Self-Reported Psychopathic-Like Traits as Predictors of Recidivism in Detained Male Adolescents

Abstract

Few studies have examined whether psychopathic-like traits predict recidivism in detained youth. The current study investigated the predictive validity of self-report of psychopathic-like traits for official recidivism. Participants were 223 detained male adolescents from all three Flemish youth detention centers, who were assessed with the Youth Psychopathic Traits Inventory (YPI). Two to four years later, official information on recidivism was collected. Recidivism was defined in four ways. Neither the affective dimension nor the broader construct of psychopathy (i.e., total YPI score) was predictive of recidivism above and beyond criminal history. The behavioral and interpersonal dimensions, however, significantly added to the prediction of substance-related recidivism. Overall, psychopathic-like traits in detained adolescents assessed by means of the YPI did not convincingly predict recidivism. Implications of the findings are discussed.

Keywords

juvenile psychopathy callous–unemotional violence delinquents

Psychopathy is a multidimensional construct defined by affective, interpersonal, and behavioral characteristics, such as lack of empathy, shallow affect, lack of guilt, manipulation, lying, grandiose self-worth, sensation-seeking behavior, impulsivity, and irresponsibility. To better understand the developmental origins of psychopathy, researchers have begun to examine psychopathic-like traits in children and adolescents. Similar to their adult counterparts (Walters, 2003), minors high in psychopathic-like traits are among the most severe antisocial individuals (Andershed, Kerr, Stattin, & Levander, 2002), have difficulties processing emotions (Marsh et al., 2008), and are prone to externalizing rather than internalizing problems (Salekin, Neumann, Leistico, DiCicco, & Duros, 2004). In addition, the few available longitudinal studies on the topic suggest that psychopathic-like traits are relatively stable during the transition from childhood to adolescence and from youth to adulthood (see Andershed, 2010, for a review).

Yet although research on adult criminals has convincingly demonstrated that psychopathy predicts recidivism (e.g., Leistico, Salekin, DeCoster, & Rogers, 2008; Salekin, Rogers, & Sewell, 1996), the relation between psychopathic-like traits in delinquent minors and recidivism has received little attention (Douglas, Epstein, & Poythress, 2008) and yielded mixed findings. For example, although some prospective studies in detained juveniles reported that psychopathic-like traits as measured by the Psychopathy Checklist: Youth Version (PCL:YV; Forth, Kosson, & Hare, 2003) predicted later violent or general recidivism (Corrado, Vincent, Hart, & Cohen, 2004; Salekin et al., 2008), other prospective studies failed to observe such relationships (Edens & Cahill, 2007). The few available prospective studies and inconsistent results are surprising given researchers have started to apply the psychopathy construct to minors to identify youth at high risk for violence and chronic offending (Salekin, 2008).

In the current article we examine to what extent psychopathic-like traits in detained male adolescents predict recidivism 2 to 4 years later. Moreover, we specifically address two issues that rarely have been considered simultaneously in previous studies. The first issue relates to the question of whether one should focus on the affective dimension or use the broader psychopathy construct to predict recidivism. The second issue relates to the importance of psychopathic-like traits to predict recidivism relative to other relevant predictors (e.g., criminal history).

Using One Dimension or All Dimensions?

Although psychopathy is a multidimensional construct, some researchers explicitly focus on the affective dimension (AD) (e.g., Frick, 2006; Frick & White, 2008). Their argument is that only the AD differentiates between subgroups of adolescent offenders (Caputo, Frick, & Brodsky, 1999) and children with disruptive behavior disorders (Christian, Frick, Hill, Tyler, & Frazer, 1997), whereas the interpersonal dimension (ID) and behavioral dimension (BD) do not. These researchers therefore consider the AD to be the most important dimension to identify psychopathic-like adolescents within the total population of antisocial youth (for a review, see Frick, 2006; Frick & White, 2008). Consequently, the AD is likely to be the best dimension to predict recidivism.

Yet enthusiasm for an explicit AD focus may be tempered for at least three reasons. First, the argument for focusing on AD is mainly based on two relatively small cross-sectional studies in children (Christian et al., 1997) and adolescents (Caputo et al., 1999) that used the Antisocial Process Screening Device (APSD; Frick & Hare, 2001). Scores on different measures of the dimensions underlying psychopathy correlate only very modestly with each other (e.g., Cauffman, Kimonis, Dmitrieva, & Monahan, 2009). Therefore, it may be rash to conclude on the basis of a few studies with only a single measure of psychopathy that the AD dimension is the most important of all psychopathy dimensions. Second, in the four studies that examined the predictive value of the AD in detained youth when controlling for the other psychopathy dimensions, only one study showed that the AD/ID and the BD of the Psychopathy Checklist–Revised (Hare, 2003) predicted official recidivism (Brandt, Kennedy, Patrick, & Curtin, 1997). In the three other studies, however, the BD and/or the antisocial dimension of the PCL:YV (Cauffman et al., 2009; Vincent, Odgers, McCormick, & Corrado, 2008; Vitacco, Neumann, & Caldwell, 2010) and the ID of the Youth Psychopathic Traits Inventory (Cauffman et al., 2009) predicted recidivism. Also in adults, the antisocial dimension demonstrated incremental validity above and beyond the other three PCL–R dimensions in predicting recidivism (Walters, Knight, Grann, & Dahle, 2008). Third, by focusing on one particular dimension one may lose sight of the original construct, an issue that is addressed in depth below.

Taken together, the empirical evidence for the predictive value of the AD to predict recidivism appears limited. In fact, the absence of the superior prediction for scores on the AD in severe antisocial adolescent and adults relative to the other dimensions argues against a sole focus on features of the AD. Because this focus is likely to influence future research and clinical work with antisocial youth (e.g., current proposal to use the AD as specifier within conduct disorder of the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders; Frick & Moffitt, 2010), future studies that examine the usefulness of this sole focus on the AD with other psychopathy measures are urgently warranted.

