Is there a meaningful subgroup of youths displaying both psychopathic traits and ADHD?
Alexandra Laini Bovellan & Fanny Aronsson
The institution of law, psychology and social work, Örebro University Course: Psychology, advanced level, Master’s Thesis, 30hp, VT21
Abstract
In this study, we examined subgroups of adolescents based on their levels of psychopathic traits and ADHD symptoms. Participants were 982 adolescents from a community sample, with a mean age of 14.28 (SD= .94) years. We used youths’ self-reports of psychopathic traits and their legal guardians’ reports of the adolescent’s ADHD symptoms to identify distinct subgroups of youths. We identified four groups that varied in levels of psychopathic traits and ADHD by using the Hierarchical clustering analysis. One group was characterized by high levels of psychopathic traits and high levels of ADHD (high combination group). The subgroups differed significantly from each other in several theoretically meaningful ways. The high combination group reported higher levels of psychopathic traits, impulsivity and hyperactivity, as well as higher levels on external variables such as aggression, delinquency and violence compared to the other subgroups. The high combination group also differed in terms of anxiety levels from the subgroup with high psychopathic traits only. These findings are in line with previous research and confirms that the construct of psychopathy is
heterogenous. We identified an especially vulnerable subgroup that resembles the characteristics of the secondary psychopath.
Keywords: Psychopathy, ADHD, adolescents, cross-sectional, cluster analysis, psychopathic traits.
Table of contents
Is there a meaningful subgroup of youths displaying both psychopathic traits and ADHD? .... 5
What is psychopathy? ... 6
Factor structure ... 8
Prevalence and associated outcomes... 8
Etiology ... 9
Psychopathic traits amongst children and youths ... 10
How is psychopathy different from other externalizing problems? ... 12
The development of psychopathic traits is heterogenous ... 13
Are there unexplored subgroups? ... 15
The connection between ADHD and Psychopathy ... 15
Secondary psychopathy could be the combination of CU-traits and ADHD ... 17
Aim of the study ... 18
Method ... 18 Participants ... 18 Procedure ... 19 Ethical considerations ... 19 Measurements ... 20 Clustering Variables ... 21 CU-traits ... 21 ADHD symptoms ... 22
External Criterion Variables ... 22
Delinquency ... 22 Aggression ... 23 Bullying others ... 23 Perpetrator of violence ... 24 Anxiety ... 24 Impulsive-irresponsible traits ... 24 Grandiose-manipulative traits ... 25 Statistical analysis ... 25 Results ... 27 Descriptive statistics ... 27 Identified subgroups ... 28 Cluster validation ... 31 Delinquency ... 31
Proactive aggression ... 31 Reactive aggression ... 32 Bullying others ... 32 Perpetrator of violence ... 32 Anxiety ... 33 Grandiose-manipulative traits ... 33 Impulsive-irresponsible traits ... 33 Discussion ... 34 Limitations ... 38 Strengths ... 40 Future implications ... 42 References ... 43
Is there a meaningful subgroup of youths displaying both psychopathic traits and ADHD?
Psychopathy is a personality structure commonly associated with negative outcomes leading to serious consequences for the individual and the society as a whole. Individuals scoring high on psychopathic traits are described as manipulative, callous, impulsive and dishonest. Combined, these traits have shown to increase the risk for developing a lifestyle involving criminality and violence, including antisocial and other deviant behaviours (Hicks, Markon, Patrick, Krueger & Newman, 2004; Sevecke, Kosson & Krischer 2009; Moore, Blair, Hetteman & Robertson-Nay, 2019; Baliousis, Duggan, McCarthy, Huband & Völlm, 2019). Findings indicate that psychopathy can be identified in younger populations (Docherty, Beardslee, Byrd, Yang & Pardini, 2019). Young individuals displaying psychopathic traits have shown to be over-represented in the criminal justice system (Semiz et al., 2008;
Docherty et al., 2019), and display similar deficits in emotional, behavioural and interpersonal dimensions as adult counterparts (Frick, O´Brian, Wootton & McBurnett, 1994). Ample evidence suggests thatwhen these traits are manifested during the adolescent developmental period, it significantly increases the risk for negative outcomes associated with the
psychopathic lifestyle, both concurrently and prospectively (Forth & Book, 2010; Salekin & Frick, 2005). Therefore, the etiology of serious antisocial behaviours and the identification of unique subgroups at particular risk for developing psychopathy in adulthood is crucial
(Salihovic, Kerr & Stattin, 2014). Research has established a link between ADHD in childhood and psychopathic personality in adulthood. For example, findings indicate an elevated prevalence of antisocial behaviours and engagement in criminal activities amongst children and adolescents with ADHD (Dalsgaard, Mortensen, Frydenberg & Thomsen, 2002). It is therefore in line with existing research to examine whether there is a meaningful
subgroup of youths displaying both high levels of psychopathic traits and high levels of ADHD symptoms. Could this potential subgroup be at greater risk for negative development? What is psychopathy?
For decades, researchers have tried to understand why some individuals engage in deviant behaviours. Personality traits that resemble what we call psychopathy, were described as early as the 18th century (Patrick, 2007). Cleckley (1941; 1988) was one of the first to conceptualise psychopathy based on personality traits such as high intelligence, superficial charm, unreliability, lack of emotions and lack of guilt. Today, the growing body of research has established psychopathic traits as dimensional and psychopathy can therefore be
investigated in the general population (Fisher & Brown, 2018). The current and empirically supported definition of psychopathy is that it is a multidimensional construct, composed of maladaptive behaviours, emotions and interpersonal functioning (Cooke & Michie, 2001).
The behavioural dimension is composed of thrill-seeking, parasitizing lifestyle, irresponsibility and impulsivity. The dimension is characterized by observable external behaviours, with impulsivity as a core component. The impulsivity of the individual often drives the person to rash decision-making without the ability to consider different potential consequences of their behaviour. Individuals high on the behavioural dimension have
difficulties retaining a job, using their irresponsibility to live on other people’s resources and lacking personal goals for the future (Cooke & Michie, 2001; Salekin, Brannen, Zalot, Leistico & Neumann, 2006).
The emotional dimension is composed of emotions of shallow affect, lack of remorse, lack of responsibility and emotionlessness. Individuals high on this dimension therefore often hurt others due to their incapacity to feel strong sensations of guilt and regret. This dimension is also involved when engaging in violence and criminality without affecting the individual
emotionally. Their deficits regarding taking other individuals’ perspective, inhibits them to mentalize and feel the emotions of others (Cooke & Michie, 2001).
The interpersonal dimension is composed of behaviours towards others and include traits such as pathological lying, superficial charm, manipulation and grandiosity. Individuals high on the interpersonal dimension are often described as charming however are driven by manipulation and lack of genuineness. Individuals high on this dimension consider
themselves to be better than others and are often enhancing this towards other people by stating their superiority (Cooke & Michie, 2001).
