You are viewing a free preview of this lesson.
Subscribe to unlock all 9 lessons in this course and every other course on LearningBro.
Jeffrey Hancock and colleagues' study of "the language of psychopaths" is the contemporary study for the individual-differences theme of measuring differences, and its central idea is at once simple and unsettling: that the way a person talks — the small, largely automatic choices of grammar and vocabulary they make without thinking — can reveal deep features of their psychology, including psychopathy. The study asks whether psychopaths (a specific, measurable personality profile associated with callousness, lack of empathy and instrumental use of others) speak differently from other offenders when describing the same kind of event — their own homicide — and it answers the question not by human impression but by feeding the offenders' words through computerised text-analysis tools that count and categorise linguistic features at a scale and objectivity no human coder could match. The playful title, "Hungry like the wolf", captures the study's most memorable finding: that psychopaths, when recounting their crimes, dwelt strikingly on their basic bodily needs — food, drink, money — as if their violence were as unremarkable as satisfying an appetite.
Where Gould's classic partner study is a critique of measurement gone wrong, Hancock is a demonstration of measurement done in a modern, computational key — objective, automated, and applied to a subtle difference (patterns of language) rather than to the crude, culturally-loaded "general intelligence" of the army tests. This makes the pairing especially rich, because it invites the question of whether the measurement of individual differences has become more careful and more valid in the century since Yerkes. This lesson tells the study in the OCR "tell the story" format — background, aim, method, results, conclusions, evaluation, and links — and attends throughout to the theme of how a difference is measured, since that is what unites it with Gould.
| This lesson covers | OCR H567 Component 02 element | AO focus |
|---|---|---|
| Background: psychopathy, its assessment (PCL-R), and language as a window on the mind | Section A — Individual differences; theme: measuring differences (contemporary) | AO1 knowledge |
| Method: analysis of murderers' narratives with text-analysis tools (Wmatrix, DAL) | Section A — Core study (Hancock) | AO1; AO2 |
| Results (subordinating conjunctions, past tense, disfluencies, basic-need words, less emotion) | Section A — Core study | AO1 |
| Conclusions (psychopaths' instrumental, detached, need-focused language) | Section A — Core study | AO1; AO3 |
| Evaluation: method, data type, ethics, validity, reliability, sampling | Section A; Section B debates | AO3 |
| Links to theme, area (Individual differences), perspective and debates | Section B — Areas, perspectives, debates | AO1; AO3 |
The specification is referenced descriptively; consult the official OCR H567 specification document for its exact published wording. This lesson develops AO1 (the aim, the tools, the linguistic findings and conclusions), AO2 (applying the idea of language-as-measurement to novel material) and AO3 (evaluating the study's validity, objectivity, sampling and ethics).
The study rests on two bodies of prior knowledge: the concept and assessment of psychopathy, and the idea that language reflects underlying psychology.
Psychopathy is a personality construct describing a cluster of traits including superficial charm, grandiosity, pathological lying, lack of remorse or guilt, shallow emotions, callousness and lack of empathy, impulsivity, and a manipulative, instrumental approach to other people — treating others as means to one's own ends. Crucially, psychopathy is not the same as "being a criminal"; it is a measurable individual difference that some offenders have and many do not. The standard assessment tool is the Psychopathy Checklist–Revised (PCL-R), developed by Robert Hare, which scores an individual (on the basis of interview and file review) against a set of criteria to yield a psychopathy score; a threshold on this scale is used to classify an offender as psychopathic. This measurement framework is what allowed Hancock's team to divide their offenders into psychopathic and non-psychopathic groups objectively, rather than impressionistically.
The second background idea is that the words people use are not random but reflect how they think and feel. This is the premise of the whole field of psychological text analysis: that features of language — which pronouns someone uses, whether they speak in the past or present tense, how much emotion words they include, how they connect events causally — are traces of underlying psychological states and dispositions, often produced automatically and below conscious control. If psychopaths differ from other people in fundamental ways — in their emotional shallowness, their instrumental view of others, their detachment — then, the reasoning goes, those differences should leave measurable fingerprints in their language, especially when they describe something emotionally significant.
Hancock's team reasoned that the ideal test of this would be to have psychopathic and non-psychopathic offenders describe the same kind of emotionally-charged event — and few events are more charged than one's own act of homicide. Several specific predictions followed from the theory of psychopathy. Because psychopaths view the world instrumentally (in terms of cause, effect and personal gain), they should use more language of causation and means-to-ends (subordinating conjunctions such as "because", "since", "so that"). Because they are emotionally shallow and detached, they should use fewer and less intense emotional words, and more past-tense forms (suggesting psychological distance from the event). And because they are focused on their own basic physical needs rather than higher social or emotional concerns, they should refer more to concrete needs like food, drink and money. The study set out to test these predictions with computerised tools.
The aim of the study was to test whether psychopathic offenders use language differently from non-psychopathic offenders when describing their crimes, and specifically to examine whether their language reflects the theorised features of psychopathy: an instrumental, cause-and-effect worldview; emotional detachment and shallowness; psychological distancing from the event; and a focus on concrete physiological needs. A further, methodological aim was to demonstrate the value of automated statistical text-analysis tools for the objective measurement of such linguistic differences, in a way that human judgement could not achieve at scale.
The study was a quasi-experiment using an independent-measures design and combining it with content (language) analysis. The independent variable — whether an offender was psychopathic or non-psychopathic (determined by PCL-R score) — was a pre-existing characteristic, not manipulated, which makes it quasi-experimental. The dependent variables were the various linguistic features measured in the offenders' spoken narratives. The core innovation was the use of computerised text-analysis software to quantify those features objectively.
