley. Plutchik, R., & Platman, S. (1977) . Personality connotations of psychiatric diagnoses. Journal of Nervous and Mental Disease, 165, 418-422. Raine, A. (1986). Psychopathy, schizoid personality and borderline/schizotypal personality disorders. Personality and Individual Differences, 7_(4), 49 3-501. Ray, J., & Ray, J.A. (1982). Some apparent advantages of subclinical psychopathy. Journal of Social Psychology, 117, 135-142. Robins, L.N. Waver ley.  (1966). Deviant children grown up.  Baltimore:  Robins, L.N. (1978). Aetiological implications in studies of childhood histories relating to antisocial personality. In R. Hare & D. Schalling (Eds.), Psychopathic behaviour: Approaches to research, (pp. 255-271). Chichester: Wiley. Rorer, L.G., & Widiger, T.A. (1983). Personality structure and assessment. Annual Review of Psychology, 34, 431-463. R o s e n b e r g , M. (1965). Society and the adolescent New Jersey: Princeton University Press.  self-image.  Sarason, I.G. (1978). A cognitive social learning approach to juvenile delinquency. In R. Hare & d. Schalling (Eds.), Psychopathic behaviour: Approaches to research, (pp. 299317). Chichester: Wiley. Schroeder, M., Schroeder, K., & Hare, R. (1983). Generalizabi1ility of a checklist for the assessment of psychopathy. Journal of Consulting and Clinical Psychology, 5 U 4 ) , 511-516.  Silber, E., & Tippett, J. (1965). Self-esteem: Clinical assessment and measurement validation. Psychological Reports, 16, 1017-1071. Simmons, J.G., Johnson, D.L., Gouvier, W.D., & Muzyczka, M.J. (1981). The Myer-Megargee typology dynamic or unstable? Criminal Justice and Behaviour, 8_(1), 49-54. Spence, J., Holmreich, R.L., & Holahan, C. (1979). Negative and positive components of psychological masculinity and femininity and their relationships to self-reports of neurotic and acting-out behaviors. Journal of Personality and Social Psychology, 37, 1673-1682. Still, L. (1987, May 21). Rapist after revenge, sentencing judge told. The Vancouver Sun, Vancouver, Canada. Sutker, P., De Santo, N., & Allain, A. (1985). Adjective self-descriptions in antisocial men and women. Journal of Psychopathology and Behavioral Assessment, 7^(2), 175-181. Trasler, G. (1978). Relations between psychopathy and persistent criminality--methodological and theoretical issues. In R. Hare & D. Schalling (Eds.), Psychopathic behaviour: Approaches to research, (pp. 273-298) . Walters, G.D. (1985). Scale 4 (Pd) of the MMPI and the diagnosis Antisocial Personality. Journal of Personality A s s e s s m e n t , 4 9 ( 5 ) , 474-476. Weiss, J. (1986). The nature of psychopathy, Parts I, II, and III. Directions in Psychiatry, 5_, Whole Nos. 25, 26, 27. Widiger, T. (1984 , June) . Revis ions of Axis 11 from the perspective of research in psychology. Paper presented to the AmericanHPsychiatric Association Work Group to revise DSM-III, New York, New York. Widiger, T., & Frances, A. (1985). The DSM-III personality disorders. Archives of General Psychiatry, 42, 615-623. Widiger, T.A., & Kelso, K. (1983). Psychodiagnosis of Axis II. Clinical Psychology Review, 3_, 491-510. Widom, C.S. (1978). A methodology for studying noninstitutionalized psychopaths. In R. Hare & D. Schalling (Eds.), Psychopathic behaviour: Approaches to research (pp. 71-84). Chichester: Wiley.  Wiggins, J.S. (1979). A psychological taxonomy of traitdescriptive terms: The interpersonal domain. Journal of Personality and Social Psychology, 37, 395-412. Wiggins, J.S. (1980). Circumplex models of interpersonal behaviour. In L. Wheeler (Ed.), Review of personality and social psychology, (Vol. I). Beverley Hills: Sage. Wiggins, J.S. (1982). Circumplex models of interpersonal behaviour in clinical psychology. In P.C. Kendall & J. Butcher (Eds.), Handbook of research methods in clinical psychology. New York: Wiley. W i g g i n s , J.S., & Broughton, R. (1985). The interpersonal circle: A structural model for the integration of personality research. In R. Hogan & W.H. Jones (Eds.), Perspectives in personality: Theory, measurement and interpersonal dynamics. Vol. I, (pp. 1-47). Greenwich, CT: JAI Press. Wiggins, J.S., Phillips, N., & Trapnell, P. (in press). Circular reasoning about interpersonal behaviour: Evidence concerning some untested assumptions underlying diagnostic classification. Journal of Personality and Social Psychology. W i g g i n s , J.S., Trapnell, P., & Phillips, N. ( in press). Psychometric and geometric characteristics of the Revised Interpersonal Adjective Scales (IAS-R). Multivariate Behavioral Research. Wong, S. (1985) . The criminal and institutional behaviours of psychopaths. Program Branch User Report, Ministry of the Solicitor General of Canada, Ottawa, Canada.  APPENDIX  162  SUBJECT CONSENT  A title for this research could be "Self-perception in adult male offenders."  This research is being conducted by persons affiliated with the  Department of Psychology at UBC.  