From a clinical framework, psychopathy is considered to be a personality disorder (PD). PDs refer to inflexible and maladaptive ways of thinking, acting and feeling, usually manifested during childhood and adolescence. These patterns are considered stable over time and can lead to significant functional impairments in many areas in life for the individual, but also for other people surrounding the individual (American Psychiatric Association, 2013). Similar to other PDs, psychopathy is also assumed to be stable over time and resistant towards treatment, including rehabilitation efforts (Lynam et al., 2009). However, psychopathy is not found in common psychiatric diagnostic manuals such as the DSM-IV or ICD-10 (American Psychiatric Association, 2013; World Health Organization, 1993). Instead, the term Antisocial personality disorder (ASPD) as specified in DSM-IV, is often associated with psychopathic personality traits. ASPD shares similar features to psychopathy, although studies have criticised the diagnostic criteria of ASPD for focusing mainly on the behavioural dimension (Hare, 1991; 2003). For example, psychopathy is considered to easily meet the DSM-IV criteria for ASPD because of the dominating behavioural features, while the reverse is not the case (Hare, 2003).
Factor structure
The existing research findings implicate psychopathy as a multidimensional
personality disorder consisting of interpersonal, emotional and behavioural dimensions based on Cleckley’s (1941;1988) conceptualisation of psychopathy (McCord & McCord, 1964; Harris & Rice, 2006). However, there is an ongoing debate regarding the heavy focus on criminality as a part of the behavioural dimension. Critics claim that criminality should be seen as a secondary behavioural consequence of the psychopathic personality, rather than a requirement to fulfill the criteria for psychopathy (Andershed, Kerr, Stattin & Levander, 2002; Cooke & Michie, 2001; Van Baardewijk et al., 2008).
Therefore, Cooke & Michie, (2001) proposed an updated model with an enhanced focus on the emotional and interpersonal dimensions. The model is underpinning psychopathy as a personality style comprising the factors Arrogant and Deceitful Interpersonal Style (manipulating, grandiosity, lying, shallow charm), Deficient Affective Experience
(callousness, lack of guilt, shallow emotions), and Impulsive and Irresponsible Behavioural Style (impulsiveness, thrill seeking). Based on the previous mentioned model, Andershed et al., (2002) developed the self-report tool Youth Psychopathic Traits Inventory (YPI) to capture young individuals with psychopathic traits. The construct of psychopathy, as
described in YPI, consists of the dimensions Grandiose–Manipulative (grandiosity, dishonest charm, lying, manipulation), Callous–Unemotional (remorselessness, unemotionality,
callousness), and Impulsive–Irresponsible (thrill seeking, impulsiveness, irresponsibility) (Andershed et al., 2002; Van Baardewijk et al., 2008; Borroni, Somma, Andershed, Maffei & Fossati, 2014; Colins et al., 2017).
Prevalence and associated outcomes
The prevalence of individuals diagnosed with psychopathy is hard to estimate due to the ongoing debate regarding diagnostic features and the complexity of the disorder.
However, around 1 - 3 % are estimated to fulfill the criteria for psychopathy in the general population (Hare et al., 1990; Hare, 1991; 2003).
Psychopathy may lead to serious consequences for the individual, other people and the larger society. Many studies have found that psychopathic personality predicts future
criminality, aggression towards intimate partners and substance use (Kennealy, Skeem, Walters & Camp, 2010; Czar, Dahlen, Bullock & Nicholson, 2011; Coid, Yang, Ullrich, Roberts & Hare, 2009). Overall, criminality is a large societal issue for many reasons. Most importantly, it causes great harm and pain to the people involved and affected by the criminal actions. It also leads to economic consequences, as well as being a threat towards the justice system, right to legal security and principles of democracy that a society is built upon (Häkkänen-Nyholm & Nyholm, 2012; Savage, 2009). For example, in Sweden, a severe assault leading to long term injuries for the victim is costing the society approximately 50 million SEK (Nilsson & Wadeskog, 2012). An extensive line of research is pressing that a society can save large resources by focusing on prevention in an attempt to avoid young individuals ending up in a criminal lifestyle (Brottsförebyggande rådet, 2017).
Etiology
The explanation for why some individuals develop psychopathy whilst others do not has been an ongoing debate amongst researchers for many years. The growing body of research has established that the development of psychopathy is a complex phenomenon involving several factors (Hare, 1991; 2003; Patrick, 2007). Theoretically, it has been suggested that the development of psychopathy can originate from innate biological dispositions as well as environmentally influenced factors (e.g., Cleckley, 1988; Lykken, 1995; Karpman, 1941).
Research argues that there is a neurological component involved related to certain brain regions, linking psychopathy to several neurological deficits (Fisher & Brown 2018;
Anderson & Kiehl, 2014). For example, individuals displaying ASPD and psychopathy may differ in their etiology based on their reactivity to threat, which seems to be connected to dysfunctions in the amygdala (Hyde, Byrd, Votruba-Drzal, Hariri & Manuck, 2014).
Furthermore, some existing research argues that psychopathy is highly inheritable (e.g., Cleckley, 1988; Lykken, 1995; Karpman, 1941). Studies have especially focused on the biological etiology of psychopathic personality based on twin-adoption studies (Beaver, Vaughn & DeLisi, 2013; Forsman, Lichtenstein, Andershed & Larsson, 2010; Larsson, Andershed, & Lichtenstein, 2006; Viding & Larsson, 2010; Waldman & Rhee, 2006). One twin study found a consistent heritability of psychopathic traits, where 58 % of the
development and stability was explained by genetic influences (Blonigen, Hicks, Krueger, Patrick & Iacono, 2006).
Finally, environmental and social factors also play an important role in the developmental path of psychopathy (Farrington, Ullrich & Salekin, 2010; Larsson et al., 2006). For example, the lack of parental warmth has shown to have a direct impact on psychopathy. Youths who are exposed to parental warmth, strengthening the relationship between the child and parent have shown to decrease psychopathic traits prospectively (Silva & Stattin, 2016; Fisher & Brown, 2018). Youths who instead perceived their parents as warm and understanding decreased in levels of psychopathic traits over time (Salihovic, Kerr, Özdemir, & Pakalniskiene, 2012). In sum, the current literature therefore suggests that the etiology of psychopathy is based on heritable genetic factors. However, psychopathic traits can be influenced by environmental and social factors which contributes to the development of psychopathy in adulthood.
Psychopathic traits amongst children and youths
The presence of psychopathic features has been proven to be manifested amongst children and young adults, which increases the vulnerability to develop psychopathy and
antisocial behaviours later in life (Moore et al., 2019; Viding, Blair, Moffitt, & Plonin, 2005). Furthermore, longitudinal research has found consistent indications that the roots of
psychopathy stems from childhood. Frick et al., (1994) measured psychopathy amongst youths and found the same dimensions as for adult counterparts. In their study, children with elevated psychopathic traits scored high on behavioural dimensions including poor impulse control and delinquency, the emotional dimension including lack of guilt, lack of empathy and superficial charm, as well as on the interpersonal dimension including grandiosity and
manipulative behaviours. This indicates that youths with psychopathic traits show similar deficits as adults (Lynam et al., 2009). Many studies of youth psychopathy have reported more serious forms of antisocial behaviours amongst adolescents with psychopathic traits. Furthermore, the traits seem to predict elevated rates of violence and aggression amongst these young individuals (Silva & Stattin, 2016).
However, the general understanding of the psychopathic personality is that only adults are diagnosed with the disorder. American Psychiatric Association (2013) advises against diagnosing individuals under the age of 18 with PD’s. Psychopathic-like traits could be seen as a part of the normal developmental period. For example, adolescents are more likely to act in an impulsive manner and are often emotionally unstable, which might explain why they score high on measures for psychopathy (Cleckley, 1941; 1988, Seagrave & Grisso, 2002). In contrast, previous studies provide ample evidence that the caution might be too strong. A wealth of studies has investigating psychopathic traits in childhood and adolescence to date (for reviews see, Salekin & Frick, 2005). Lynam et al., (2009) are claiming that concerns about describing youths as psychopathic or labelling them with other PD’s is due to stigmatising factors, labelling effects and theoretical expectations, rather than based on empirical data.
