Validation of the Portuguese Version of the Antisocial Process Screening Device–Self-Report With a Focus on Delinquent Behavior and Behavior Problems

Abstract

The main objectives of the present study were to validate a Portuguese version of the Antisocial Process Screening Device–Self-Report and to evaluate the predictive importance of some constructs in discriminating between inmate delinquent youth and community youth. With a total of 760 participants, male (n = 543) and female (n = 217), divided in an inmate forensic sample (n = 250) and a community sample (n = 510), the authors were able to demonstrate psychometric properties that justify its use with the Portuguese juvenile population, in terms of factor structure, internal consistency, temporal stability, convergent validity, divergent validity, concurrent validity, and cutoff score. The predictive importance of psychopathic traits, self-reported delinquent behavior, and behavior problems on the prediction of sample membership (forensic vs. community) was established by binary logistic regression.

Keywords

assessment Antisocial Process Screening Device–Self-Report juvenile psychopathy validation

The notion that some youth transiently engage in criminal activities while others persistently make it a way of life into adulthood has been known for some time (Moffitt, 1993). Studies have systematically demonstrated that a small minority of youths are responsible for most of the serious and violent crimes (e.g., Baron, 1995) and that early offenders have significantly increased probabilities of becoming lifelong offenders (e.g., Farrington, Loeber, & Kalb, 2001). Such knowledge has led researchers to identify important variables that can help explain the phenomena of serious and persistent juvenile delinquency. One such variable is juvenile psychopathy, which until recently was almost ignored by psychopathologists and forensic psychologists (Verona, Sadeh, & Javdani, 2010). Borrowing from the established nomological network of psychopathy in adult male offenders, researchers have modified adult psychopathy assessment instruments to make them developmentally appropriate for use with youth (Frick & Hare, 2001; Hare, 1991). The presence or absence of psychopathic traits may help to identify unique etiological pathways in the development of antisocial behavior (Kotler & McMahon, 2005), and the importance of psychopathic traits in distinguishing serious and persistent antisocial youths has been gaining increasing support. Research suggests that juveniles with psychopathic-like traits begin their criminal activities earlier in life and commit more violent and nonviolent crimes (Caputo, Frick, & Brodsky, 1999; Forth, 1995; Kruh, Frick, & Clements, 2005). They also present with more severe conduct problems (CP; Frick, O’Brien, Wootton, & McBurnett, 1994) and higher levels of narcissism (Barry, Grafeman, Adler, & Pickard, 2007). The main purpose of the present study is to validate a Portuguese version of one of the most commonly used youth psychopathic traits assessment instruments (Patrick, 2010), namely, the Antisocial Process Screening Device (APSD; Frick & Hare, 2001), to facilitate and promote the investigation of this important construct in the Portuguese ethnic/cultural reality.

The APSD (Frick & Hare, 2001) is a measure specifically designed to evaluate psychopathic traits in children and adolescents. The self-report version (APSD-SR) is only used with adolescents. It was originally modeled after the Psychopathy Checklist–Revised (PCL-R; Hare, 2003) as a screening measure. There has, however, been some disagreement regarding the factor structure of the measure. While some authors (e.g., Frick et al., 1994; Pardini, Lochman, & Frick, 2003) argued for a bidimensional structure constituted by callous–unemotional (CU) traits and impulsivity–conduct problems (I-CP), studies undertaken with larger samples (e.g., Dadds, Fraser, Frost, & Hawes, 2005; Frick, Barry, & Bodin, 2000; Frick, Bodin, & Barry, 2000; Fung, Gao, & Raine, 2010) showed that the APSD can also be conceptualized as having a tridimensional structure composed of CU traits, impulsivity, and narcissism. Other authors (e.g., Fite, Greening, Stoppelbein, & Fabiano, 2009) found evidence that could support both a bidimensional and a tridimensional structure.