The coordinator for this research project is  Mike Foreman who can be reached at 228-5581. The purpose of this research is to investigate how individuals who have had problems with the law see themselves and see themselves in relation to others.  In order to do this, we wish to have you complete three different  questionnaires. All information which you provide is confidential and will not be identified to anyone nor entered in to any file or record kept by the police or the institution.  Our files will be coded and not identified by your name.  The total time to be expected of you is about 3 hours which may not happen all at once.  You will be paid $5.00 (five dollars) for the completion  of the questionnaires.  If you have any questions about the procedures please  ask, and I will explain further to be sure you fully understand what is expected of you and what the program involves. If you do decide to participate and change your mind later, you may withdraw from the program at any time without consequence.  Refusal to  participate or withdrawal from the program will in no way affect your status or standing in the Correctional centre. I have read this form and agree to participate in this study and understand that I may withdraw at any time without consequence.  Signature . Date  Witness  163 INTRODUCTION TO THE QUESTIONNAIRE  The following questionnaires are being used as a means of measuring how individuals such as yourself in Matsqui think about themselves.  The results  of this study will be used to make suggestions for future research into how psychological services may make a better contribution to the Corrections system. Since the questionnaires rely on your self-report, they can only be as accurate as you will allow.  The information gained is confidential  (identification is coded and will be used only for the purposes of this research) and will not reflect on you as an individual.  Our interest is in  the overall average response of a large group of people to which you are one contributor. Please respond to the items as best you can so as to provide an accurate representation of what you actually believe. Altogether, the questionnaires take about 1 hour to complete, although you may be asked to do another questionnaire which takes about 1 to 1 1/2 hours. You may withdraw from this program at any time without any consequence or effect on your status in the institution.  If the questionnaires are  completed it will be assumed that you have agreed to participate in the study. When completing the word lists, if you are uncertain of the meaning of a word ask to have it explained to you.  Please be sure to provide a response  to all the words listed as directed in the instructions which follow. Thank you for your cooperation.  16 4 IAS SET 1  On the page that follows, you will find a list of words that are used to describe people's personal characteristics. explanations of the words in the list.  The next two pages provide  If you are uncertain about the meaning  of a word in the list, look up the explanation for that word to see if that makes it more clear.  If you are still unsure about the meaning, ask to have  the word explained to you. For each word in the list, indicate how accurately you think the word describes you. The accuracy with which a word describes you is to be judged on the following scale:  1  2  3  4  5  6  7  8  extremely very quite slightly slightly quite very extremely inaccurate inaccurate inaccurate inaccurate accurate accurate accurate accurate  For example, consider the word BOLD.  How accurately do  you think BOLD describes you as a person? If you think this word is a quite accurate description of you, write the number "6" to the left of the word:  6 BOLD  If you think this word is a slightly accurate description of you, write the number "5" next to it, if very inaccurate, write the number "2", and so forth. Rate the accuracy of all the words in the list as to how well you think they describe you.  165  IAS SET 2  On the pages that follow, you will find the same list of words used to describe people's personal characteristics, and the glossary which helps to explain them.  This time, rate the words as to how accurately they describe your ideal self - the person you would best like to be. Rate the accuracy of the words as before with the following scale:  1  2  3  4  5  6  7  8  extremely very quite slightly slightly quite very extremely inaccurate inaccurate inaccurate inaccurate accurate accurate accurate accurate  Rate all the words in the list as to how well they describe your ideal self - the person you would like to be.  166 IAS SET  2  The following pages provide the same list of words and their explanation. This time, for each word in the list indicate how accurately the word describes you as you think your friends would describe you.  