Chapter 5: Facets of Clinical Psychopathy: Toward Clearer Measurement
Consequence Hypothesis – theoretical support I: 1-of-4 arguments pro-consequence hypothesis
Four theoretical arguments supporting the view that AB is a consequence (i.e., an effect), rather than a symptom, of psychopathy; we term this the “consequence hypothesis.” 1) Classical clinical descriptions of psychopathy (e.g., Arieti, 1963; Gough, 1948; Karpman, 1961; McCord & McCord, 1964) do not include AB as a central feature or symptom of the psychopathy. As the McCord’s argued, “Much of psychology’s confusion over the psychopath can be traced to a basic mistake: equating deviant behavior with the psychopathic personality. … Deviant behavior, then, is an inadequate criterion of psychopathy.” Diagnostic criteria for psychopathy that overfocus on AB have been heavily criticized (Hare, 1991; Millon, 1981; Rogers & Dion, 1991).
Consequence Hypothesis – theoretical support II: 2-of-4 arguments pro-consequence hypothesis
2) there are rational reasons to argue that psychopathy may play a causal role in regard to AB: “the personality features that are under psychopathy seem likely to result in and drive AB” (McDermott et al;., 2000, p. 185). For example, interpersonal symptoms of psychopathy, such as grandiosity, predispose individuals with the disorder to engage in sadistic criminal acts, motivated by a desire to control, demean, or humiliate a victim; affective deficits, such as lack of empathy & anxiety, result in a failure to inhibit antisocial (especially violent) thoughts & urges; and impulsivity increases the likelihood of engaging in criminal acts without considering their consequences (e.g., Baumeister, Smart & Boden, 1996; Blackburn, 1993, 1998; Blair, Jones, Clark & Smith, 1995; Fowles & Missel, 1994; Hare, Cooke & Hart, 1999; Hart 1998; Kernberg, 1998; Malamuth & Brown, 1994; Meloy, 1988, 1988; Serin, 1991).
Consequence Hypothesis – theoretical support III: 3-of-4 arguments pro-consequence hypothesis
3) Antisocial & socially deviant behavior can be considered to be qualitatively different from other symptoms of psychopathy insofar as they reflect specific acts rather than general personality traits. Blackburn (1988) emphasized that to define or diagnose personality disorder in terms of both traits and acts is to mix criteria from distinct conceptual domains (see also Widiger & Lynam, 1998). In a similar vein, Lilienfeld (1994) criticized diagnostic criteria for psychopathy (e.g., PCL-R criteria) that conflated both “basic tendencies” (traits) and “characteristic adaptions” (acts).
Consequence Hypothesis – theoretical support IV: 4-of-4 arguments pro-consequence hypothesis
4) theoretical views of crime and violence suggest that AB results from the influence of a wide range of biological, psychological and social factors (e.g., Gottfredson & Hirschi, 1990). In general theories of crime, psychopathic personality disorder is only one of many important causal factors. Other mental disorders that have been linked to antisocial behavior include psychopathic disorders, mental retardation, substance use and dependence, and other personality disorders.
Consequence Hypothesis – theoretical support IX: PCL-R items too broad
In the final model, the Arrogant & Deceitful Interpersonal Style facet of psychopathy was linked strongly to Relationship lability. The Impulsive & Irresponsible Behavioral Style facet of psychopathy was linked to Criminal behavior both directly, and also indirectly, through Early behavioral problems. These findings challenge current conceptualizations of psychopathy as operationalized by the PCL-R and suggest that the domain tapped by the by the PCL-R items is too broad, embracing both the disorder and certain consequences of the disorder.
Consequence Hypothesis – theoretical support V: 3-factor model & 7 PCL-R items
We (Cooke & Mitchie) examined the link between the 3-factor model of psychopathy (SEE: Psychopathy Symptoms form Coherent Construct? VIII: 3-factor model article) and the 7 PCL-R items (Psychopathy Symptoms form Coherent Construct? VII: IRT & CFA on 20 PCLR items article) that did not aggregatge into the coherent latent trait (SEE: latent trait article article)we call psychopathy in large items that did not aggregate into the coherent trait we call psychopathy in sets of North American, UK and Continental Europpean data (Cooke et al., 2004).