In contrast to the single focus on the AD, some researchers argue that focusing on one particular dimension runs the danger of losing sight of the original construct (Hare & Neumann, 2008; Neumann, Hare, & Newman, 2007). For instance, in a large study of adolescent offenders, a super-ordinate factor accounted for the majority of the variance in all the psychopathy dimensions (Neumann, Kosson, Forth, & Hare, 2006). Also, a psychopathy total score predicts recidivism better than the AD score (Cauffman et al., 2009; Vincent et al., 2008). The superordinate factor thus seems to capture something essential that is missed when focusing on one particular dimension. This point of view suggests that a psychopathy total score may be a better predictor of recidivism than any single dimension score. Yet few studies present results from analyses in which a total psychopathy score was used to predict recidivism together with results from analyses in which the predictive value of each psychopathy dimension relative to the others was examined (Cauffman et al., 2009; Vincent et al., 2008).

Incremental Contribution of Psychopathic-Like Traits

Decades of research and clinical work repeatedly demonstrate that factors related to the past best foretell the future. It is not surprising that a meta-analysis on many different risk factors in already delinquent youth demonstrated that criminal history is the strongest predictor of recidivism (Cottle, Lee, & Heilbrun, 2001). A crucial question, thus, is whether the AD or the total psychopathy score predicts recidivism, especially after controlling for criminal history. An important concern raised by several researchers (e.g., Asscher et al., 2011; Murrie, Boccaccini, McCoy, & Cornell, 2007; Seagrave & Grisso, 2002) relates to the stigma the psychopathy construct is likely to induce. Therefore, it can be argued that psychopathy measures should not be used unless they have some added value beyond other less prejudicial and easy to obtain variables such as past antisocial behavior (Edens, Campbell, & Weir, 2006; Walters et al., 2008). It is surprising that few studies have evaluated the incremental contribution of psychopathy measures once criminal history is controlled. Again, these studies have yielded mixed findings. For example, although in one study psychopathic-like traits as measured by the APSD had incremental predictive validity over and above a host of other control variables (e.g., criminal history) to predict general recidivism (Salekin, 2008), this finding was not replicated in another study using the same instrument (Douglas et al., 2008).

The Current Study

The current study was designed to examine whether psychopathic-like traits as measured by the Youth Psychopathic Traits Inventory (YPI; Andershed et al., 2002) predict recidivism in detained male adolescents. Several authors have described the YPI as a promising self-report instrument to measure psychopathic-like traits in adolescents (e.g., Kotler & McMahon, 2005; Skeem & Cauffman, 2003). Yet the predictive utility of the YPI to predict recidivism has been examined in only one study (Cauffman et al., 2009). This interesting article, however, had some features that may restrict a thorough examination of the YPI’s utility to predict recidivism in detained adolescents. First, only about one fifth of the adjudicated participants in the Cauffman et al. study were detained or in jail (Schubert et al., 2004), making it difficult to conclude whether the YPI is useful to predict recidivism in “pure” samples of detained adolescents.¹ Second, the YPI data were used to predict short-term official recidivism (i.e., 6 and 12 months later; Cauffman et al., 2009).² To examine psychopathic-like traits in adolescence and their prospective relation to offending into young adulthood, longer follow-up periods are needed (e.g., Edens & Cahill, 2007). Third, in line with most studies on youth psychopathy, both official nonviolent and official violent crimes were collapsed into one general official recidivism outcome (Cauffman et al., 2009). Given heightened interest in severe forms of interpersonal crime, it is important to know whether the YPI is predictive of official violent recidivism in particular (Edens et al., 2006).

Examining the usefulness of self-report measures to assess psychopathic-like traits and recidivism is important for several reasons. First, in youth detention centers it is difficult to include parents as informants for screening and assessment purposes (Colins, Vermeiren, Schuyten, Broekaert, & Soyez, 2008). Therefore, self-report instruments may be an appropriate alternative to instruments that rely on parents to measure psychopathic-like traits (e.g., the APSD). Second, because some studies recently demonstrated that self-report measures of psychopathic-like traits predicted recidivism (e.g., Cauffman et al., 2009; Salekin, 2008), such self-report instruments may be of considerable importance for risk assessment purposes. Third, most studies have used the PCL:YV to predict recidivism. Although the PCL:YV is currently the most reliable and valid measure of assessing psychopathic-like traits among detained adolescents (Andershed, Hodgins, & Tengstrom, 2007), it requires extensive training and is time-consuming and costly to use with all detained youth. Brief and easy-to-complete self-report measures therefore may provide an economic advantage (Lilienfeld & Fowler, 2006).

For the reasons mentioned above, the current study examines whether self-reported psychopathic-like traits in detained male adolescents are predictive of recidivism 2 to 4 years later. The following hypotheses were tested:

The YPI AD predicts recidivism after controlling for the other YPI psychopathy dimensions, time at risk, and criminal history (i.e., past charges and age of first arrest).

The total YPI score is a stronger predictor of recidivism than the AD of the YPI.

Method

Participants

Between January 2005 and February 2007, 305 recently detained minors (younger than 18) from all three youth detention centers (YDCs) for boys in Flanders (i.e., the Dutch-speaking part of Belgium) were randomly selected from the entire population of detained youth. Criteria for inclusion were being of Belgian or Moroccan origin, being detained for at least 1 month, and having sufficient knowledge of Dutch. The first two inclusion criteria were required for two major reasons. First, many youth in the YDC are of non-Belgian origin (e.g., Turkish, Polish, Sudanese). Given time restrictions, it was not possible to include enough youth from various origins that would allow meaningful group comparisons (e.g., Turkish vs. Belgian). Therefore, we decided to focus on the largest “non-Belgian” subgroup in the YDC, namely, youth of Moroccan origin. Second, to recruit the most deviant youth and to allow enough time for approaching and investigating the participant, an expected detention duration of 1 month at minimum was required. Eventually, we were able to interview all participants between 3 days and 3 weeks after their detention intake. In Flanders, youth can be referred to one of these three YDC by a juvenile judge in the preliminary phase (i.e., pretrial detention) or after being sentenced (posttrial incarceration). To the best of our knowledge, all participants in the current study were pending final trial.