The accuracy with which a word may describe how you think your friends see you is to be judged on the same scale as before:  1  2  3  4  5  6  7  8  extremely very quite slightly slightly quite very extremely inaccurate inaccurate inaccurate inaccurate accurate accurate accurate accurate  167  IAS SET  2  On the following pages you will find the same list of words again.  This time, rate each word in the list as to  how accurately the word describes you as^ you think a^ member of the institutional staff that knows you (for example, your case manager) would describe you.  1  2  3  4  5  6  7  8  extremely very quite slightly slightly quite very extremely inaccurate inaccurate inaccurate inaccurate accurate accurate accurate accurate  Indicate the staff member that you are thinking of:  168 1  2  3  4  5  6  7  8  Extremely Very Quite Slightly Slightly Quite Very Extreme Inaccurate inaccurate inaccurate inaccurate accurate accurate accurate accurat  (001 )  introverted  (033)  unargumentative  (002)  undemanding  (034)  tender  (003)  assertive  (035)  unsympathetic  (004)  unauthoritative  (036)  timid  (005)  uncalculating  (037)  unbold  (006)  accommodating  (038)  forceful  (007)  kind  (039)  unwily  (008)  charitable  (040)  extroverted  (009)  shy  (041 )  gentle-hearted  (010)  uncunning  (042)  persistent  (011 )  cold-hearted  (043)  perky  (012)  ruthless  (044)  friendly  (013)  dissocial  (045)  unneighbourly  (014)  tender-hearted  (046)  self-confident  (015)  soft-hearted  (047)  outgoing  (016)  cheerful  (048)  boastful  (017)  dominant  (049)  bashful  (018)  antisocial  (050)  firm  (019)  iron-hearted  (051)  uncrafty  (020)  enthusiastic  (052)  unsociable  (021)  self-assured  (053)  hard-hearted  (022)  cruel  (054)  wily  (023)  unsparkling  (055)  calculating  (024)  cunning  (056)  uncheery  (025)  meek  (057)  sly  (026)  uncharitable  (058)  neighbourly  (027)  unsly  (059)  warmthless  (028)  unaggressive  (060)  distant  (029)  jovial  (061 )  cocky  (030)  crafty  (062)  sympathetic  (031 )  boastless  (063)  forceless  (032)  domineering  (064)  tricky  169 Glossary for the 64 item IAS-R  01  introverted - feel more comfortable by yourself; are less interested in other people  02  undemanding - don't demand or expect much from others  03  assertive - tend to be aggressive and out-spoken with others  04  unauthoritative - don't try to influence others; go with other's opinions  05  uncalculating - don't try to manipulate others or maximize your own gain  06  accommodating - obliging, tend to do favours for others  07  kind - thoughtful and caring for others  08  charitable - generous, like to help others  09  shy - lack of self-confidence, tend to be uncomfortable around others  10  uncunning - not crafty or sly, tend to be straightforward with others  11  cold-hearted - have little warmth or feeling for others  12  ruthless - pursue your own interests regardless of the effect on others  13  dissocial - don't care for the company of others  14  tender-hearted - easily feel love, pity, or sorrow for others  15  soft-hearted - tend to be easy-going or gentle with others  16 cheerful - happy, usually in good spirits 17 dominant - tend to lead or control others 18  antisocial - dislike the company of others; behaviour not affected by social conventions  19  iron-hearted - tend to be stern or harsh with others  20  enthusiastic - enjoy active involvement with others  21  self-assured - confident, know yourself to usually be right  22  cruel - can cause pain and suffering to others  23  unsparkling - not lively or entertaining with others  24  cunning - crafty or sly, skillful in manipulating others  25  meek - show little spirit or courage; mild mannered  26 uncharitable - don't like to help others; judge others severely 27  unsly - not tricky or cunning; tend to be honest and sincere  28  unaggressive - not forceful  29  jovial - happy, good sense of humour  30  crafty — can mislead or manipulate others for your own purpose  31  boastless — don't like to brag  32  domineering - tend to control or manipulate others  33  unargumentative - tend to avoid arguments or fights  170  34  tender - warm and loving with others  35  unsympathetic - unable to understand or uninterested in the feelings of others  36  timid - tend to be fearful or uncomfortable around others  37  unbold - not daring or courageous  38  forceful - tend to take charge or assert control  39  unwily - not tricky or crafty  40  extroverted - enjoy the company of others  41  gentle-hearted - kind or warm with others  42  persistent - don't give up even if others think you are wrong  43  perky - lively or vigorous, enthusiastic with others  44  friendly - pleasant toward others  45  unneighbourly - unfriendly, avoid contact with others around you  46  self-confident - self-assured, trust your own feelings or opinions  47  out-going - enjoy meeting other people  48  boastful - tend to brag  49  