SEE: latent trait article
Consequence Hypothesis – theoretical support VI: 7 PCL-R items to 2-factors
We (Cooke & Mitchie) 1st examined the structure of these 7 PCL-R items. Using PCL-R ratings from a sample of 1,316 British prisoners, we carried out exploratory factor analysis and extracted 2-factors: 1) interperpreted as Criminal Behavior, was marked by Juvenile Delinquency (Item 18), Revocation of conditional release (Item 19), and Criminal versatility (Item 20); 2) Relationship lability, marked by Promiscuous sexual behavior (Item 11) and Many short-term maritial relationships (Item 17). CFA indicated that a good fit could be achieved (Cooke et al., 2004).
Consequence Hypothesis – theoretical support VII: 2-dimensions v 3-factor model
We (Cooke & Mitchie) examined how the 2-dimensions (SEE: Consequence Hypothesis – theoretical support VI: 7 PCL-R items to 2-factors article) related to the 3-factor model of psychopathy (SEE: Psychopathy Symptoms form Coherent Construct? VIII: 3-factor model article). Are the 2-dimensions additional factors that parallel the interpersonal, affective and behavioral factors identified as Cooke & Michie (2001) 3-factors, or are they better conceptualized as consequences of psychopathy? SEM (structural equation model) methods to examine these questions.
Consequence Hypothesis – theoretical support VIII: SEM 2-advantages
SEM methods has 2-major advantages over traditional correlational methods (Hull, Lehn & Tedlie, 1991). 1) it allows the estimation of the unique effects of one symptom factor controlling for the other symptom factors. 2) estimates and controls for measurement error: Thus, the relative importance of symptom factors can be compared directly, w/out concern that the effects are attenuated by differential measurement reliability.
Consequence Hypothesis – theoretical support X: PCL-R includes AB & socially deviant behavior
We (Cooke & Michie) conclude that the PCL-R construct of psychopathy had drifted from the traditional conceptualization of the disorder (e.g., Cleckley, 1976; Karpman, 1961; McCord & McCord, 1964; Schneider, 1950), and that it may be time to correct this course. It may be time to “reconstruct” psychopathy by reducing or eliminating reliance on criteria that are overly saturated w/AB & socially deviant behavior, thus putting personality back at the heart of this personality disorder.
Different Symptoms Reflect Different Severities of Disorder? I: PCL-R = 3-levels of severity – low, medium, high
In a series of IRT (item response theory) analyses, we (Cooke & Michie) have demonstrated that the PCL-R items reflect different levels of severity of the disorder; some tend to be present a low level of severity, other tend to be present only at high levels of severity (Cooke & Michie, 1997, 1999; Coke et al., 2004; Cooke, Michie, Hart & Hare, 1999). Broadly speaking, the symptom related to the1) Impulsive & Irresponsible Behavioral Style factor are most diagnostic at low levels of psychopathy; 2) moderate levels; and 3) symptoms related to the Arrogant and Deceitful Interpersonal Factor are diagnostic at high levels. The ordering suggests that if patients present w/features such as grandiosity & superficial charm, then they are very likely also to have symptoms such as shallow emotions & poor impulse control; but, many patients w/poor impulse control will not exhibit interpersonal or affective symtoms of psychopathy.
Different Symptoms Reflect Different Severities of Disorder? II: selection of symptoms for different purposes
This finding allows the rational selection of symptoms for different purposes. For example, if the goal is to make a categorical diagnosis of psychopathy, then the emphasis should be on the assessment of symptoms that discriminate around the diagnostic cutoff on the latent trait; but if the goal is to assess the severity of the disorder in dimensional terms, then the emphasis should be on assessing symptoms that reflect the entire range of the latent trait. From the perspective of theory, understanding which symptoms discriminate at different levels of the trait assists in identifying the conceptual core of the disorder.