The combined evidence suggest that psychopathic-like traits are identifiable in a younger population (Kotler & McMahon, 2005; Stellwagen & Kerig, 2010), and that the earlier an individual develops emotional and behavioural difficulties, the more likely it is to persist later in life (Docherty et al., 2019). Studies examining the development of
psychopathic traits amongst children and adolescents have found stability in the personality traits across adolescence (Lynam et al., 2009; Salihovic, 2013). Additional studies have found that psychopathic traits amongst adolescents predict antisocial behaviours above and beyond other risk factors, such as conduct problems, aggression, impulsivity and earlier convictions (Lynam et al., 2009; Frick, Cornell, Barry, Bodin & Dane, 2003). Due to the serious
consequence’s psychopathy can lead to, ample evidence shows the need to target young individuals at risk to prevent adult psychopathy and other antisocial behaviours to develop (Salihovic et al., 2014). Therefore, it is important to extend the construct of psychopathy to youths.
How is psychopathy different from other externalizing problems?
The behavioural and impulsive components of psychopathy tend to overlap with other externalising disorders, such as ASPD, CD and ADHD. The overlap can lead to difficulties distinguishing one diagnosis from another, affecting the reliability and validity of the
measures. Therefore, researchers extending adult psychopathy to youths are heavily focusing on the interpersonal and affective dimensions in an effort to understand the development of the disorder (Docherty et al., 2019).
The callous-unemotional (CU) traits, also referred to as the core affective dimension of the psychopathic personality, are composed of remorselessness, unemotionality and
callousness (Smith & Hung, 2012). This emotional dimension distinguishes psychopathy from other externalising disorders based on the lack of affective experiences and emotional
to punishment, especially cues for stress and fear triggered in individuals without the same set of characteristics (Moore et al., 2019). Children with high levels of CU-traits in combination with antisocial behaviour seems to be most vulnerable for developing a more stable pattern of ASPD and psychopathy (Moore et al., 2019; Viding & McCrory, 2012). Another study found that individuals demonstrating high levels of CU-traits in combination with high levels of the interpersonal dimension, have an increased risk for engaging in criminality. This seems to be the case even in the absence of conduct problems (Docherty et al., 2019). Longitudinal studies have shown that CU-traits are more stable during the developmental period, in contrast to what can be said about the whole construct of psychopathy. High CU-traits for children increases the risk for adult psychopathy, while behavioural and interpersonal problems do not (Obradović, Pardini, Long & Loeber, 2007). Therefore, research concludes that CU-traits is an important core component of the psychopathic personality associated with severe
consequences. Additionally, CU-traits deviate from other externalising disorders that may appear similar due to the observable external outcomes.
The development of psychopathic traits is heterogenous
It is well known that the psychopathic personality is dimensional and heterogeneous (Karpman, 1941; Cleckley, 1941; 1988; Lykken, 1995; Skeem, Poythress, Edens & Lilinfeld & Kale, 2003; Hicks et al., 2004). To fulfill the criteria for psychopathy, an individual needs to score high on all three dimensions; interpersonal, emotional and behavioural, separately (Cooke & Michie, 2001). However, individuals could demonstrate a variation of psychopathic traits without meeting the requirements for the psychopathic personality disorder (Andershed & Skeem, 2004). This also means that the scoring on the three different dimensions can differ amongst individuals but still lead up to the same diagnosis. Researchers have focused on investigating the heterogeneity of the psychopathic personality in an attempt to try to
understand these differences leading up to different types of psychopathy and related outcomes (Salihovic et al., 2014; Karpman, 1941).
Theoretically, Karpman (1941), was one of the first to claim that psychopathic personality may be the result of two different trajectories, described as two unique subtypes; the primary and secondary subtypes of psychopathy. The primary subtype is representing individuals where the deviant behaviour is due to biological, cognitive and emotional deficits, while the secondary subtype is the result of an environmentally emotional disturbance due to upbringing. In other words, both subtypes display similarities in terms of elevated antisocial behaviours, but different etiological trajectories (Hicks et al., 2004). Studies have focused on the factors distinguishing the two types, but less is known about what makes this discrepancy possible within the spectrum of psychopathy (Salihovic et al., 2014).
Researchers have also tried to understand these potential subgroups when studying adolescents. One study examined psychopathic like traits amongst adolescents and found that there were subgroups of adolescents with high levels of psychopathic traits similar to the existing literature on adult psychopathy. One of the identified groups showed features comparable to the primary subtype while the other group of individuals showed a more impulsive, aggressive and anxious profile, one that could be compared to the secondary subtype. These findings have raised the question whether this developmental difference could be linked to levels of anxiety (Salihovic et al., 2014).
In sum, the empirical results highlight the importance of applying existing research in the heterogeneous development of psychopathic traits in a younger population (Frick et al., 1994). It is also important to consider that the majority of young individuals who display psychopathic traits during adolescence will not develop serious maladaptive behaviours and emotional deficits associated with psychopathy as an adult. However, psychopathy is a very serious disorder and research show the significance of targeting the individuals in elevated
risk before adult psychopathy can develop (Salihovic et al., 2014). Therefore, it is of high importance to investigate different subtypes of young individuals displaying psychopathic traits. Some groups might be especially vulnerable to develop a more severe pattern of antisocial behaviours and emotional deficits. Considering the variation of the etiology and heterogeneity of the disorder, preventive interventions might differ significantly.
Are there unexplored subgroups?
Based on the above-mentioned research, there is empirical evidence that psychopathy is heterogeneous involving relevant subgroups. According to Karpman's (1941) theory regarding primary and secondary psychopaths, the secondary psychopath is hypothesized to have an overly active behavioural activation system, consequently leading to impulsive behaviours. In combination with the limited affective capacity, the person might experience negative emotions in regards to not being able to control impulsive and potentially dangerous actions (Hicks et al., 2004). When reviewing existing research regarding the secondary psychopath, many features such as lack of impulse control and high levels of behavioural activation are similar to the criteria of the diagnosis ADHD, indicating that the behavioural dimension of the secondary psychopath and ADHD is overlapping. This raises the question whether the impulsivity, hyperactivity and ability to feel anxious amongst secondary psychopaths is better explained by a combination of the psychopathic personality and ADHD.
The connection between ADHD and Psychopathy
ADHD is one of the most common child psychiatric disorders with a prevalence around 3 % - 9 % in the general population (Scahill & Schwab-Stone, 2000; Sevecke et al., 2009; American Psychiatric Association, 2013; Willcutt, 2012). ADHD is characterized by impulsiveness, attention deficits, hyperactivity and difficulties in several environmental settings, for example in school, social interactions with peers, parents and teachers (Storebø &
Simonsen, 2016).Due to the nature of the neurological deficits related to the diagnosis, it is often considered to be a lifelong condition. When fulfilling the criteria for ADHD in
adulthood, studies have shown that these individuals have a higher risk of poorer long-term outcomes in terms of academic performance, occupational rank, risky sexual behaviour, relationships and matrimonial problems, as well as higher rates of criminal behaviours and psychiatric comorbidities (Biedeman et al., 1996). ADHD has also shown to have high rates of comorbid disorders. Around 70% of children with ADHD have at least one other comorbid disorder (Barkley, 2006; Cherkasova, Sulla, Dalena, Pondé & Hechtman, 2013; Dalsgaard, Nielsen, & Simonsen, 2013). According to the literature, conduct problems, criminality and other antisocial behaviours are more prevalent in children with ADHD compared to children without the diagnosis (Dalsgaard et al., 2002).