bashful - tend to shy away from public attention  50  firm - steady or steadfast; have others do things your way  51  uncrafty - not tricky or sly when dealing with others  52  unsociable - don't enjoy meeting people or being in the company of others  53  hard-hearted - have no feeling for others  54  wily - crafty, cagey, or tricky  55  calculating - tend to use or manipulate others to your own advantage  56  uncheery - not lively or bright with others  57  sly - crafty, secretive, or cunning when dealing with others  58  neighbourly - friendly, get involved with people around you  59  warmthless - have no feelings of affection or pleasure for others  60  distant - tend to be cold toward others; avoid relationships  61  cocky - conceited, self-centred; think highly of your own abilities  62  sympathetic - able to share or understand the interests or feelings of others  63  forceless - tend to be timid or weak; prefer the leadership of others  64  tricky - able to fool or deceive others  The enclosed form is being forwarded to you as part of a research project regarding self-perception among male inmates. One aspect of self-perception involves how one believes oneself to be seen by others, which the inmate has been asked to complete with respect to a particular member of the institutional staff whom they may choose. As a basis for comparison it is important to have a rating completed by that staff member of the inmate.  I would very much  appreciate your completing this form at your earliest opportunity and returning it to me, Mike Foreman, c/o the institution hospital. All information obtained is confidential and will not be made available to the inmate involved.  He is; however, aware of  this questionnaire being sent to you and has given his consent to have it completed. If you have any questions please contact me by leaving a message with the nurse's station at the institution Hospital. Thank you for your cooperation.  172  RATING BY OTHERS  On the pages that follow, you will find a list of words that are used to describe people's personal characteristics, and a glossary to help explain their meaning.  For each word  in the list, please indicate how accurately you think the word describes  using the following  scale:  1  2  3  4  5  6  7  8  extremely very quite slightly slightly quite very extremely inaccurate inaccurate inaccurate inaccurate accurate accurate accurate accurate  For example, consider the word BOLD. that word describe him as a person?  How accurately does  If you think that this word  is a quite accurate description of him, write the number "6" to the left of the word:  BOLD  If you think that this word is a slightly inaccurate description of him, write the number "4" next to it; if a very inaccurate description write the number "2", etc.  If a word seems odd (some are) or if you are unsure of its meaning, please look it up in the list on pages 3 and 4.  f  173  I  APPENDIX B  I i  Harpur, Hare, and Hakstian demonstrating differential (characterological  (1988b) have analyzed  relationships  vs. behavioural  external variables.  data  of Factors 1 and 2  item composition)  The pattern of differences  underscores the integration of both personality  to  seen and  behavioural aspects o f psychopathy as assessed by the PC, and points to the tendency of most other measures to emphasize the behavioural aspects without capturing features considered  the personality  central to the concept of  psychopathy.  Correlations of the Factors with the variables  relevant  to  this research have been excerpted with permission, for a complete presentation and Hakstian  (1988b).  the reader is referred  to Harpur, Hare,  CORRELATIONS OF THE FACTORS AND PC TOTAL S C O R E S  WITH SELECTED SELF-REPORT SCALES  1  2  PCL TOTAL  38 06 38  .11 .05 .16  .31 .3 .28 .2 .32 .3  .25 .19 .27  2.55 2.65* 2.05  Pd + Mab Pd - Soc  l 06 1 17  .10 .08  .30 .3 .49 .4  .23 .33  2.63* 5.07**  Soc  223  -.06  -.44  -.31  6.38**  -.01 -.29  .19 -.30  4.19** 0.32  .32 -.42  .45 -.46  2.40 0.08  .66 .63  .61 .54  FACTORS N  SCALE MMPI Pd  Pdb  Ma  SS  CP I  :18 :J  IAS-R - SELF AND OTHER REPORTS SELF-RATING .35 DOM 113 -.26 LOV 113 STAFF RATING .53 DOM 79 -.41 LOV 79 WITH APA DIAGNOSTIC CRITERIA APDC 80 .34 APD-Rd 176 .32  NOTE  di  Z-stati sti c a  §:!!*  3.36** 5.06**  a.  Test of the difference between the correlations of Factor 1 and Factor 2 with the given scale.  b.  Data from an independent sample.  c.  APD diagnosis was decided by joint agreement of two raters.  d.  Approximately 55% of the sample were assessed by two raters, the remainder by one rater only.  * p < .005. ** p < .0001. MMPI = Minnesota Multiphasic Personality Inventory. Pd = Psychopathic deviate scale. M a = Hypomania scale. CPI = California Psychological Inventory. Soc Socialization scale.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0107157/manifest

Comment

Related Items