Of the 305 boys, 14 could not be approached because of practical circumstances (e.g., daily activities, confinement), 45 boys refused participation, and 1 did not have sufficient knowledge of Dutch, resulting in a participation rate of 80% (N = 245). Of the 245 participants, 22 were excluded because of incomplete data collection or because one public prosecutor did not deliver recidivism data, resulting in a final sample of 223 boys.

Procedure

This study was approved by the institutional review board (IRB) of the Faculty of Psychology and Educational Sciences, Ghent University. Because the YDC has been ordered by decree to observe the juveniles assigned to them (e.g., screening for emotional problems), the IRB waived the requirement to get parental consent. Selected detainees were approached individually and given oral and written information about the aims, content, and duration of the interviews. They were assured that their information was confidential and that refusal to participate would not affect their judicial status or stay in the YDC. The boys then could consult their primary caregivers or other adults about participation. Participants were required to give written informed consent before starting the study. Participants were approached and assessed following a standardized protocol. Participants were assessed in a private area in the YDCs by the first author or by one of two final-year university students who did not belong to the YDC staff. Participating youth did not receive compensation.

Measures

Psychopathic-like traits

The current study used the YPI to assess psychopathic personality traits (Andershed et al., 2002). The YPI is a self-report questionnaire that consists of 50 items, organized into 10 subscales of 5 items. Each item is scored on a 4-point Likert-type scale with scores ranging from 1 (does not apply at all) to 4 (applies very well). The YPI is based on a three-factor model of psychopathy (Cooke & Michie, 2001). Factor 1 of the YPI represents an arrogant and deceitful ID, Factor 2 a callous unemotional or deficient AD, and Factor 3 an impulsive and irresponsible BD (Andershed et al., 2002). In the present study alpha coefficients for the YPI three factors (i.e., by using all items in each factor) and the YPI total score (i.e., by using all YPI items) were as follows: α = .88 for the ID, α = .73 for the AD, α = .85 for the BD, and α = .92 for the YPI total score. The three YPI factors were moderately correlated (r = .56-.60), with the greatest association between the ID and BD. In line with previous YPI studies (Andershed et al., 2007), the three factors were created by computing the mean scores for the subscales (by summing up the mean subscale scores in the respective factor and then dividing by the number of subscales in that factor). The total score of the YPI was created by computing the mean scores for all the 10 subscales (by summing up the 10 subscales and then dividing by 10).

Criminal history and official recidivism

We collected participants’ arrest data from the registration system of the public prosecutor. This registration system provides information about the number and type of arrests before and after their 18th birthday. Criminal history is the age at which participants were first charged for any type of crime (e.g., shoplifting, violence, insults) and the total number of arrests before participation in our study. Violent rearrest refers to murder, manslaughter, sex offenses, assault, and theft with violence. Severe nonviolent rearrest refers to burglary; minor nonviolent rearrest refers to theft, shoplifting, vandalism, fire setting, and other minor offending, such as traffic offenses (e.g., driving without a license), threats (e.g., threaten to hurt someone), and insults (e.g., the speaking of improper and defamatory words tending to seriously insult someone). Substance-related rearrest refers to use, possession, and dealing of drugs (alcohol not included). General rearrest refers to violent, severe nonviolent, minor nonviolent, and/or substance-related rearrests. The mean follow-up period was 1,206.20 days (i.e., 3.3 years; SD = 191.42 days; range = 718-1,478 days).

Time at risk

Time at risk was defined as the number of days between the day the adolescent participated and February 1, 2009, minus the number of days participants were detained (as a minor) or in provisional custody (as an adult).

Sociodemographic characteristics

Standardized information about age, origin, and parental or caretaker’s occupation was assessed by means of a questionnaire designed by the authors. Socioeconomic status (SES) was made operational by dichotomizing parental or caretaker’s occupation. Parental SES was placed in the low category (vs. high) if both were unemployed or held a low-level job (unskilled and skilled labor).

Seriousness of recidivism (dependent variable)

Although frequent offenders are generally serious offenders, some frequent offenders are not (Tolan & Gorman-Smith, 1998). To allow an accurate definition of “serious offenders,” both frequency and severity of offending should be taken into account (Hoeve et al., 2008). We therefore created three recidivism groups using the following steps.

In the first step, we assigned different levels of seriousness to each type of crime committed in the follow-up period. In line with Cohen (1986), we considered violent offenses as the most serious offenses, followed by substance-related offenses, burglary, and minor nonviolent offenses. We chose this classification for two reasons. First, the Cohen classification resembles our view that violent offenses in general have a greater impact on the physical integrity of the victim and public safety than other types of offenses. Second, substance-related offending is related to acquisitive crimes (e.g., to maintain substance use; Anglin & Perochet, 1998) as well as violent crimes (e.g., to enforce rules in the illegal drug market; White, Loeber, Stouthamer-Loeber, & Farrington, 1999). As Cohen considered the level of seriousness of substance-related offending in between nonviolent and violent offenses, we preferred this classification over classifications in which substance-related offending for example was considered only to be more serious than a misdemeanor (Mulder, Brand, Bullens, & van Marle, 2010) or more serious than hitting a teacher (e.g., Chung, Hill, Hawkins, Gilchrist, & Nagin, 2002).

In the second step, we weighted each of the four types of offenses by coding minor nonviolent rearrests as Level 1, burglary as Level 2, substance-related rearrests as Level 3, and violent rearrests as Level 4. Specifically, for each participant the total number of minor nonviolent rearrests was multiplied by 1, the total number of burglaries by 2, the total number of substance-related rearrests by 3, and the total number of violent rearrests by 4.

In the third step, we created a combined continuous measure (i.e., seriousness of recidivism) by summing up participants’ multiplied total number of minor nonviolent rearrests, multiplied total number of burglaries, multiplied total number of substance-related rearrests, and multiplied total number of violent rearrests. This combined continuous measure could not be used in linear regression because assumptions for this kind of analysis were violated.

Therefore, and based on this combined continuous measure, participants were assigned to one of three groups in a final step. Participants who scored below the 25th percentile were assigned to the minor recidivism group. Participants who scored between the 25th and 75th percentile were assigned to the moderate recidivism group. Participants who scored above the 75th percentile were assigned to the serious recidivism group.