Research findings support that there could be a connection between the diagnosis of ADHD and psychopathic personality (Dalsgaard et al., 2002). Studies have shown that both ADHD and conduct problems in children are risk factors for the development of psychopathy in adulthood, especially since ADHD is a common psychiatric disorder in the general
population (Sevecke et al., 2009; Smith & Hung, 2012). Researchers have mainly focused on the co-morbidity of ADHD and CD when studying children and adolescents and potential existing subgroups (Faraone, Biederman, Mennin, Russell & Tsuang, 1998; Smith & Hung, 2012; Storebø & Simonsen, 2016; Mordre, Groholt, Kjelsberg, Sandstad & Myhre, 2011; Satterfield et al., 2007; Satterfield & Schell, 1997). The combination of the two diagnoses seem to display an especially problematic subgroup. This group of individuals are at elevated risk to develop more severe externalising disorders such as ASPD and psychopathy, compared to children with solely ADHD or CD (Sevecke et al., 2009; Rösler & Retz, 2008). Therefore, it is in line with existing research to examine if there is a potential subgroup displaying psychopathic traits and ADHD.
Secondary psychopathy could be the combination of CU-traits and ADHD
It is interesting to think about ADHD in regards to the two subtypes of psychopathy explained by Karpman (1941), where the secondary subtype could be a result of the
combination of psychopathic traits and ADHD. As previously mentioned, the primary psychopath is explained as a subgroup of individuals representing deviant behaviours due to biological, cognitive and emotional deficits. The antisocial behaviour of the primary
psychopath is often due to their lack of conscience which also thrives on a fearless
temperament and weak behavioural inhibition system. The antisocial behaviour of the primary psychopath is unlikely to be accompanied by emotional arousal since they have a neurological emotional deficit (Karpman, 1941; Fowles, 1980). The secondary psychopath on the other hand is the result of an environmentally emotional disturbance due to upbringing. The secondary psychopath’s antisocial behaviour is seen as an expression of an interpersonal neurotic conflict. The secondary psychopathy is capable of feeling emotional arousal, such as anxiety. The antisocial behaviour is often interpreted as a part of their overactive behavioural activation system, resulting in impulsivity (Hicks et al., 2004).
Both subtypes of psychopaths are likely to engage in antisocial behaviours, but are driven by different neurological underpinnings and motivation systems. The secondary psychopath, who has a normal threshold for emotional arousal, may actually experience greater emotional distress than primary psychopaths because of the consequences of his or her impulsive behavioural style (Lykken, 1995). When separating the two subtypes of primary and secondary psychopaths, the secondary psychopath seems to share several deficits with the diagnosis ADHD, such as hyperactivity, impulsivity and anxiety (Storebø & Simonsen, 2016). Studies have found that 70 % of individuals with ADHD experience anxiety as a comorbid disorder (Barkley, 2006; Cherkasova et al., 2013). No studies have examined if there could be a unique subgroup of youths displaying high levels of psychopathic traits and
ADHD, resulting in a personality profile similar to the secondary psychopath according Karpman's (1941) theory. Could there be other factors, such as a co-occurring ADHD,
accounting for the discrepancies between the primary and secondary subtype of psychopathy? Aim of the study
The aim of this paper is to investigate if there are relevant subgroups of youths displaying high levels of CU-traits and high levels of ADHD symptoms (high combination group). We expect to find groups with varying levels of CU-traits and ADHD symptoms given the normative nature of the sample. We hypothesize that the subgroups will differ significantly from each other on variables of interest. Specifically, we are expecting that the individuals in the high combination group will be more delinquent, aggressive and impulsive in comparison to the other subgroups. We also expect that the high combination group will show elevated levels of core psychopathic traits by scoring higher on the
grandiose-manipulative and impulsive-irresponsible dimensions in comparison to the other subgroups. Method
Participants
We used data from a longitudinal community-based project in a mid-sized town in Sweden with a population of 26 000 inhabitants. The aim of the project, called “10-18”, was to collect information about the developmental patterns of adolescents' individual
characteristics regarding adjustment problems and delinquency. The project was conducted through Örebro University led by professors Håkan Stattin and Margaret Kerr. The study, including measurements and procedures used, was approved by the Örebro University ethics board. The average income in the town at the time was somewhat lower compared to the average Swedish income. There were no differences in terms of unemployment rates compared to the rest of the country (4 %). Twelve percent of the inhabitants had a foreign background, similar to the 14 % average in other communities in Sweden. The target sample
for this study was based on the third wave of the data collection. The participants consisted of adolescents in all 7th through 9th grades in the community (N = 982), making the youths between 13 to 15 years of age (M = 14.28, SD = .94). Of all the participating adolescents, 52 % were boys and 48 % were girls.
Procedure
To fulfill the aim of the project and to collect information about developmental patterns of adolescents' individual characteristics, all youths between 10 to 18 years and their parents in the town was targeted as potential participants. The youths were recruited in classroom settings during school hours, while the parents received information by postage. The information included details about the project, such as duration and topics, as well as requirement procedures. Since the target sample consisted of minors, a passive consent from the guardians was necessary for participation. A passive consent infers that the parents that did not actively decline participation of the study was included, but the adolescents were required to actively confirm their participation despite their parents’ approval. Approximately 1 % declined to participate in the study. The adolescents and guardians were informed that the participation was voluntary and that they could withdraw their consent at any time throughout the study. They were also guaranteed that their answers would be anonymous and treated with confidentiality. No payment was offered, although movie tickets were drawn in the classes.
The students were asked to fill out self-reports each year over a 5-year period (2001 - 2006). This occurred during school hours and was administered by trained research assistants. Guardians were included in the study by filling out questionnaires about their adolescent every second year and returning them in pre-paid postage envelopes. Around 70 % of the guardians participated during each data collection.
The study was approved by the Örebro University ethics board. To ensure ethical principles of the study, requirements according to HSFR (Etikregler för Humanistisk-Samhällsvetenskaplig Forskning) were followed and met. All participants were informed about the purpose of the study, the possibility not to participate and to withdraw their
participation at any time. The requirements of confidentiality were met since no unauthorized individuals had access to personal data. When accessing the collected data, a signature regarding professional secrecy was required. The requirement of consent and access was met since it was a requirement for participation. The collected data was used for research only.
The application of ethical approval was based on a passive consent from the youths’ parents and an active consent from the adolescents. This method was chosen because of the target sample of the study. The project aimed to learn more about youths' developmental patterns, with a certain interest in deviant behaviours. Therefore, the passive parental approval was considered appropriate to eliminate attrition amongst these individuals. Even though the youths' guardians had accepted the participation of their youth in the study, the adolescents were informed about their own choice to participate. Adolescents who did not want to participate in the study could spend the time with their regular schoolwork. Post addresses to the guardians involved were collected from the different schools to assure that all parents were informed about the study and the possibility of not participating. To ensure the integrity of the participants, strict procedures and rules were applied when collecting the data. The participants were unidentifiable individually since the data was reported at a group level involving all high school adolescents in the region.