One of the first studies to report a three-factor structure was undertaken by Frick, Bodin, et al. (2000). Using a large community sample (n = 1,136) and a clinical sample (n = 160), confirmatory factor analysis (CFA) provided some support for the tridimensional model, χ² = 198.42, p ≤ .001, CFI = .92, NNFI = .91. However, support for the three-factor model is far from unanimous. Fritz, Ruchkin, Kaposov, and Klinteberg’s (2008) study, which involved a Russian youth inmate sample (n = 250) consisting of members who voluntarily completed the APSD-SR, found poor fit for the three-factor model, χ² = 547.8 (149), NFI = .681, CFI = .742, RMSEA = .076 (.069-.083), and two-factor model, χ² = 599.4 (103), NFI = .526, CFI = .566, RMSEA = .102 (.094-.110)[AQ: Please provide expansions for NFI, CFI, and RMSEA if required.]. As a second option, these authors used a principal components analysis (PCA) that offered some support for the three-factor model.

Muñoz and Frick (2007) assessed other psychometric properties of the APSD-SR. Whereas the internal consistency of the total APSD scale ranged from .78 to .81, the internal consistency of the subscales was generally below acceptable values for confirmatory research (.50-.68). The stability of the APSD-SR was also assessed in their study. For the total score, 1-year stability was .70 to .72, and the 2-year stability was .64. The subscales of the APSD showed less stability, ranging from .49 to .63 across 1 year and .43 to .48 across 2 years. These authors also found associations between the APSD and other measures of antisocial behavior, both concurrently and predictively, such as Self-Report of Delinquency Scale (Elliott & Ageton, 1980), in particular on the impulsivity dimension.

Frick, Barry, et al. (2000) found that the association between the APSD and conduct disorder symptoms was .48 for the total score, .65 for narcissism, .58 for impulsivity, and .52 for the CU factor. According to Frick, Barry, et al., these correlations with conduct disorder provide good evidence regarding the concurrent validity of the APSD-SR. In another study, Lee, Vincent, Hart, and Corrado (2003) examined the concurrent validity of the APSD-SR and Psychopathy Checklist–Youth Version (PCL-YV; Hare, 2003). They reported a moderate, significant correlation between the two measures. However, a low predictive efficiency for the APSD-SR predicting the PCL-YV was found using a cut score of 25. The authors reexamined their results using a median split for age and found the predictive efficiency for the 17-to-19 age group to be good, and for the 14-to-16 age group to be no different than chance.

It is unquestionable that the concept of juvenile psychopathy has progressively been gaining importance in forensic theory and practice (Salekin & Lynam, 2010). The research effort that has been put into developing measures that tap the psychopathy construct in children and adolescents is worthwhile if we have in mind the possibilities of early identification and treatment (Salekin, 2010). However, the use of the psychopathy construct, originally developed for adult male populations, still spawns some controversies when considered in children and adolescents (Seagrave & Grisso, 2002), so further investigation is needed. The purposes of the present study, integrated in a larger exploratory investigation with the aim of researching Portuguese institutionalized juvenile delinquents, were multiple: to validate a Portuguese version of the APSD-SR, to promote the investigation of the generalizability of the construct of psychopathy to the Portuguese ethnic/cultural reality, and to evaluate the predictive importance of some constructs, namely, psychopathic traits, self-reported delinquent behavior, and behavior problems in discriminating between inmate delinquent youth and community youth.

Method

Participants

The forensic sample consisted of inmates recruited from six nationwide juvenile detention centers belonging to the Portuguese Ministry of Justice. A total of 250 participants (age range = 13-20 years, mean age = 15.81 years, SD = 1.32 years), male (n = 221, age range = 13-20 years, mean age = 15.86 years, SD = 1.31 years) and female (n = 29, age range = 13-18 years, mean age = 15.45 years, SD = 1.35 years), agreed to voluntarily participate in the study. They were detained by the court’s decision.

The community sample was randomly recruited from public schools. A total of 510 participants (age range = 12-20 years, mean age = 15.92 years, SD = 1.48 years), male (n = 322, age range = 12-20 years, mean age = 16.03 years, SD = 1.62 years) and female (n = 188, age range = 13-20 years, mean age = 15.87 years, SD = 1.41 years), agreed to participate after being informed that it was voluntary and completely confidential.

The forensic and community participants statistically differed on some moderator variables. The forensic sample had fewer females, χ² = 5.484, p ≤ .001; fewer White Europeans, χ² = 38.776, p ≤ .001; fewer urban-background participants, χ² = 18.580, p ≤ .001; fewer years of education, F = 1194.506, p ≤ .001; lower parent’s socioeconomic status, U = 33514, p ≤ .001; and more divorced or deceased parents, χ² = 127.898, p ≤ .001. No statistical differences between the forensic and community participants were found regarding age and nationality.