Other recidivism definitions (dependent variable)

To allow a comparison with other studies, we also considered violent recidivism separate from severe nonviolent recidivism (i.e., burglary) and substance-related recidivism in secondary analyses. Because almost all participants (83%) were rearrested for minor nonviolent crimes, we did not focus on this type of nonviolent reoffending.

Analyses

Differences between seriousness of recidivism subgroups in variables of interest were examined with χ² for categorical variables and one-way analyses of variance for continuous variables. Bonferroni correction was used for multiple comparisons.

The association of psychopathic-like traits with seriousness of recidivism (minor vs. moderate vs. serious) was tested by ordinal logistic regression (OLR; link function: logit). The proportional odds assumption underlying OLR was verified by means of the test of parallel lines. We present proportional odds ratio (OR) and 95% confidence interval (CI) values when time at risk, age of first arrest, and total past arrests were the only predictors in the model (Model 1) and when all three YPI dimension scores or the YPI total score were included as predictors as well (Model 2). To test whether psychopathic-like traits add to the prediction of recidivism above and beyond age of first arrest and total past arrests, we present the block χ² statistic for each block and changes in −2 log likelihood ratio between Model 1 and Model 2.

As violent, serious nonviolent, and substance-related rearrests were nonnormally distributed, each type of recidivism was dichotomized (yes vs. no) and entered as a dependent variable in binary logistic regression analyses (method = enter). The same two sets of predictors as used in the OLR analyses were included in these analyses. All tests were two-tailed, with .05 as the standard for statistical significance. Statistical analyses were performed using SPSS 17.0.

Results

In addition to the information presented in Table 1, our total sample showed very high rates of recidivism, with 92.4% (n = 206) having at least one new arrest during the follow-up period. Specifically, 83.0% (n = 185) were minor nonviolent recidivists, 41.3% (n = 92) were serious nonviolent recidivists (e.g., burglars), 41.3% (n = 92) were substance-related recidivists, and 66.4% (n = 148) were violent recidivists. Of the 223 participants, 90.6% were no longer underage (i.e., <18 years) when recidivism data were collected.

Table 1:

Sociodemographics, Crime-Related Characteristics, and Psychopathic-Like Traits for the Total Sample and by Recidivism Group Membership

			Seriousness of Recidivism Group Membership
	Total Sample		(1) Minor Recidivists (n = 43)^a		(2) Moderate Recidivists (n = 123)		(3) Serious Recidivists (n = 57)
	M	SD	M	SD	M	SD	M	SD	Bonferroni-Corrected Post Hoc Test p < .05
Sociodemographics
Belgian origin (n, %)	172	77.1	38	88.4	102	82.9	32	56.1	1,2 > 3
Low SES (n, %)	143	64.1	28	66.7	73	63.5	42	76.4
Age at baseline	16.49	1.17	16.52	1.17	16.58	1.09	16.27	1.31
Age at follow-up	19.79	1.27	19.60	1.39	19.86	1.17	19.77	1.38
Criminal history
Age of first arrest	13.08	2.52	13.65	2.44	13.27	2.46	12.24	2.56	1,2 > 3
Total past arrests	12.64	12.51	6.74	5.16	11.23	12.12	20.14	13.85	1 < 2,3
Recidivism
Total rearrests	9.80	11.07	0.95	0.95	6.04	3.58	24.58	11.99	1 < 2,3; 2 < 3
Total violent rearrests	2.32	3.19	0	0.00	1.53	1.31	5.81	4.32	1 < 2,3; 2 < 3
Total minor nonviolent rearrests	4.64	6.11	0.81	0.96	3.08	2.73	10.88	8.65	1 < 2,3; 2 < 3
Total burglary rearrests	1.72	3.79	0.07	0.26	0.64	1.22	5.28	6.00	1,2 < 3
Total substance-related rearrests	1.10	1.97	0.07	0.26	0.79	1.24	2.56	2.97	1,2 < 3
Time at risk (in days)	1033.74	247.63	984.70	271.64	1050.87	235.87	1033.79	253.07
Youth psychopathic-like traits
Interpersonal dimension	1.71	0.51	1.76	0.54	1.67	0.45	1.78	0.61
Affective dimension	2.03	0.45	1.98	0.45	1.99	0.42	2.14	0.49
Behavioral dimension	2.48	0.60	2.47	0.57	2.44	0.57	2.55	0.70
YPI total score	2.04	0.44	2.04	0.45	2.00	0.40	2.12	0.51

Note. YPI = Youth Psychopathic Traits Inventory.

Of the 43 participants in the minor recidivism group, 17 individuals (39.5%) were not rearrested at all.

Taking both frequency and severity of rearrests into account, three recidivism groups with different levels of seriousness were created: a minor recidivism group (n = 43), a moderate recidivism group (n = 123), and a serious recidivism group (n = 57). The serious recidivism group was less often from Belgian origin, was arrested for the first time at a younger age, and had a higher number of rearrests for all types of crime compared to minor and moderate recidivism groups. In addition, youth in the serious recidivism group committed more crimes in the past than those categorized in the minor recidivism group (see Table 1).

Table 2 (Block 1) shows that the total number of past arrests significantly predicted serious recidivism group membership (vs. minor and moderate recidivism group membership). However, the YPI dimensions did not significantly add to the prediction of serious recidivism group membership (vs. minor and moderate recidivism group membership) as demonstrated by nonsignificant Block 2 χ² and −2LL statistics (see Table 2). Using the YPI total score as a predictor also did not add to the prediction of serious recidivism group membership when compared to minor and moderate recidivism group membership, Block 2: −2LL = 409.04, χ²(4) = 34.39, p = .00; OR and CI for total YPI score: OR = 1.12, CI = 0.62 to 2.01).