Clustering Variables
To explore the potential subgroup displaying high levels of CU-traits and high levels of ADHD, we used the adolescents' self-reports of CU-traits and guardian-reported measures for ADHD symptoms as clustering variables in our study.
CU-traits
To measure the callous-unemotional dimension of psychopathy, including the adolescents' lack of affective experiences and emotional understanding, we used the Youth Psychopathic Traits Inventory (YPI: Andershed et al., 2002). The YPI is a self-report tool, designed to capture psychopathic traits in the normal population amongst youths from the age of 12. The whole instrument consists of 64 items and is based on the three-factor model comprising the grandiose-manipulative, callous-unemotional and impulsive-irresponsible dimensions of psychopathy. The measurements are divided into 10 subscales, representing the three different dimensions (Andershed et al., 2002; Van Baardewijk et al., 2008; Borroni et al., 2014; Colins et al., 2017). The reliability and validity have been reported in many studies (e.g., Andershed et al., 2002; Declercq, Markey, Vandist & Verhaeghe, 2009; Poythress, Dembo, Wareham & Greenbaum, 2006; Skeem & Cauffman, 2003). One study examined the construct validity of YPI in comparison to other measurements of psychopathy amongst adults (Psychopathic Personality Inventory Revised: PPI-R). The findings provided supporting evidence for the construct validity, as the YPI was highly correlated with the construct of psychopathy as measured by PPI-R (Campbell, Doucette & French, 2009). Cronbach’s alpha for the total YPI scale indicated excellent reliability (a = .92).
The CU-trait dimension is based on 21 items, with a rating scale from 1 (does not apply at all) to 4 (applies very well). The dimension is divided into three subscales with varying questions: remorselessness (“When someone finds out about something I’ve done wrong, I feel more angry than guilty”), unemotionality (“To be nervous or worried is a sign of
weakness”) and callousness (“I don’t care very much about that others can be hurt by things that I do”). Cronbach’s alpha showed good reliability (a = .87).
ADHD symptoms
A revision of Swanson, Nolan & Pelham Rating scale (SNAP-IV) was used to
measure ADHD symptoms. It is a parent-reported measure which has shown good reliability (Bussing et al., 2008). In this study, the guardians were asked to report the behaviours that best described their youth. The instrument consists of 17 items, with a rating scale from 1 (does not apply at all) to 4 (applies very well). The items are based on the DSM-IV criteria for ADHD, divided into the subscale’s hyperactivity (“Doesn’t like doing things that demands calmness and stillness”), impulsivity (“Often has difficulties with waiting for his or her turn”) and attention problems (“Very often has difficulties with organising tasks and activities”). The Cronbach’s alpha showed good reliability for the subscale hyperactivity/impulsivity (a = .87) and excellent reliability for the subscale attention problems (a = .91).
External Criterion Variables
In an attempt to support the subgroups in our sample, we compared the different groups on variables that are theoretically relevant in regard to psychopathy and ADHD (Bailey & Shelton, 2014; Hicks et al., 2004; Lynam et al., 2009). These variables were not included in the making of the cluster analysis.
Delinquency
Adolescent delinquency was measured through self-reports based on Magnusson, Dunér & Zetterblom (1975) longitudinal study on adjustment. The measure has been updated (Kerr & Stattin, 2000), and is similar to other commonly used self-reports that have shown good reliability (e.g., Haynie, 2001). The measurement consists of 22 items with a rating scale from 1 (no, it has not happened) to 5 (more than 10 times) based on the adolescents' behaviour during the last year. The delinquency measure is divided into the subscales violence (“Have
you taken part in a street fight in town?”), serious property offences (“Have you participated in breaking into a home, shop, stand, storage building or other buildings, with the intention of taking things?”), minor property offences (“Have you taken things from a store, stand or shop without paying?”) and substance use (“Have you smoked hashish, eg. marijuana/cannabis?”). Cronbach’s alpha showed excellent reliability for the total scale (a = .93). Cronbach's alpha for violence (a = .80), serious property offences (a = .87) and substance use (a = .81) showed good reliability. The subscale minor property offences showed excellent reliability (a = .91).
Aggression
The measure of adolescent aggression was created for the project, based on other similar measurements of aggression (eg. Raine et al., 2006). The self-reported instrument consists of 12 items with a rating scale from 1 (does not apply at all) to 4 (applies exactly). The measure includes two subscales: proactive aggression (“I pick a fight with people I don't like, even though they haven't done anything to me”) and reactive aggression (“If I am really angry, I usually throw, hit or kick things”). Cronbach's alpha showed good reliability for both subscales: proactive aggression (a = .82) and reactive aggression (a = .86) respectively.
Bullying others
To measure to what extent adolescents are bullying others, a measurement based on three statements was created for the project. The self-reported instrument was created based on an existing validated instrument (Alasker & Brunner, 1999). The three statements included ostracism, demeaning/spiteful comments or physical abuse with a rating scale between 1 (no, it has not happened this semester) to 4 (yes, it has happened several times a week). The items consisted of statements such as “Have you mocked, or said nasty things to anybody in school or on the way to or from school?, “Have you hit, kicked or attacked anybody in school or on the way to or from school?” and “Have you been part of ostracizing anybody?”. Cronbach's alpha showed good reliability (a = .82).
Perpetrator of violence
To measure to what extent adolescents were engaged in violence and threats towards others as a perpetrator, a measurement based on four statements was created for the project. The self-reported instrument was created based on an existing validated instrument
(Andershed et al., 2002). The four statements included ostracism, demeaning/spiteful
comments or physical abuse with a rating scale between 1 (no, it has not happened this spring semester) to 4 (Yes, it has happened 4 or more times). The items consisted of statements such as “Have you taken part in attacking others at night without them having threatened or
attacked you or your friends’ first?” and “Have you taken part in threatening or forcing anyone to give you money, a cell phone, cigarettes or anything else? ”. Cronbach's alpha showed acceptable reliability (a = .70).
Anxiety
A parent-reported measure was created for the project to measure the adolescent’s anxiety levels. In this study, the guardians were asked to answer questions about anxiety that best described their youth. The parent-reported instrument consists of six items with a rating scale from 1 (does not apply at all) to 4 (applies exactly). The measure included statements such as “Is often more nervous than others in his/her age”, “Worries a lot unnecessarily” and “Often has problems with falling asleep because of his/her thoughts about difficult things”. Cronbach's alpha showed good reliability (a = .82).
Impulsive-irresponsible traits
To measure the impulsive-irresponsible dimension of psychopathy, including the adolescents' levels of thrill seeking, impulsivity and irresponsibility, the Youth Psychopathic Traits Inventory (YPI: Andershed et al., 2002) was used. The reliability and validity have been reported in many studies (e.g., Andershed et al., 2002; Declercq et al., 2009; Poythress et al., 2006; Skeem & Cauffman, 2003). The dimension is measured based on 15 items, with a
rating scale from 1 (does not apply at all) to 4 (applies very well). The dimension is divided into three subscales with varying questions: thrill seeking (“I get bored quickly when there is too little change”), impulsivity (“It often happens that I talk first and think later”) and
irresponsibility (“It has happened several times that I've borrowed something and then lost it”). Cronbach’s alpha showed excellent reliability (a = .90).