The participants were 52 male offenders held in Canadian correctional facilities, all of whom had been convicted of murder and had admitted their crime. Using the PCL-R, they were divided into two groups: 14 classified as psychopaths (scoring above the relevant threshold) and 38 non-psychopaths. All were describing a genuine, serious crime they had committed. The use of a real, forensically-assessed offender sample is a notable strength of the study's ecological relevance, though it also constrains generalisability (below).
Each offender took part in an interview in which he was asked to describe his homicide in detail, in his own words. These narratives were audio-recorded and transcribed verbatim, producing the raw language data. The transcripts were then analysed using two computerised text-analysis tools:
| Tool | What it does |
|---|---|
| Wmatrix | A corpus-analysis tool that automatically tags each word grammatically and semantically (by part of speech and by meaning category), allowing the frequency of different word-types and semantic categories to be counted and compared between groups |
| DAL (Dictionary of Affect in Language) | A tool that rates words on emotional dimensions — in particular their pleasantness and their emotional intensity/activation — allowing the emotional tone of the narratives to be quantified |
The procedure, then, was: interview → transcription → automated analysis with Wmatrix and DAL → statistical comparison of the psychopathic and non-psychopathic groups on the measured linguistic features. Because the analysis was done by software applying fixed rules, it was highly objective and reliable — the same text would be tagged and scored identically every time — a marked contrast with the subjective coding that human analysis of such narratives would require.
The findings largely confirmed the predictions derived from the theory of psychopathy, and the candidate should know the key differences between the two groups.
More subordinating conjunctions (instrumental, cause-effect language). Psychopaths used more subordinating conjunctions — words such as because, since and so that — than non-psychopaths. These words frame events in terms of cause and effect and means to an end, consistent with the instrumental, "I did X because I wanted Y" worldview theorised for psychopathy. Their accounts read more as goal-directed explanations than as emotional narratives.
More focus on basic physiological needs. In keeping with the study's title, psychopaths referred more to basic needs — food, drink and money — in describing their crimes, and correspondingly less to higher-level social needs such as family, religion and spirituality. This is the "hungry like the wolf" finding: their language treated the crime in the register of appetite and material want rather than of relationship or meaning.
Less emotional and less positive language. The DAL analysis showed that psychopaths' language was, overall, less emotionally intense and less pleasant — they used fewer and less positive emotion words. This fits the theorised emotional shallowness and lack of remorse; their descriptions were emotionally flatter than those of the non-psychopathic murderers.
More past-tense and more psychological distance. Psychopaths used more past-tense verbs (and relatively fewer present-tense forms) than non-psychopaths, suggesting greater psychological detachment from the events they were describing — recounting the crime as something finished and removed from the self, rather than reliving it.
More disfluencies. Psychopaths' speech contained more disfluencies — filler sounds such as "uh" and "um". The researchers linked this, tentatively, to the greater cognitive effort psychopaths may need to construct a convincing narrative, consistent with the manipulative, impression-managing aspect of psychopathy (though this interpretation is more speculative than the others).
| Linguistic feature | Psychopaths vs non-psychopaths | Theorised link |
|---|---|---|
| Subordinating conjunctions (because, since) | More | Instrumental, cause-and-effect worldview |
| References to food, drink, money | More | Focus on basic physiological needs |
| Higher social needs (family, religion) | Fewer | Detachment from social/emotional concerns |
| Emotional intensity / pleasantness (DAL) | Lower | Emotional shallowness, lack of remorse |
| Past-tense verbs | More | Psychological distancing from the event |
| Disfluencies ("uh", "um") | More | Greater effort constructing the narrative |
Taken together, these differences paint a coherent picture: the psychopathic murderers' language was more instrumental, more detached, more focused on concrete self-interest, and emotionally flatter than that of the other offenders describing comparable crimes.
Hancock and colleagues concluded that psychopaths' language reflects their distinctive psychology in measurable, patterned ways. Their instrumental, predatory worldview shows up in a greater use of cause-and-effect (subordinating) language; their emotional shallowness and lack of remorse show up in less intense, less positive emotional language; their detachment shows up in more past-tense distancing; and their focus on their own material and bodily wants — rather than on higher social or emotional concerns — shows up in more references to basic physiological needs such as food, drink and money. The crime, in a psychopath's telling, is framed less as a moral or emotional event and more as the satisfaction of a want — hence "hungry like the wolf".
The researchers concluded, further, that computerised text analysis is a powerful and objective tool for measuring such individual differences. By using software (Wmatrix and DAL) that applies fixed rules to every transcript, they could detect subtle linguistic patterns reliably and without the subjectivity of human coding, at a scale and consistency human analysts could not achieve. This demonstrates a modern, objective approach to the measurement of a difficult-to-measure individual difference.
It is important to read these conclusions with appropriate care, and several qualifications matter. First, the differences, though statistically detectable, are differences in degree between groups, not absolute markers: not every psychopath will show every feature, and the features overlap between groups — so the study identifies a statistical linguistic signature, not a "psychopath detector" that could diagnose an individual from their speech. Second, as a quasi-experiment, the study shows that psychopathy is associated with these language patterns, but cannot prove that psychopathy causes them in a simple way, nor rule out that features of the interview context contributed. Third, some interpretations (notably the disfluency finding) are more tentative than others. Keeping these qualifications in view — a group-level statistical signature rather than an individual test, association rather than simple cause, and varying confidence across the findings — is what separates an accurate account from an overstated one.
Subscribe to continue reading
Get full access to this lesson and all 9 lessons in this course.