Measures

The Antisocial Process Screening Device–Self-Report (APSD-SR; Frick & Hare, 2001) is a 20-item measure designed to assess psychopathic-like traits in adolescents. Originally called the psychopathy screening device, it was modeled after the PCL-R (Hare, 2003). Each item is scored on a 3-point ordinal scale labeled 0 (not at all true), 1 (sometimes true), or 2 (definitely true). Higher scores signify an increased presence of the traits in question. The total score, as well as each dimension score, is obtained by adding the respective items. Some studies (e.g., Frick et al., 1994) reported two main factors: a CU factor comprising 6 items (tapping interpersonal and affective dimensions of psychopathy such as lack of guilt and absence of empathy) and an I-CP factor (containing 10 items tapping overt behavioral dimensions of CP and poor impulse control). Other studies (e.g., Frick, Barry, et al., 2000) reported three main factors—the CU factor (which remained almost the same) and the I-CP factor, which appeared to subdivide into two further factors: narcissistic and impulsive traits. Higher scores indicate an increased presence of the characteristics associated with each factor. Internal consistency reliability for the present study, estimated by Cronbach’s alpha, was as follows: APSD total = .75, I-CP = .77, CU = .56, Nar = .68, and Imp = .47.

The Adapted Self-Reported Delinquency Scale (ASDS; Carroll, Durkin, Houghton, & Hattie, 1996) is a 38-item measure that assesses juvenile criminal behaviors scored on a 3-point scale 0 (never), 1 (sometimes), or 2 (frequently). The total score is obtained by adding the items. Higher scores mean higher frequency of criminal activity. Internal consistency reliability for the present study, estimated by Cronbach’s alpha, was very high at .96.

The Strengths and Difficulties Questionnaire (SDQ; Goodman, Meltzer, & Bailey, 1998)–Self-Response Version is a short behavioral problems questionnaire that assesses children and adolescents aged between 11 and 16 years. The 25 ordinal items reflect five dimensions: emotional symptoms (ES), CP, hyperactivity (H), peer problems (PP), and prosocial behavior (P). Responses are scored as 0 (not at all true), 1 (sometimes true), or 2 (definitely true). Each dimension score is obtained by adding the respective items. Internal consistency reliability for the present study, estimated by Cronbach’s alpha, was ES = .51, CP = .46, H = .52, PP = .43, P = .61. These values are low but still acceptable for research purposes (DeVellis, 1991).

The Marlowe–Crowne Social Desirability Scale–Short Form (MCSDS-SF; Ballard, 1992) was developed from the original MCSDS (Crowne & Marlowe, 1960) scale and is used to assess the tendency to give socially desirable responses on self-reports. It has 13 items scored either 0 (no) or 1 (yes) with the total score obtained by adding the items. Higher scores reflect the tendency to provide more socially desirable responses. Internal consistency for the present study, estimated by Kuder–Richardson, was .60.

The Rosenberg Self-Esteem Scale (RSES; Corcoran & Fischer, 2000; Rosenberg, 1979, 1989) is a self-report, one-dimensional measure that assesses self-esteem in adolescents and adults. The RSES has 10 items scored on a 4-point ordinal scale on a range from 0 (strongly disagree) to 3 (strongly agree). The total score is obtained by adding the items with higher scores reflecting higher levels of self-esteem. Internal consistency reliability for the present study, estimated by Cronbach’s alpha, was .79.

The Child and Adolescent Taxon Scale (CATS; Harris, Rice, & Quinsey, 1994; Quinsey, Harris, Rice, & Cormier, 2006) is an actuarial rating scale developed from variables related to childhood and adolescent antisocial and aggressive characteristics (e.g., childhood aggression problem, arrested below the age of 16). This scale has eight items scored either 0 (no) or 1 (yes). The total score is obtained by adding the items. Higher scores mean higher psychopathic characteristics. Because this is an actuarial scale, no internal consistency reliability was estimated.

Procedure

Authorization to translate and validate the APSD-SR for a Portuguese population was obtained from the first author of the scale. Appropriate procedures (Hambleton, 2001; Van de Vijver & Hambleton, 1996) were followed during the translation and retroversion. The study was approved by the ethics committee of the Portuguese Ministry of Justice. The forensic sample was recruited from six nationwide juvenile detention centers under the management of the Portuguese Ministry of Justice. The measures were administered by means of individual face-to-face interviews in an appropriate setting. The community sample was randomly recruited from schools in Lisbon (Portugal). Participants who were unable or unwilling to complete the measures were excluded.