Grandiose-manipulative traits
To measure the grandiose-manipulative dimension of psychopathy, including the adolescents' levels of dishonest charm, grandiosity, pathological lying and manipulation, the Youth Psychopathic Traits Inventory (YPI: Andershed et al., 2002) was used. The reliability and validity have been reported in many studies (e.g., Andershed et al., 2002; Declercq et al., 2009; Poythress et al., 2006; Skeem & Cauffman, 2003). The dimension is measured based on 28 items, with a rating scale from 1 (does not apply at all) to 4 (applies very well). The
dimension is divided into four subscales with varying questions: dishonest charm (“I have the ability to con people by using my charm and smile”), grandiosity (“I'm more important and valuable than other people”) pathological lying (“Sometimes I find myself lying without any particular reason”) and manipulation (“To get people to do what I want, I often find that it works well to con them “). Cronbach’s alpha showed good reliability (a = .86).
Statistical analysis
Descriptive analysis was conducted to provide information about the mean levels and standard deviations of criterion and external variables for the total sample. To explore linear associations between the criterion and external variables, Pearson’s two-tailed correlation test was used. To identify potential subgroups of youths based on their self-reported levels of CU-traits and parent-reported ADHD symptoms, we used Hierarchical clustering analysis
according to SPSS 22. This cluster analysis is a data driven and exploratory method that seeks to identify subgroups of individuals within a heterogeneous sample. The model is designed to
define subgroups by categorising individuals into different groups and providing various indications of the “best-fit” model. Individuals belonging to a single subgroup are then considered to be homogenous. When using a data driven method, four criteria have been developed to assist in ensuring validity and robustness of the analysis. First, the number of clusters must be evaluated based on theory and previous research. Second, the clusters should be homogenous with an error sum of squares (ESS) preferably over 67 %, but at least 50 %. Third, the cluster coefficient should not be less than 10 %. Fourth, each subgroup of clusters should contain at least 10 participants. Furthermore, it is recommended to validate the clusters based on external variables to investigate if the subgroups are meaningful or not. The
assumption of meaningful subgroups is met when the identified groups differ significantly on the external variables of interest (Bergman, Magnusson & El-Khouri, 2003).
Before conducting the analysis, we standardized all variables. The analysis process was conducted in two steps. The first step involved completing two different hierarchical cluster analyses to come up with the most theoretically relevant fit, which is also
recommended according to previous research (Field, 2014). The first analysis was conducted with Ward's method. A second analysis, Quick Cluster Analysis, was used to confirm the first analysis conducted with Ward's method. The agglomerations-estimate found four groups based on our cluster variables. We also investigated a five-group solution that contributed to a less theoretically meaningful result, and therefore the four-group solution was used.
In the second step, we used One-way Analysis of Variance (ANOVA) to validate the clusters and investigate differences between the subgroups on our relevant external variables (delinquency, aggression, bullying others, perpetrator, anxiety, impulsive-irresponsible and grandiose-manipulative). ANOVA provides significant group differences based on the mean scores of the external variables. To further investigate between which groups the significant differences occur, Student-Newman Keul (SNK) post hoc test was used. The SNK post hoc
test provides more detailed information about group differences between the identified subgroups. For variables where the homogeneity of variance assumption was not met (hyperactivity/impulsivity, inattention, impulsive-irresponsible dimension, proactive
aggression, reactive aggression, perpetrator of violence, bullying, delinquency and grandiose-manipulate dimension), the Welch test was used to adjust the degrees of freedom.
Results Descriptive statistics
The results from the correlation analysis are presented in Table 1. The variables correlated as expected. The means and standard deviations for all groups on the external criterion variables is presented in Table 2.
Table 1
Correlation between clustering variables and external validation variables.
Variables CU-traits Hyperactivity/impulsivity Inattenation
Delinquency (total score) .33** .18** .23**
Proactive agression .41** .18** .21** Reactive agression .36** .17** .19** Bullying others .30** .15** .17** Perpetrator of violence .26** .13** .20** Anxiety -.05 .45** .49** Impulsive-irresponsible .23** .26** .35** Grandiose-manipulative .78** .21** .25** ** p<.001. Table 2 Descriptive statistics. Variables M SD Inattention 1.88 .59 Hyperactivity/impulsivity 1.76 .54 CU-traits 1.87 .55
Delinquency (total score) 1.16 .34
Proactive agression 1.50 .54 Reactive agression 1.97 .45 Bullying others 1.18 .45 Perpetrator of violence 1.08 .32 Anxiety 1.80 .58 Impulsive-irresponsible 2.20 .51 Grandiose-manipulative 1.85 .58
Note: the means and standard deviation are based on raw scores.
CU-traits were positively and significantly associated with delinquency, proactive aggression, reactive aggression, bullying others, perpetrator of violence and the two other dimensions of psychopathy (impulsive-irresponsible and grandiose-manipulative). CU-traits were negatively associated with anxiety. Hyperactivity/Impulsivity was positively and significantly associated with delinquency, proactive aggression, reactive aggression, bullying others, perpetrator of violence, anxiety and the two other dimensions of psychopathy
(impulsive-irresponsible and grandiose-manipulative). Inattention was positively and
significantly associated with delinquency, proactive aggression, reactive aggression, bullying others, perpetrator of violence, anxiety and the two other dimensions of psychopathy
(impulsive-irresponsible and grandiose-manipulative). The correlation between CU-traits and anxiety was not significant (r = -0.5, p = .25).
Identified subgroups
Is there a meaningful subgroup of youths displaying high levels of CU-traits and high levels of ADHD symptoms (high combination group)? To answer this question and to identify subgroups with varying levels of CU-traits and ADHD symptoms, we conducted a
Hierarchical cluster analysis. We have ensured to meet the four criteria that a data driven method should fulfill during the analysing process. First, we based our number of clusters on previous theory and research. The most relevant estimation of clusters, according to this criterion, was the four-group solution. Second, the clusters showed homogeneousness with an error sum of squares (ESS) of 62 %. The result is below the preferable value, but still
indicating acceptable differences between the clusters. Third, the cluster coefficient was 52 % and above the 10 % cut off. Fourth, our identified subgroups contained a large sample of participants, varying from 122 to 210 participants in each group. Furthermore, we validated the clusters based on external variables to investigate if the subgroups are meaningful or not.
We fulfilled the recommendations and assumptions of meaningful subgroups since our identified groups differed significantly on the external variables of interest. The standardized mean level differences between the groups are presented in Table 3.
Table 3
Standardized mean level differences between the subgroups.
Group 1 Group 2 Group 3 Group 4 F
LOW ADHD CU ADHD + CU
(n =210) (n =133) (n =128) (n =122) M M M M Clustering variables CU-traits -.78a -.57b .61c 1.3d 490.912** Hyperactivity/impulsivity -.79a .85b -.43c .75b 229.114** Inattention -.77a .76b -.58c .86b 259.475** Extrernal criterion variables
Delinquency (total score) -.35a .005b .04b .52c 19.490**
Proactive agression -.46a -.11b .13c .54d 37.264** Reactive agression -.45a .006b .17b .42c 27.342** Bullying others -.29a -.07b .08b .41c 16.504** Perpetrator of violence -.23a -.06a -.07a .34b 11.904** Anxiety -.30a .64b -.47a .22c 42.438** Impulsive-irresponsible -.62a -.15b .17c .83d 77.708** Grandiose-manipulative -.64a -.37b .37c 1.0d 145.650**
Note: The letters a, b, c, d, indicates significant differences between the subgroups. The difference is not significant when two or more subgroups share the same
letter.