Equifinality is the principle that in open systems a given end state can be reached by many potential means. The term is due to Ludwig von Bertalanffy, the founder of General Systems Theory. He prefers this term, in contrast to “goal”, in describing complex systems’ similar or convergent behavior. It emphasizes that the same end state may be achieved via many different paths or trajectories. In closed systems, a direct cause-and-effect relationship exists between the initial condition and the final state of the system: When a computer’s ‘on’ switch is pushed, the system powers up. Open systems (such as biological and social systems), however, operate quite differently. The idea of equifinality suggests that similar results may be achieved with different initial conditions and in many different ways.  This phenomenon has also been referred to as isotelesis (Greek: /isos/ “equal”, /telesis/ “the intelligent direction of effort toward the achievement of an end.”) when in games involving superrationality.
Read more: http://www.answers.com/topic/equifinality#ixzz1dvXSB6V0
Exploring Psychopathy Facets w/Modern Psychometrics I: 7-questions it can answer
Modern psychometrics can answer these questions: 1) Do the symptoms of psychopathy form a coherent system? 2) Is AB a primary or secondary symptom of psychopathy?; 3) Which symptoms of psychopathy are most diagnostic? 4) Does the diagnostic efficiency of symptoms of psychopathy vary according to the severity of the disorder? 5) Is the expression of psychopathic symptomatology affected by culture, race, gender and the presence of other mental disorders, such as schizophrenia or mental retardation? 6) Can the precision of measurement of psychopathic symptoms be improved by using different definitions of the disorder? 7) How well can clinicians distinguish the intensity of psychopathic symptoms?
Exploring Psychopathy Facets w/Modern Psychometrics II: disorder structure to disorder facets
Questions 1 & 2 (SEE: Exploring Psychopathy Facets w/Modern Psychometrics I: … article) can be explored using confirmatory factor analysis (CFA) & structural equation modeling (SEM); questions 3-7 can be explored using item response theory (IRT) methods. We start by looking at the overall structure of the disorder. We then consider the individual symptoms of the disorder and how they are best measured. We finish by considering the special challenges that confront us when we try to assess change in facets of this disorder.
Facets of Clinical Psychopathy – Intro I: anthropological, mythological & cinematic
Murphy (1976) in a anthropological study described the Intuit of NW Alaska & the Yoruba tribe of Nigeria that called psychopathy “kunlangeta” & “aranakan” respectively. Greek mythology to contemporary cinema (Medea, Salome, Manon Lescaut), les femmes fatales harmed others by being seductive, manipulative, cruel, egocentric, callous, affectionless & unfaithful. Using their sensuality and sexuality, they controlled and dominated others (Forouzan & Cooke, 2004).
Facets of Clinical Psychopathy – Intro II: 20th-century first-half – cluster of symptoms emerged
The first half of the 20th-century a consensus emerged that an important cluster of symptoms related to aggression, impulsivity & AB (Berrios, 1996). Henderson described the “predominantly aggressive psychopath”; Kahn, the “impulsive,” “weak” & sexual psychopath: Schneider, the “labile,” “explosive” and “wicked’ psychopath (Berrios, 1996).
Facets of Clinical Psychopathy – Intro II: alienists – 19th-century forensic psychiatrist
Clinical description of psychopathy can be traced back to case studies of Pinel & Pritchard (Berrios, 1996), their primary contribution being the understanding that metal disorder can exist even when reasoning is intact. These and other alienists (19th-century term for forensic psychiatrist) gave names such as manie sans délire, monomanie, moral insanity and folie lucide (Millon, 1981). To ensure that their testimony would be relevant, the alienists had to extend their expertise beyond the realm of “total insanity.”
Facets of Clinical Psychopathy – Intro III: 20th-century second-half -> interpersonal, affective, behavioral
The second half of the 20th-century saw a further narrowing of the construct of psychopathy, to a disorder defined by specific interpersonal, affective & behavioral features. Rich clinical descriptions: Arieti (1963), Cleckley (1976), and McCord & McCord (1964) – have provided a framework for describing these individuals. There is broad agreement that interpersonally, psychopathic individuals are dominant, forceful, arrogant and deceptive; affectively they lack appropriate emotional responses being limited & short lived; behaviorally, they are impulsive and lack planfulness. Current work suggests that clinicians experienced with this group of patients have a more sophisticated view of the disorder, adding further domains including problems of self, attachment and cognitive processes (Cooke, 2004).
