Hancock - Psychopaths

• background

  language as diagnosis

  • research shown that words you choose to use can reveal your inner thoughts and show psychological state e.g. lying

  language of psychopaths

  • psychopaths typically skilled in conversations - to charm or to lie
  • would lead us to expect their language as cohesive but language of a psychopath is actually less cohesive than non-psychopaths
• aims

  • to use text analysis to analyse the language characteristics of psychopaths in describing violent crimes
  • 3 major characteristics:

  -- an instrumental/predatory world view (use lang. that connects main clauses to dependent clauses e.g. so that

  -- unique socioemotional needs (lang. focuses on basic needs)

  -- poverty of affect (lack of emotional words, more disfluencies and psychological distancing e.g. past tense verbs)
• design

  • quasi experiment -> IV was psychopath or non-psychopath
  • DV was measures of lang. from text analysis
  • independent measures design
  • data collected using semi-structured interviews
• sample

  • 52 male murderers in Canadian correctional facility
  • self-selected (volunteered)
  • 14 = psychopaths, 38 = non-psychopaths
  • no sig. difference between sample in type of murder (manslaughter, first and second degree murder) or age
• materials

  • PCL-R checklist, maximum score of 40 + taps on 2 factors:

  -- affective/interpersonal traits (superficial charm, lack of remorse)

  -- impulsive and antisocial traits (parasitic lifestyle)

  • score of 25 or above = psychopathy
  • text analysis, two methods used:

  -- Wmatrix - classify words

  -- DAL - classify emotional content
• procedure

  • all ppts told purpose was to assess the way they committed the murders
  • tested for psychopathy using PCL-R completed by trained prison psychologists
  • ppts asked to recall the murder - beginning to end
  • all ppts interviewed by 2 senior psych. students and researcher
  • interviews were blind and semi-structured
  • approx. 25 minutes long, videotaped + put into written form
  • analysed using Wmatrix e.g. parts of speech (nouns, verbs)
• results

  • no sig. difference in number of words used
  • psychopaths produced more subordinating conjunctions (e.g. because) -> more cause and effect relationship between events
  • psychopaths used twice as many of basic needs (e.g. food, clothing) - Maslow's hierarchy of needs, NPs sig.ly more 'higher' need words (e.g. family, religion)
  • psychopaths used more past tense verbs (e.g. stabbed not stab)
  • no difference in terms of their emotional lang. (DAL scores)
  • PCL-R (affective traits) showed difference -> neg. correlation with emotionally intense lang.
  • psychopaths used more disfluencies
• research methods

  • semi-structured interviewing - psychopaths had more leeway to reveal unconscious thoughts and feelings
  • some ppts might have forgotten specific details about crime and structured questions could have recalled more
  • interview only concerned one event - murder, and all events were different so descriptions may not be fully comparable
• sample

  • large number of psychopaths fairly easily identified because 15-25% of prison pop. likely to be psychopaths compared to 1% of general pop.
  • psychopathic murderers may have different personality profiles from psychopaths who do not commit murders
  • ppts were volunteers (self-selected) who may differ from psychopaths who were less 'proud' of the murder -> not typical psychopathic murderers
• ethnocentrism

  • Canadian prisoners - rep. one culture (individualist)
  • psychopathy might not manifest itself in the same way as in other cultures (e.g. collectivist)
• data

  • qualitative -> audio-taped transcripts (100,000 words collected)

  -- provides a rich record of ppts thoughts, feelings and attitudes

  • to analyse data, researchers needed to reduce qual. into categories of speech e.g. past tense verbs, disfluencies, giving them quantitative data

  -- making it possible to compare speech used by psychopaths and NPs
• validity

  • results reasonably high in internal validity (face)
  • low external (population) validity -> can't generalise to other pops. female, non-murderers, other cultures
• reliability

  • PCL-R checklist has a high level of reliability
  • inter-rater check was conducted with 2 raters
• ethics

  • informed consent -> ppts weren't told the full extent of the purpose of the study
  • social sensitivity - a sample of psychopaths and NPs and potential negative impact
  • researchers should be aware of distressing nature of results