** p<.001.
Group 1 was our largest group (N = 210). In line with our expectations, given the normative nature of the sample, this group scored low on both CU-traits and ADHD
symptoms. This group engaged less in problematic behaviours such as bullying and criminal actions, as well as displaying low levels of hyperactivity/impulsivity and attention problems associated with ADHD in comparison to the other subgroups. Group 1 showed less emotional difficulties associated with CU-traits, as well as scoring low on the grandiose-manipulative and impulsive-irresponsible dimensions of psychopathy in comparison to the other
subgroups. Furthermore, this group had lower levels of anxiety compared to the other subgroups besides group 3.
Group 2 (N = 133) scored low on CU-traits and high on ADHD symptoms. Similar to group 1, this group showed low levels of emotional difficulties associated with CU-traits, as well as low levels of the grandiose-manipulative and impulsive-irresponsible dimensions of psychopathy. However, this group had elevated levels of ADHD symptoms in comparison to group 1 and 3. Group 2 displayed high levels of hyperactivity/impulsivity and attention problems associated with ADHD, such as difficulties sitting still and organizing tasks. Furthermore, this group had the highest levels of anxiety in comparison to the other subgroups.
Group 3 (N = 128) scored high on CU-traits and low on ADHD symptoms. Similar to group 1, this group had less difficulties with hyperactivity/impulsivity and attention problems associated with ADHD in comparison to group 2 and 4. However, this group had elevated levels of emotional difficulties associated with CU-traits, as well as elevated levels of the grandiose-manipulative and impulsive-irresponsible dimensions of psychopathy. Group 3 engaged in more serious deviant behaviours, such as criminal actions and violence towards others in comparison to group 1 and 2. Furthermore, group 3 had the lowest level of anxiety in comparison to the other subgroups. However, the difference between group 3 and 1 was not significant.
Group 4 was our smallest group (N = 122) were the identified individuals scored high on both CU-traits and ADHD symptoms (high combination group). As expected, this group was more frequently involved in delinquency and other serious harmful behaviours, such as bullying and violence towards others in comparison to the other subgroups. Similar to group 3, this group displayed high levels of emotional difficulties associated with CU-traits, as well as high levels of the grandiose-manipulative and impulsive-irresponsible dimensions of psychopathy in comparison to the other subgroups. Similar to group 2, this group showed high levels of hyperactivity/impulsivity and attention problems associated with ADHD in
comparison to group 1 and 3. In line with our hypothesis, the high combination group had higher scores on our external variables, implying that this group has developed more problematic behavioural and emotional patterns in comparison to the other
subgroups. Furthermore, the high combination group had elevated levels of anxiety in comparison to group 1 and 3.
Cluster validation
To validate the clusters and investigate meaningful differences between the subgroups, we compared the mean scores of the groups on external variables with One-Way ANOVA. Student-Newman Keul (SNK) post hoc test was used to further investigate significant group differences.
We hypothesize that the high combination group will score higher in comparison to the other subgroups on delinquency, aggression, bullying others, perpetration of violence, anxiety, grandiose-manipulative traits and impulsive-irresponsible traits.
Delinquency
The findings from the One-Way ANOVA showed that there was a significant statistical difference across the groups in delinquency (F(3,589) = 19.490, p = <.001). A follow up post-hoc (SNK) test showed significant differences between all four subgroups. The results indicated that the high combination group engaged in more delinquent behaviour in comparison to the other subgroups.
Proactive aggression
The findings from the One-Way ANOVA showed that there was a significant statistical difference across the groups on the variable proactive aggression (F(3, 586) = 37.264, p = <.001). A follow up post-hoc (SNK) test showed significant differences between all groups, except between the groups with elevated ADHD symptoms and the group with
elevated CU-traits. The results indicated that the high combination group showed more proactive aggression in comparison to the other subgroups.
Reactive aggression
The findings from the One-Way ANOVA showed that there was a significant
statistical difference across the groups on the variable reactive aggression (F(3, 586) = 27.342, p = <.001). A follow up post-hoc (SNK) test showed significant differences between all groups, except between the groups with elevated ADHD symptoms and the group with elevated CU-traits. The results indicated that the high combination group showed more reactive aggression in comparison to the other subgroups.
Bullying others
The findings from the One-Way ANOVA showed that there was a significant
statistical difference across the groups on the variable bullying others (F(3, 583) = 16.504, p = <.001). Follow up post hoc test (SNK) showed significant differences between all groups, except between the groups with elevated ADHD symptoms and the group with elevated CU-traits. The results indicated that the high combination group engaged in bullying behaviours to a higher extent in comparison to the other subgroups.
Perpetrator of violence
The findings from the One-Way ANOVA showed that there was a significant statistical difference across the groups on the variable perpetrator of violence (F(3, 584) = 11.904, p = <.001). Follow up post hoc test (SNK) showed that the high combination group were significantly different in comparison to the other groups in terms of violence towards others. There were no significant differences between the other subgroups. The results indicated that the high combination group significantly engaged in more violence towards others in comparison to the other subgroups.
Anxiety
The findings from the One-Way ANOVA showed that there was a significant statistical difference across the groups on levels of parent-reported anxiety (F(3, 589) = 42.438, p = <.001). A follow up post-hoc (SNK) test showed significant differences between all groups except between the group with low levels of ADHD symptoms and CU-traits, and the group with high levels of CU-traits only. The individuals with low levels of both ADHD symptoms and CU-traits and the group with elevated levels of CU-traits showed less levels of anxiety in comparison to the other groups. However, the difference between the two
subgroups was not significant. Furthermore, the results indicated that individuals with high levels of ADHD symptoms and individuals in the high combination group had higher levels of anxiety in comparison to the other groups.
Grandiose-manipulative traits
The findings from the One-Way ANOVA showed that there was a significant statistical difference across the groups on the grandiose-manipulative dimension of psychopathy (F(3, 579) = 490.912, p = <.001). A follow up post-hoc (SNK) test showed significant differences between all four subgroups. The results indicated that the high combination group also had the highest score on the grandiose-manipulative dimension of psychopathy.
Impulsive-irresponsible traits
The findings from the One-Way ANOVA showed that there was a significant statistical difference across the groups on the impulsive-irresponsible dimension of psychopathy (F(3, 589) = 146.650, p = <.001). A follow up post-hoc (SNK) test showed significant differences between all four subgroups. The results indicated that the high combination group also had the highest score on the impulsive-irresponsible dimension of psychopathy.
In sum, our findings show that adolescents in the high combination group with high levels of psychopathic traits and ADHD symptoms can be classified as an especially vulnerable and problematic subgroup. Consistent with our expectations and previous empirical findings, this group showed more delinquency, aggression, violent and bullying behaviours, grandiosity and impulsivity compared to the other subgroups. However, the results showed that the high combination group had elevated levels of anxiety, although not higher than the group with high levels of ADHD symptoms only.