Results

The first step when attempting to validate the APSD-SR in a different culture is to confirm the factor structure obtained by Frick, Barry et al. (2000) in previous studies. So the initial attempt to replicate the factor structure of the APSD-SR was undertaken using EQS 6.1 (Bentler, 2004). However, no support was found for either the original two-factor structure, χ² = 865.12, p ≤ .001; χ²/df = 5.12; GFI = .88; CFI = .72; RMSEA = .074 (.069-.079), or the more recent three-factor structure, χ² = 552.85, p ≤ .001; χ²/df = 4.19; GFI = .92; CFI = .77; RMSEA = .065 (.059-.070), by means of CFA due to poor fits, although the fit regarding the three-factor structure was much more tolerable. The removal of participants who scored higher on social desirability had virtually no effect on the factor structure. Because the confirmatory approach was not viable using the present sample, it was then decided to use IBM SPSS v19 (IBM SPSS, 2010) CPA to explore the empirical dimensions of the present sample. The use of an exploratory procedure was also considered appropriate because of the different factor structures obtained by Frick et al. 1994; Frick, Barry et al. 2000.

The Kaiser–Myer–Olkin measure of sampling adequacy (.83) and Bartlett Test of Sphericity (p ≤ .001) indicated the suitability of the data for factor analysis. Preliminary PCA without rotation was undertaken using a criterion of greater than or equal to .30 as the level of loading significance (Nunnally & Bernstein, 1994). The PCA suggested a two-factor solution by both the eigenvalue and scree test criteria. A two-component solution was subsequently forced with the components accounting for 28.13% of the common variance in scale items. Loadings for each component are set out in Table 1. The correlation matrix regarding these factors revealed positive correlations, some of which were strong (see Table 2).

Table 1.

APSD-SR Item Loadings

		Factor 1	Factor 2
Item 1	Blames others for mistakes	.57
Item 2	Engages in illegal activities	.54
Item 3	Concerned about schoolwork (R)		.35
Item 4	Acts without thinking	.35
Item 5	Shallow emotions	.45
Item 6	Lies easily and skillfully	.53
Item 7	Keeps promises (R)		.54
Item 8	Brags about accomplishments	.52
Item 9	Gets bored easily	.36
Item 10	Uses or cons others	.74
Item 11	Teases other people	.59
Item 12	Feels bad or guilty (R)		.59
Item 13	Risky and dangerous behaviors	.55
Item 14	Charming in insincere ways	.56
Item 15	Becomes angry when corrected	.30
Item 16	Thinks he is more important	.56
Item 17	Does not plan ahead	.46
Item 18	Concerned about feelings of others (R)		.59
Item 19	Shows emotions (R)		.32
Item 20	Keeps same friends (R)		.55
Eigenvalue		3.90	1.73
Variance		19.50%	8.63%

Note: APSD-SR = Antisocial Process Screening Device–Self-Report; R = Portuguese version reversible items, Loadings absent if <.30.

Table 2.

Correlation Matrix

	APSD total	N-I	CU	Nar	Imp
APSD total	1
N-I	.91**	1
CU	.57**	.18**	1
Nar	.77**	.87**	.10**	1
Imp	.75**	.83**	.14**	.52**	1

Note: APSD = Antisocial Process Screening Device; N-I = narcissism–impulsivity; CU = callous–unemotional; Nar = narcissism; Imp = impulsivity.

Significant at .01 level Pearson’s r.