Facets of Clinical Psychopathy – Intro IV: Cleckley psychopath = PCL-R
The PCL-R operationalized a particular conceptualization of psychopathy, that of Hervey Cleckley: “the ‘Cleckley psychopath’ is the clinical basis for the PCL-PCL-R – and other tests derived from it, PCL: SV; PCL:YV; these include some clinical features not considered by Cleckley and exclude others to which Cleckley did not refer (Blackburn, in press; Forouzan & Cooke, 2004).
Facets of Clinical Psychopathy – Intro IX: chapter purpose
Reviewing the psychometric properties of the PCL-R, identifying important limitations of the test, and clarifying what is know about the construct of psychopathy are the purpose of this chapter.
Facets of Clinical Psychopathy – Intro V: clinical construct: bottom up v top down
Keywords: bottom up approach & domain of interest, bottom up v top down, clinical construct: bottom up v top down, domain of interest & bottom up approach, Handbook of Psychopathy. There are 2-broad approaches to the exploration of the clinical construct & its facets: 1) bottom up 2) top down. In the realm of personality disorder, the bottom up approach (e.g., Clark, 1992; Livesley, Jackson & Schroeder, 1992) entails selecting a large number of traits that provide a broad & systematic representation of the domain of interest.
Facets of Clinical Psychopathy – Intro VI: top down approach pros & cons
The top down approach is perhaps more efficient in that a restricted number of traits are selected for study based on an established conceptualization of a disorder; this has the limitation that any resulting model can only be as good as the original conceptualization. If the original experience is based on experience in only a limited number of subjects or settings, as was the case of Cleckley’s concept, then there is a significant danger of both construct under-representation and the inclusion of “surplus construct” irrelevancies,” that is, traits that are irrelevant to the target construct (Cook & Campbell, 1979; Lillienfeld, 1994; Shadish, 1999).
Facets of Clinical Psychopathy – Intro VII: PCL-R missed key or included irrelevant features
The PCL-R suffers from the fundamental limitations inherent in any psychological test developed using a top down approach to the measurement of the construct: If the conceptualization on which the test is based missed key features or included irrelevant features of the disorder, then it will provide a biased measure of the latent construct (Cook & Campbell, 1979; Pedhazur & Schmelkin, 1991; Shadish, Cook & Campbell, 1999).
Facets of Clinical Psychopathy – Intro VIII: develop new tests beyond PCL-R
Using the PCL-R, psychopathy has been assessed in a standardized manner by numerous investigators, in a wide range of settings, and in both laboratory and applied research contexts (Cooke, Forth & Hare, 1998). This led to a body of research findings that can be contrasted and even cumulated across studies. Yet, after almost 2-decades of use, it is time to stand consider what has been learned about the limitations of the PCL-R. In doing so, we can develop new measures of psychopathy, thereby avoiding mono-operation bias or monomeasure bias (Pedhazur & Schmelkin, 1991; Smith, Fisher & Fisher, 2003).
Is AB a Primary or Secondary Symptom of Psychopathy? I: psychopathy & AB association unclear
The link between psychopathy & AB is formalized in the diagnostic criteria of nosological systems such as the ICD-10 Classifications of Mental & Behavioral Disorders (ICD-10; World Health Organization, 1992) & the 4th-edition of Diagnostic & Statistical Manual of Mental Disorders (DSM-IV; American Psychiatric Association,1994). Yet, the nature of the association between psychopathy & AB is unclear. Is AB a symptom of psychopathy or is it simply a consequence of psychopathy?
ICD = International Classification of Diseases
Is AB a Primary or Secondary Symptom of Psychopathy? II: symptom v consequence
A symptom is generally considered to be direct manifestation of a disease process or disorder, and therefore it plays an important role in assessmet and diagnosis. In contrast, a consequence – also referred to as a secondary symptom or sequela – is indirectly associated w/the disease. A consequence may be a marker or predictor of the disease but is likely to have low sensitivity or specificity. This could be the result of equifinality, in that the same consequences may result from several different diseases or disorders.