Table 2:

Ordinal Logistic Regression Models to Test the Incremental Utility of Psychopathic-Like Traits to Predict Serious Recidivism Group Membership

	Model 1		Model 2
	OR	CI	OR	CI
Age of first arrest	0.95	0.80-1.07	0.95	0.85-1.07
Total past arrests	1.07	1.03-1.10***	1.06	1.03-1.10***
Interpersonal dimension	—		0.88	0.44-1.73
Affective dimension	—		1.57	0.60-3.41
Behavioral dimension	—		0.95	0.53-1.69
−2 LLR	409.19		407.87
Block χ²	34.24***		35.56

Note. CI = confidence interval; OR = odds ratio; −2 LLR = −2 log likelihood ratio. Time at risk was also included as a continuous covariate in Model 1 and Model 2.

***

p < .001.

In secondary analyses, none of the YPI dimensions (see Table 3) nor the total YPI score significantly added to the prediction of violent recidivism, Block 2: −2LL = 243.57, χ²(1) = 0.07, p = .79; total YPI score: OR = 0.91, CI = 0.43 to 1.89, or serious nonviolent recidivism (e.g., burglary), Block 2: −2LL = 291.83, χ²(1) = 0.69, p = .41; total YPI score: OR = 0.76, CI = 0.41 to 1.44. Including the YPI dimensions (Block 2) significantly added to the prediction of substance-related recidivism (see Table 3). The ID was negatively related whereas the BD was positively related with substance-related recidivism. The YPI total score did not contribute to the prediction of having any new substance-related arrest, Block 2: −2LL = 300.12, χ²(1) = 0.22, p = .64; total YPI score: OR = 1.16, CI = 0.63 to 2.14.

Table 3:

Binary Logistic Regression Models to Test the Incremental Utility of Psychopathic-Like Traits to Predict Violent, Serious Nonviolent, and Substance-Related Recidivism

	Violent Recidivism				Serious Nonviolent Recidivism				Substance-Related Recidivism
	Block 1		Block 2		Block 1		Block 2		Block 1		Block 2
	OR	CI	OR	CI	OR	CI	OR	CI	OR	CI	OR	CI
Age of first arrest	0.92	0.80-1.06	0.92	0.80-1.06	0.91	0.81-1.03	0.91	0.81-1.02	1.00	0.89-1.12	0.99	0.87-1.12
Total past arrests	1.11	1.05-1.16***	1.10	1.05-1.16***	1.01	0.99-1.04	1.01	0.99-1.04	1.01	0.99-1.04	1.01	0.99-1.04
Interpersonal dimension	—		0.74	0.34-1.65	—		0.72	0.35-1.48	—		0.30	0.14-0.64**
Affective dimension	—		1.43	0.56-3.65	—		1.33	0.59-2.99	—		0.90	0.40-2.02
Behavioral dimension	—		0.96	0.49-1.86	—		0.87	0.48-1.60	—		3.53	1.80-6.92***
−2 LLR	243.64		242.80		295.52		290.91		300.34		282.41
Block χ²	41.16***		0.84		9.76*		1.62		1.94		17.94***

Note. CI = confidence interval; OR = odds ratio; −2 LLR = −2 log likelihood ratio. Time at risk was also included as a continuous covariate in Block 1 and Block 2.

p < .05. **p < .01. ***p < .001.

Discussion

This article examined whether self-reported psychopathic-like traits predicted recidivism in detained youth. The AD of the YPI and the YPI total score did not add to the prediction of recidivism beyond criminal history. The ID and BD of the YPI, however, were related with substance-related recidivism. In general, the present study suggested that self-reported psychopathic-like traits as measured by the YPI are not useful for predicting other than substance-related recidivism, among detained male adolescents.

The current study does not support the suggestion that the AD is the most predictive dimension of recidivism for antisocial youth. This finding converges with those of recent studies on serious delinquent adolescents (Cauffman et al., 2009) as well as with adult psychiatric patients, offenders, and inmates (Walters et al., 2008). Of course, given that psychopathy involves a deficit in recognizing emotions (Marsh et al., 2008), it can be argued that juveniles are unaware of or unable to report these traits themselves (Rutter, 2005). It is interesting that APSD parent and/or teacher ratings have been found to predict criminal behavior (e.g., Dadds, Fraser, Frost, & Hawes, 2005; Frick, Stickle, Dandreaux, Farrell, & Kimonis, 2005; McMahon, Witkiewitz, & Kotler, 2010). Future studies are needed to test whether the AD as rated by parents and/or teachers is more strongly related to recidivism than other dimensions. If so, then parent and teacher ratings of the AD in detained male adolescents may be preferred to self-report when trying to predict recidivism. Finally, the total YPI score was also not predictive of serious recidivism group membership. In line with previous research in seriously delinquent adolescents using expert-rater based measures (Cauffman et al., 2009; Vincent et al., 2008), we also could not support the assumption that the broader construct of self-reported psychopathy is a better predictor of serious recidivism than the AD.

The current study showed that the assessment of psychopathic-like traits in detained male adolescents by means of the YPI does not predict official violent recidivism. This finding converges with two recent studies in juvenile justice youth that showed that two other well-known self-report measures of psychopathic-like traits (i.e., APSD and Childhood Psychopathy Scale; Lynam, 1997) were not related to violent recidivism after controlling for relevant variables (Douglas et al., 2008; Salekin, 2008). Yet our finding is not in line with those regarding three less frequently used self-report measures (i.e., NEO Psychopathy Resemblance Index [Lynam & Widiger, 2007]; Self-Report Psychopathy Scale–II [Hare, 1991]; Personality Assessment Inventory–Antisocial Scale [Morey, 1991]) of psychopathic-like traits being predictive of violent recidivism when relevant control variables were included in the analyses (Cauffman et al., 2009; Salekin, 2008). As many youth are identified as psychopathic on one measure but not on another (Cauffman et al., 2009), it is possible that finding a relation between psychopathic-like traits and violent recidivism may depend on the instrument used. Given the scarcity of research, this issue needs further study as the construct of psychopathy is increasingly being used in clinical practice and for legal decision making (Edens et al., 2006). In addition, it is also possible that an instrument has predictive utility in one specific setting but not in another. For example, although the YPI total score has been related with official recidivism in a sample of adjudicated youth (Cauffman et al., 2009), the current study showed that this may not be the case in detained youth.