Discussion
There is ample evidence that the psychopathic personality is dimensional and heterogeneous (Karpman, 1941; Cleckley, 1941; 1988; Lykken, 1995; Skeem, Poythress, Edens & Lilinfeld & Kale, 2003; Hicks et al., 2004). Scholars agree that psychopathy can be divided into the two subgroups of primary and secondary psychopaths, whose differences can be observed and measured in terms of anxiety, aggression and impulsivity (Cleckley, 1988; Karpman, 1941; Lykken, 1995; Skeem et al., 2003). The research also offers a theoretical developmental explanation regarding etiological trajectories. Both subtypes engage in similar antisocial behaviours, however, the primary subtype is due to biological factors whilst the secondary subtype is due to environmental influences (Hicks et al., 2004). Put together, the available research offers an explanation for the differences between the subtypes, but less is known about the underlying construct that contributes to the above-mentioned differences. Therefore, we have analysed subgroups on relevant external variables to examine potential meaningful differences. The results showed that the differences could be interpreted as an explanation for the discrepancies between the primary and secondary subtypes of
psychopathy.
Based on previous research regarding potential subgroups and trajectories of the psychopathic development, we used a normative sample of youths to test whether there is a
meaningful group displaying high levels of CU-traits and high levels of ADHD symptoms (high combination group). Furthermore, we wanted to investigate to what extent this group distinguishes from other groups with varying levels of CU-traits and ADHD symptoms. We identified four groups of adolescents, where one group displayed a combination of high levels of CU-traits and ADHD symptoms. These individuals differed from the other subgroups in several theoretically meaningful ways. Adolescents in the high combination group were more violent, aggressive, impulsive and delinquent in comparison to the other identified subgroups. These results suggest that there could be a meaningful subgroup of youths displaying early signs of psychopathy that resembles the characteristics of the secondary subtype described by Karpman (1941).
Our results showed that the individuals in the high combination group displayed an elevated problematic developmental pattern in comparison to the other subgroups. The high combination group showed significantly higher levels of delinquency, aggression, bullying and violence, as well as higher scores on the other two dimensions of psychopathy
(impulsive-irresponsible and grandiose-manipulative). These findings indicate that the high combination subgroup engaged in more maladaptive behaviours associated with the
psychopathic lifestyle, potentially at an elevated risk of developing psychopathy in adulthood (Hicks et al., 2004; Sevecke et al., 2009; Moore et al., 2019; Baliousis et al., 2019).
The psychopathic personality, involving CU-traits as a core dimension, are associated with low emotional capacity. This suggests that individuals scoring high on CU-traits should display lower levels of emotional arousal, such as anxiety, in comparison to the general population (Cooke & Michie, 2001). In line with previous research, our findings show that the group scoring high on CU-traits only, have reported less anxiety in comparison to the other identified groups. On the other hand, the high combination group with both CU-traits and ADHD symptoms, had elevated levels of anxiety. This result occurred even though the high
combination group had higher levels of CU-traits in comparison to the CU-trait only group. We expected that the high combination group would have lower levels of anxiety due to their high levels of CU-traits. Furthermore, the group with ADHD symptoms only, scored the highest on reported levels of anxiety, but there was no significant difference between the high combination group and the group with high levels of ADHD symptoms only. Research shows that many individuals with ADHD have at least one other comorbid disorder, often related to anxiety (Barkley, 2006; Cherkasova et al., 2013). Therefore, these findings could provide a theoretical explanation of why our high combination group displayed higher levels of anxiety than expected. It seems that the component of ADHD contributes significantly to higher levels of anxiety.
According to Karpman's (1941) theory, the secondary subtype distinguishes from the primary subtype in anxiety levels. The lack of empathy that characterizes the primary
psychopath is hypothesized to result in low levels of anxiety. In contrast, the secondary subtype tends to be highly anxious (Salihovic et al., 2014, Hicks et al., 2004). Our findings showed that the high combination group was more similar to the secondary subtype, whilst the group with elevated CU-traits only, was more similar to the primary subtype in terms of anxiety levels. In addition to anxiety levels, the secondary subtype tends to be more
aggressive and impulsive in comparison to the primary subtype (Salihovic et al., 2014). Our findings suggest that the high combination group is more aggressive in terms of both
proactive and reactive aggression. The group also showed higher levels of delinquent and violent behaviour, which might be the result of a more aggressive temper.
The measurements used to capture the adolescents’ levels of aggression was created for the project, based on other similar measurements of aggression (eg. Raine et al., 2006). The measurement was divided into proactive and reactive aggression. Proactive aggression is described as a deliberate, compelling behaviour that is driven by external reinforcements. The
proactive aggression is often used as a means of obtaining a desiderate goal (Bandura, 1977). On the other hand, reactive aggression is described as an angry, defensive response that often emerges when an individual feels threatened, frustrated or provoked (Dodge, Lochman, Harnish, Bates & Pettit, 1997). Based on previous research and theory, we hypothesised that the high combination group would score higher on reactive aggression and lower on proactive aggression, due to their elevated levels of anxiety and impulsivity. However, our findings indicated that the high combination group scored higher on proactive aggression. These results therefore contradict our theoretical understanding of the similarities based on
impulsivity and aggression between the secondary subtype and the high combination group. The confusing results could be explained by the questionable validity and reliability of the measurement used. There is a possibility that the measurement created was not sensitive enough and failed to capture the differences between the two types of aggression. Therefore, the results of the two different kinds of aggression should be interpreted with caution. Importantly, our results showed that the high combination group scored higher on their total levels of aggression, including both reactive and proactive aggression, compared to the other subgroups. This indicates that the combination of CU-traits and ADHD results in a more aggressive and impulsive behaviour, similar to the theoretical explanation of the secondary subtype of psychopathy (Karpman, 1941).
In sum, our results confirm previous research stating that psychopathy is a heterogeneous construct. Furthermore, our study contributes to an alternative empirical explanation of the secondary psychopath. The results show that the high combination group shares similar features of the secondary subtype. In line with Karpman's theory (1941), our high combination group is demonstrating higher levels of anxiety, aggression and impulsivity in comparison to the other identified subgroups with varying levels of CU-traits and ADHD symptoms. This indicates that the combination of high levels of CU-traits and ADHD
symptoms might be associated with a more serious form of maladaptive behaviours, resulting in a less favourable lifestyle. Since the group with high CU-traits only also show a
problematic pattern of behaviour, the explanation of the developmental pattern is probably due to the high levels of psychopathic traits amongst these individuals. However, considering the high combination group are even more problematic, ADHD seems to have an important role in explaining the differences, raising the question of whether the discrepancy between the groups with elevated levels of CU-traits could be explained by a co-occurring ADHD. The high combination group represents a subgroup with elevated risk of developing a more problematic antisocial personality. For example, the prevalence of ADHD in adolescents displaying antisocial behaviours is high. Studies have shown that as many as 20-72 % of incarcerated adolescents have both psychopathic traits and ADHD (Sevecke et al., 2009). The characteristics of aggression and impulsivity due to the ADHD symptoms seem to have a problematic impact, as it inhibits the individual to control impulses.The combination of high emotional neuroticism and impulsive behaviours associated with the secondary subtype, might lead to negative emotions regarding the inability to control potentially dangerous actions (Hicks et al., 2004; Salihovic et al., 2014). This indicates that the secondary subtype might be better explained as a combination of psychopathic traits and ADHD.
Limitations
There are some limitations of the current study that need to be acknowledged. First, the cross-sectional design precludes us from understanding the development of psychopathic traits amongst our sample of youths. The design provides a momentary overview of the adolescent’s characteristics, but we cannot infer conclusions regarding stability over time for each individual in their respective clusters regarding developmental patterns. Furthermore, the design prohibits us from concluding whether the identified subgroups display distinct