Item Response Theory I: indirect measurement of psychological characteristics
Measurement of psychological characteristics is indirect: an individual’s level of a psychological characteristic (e.g., intelligence, depression or psychopathy) cannot be observed directly but has to be inferred from observable behavior, such as responses to test items of verbal accounts of symptoms. In the language of test theory, a person’s standing on the unobservable latent trait (e.g., intelligence) is inferred from manifest variables (e.g., scores on tests of abstract reasoning; Waller, Thompson & Wenk, 2000).
Item Response Theory II: IRT (item response theory) & ICCs (item characteristic curves)
IRT (item response theory) methods provide a comprehensive description of the performance of individual items with an understanding of the performance of tests, or facets of tests, being achieved by a bottom up approach. ICCs (item characteristic curves) provide graphical representation of key item qualities. By focusing at the individual item level we can gain greater understanding of the key features of the disorder and also gain of how to improve the approach to measurement.
SEE: Facets of Clinical Psychopathy – Intro V: clinical construct: bottom up v top down article
latent trait: The “latent trait” view holds that there is some personal attribute or characteristic present in all people which controls their inclination or propensity to commit crime.
Psychopathy Symptoms form Coherent Construct? I: structural disorder research = 1-of-3 advantages
Obtaining greater understanding of the structural properties of the disorder can yield many advantages (Watson, Clark & Harkness, 1994). 1) it can serve as a starting pt for the identification of fundamental psychological structures or processes. Watson & colleagues (1994) argued that structural research on intelligence tests revealed distinct verbal & spatial factors, with subsequent research indicating that these factors measured separate neuro-psychological subsystems.
Psychopathy Symptoms form Coherent Construct? I: structural disorder research = 2-of-3 advantages
2) explication of the structure of a disorder can clarify measurement; failure to group items into unidimensional constructs may result in a lack of clarity in the nomological net linking constructs to other variables. For example, Verona, Patrick & Joiner (2001) clarified Cleckley’s view that psychopaths are relatively immune from suicide by demonstrating that although such behavior was associated with chronic antisocial deviance features of the disorder, it was not associated with affective or interpersonal features of the disorder.
Psychopathy Symptoms form Coherent Construct? I: structural disorder research = 3-of-3 advantages = improved construct design
3) an appreciation of structure can improve scales, by providing direction on where new variables should be added to improve construct (design) representation, or variables should be removed to reduce construct-irrelevant variance (Lilienfeld, 1994). For example it has been argued elsewhere (Cooke & Michie, 2001) that the measurement of psychopathy could potentially be improved by the addition of items to measure features implicated in other descriptions of psychopathy – for example, emotional coldness, incapacity for love, egocentricity, fearlessness & absence of anxiety (Harris, Rice & Quinsey, 1994; Poythress, Edens & Lilienfeld, 1999) – & by the removal of items that are essentially counts of antisocial acts (Blackburn, 1992; Cooke, Michie, Hart & Clark, 2004; see later).
Psychopathy Symptoms form Coherent Construct? II: lack of clarity about psychopathy structure
The very diversity of the characteristics attributed to psychopathy challenges the prima facie notion that the construct of psychopathy is a coherent syndrome: The empirical evidence reflects this. Hare (1980) made an early attempt to evaluate the structure of psychopathy by carrying out a principal component analysis of the 16-criteria for psychopathy described by Cleckley (1976) and extracted 5-components. Raine (1985) used oblique factor analysis and extracted 7-factors.
Psychopathy Symptoms form Coherent Construct? III: PCL-R lack of clarity -> psychopathy structure
With the publication of the original PCL-R manual (Hare, 1991), the lack of clarity about the structure of psychopathy disorder remained. The manual contained, implicitly or explicitly, 3-putative models: a 3-factor model, a two-factor model and a hierarchical model. Hare (1991), following the clinical tradition, identified and specified 3-distinct domains – the interpersonal, affective and behavioral – thought to underpin the expression of psychopathy. He implied a hierarchical structure: “There is heuristic value in viewing psychopathy as a higher-order construct composed of two correlated factors, one reflecting the personality traits widely considered to be descriptive of this syndrome, and the other reflecting socially deviate behaviors. Together, the factors provide a useful description of the syndrome” (Hare, 1991, p. 37).