If neither the YPI total score nor any of the three YPI dimensions provide additional predictive utility above past behavior, then there is no reason to use the YPI to assess future risk in detained male adolescents. However, in the present study the ID and BD of the YPI significantly added to the prediction of substance-related recidivism above and beyond criminal history. The ID was negatively related to this type of recidivism, a surprising finding given that this YPI dimension has previously been found to be related to substance-related offending (Poythress, Dembo, Wareham, & Greenbaum, 2006). In addition, substance-related recidivism was predicted by the BD of the YPI. This latter finding converges with previous studies demonstrating a relation between impulsive and irresponsible behavior (as assessed by the BD subscale of the YPI) and drug-related offending (e.g., Coid et al., 2007). However, neither the AD (often considered as the core feature of the construct) nor the broader construct (i.e., YPI total score) was positively related with future substance-related arrests. Therefore, it can be questioned whether the psychopathy construct itself is predictive of substance-related recidivism or whether alternative explanations are more likely. For example, because substance use disorders (SUD) are associated with impulsive behavior (Verdejo-Garcia, Lawrence, & Clark, 2008) and substance-related recidivism (Colins et al., 2011), the BD may no longer be related to substance-related rearrests after controlling for a history of SUD. Unfortunately, our sample size did not allow us to include other variables that may moderate the relation between psychopathy and recidivism (e.g., SUD). This study limitation warrants future studies with larger sample sizes to address the importance of relevant covariates in predicting recidivism.

The present research has several strengths that should be noted. First, it made use of a prospective design with a substantial time span during which most participants reached young adulthood. Second, it examined the incremental predictive utility of psychopathy by including relevant control variables. Third, the current study used an alternative approach to define recidivism by taking both frequency and severity of offending into account. This allowed us to test whether self-reported psychopathic-like traits were positively associated with serious recidivism group membership. This latter association was expected given that psychopathic-like youth are considered to be among the most serious and versatile offenders.

Some further limitations should also be considered when interpreting the results. First, we defined criminal history and recidivism using official arrest data and thus may have underestimated the degree to which youth actually offend. Given the evidence that the predictive usefulness of psychopathy may depend on the source of recidivism information (Cauffman et al., 2009), more studies are needed with different methods to assess criminal behavior (e.g., arrest, conviction, self-report). Second, because we collected YPI data only at the initial assessment, we do not know whether recidivists have a high YPI score at the end of the follow-up period. Thus, future research should simultaneously assess psychopathy and offending to delineate the developmental trajectory of psychopathic personality traits and antisocial conduct across a larger time span (Salekin, 2008).

In conclusion, in the current study psychopathic-like traits in detained adolescents assessed by means of the YPI did not convincingly predict recidivism. The current findings do not support the construct of psychopathy being increasingly used for legal decision making. Future research, therefore, is warranted to test whether our findings can be replicated with other measures of psychopathic-like traits than the YPI and with other informants than detained youth themselves (e.g., parents).

Footnotes

Authors’ note:

This study was supported in part by a doctoral fellowship awarded to Dr. Colins by the Special Research Fund from Ghent University for his cross-sectional PhD study on psychiatric disorders and psychopathic-like traits in detained male adolescents. The current publication uses data from this PhD study to predict recidivism. The Special Research Fund had no further role in study design; collection, analysis, and interpretation of data; the writing of the report; or the decision to submit the paper for publication.

Notes

Olivier F. Colins, PhD, is senior researcher at the Department of Child and Adolescent Psychiatry, Curium/LUMC. He is currently involved in the implementation of forensic mental health screening and assessment, in examining the relation among psychopathology, psychopathic-like traits, and adverse outcomes, and in studying empathy-related brain functioning in adolescents with aggressive conduct disorder, adolescents with autism spectrum disorder, and general population adolescents.

Robert Vermeiren, MD, is professor of child and adolescent psychiatry (Curium/LUMC) and professor of forensic youth psychiatry at the Vu University Medical Center, Amsterdam. His research focuses on forensic screening, assessment, and treatment; routine outcome monitoring; and the development and neurobiological correlates of antisocial behavior.

Marleen De Bolle, PhD, is doctor-assistant at the Department of Developmental, Personality and Social Psychology, Ghent University. She is currently involved in the study of temperamental factors, personality development, the development of internalizing and externalizing psychopathology, and structural and longitudinal associations between general and maladaptive personality traits and psychopathology across the life span.

Eric Broekaert is head of the Department of Orthopedagogics (Special Education) at Ghent University, Belgium. He is involved in the study of substance abuse and substance abuse treatment in adults and of emotional and behavioral problems in children in many different situations (e.g., clinic-referred youth, child soldiers, and nonaccompanied refugees).

References

Andershed

(2010). Stability and change of psychopathic traits: What do we know? In Salekin

R. T.

Lynam

D. R.

(Eds.), Handbook of child and adolescent psychopathy (pp. 233-250). New York, NY: Guilford.

Andershed

Hodgins

Tengstrom

(2007). Convergent validity of the Youth Psychopathic Traits Inventory (YPI)—Association with the Psychopathy Checklist: Youth Version (PCL:YV). Assessment, 14, 144-154.

Andershed

Kerr

Stattin

Levander

(2002). Psychopathic traits in non-referred youths: Initial test of a new assessment tool. In Blaauw

S. L.

(Ed.), Psychopaths: Current international perspectives (pp. 131-158). The Hague, Netherlands: Elsevier.

Anglin

M. D.

Perochet

(1998). Drug use and crime: A historical review of research conducted by the UCLA Drug Abuse Research Center. Substance Use and Misuse, 33, 1871-1914.

Asscher

J. J.

van Vugt

E. S.

Stams

G. J. J. M.

Deković

Eichelsheim

V. I.

Yousfi

(2011). The relationship between juvenile psychopathic traits, delinquency, and (violent) recidivism: A meta-analysis. Journal of Child Psychology and Psychiatry, 52, 1134-1143.