Psychopathy Symptoms form Coherent Construct? IV: two-factor PCL-R was widely accepted
For years, a two-factor model of the PCL-R (Hare et al., 1990; see also Harpur, Hakstian & Hare, 1988) was widely accepted by investigators in the field. In this model, PCL-R items were considered to be underpinned by two distinct correlated factors: the affective and interpersonal items formed a factor termed “the selfish callous, and remorseless use of others” and remorseless use of others” and the behavioral items formed a factor termed “the chronically unstable and antisocial lifestyle; social deviance” factor (Hare, 1991, p. 76).
Psychopathy Symptoms form Coherent Construct? IX: 1-of-3 points re: 3-model factor
1-of-3 points regarding Cooke & Michie’s 3-factor model: 1) the structure is Hierarchial, w/superordinate construct that was sufficiently unidimensional to be regarded as a coherent psychopathological construct or syndrome (Cooke & Michie, 2001; Zinbarg, Barlow & Brown , 1997).
Psychopathy Symptoms form Coherent Construct? V: Factors I & II not in PCL-R data
The 2-factor model can be questioned on various statistical grounds. For example, the development of the model was over-reliant on congruence coefficients. The coefficients are used to assess the similarity among factors across samples; however, their utility has been called into question (e.g., Van de Vijver & Leung, 1997). The major function of factor analysis is to demonstrate that the variation in scores can be summarized in terms of a few major dimensions, that is, by a simple structure examination of factor loading plots indicated that simple structure based on 2-factors was not present in PCL-R data.
Psychopathy Symptoms form Coherent Construct? VI: 2-factor model is hierarchical
The 2-factor model like the 3-factor model is inherently hierarchical in that the factors can be considered to underpin a higher-order construct (psychopathy); that is, they can be regarded as distinct facets of a superordinate construct in the way that verbal and performance intelligence underpins general intelligence.
Psychopathy Symptoms form Coherent Construct? VII: IRT & CFA on 20 PCLR items
Using both IRT (item response theory) & CFA (confirmatory factor analysis) methods, 13 of the 20 PCL-R items are conceptually distinct and psychometrically nonredundant (Cooke & Michie, 2001). There is no good psychometric evidence that the remaining 7-items measure the construct of psychopathy (Cooke et al., 2004; also see later chapter).
Psychopathy Symptoms form Coherent Construct? VIII: 3-factor model
We (Cooke & Michie) developed a hierarchical structure in which the superordinate trait, psychopathy, was underpinned by 3-highly correlated symptom facets. We called these factors 1) Arrogant & Deceitful Interpersonal Style, 2) Deficient Affective Experience, and 3) Impulsive & Irresponsible Behavioral Style. Factor 1 specified by Glibness/superficial charm, Grandiose sense of self worth, Pathological Lying, and Conning/manipulative; Factor 2 by Lack of Remorse or guilt, Shallow affect, Callous/lack of empathy, and Failure to accept responsibility for own actions;
Factor 3 by Need for stimulation/proneness to boredom, Irresponsibility, Impulsivity, Parasitic lifestyle, and Lack of realistic, long-term goals.
Psychopathy Symptoms form Coherent Construct? X: 2-of-3 points re: 3-model factor
2) the 3-facets can be regarded as having reliable general variance as a consequence of the influence of the broad construct shared with the other facets, but in addition, there was reliable specific variance unique to each particular facet. The value of refining the broad construct into specific facets has advantages in that the specificity between aspects of the disorder and external variables may be clearer (e.g., Dolan & Anderson, 2003; Hall, Benning & Patrick, 2004; Raine, Lencz, Bihrle, LaCasse & Colletti, 2000; Soderstrom et al., 2002).
Psychopathy Symptoms form Coherent Construct? XI: 3-of-3 points re: 3-model factor
3) the model encompasses only 13-of-20 PCL-R items. The excluded items primarily reflect antisocial behavior rather than core traits. Some personality theorists distinguish between basic tendencies and characteristic adaptions, the former being core personality traits and the latter being overt personality manifestations (McCrae & Costa, 1995). Although this distinction is clearly one of degree rather than one of kind, Lilienfeld (1998) has argued that the items associated with Hare’s original Factor I (shallow affect, Grandiose sense-of-self-worth) are basic tendencies whereas those associated with Hare’s original Factor II (Poor behavioral controls, Revocation of conditional release) are characteristic adaptions.