Brandt

J. R.

Kennedy

W. A.

Patrick

C. J.

Curtin

J. J.

(1997). Assessment of psychopathy in a population of incarcerated adolescent offenders. Psychological Assessment, 9, 429-435.

Caputo

A. A.

Frick

P. J.

Brodsky

S. L.

(1999). Family violence and juvenile sex offending: The potential mediating role of psychopathic traits and negative attitudes towards women. Criminal Justice and Behavior, 26, 338-356.

Cauffman

Kimonis

E. R.

Dmitrieva

Monahan

K. C.

(2009). A multimethod assessment of juvenile psychopathy: Comparing the predictive utility of the PCL:YV, YPI, and NEO PRI. Psychological Assessment, 21, 528-542.

Christian

R. E.

Frick

P. J.

Hill

N. L.

Tyler

Frazer

D. R.

(1997). Psychopathy and conduct problems in children. 2. Implications for subtyping children with conduct problems. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 233-241.

10.

Chung

Hill

K. G.

Hawkins

J. D.

Gilchrist

L. D.

Nagin

D. S.

(2002). Childhood predictors of offense trajectories. Journal of Research in Crime and Delinquency, 39, 60-90.

11.

Cohen

(1986). Research on criminal careers: Individual frequency rates and offense seriousness. In Blumstein

Cohen

Roth

J. A.

Visher

C. A.

(Eds.), Criminal careers and “career criminals” (pp. 292-418). Washington, DC: National Academy Press.

12.

Coid

Yang

Ullrich

Zhang

Roberts

. . . Farrington

(2007). Predicting and understanding risk of re-offending: The Prisoner Cohort Study. Research summary. London, UK: Ministry of Justice.

13.

Colins

Vermeiren

Schuyten

Broekaert

Soyez

(2008). Informant agreement in the assessment of disruptive behavior disorders in detained minors in Belgium: A diagnosis-level and symptom-level examination. Journal of Clinical Psychiatry, 69, 141-148.

14.

Colins

Vermeiren

Vahl

Markus

Broekaert

Doreleijers

(2011). Psychiatric disorder in detained male adolescents as risk factor for serious recidivism. Canadian Journal of Psychiatry, 56, 44-50.

15.

Cooke

D. J.

Michie

(2001). Refining the construct of psychopathy: Towards a hierarchical model. Psychological Assessment, 13, 171-188.

16.

Corrado

R. R.

Vincent

G. M.

Hart

S. D.

Cohen

I. M.

(2004). Predictive validity of the Psychopathy Checklist: Youth Version for general and violent recidivism. Behavioral Sciences and the Law, 22, 5−22.

17.

Cottle

C. C.

Lee

R. J.

Heilbrun

(2001). The prediction of criminal recidivism in juveniles—A meta-analysis. Criminal Justice and Behavior, 28, 367-394.

18.

Dadds

M. R.

Fraser

Frost

Hawes

(2005). Disentangling the underlying dimensions of psychopathy and conduct problems in childhood: A community study. Journal of Consulting and Clinical Psychology, 73, 400-410.

19.

Douglas

K. S.

Epstein

M. E.

Poythress

N. G.

(2008). Criminal recidivism among juvenile offenders: Testing the incremental and predictive validity of three measures of psychopathic features. Law and Human Behavior, 32, 423-438.

20.

Edens

J. F.

Cahill

M. A.

(2007). Psychopathy in adolescence and criminal recidivism in young adulthood: Longitudinal results from a multiethnic sample of youthful offenders. Assessment, 14, 57-64.

21.

Edens

J. F.

Campbell

J. S.

Weir

J. M.

(2006). Youth psychopathy and criminal recidivism: A meta-analysis of the psychopathy checklist measures. Law and Human Behavior, 31, 53-75.

22.

Forth

A. E.

Kosson

D. S.

Hare

R. D.

(2003). The Psychopathy Checklist: Youth Version. Toronto, Canada: Multi-Health Systems.

23.

Frick

P. J.

(2006). Developmental pathways to conduct disorder. Child and Adolescent Psychiatric Clinics of North America, 15, 311-331.

24.

Frick

P. J.

Hare

R. D.

(2001). The Antisocial Process Screening Device. Toronto, Canada: Multi-Health Systems.

25.

Frick

P. J.

Moffitt

T. E.

(2010). A proposal to the DSM-V childhood disorders and the ADHD and disruptive behavior disorders work groups to include a specifier to the diagnosis of conduct disorder based on the presence of callous–unemotional traits. Retrieved from http://www.dsm5.org/ProposedRevisionAttachments/ProposalforCallousandUnemotionalSpecifierofConductDisorder.pdf. Accessed August 2011.

26.

Frick

P. J.

Stickle

T. R.

Dandreaux

D. M.

Farrell

J. M.

Kimonis

E. R.

(2005). Callous–unemotional traits in predicting the severity and stability of conduct problems and delinquency. Journal of Abnormal Child Psychology, 33, 471-487.

27.

Frick

P. J.

White

S. F.

(2008). Research Review: The importance of callous–unemotional traits for developmental models of aggressive and antisocial behavior. Journal of Child Psychology and Psychiatry, 49, 359-375.

28.

Hare

R. D.

(1991). The Self-Report Psychopathy Scale–II. Unpublished test, University of British Columbia, Vancouver, Canada.

29.

Hare

R. D.

(2003). The Hare Psychopathy Checklist–Revised. Toronto, Canada: Multi-Health Systems.

30.

Hare

R. D.

Neumann

C. S.

(2008). Psychopathy as a clinical and empirical construct. Annual Review of Clinical Psychology, 4, 217-246.

31.

Hoeve

Blokland

Dubas

J. S.

Loeber

Gerris

J. R. M.

van der Laan

P. H.

(2008). Trajectories of delinquency and parenting styles. Journal of Abnormal Child Psychology, 36, 223-235.

32.

Kotler

J. S.

McMahon

R. J.

(2005). Child psychopathy: Theories, measurement, and relations with the development and persistence of conduct problems. Clinical Child and Family Psychology Review, 8, 291-325.

33.