Psychopathy Symptoms form Coherent Construct? XII: PCL-R items 14-20 NOT= coherent syndrome
The failure of the PCL-R items (14-20) to coalesce into a coherent syndrome raises important theoretical questions – questions that can be traced back to the alienists of the 19th century – abut the relationship between antisocial behavior and psychopathy: It is to this issue that we return shortly.
Psychopathy Symptoms form Coherent Construct? XIII: symptoms variance cluster = coherent syndrome:
A recent study by Hill, Neumann & Rogers (2004) highlighted a problem emerging in the field, namely, the failure to distinguish between hierarchical and nonhierarchical models. The primary function of factor analysis in this context is to demonstrate that symptoms cluster to form a coherent syndrome. Such a model postulates a higher-order construct, psychopathy, which is underpinned by distinct but related symptom facets. In other words, the various expressions of the disorder can be construed as a consequence of the general variance “produced” by a broad construct that is shared with the lower-order facets. And, there may be reliable specific variance unique to each particular facet. This type of hierarchical model parallels models of intelligence in which general intelligence underpins specific aspects like verbal performance intelligence.
Psychopathy Symptoms form Coherent Construct? XIV: 2-3 factor models
To claim that psychopathy is a coherent syndrome it is therefore necessary to think hierarchically. Only 2-or-3 factor models are inherently hierarchical insofar as correlated factor models are mathematically equivalent (equivalence of hierarchical & correlated factor model parameters) to models that include a superordinate factor overarching superordinate factors.
Psychopathy Symptoms form Coherent Construct? XV: 4-plus factors model -> less focused
With 4-plus factors, the equivalence of hierarchical & correlated factor models parameters no longer pertains. This is because the number of parameters (a distinguishing characteristic or feature) in the hierarchical model is less than in the correlated factor model. As a result, the correlated model is less parsimonious (i.e., less focused) because fewer constraints/parameters are applied than within the hierarchical model.
OR: Four-plus hierarchical model has fewer parameters than correlated factor model parameters, this equals more constraints on 4-factor hierarchical model than on its correlated factor model, thus the correlated model is less focused.
Psychopathy Symptoms form Coherent Construct? XVI: hierarchical model v correlated factor model
Psychopathy has been observed to be associated w/body building, tattoos, steroid use and body piercing (e.g., Post, 1968), in body builders in adult male prisoners in North America. If this were to be included in the PCL-R, it is plausible that a “Body enhancement” factor would emerge. These are not primary symptoms of psychopathy, given that the frequency of these activities is influenced cultural norms that vary across time. These are characteristic adaptions v basic tendencies. If “Body enhancement” is not a core feature of the disorder, then a hierarchical model would fit poorly, but a correlated factor model would fit the data well.
Understanding Item Performance – measurement & conceptualization I: intro – PCL-R items weighted equally
Which Symptoms are most Diagnostic? The items in the PCL-R are weighted equally (i.e., unit weighted), based on the assumption that all items – the symptoms or features they reflect – are equally diagnostic.. IRT (item response theory) analysis can demonstrate whether this assumption is unwarranted, thereby resulting in loss of measurement precision (Cooke & Michie, 1997; Cooke et al., 1998, 2001, 2005a, 2005b).
Understanding Item Performance – measurement & conceptualization II: intro – ICC slopes & core features
Examination of PCL-R items in a variety of samples has revealed that items vary considerably in their slopes (ICC), and thus in their usefulness as measures of the underlying trait (Cooke & Michie, 1977). PCL-R items such as Promiscuous sexual behavior, Poor Behavioral Controls, and Criminal Versatility have shallow slopes, whereas items such as Callous/lack of empathy, Lack of Remorse/guilt, and Grandiosity have steep slopes. The latter (steep slopes) items are much more useful in the measurement of psychopathy than the former (shallow slopes). Adding poor or irrelevant items to a test can degrade measurement: Less can be more (Smith et al., 2003). This approach assists in identifying the core feature of the disorder (i.e., those that are most diagnostic).