Leistico

A. M. R.

Salekin

R. T.

DeCoster

Rogers

(2008). A large-scale meta-analysis relating the hare measures of psychopathy to antisocial conduct. Law and Human Behavior, 32, 28-45.

34.

Lilienfeld

S. O.

Fowler

K. A.

(2006). The self-report assessment of psychopathy: Problems, pitfalls, and promises. In Patrick

C. J.

(Ed.), Handbook of psychopathy (pp. 107-132). New York, NY: Guilford.

35.

Lynam

D. R.

(1997). Pursuing the psychopath: Capturing the fledgling psychopath in a nomological net. Journal of Abnormal Psychology, 106, 425-438.

36.

Lynam

D. R.

Widiger

T. A.

(2007). Using a general model of personality to identify the basic elements of psychopathy. Journal of Personality Disorders, 21, 160-178.

37.

Marsh

A. A.

Finger

E. C.

Mitchell

D. G. V.

Reid

M. E.

Sims

Kosson

D. S.

. . . Blair

R. J. R.

(2008). Reduced amygdala response to fearful expressions in children and adolescents with callous–unemotional traits and disruptive behavior disorders. American Journal of Psychiatry, 165, 712-720.

38.

McMahon

R. J.

Witkiewitz

K. A.

Kotler

J. S.

(2010). Predictive validity of callous–unemotional traits measured in early adolescence with respect to multiple antisocial outcomes. Journal of Abnormal Psychology, 119, 752-763.

39.

Morey

L. C.

(1991). Personality Assessment Inventory: A professional manual. Odessa, FL: Psychological Assessment Resources.

40.

Mulder

Brand

Bullens

van Marle

(2010). A classification of risk factors in serious juvenile offenders and the relation between patterns of risk factors and recidivism. Criminal Behaviour and Mental Health, 20, 23-38.

41.

Murrie

D. C.

Boccaccini

M. T.

McCoy

Cornell

D. G.

(2007). Diagnostic labeling in juvenile court: How do descriptions of psychopathy and conduct disorder influence judges? Journal of Clinical Child and Adolescent Psychology, 36, 228-241.

42.

Neumann

C. S.

Hare

R. D.

Newman

J. P.

(2007). The super-ordinate nature of the Psychopathy Checklist–Revised. Journal of Personality Disorders, 21, 102-117.

43.

Neumann

C. S.

Kosson

D. S.

Forth

A. E.

Hare

R. D.

(2006). Factor structure of the Hare Psychopathy Checklist: Youth Version (PCL:YV) in incarcerated adolescents. Psychological Assessment, 18, 142-154.

44.

Poythress

N. G.

Dembo

Wareham

Greenbaum

P. E.

(2006). Construct validity of the Youth Psychopathic Traits Inventory (YPI) and the Antisocial Process Screening Device (APSD) with justice-involved adolescents. Criminal Justice and Behavior, 33, 26-55.

45.

Rutter

(2005). Commentary: What is the meaning and utility of the psychopathy concept? Journal of Abnormal Child Psychology, 33, 499-503.

46.

Salekin

R. T.

(2008). Psychopathy and recidivism from mid-adolescence to young adulthood: Cumulating legal problems and limiting life opportunities. Journal of Abnormal Psychology, 117, 386-395.

47.

Salekin

R. T.

Neumann

C. S.

Leistico

A. M. R.

DiCicco

T. M.

Duros

R. L.

(2004). Psychopathy and comorbidity in a young offender sample: Taking a closer look at psychopathy’s potential importance over disruptive behavior disorders. Journal of Abnormal Psychology, 113, 416-427.

48.

Salekin

R. T.

Rogers

Sewell

K. W.

(1996). A review and meta-analysis of the Psychopathy Checklist and Psychopathy Checklist–Revised: Predictive validity of dangerousness. Clinical Psychology—Science and Practice, 3, 203-215.

49.

Schubert

Mulvey

Steinberg

Cauffman

Losoya

Hecker

. . . Knight

G. P.

(2004). Operational lessons from the Pathways to Desistance project. Journal of Youth Violence and Juvenile Justice, 2, 237-255.

50.

Seagrave

Grisso

(2002). Adolescent development and the measurement of juvenile psychopathy. Law and Human Behavior, 26, 219-239.

51.

Skeem

J. L.

Cauffman

(2003). Views of the downward extension: Comparing the youth version of the psychopathy checklist with the youth psychopathic traits inventory. Behavioral Sciences & the Law, 21, 737-770.

52.

Tolan

P. H.

Gorman-Smith

(1998). Development of serious and violent offending career. In Loeber

Farrington

D. P.

(Eds.), Serious and violent juvenile offenders: Risk factors and successful interventions (pp. 313-345). Thousand Oaks, CA: Sage.

53.

Verdejo-Garcia

Lawrence

A. J.

Clark

(2008). Impulsivity as a vulnerability marker for substance use disorders: Review of findings from high-risk research, problem gamblers and genetic association studies. Neuroscience and Biobehavioral Reviews, 32, 777-810.

54.

Vincent

G. M.

Odgers

C. L.

McCormick

A. V.

Corrado

R. R.

(2008). The PCL:YV and recidivism in male and female juveniles: A follow-up into young adulthood. International Journal of Law and Psychiatry, 31, 287-296.

55.

Vitacco

M. J.

Neumann

C. S.

Caldwell

M. F.

(2010). Predicting antisocial behavior in high-risk male adolescents contributions of psychopathy and instrumental violence. Criminal Justice and Behavior, 37, 833-846.

56.

Walters

G. D.

(2003). Predicting institutional adjustment and recidivism with the Psychopathy Checklist factor scores: A meta-analysis. Law and Human Behavior, 27, 541-558.

57.

Walters

G. D.

Knight

R. A.

Grann

Dahle

K. P.

(2008). Incremental validity of the Psychopathy Checklist facet scores: Predicting release outcome in six samples. Journal of Abnormal Psychology, 117, 396-405.

58.

White

H. R.

Loeber

Stouthamer-Loeber

Farrington

D. P.

(1999). Developmental associations between substance use and violence. Development and Psychopathology, 11, 785-803.