Reliability and Validity

The next step was the estimation of Cronbach’s alpha, mean interitem correlation, and corrected item-total correlation range (see Table 3). The 3-month stability for the forensic sample revealed a strong (r = .80, p ≤ .01) statistically significant correlation. Only 88 participants completed the survey at Time 2, mostly due to many being transferred to detention centers located elsewhere in the country or having completed their detention sentence. The convergent validity of the APSD-SR and its dimensions with CATS revealed moderately low correlations that were statistically significant; the divergent validity of the APSD-SR and its dimensions with RSES revealed mostly very weak or nonexisting associations (see Table 4). The concurrent validity of the APSD with a Diagnostic and Statistical Manual of Mental Disorders–Text Revision (DSM-IV-TR; American Psychiatric Association, 2000) conduct disorder diagnosis revealed the existence of some moderately statistically significant positive point-bisserial correlations (see Table 5). The presence of the conduct disorder diagnosis was coded as 1 and its absence as 0. The diagnosis of conduct disorder was made by the first author of this manuscript. The known-groups validity of the APSD-SR and its factors was calculated using Wilks’s Lambda. Statistically significant differences were found between the forensic sample and the community sample, Wilks’s λ = .639, χ² = 338.05 (4), p ≤ .001.

Table 3.

Cronbach’s Alpha, Mean Interitem Correlation, and Corrected Item-Total Correlation Range

	Full sample	Forensic sample	Community sample
APSD total
Cronbach’s α	.75	.70	.71
MIIC	.13	.11	.12
CITCR	−.03-.55	−.06-.48	−.09-.59
N-I
Cronbach’s α	.77	.75	.77
MIIC	.20	.19	.21
CITCR	.23-.57	.15-.52	.24-.61
CU
Cronbach’s α	.56	.54	.46
MIIC	.17	.16	.13
CITCR	.20-.42	.07-.43	.10-.30
Nar
Cronbach’s α	.68	.67	.67
MIIC	.24	.24	.24
CITCR	.21-.54	.17-.53	.22-.58
Imp
Cronbach’s α	.47	.40	.52
MIIC	.15	.12	.18
CITCR	.16-.35	.09-.25	.25-.34

Note: APSD = Antisocial Process Screening Device; MIIC = mean interitem correlation; CITCR = corrected item-total correlation range; N-I = narcissism–impulsivity; CU = callous–unemotional; Nar = narcissism; Imp = impulsivity.

Table 4.

Convergent Validity With CATS and Divergent Validity With RSES

Pearson’s r	CATS	p value	RSES	p value
APSD total	.34	≤.01	−.16	≤.01
N-I	.28	≤.01	−.12	≤.01
CU	.23	≤.01	−.15	≤.01
Nar	.20	≤.01	−.00	ns
Imp	.22	≤.01	−.23	≤.01

Note: APSD = Antisocial Process Screening Device; CATS = Child and Adolescent Taxon Scale; RSES = Rosenberg Self-Esteem Scale; Pearson’s r = Pearson’s correlation; N-I = narcissism–impulsivity; CU = callous–unemotional; Nar = narcissism; Imp = impulsivity; ns = not significant.

Table 5.

Concurrent Validity of the APSD With DSM-IV-TR CD Diagnosis

r_pb	DSM-IV-TR CD	p value
APSD total	.33	≤.01
N-I	.29	≤.01
CU	.18	≤.01
Nar	.22	≤.01
Imp	.23	≤.01

Note: APSD = Antisocial Process Screening Device; DSM-IV-TR = Diagnostic and Statistical Manual of Mental Disorders–Text Revision; CD = conduct disorder; r_pb = point-bisserial correlation; N-I = narcissism–impulsivity; CU = callous–unemotional; Nar = narcissism; Imp = impulsivity.

The optimum cutoff score was estimated using logistic regression from which sensitivity, specificity, and receiver operating characteristic (ROC) area were estimated for different cutoffs (see Table 6). These three parameters were used simultaneously to improve classification efficiency. The chosen cutoff score was 12, which has the best values on these three parameters.

Table 6.

Sensibility, Specificity, and ROC Area for the APSD-SR

	Sensitivity (%)	Specificity (%)	ROC area	p value*
CS 7	0	100	.60	≤.001
CS 8	0	99.6	.63	≤.001
CS 9	0	100	.66	≤.001
CS 10	0	99.8	.69	≤.001
CS 11	0	99.8	.71	≤.001
CS 12	74.4	68.8	.72	≤.001
CS 13	62.4	75.9	.69	≤.001
CS 14	0.4	99.6	.68	≤.001
CS 15	49.6	85.1	.67	≤.001
CS 16	43.2	89.6	.66	≤.001
CS 17	35.6	92.7	.64	≤.001
CS 18	29.6	94.7	.62	≤.001
CS 19	24	95.9	.60	≤.001

Note: ROC = receiver operating characteristic; APSD-SR = Antisocial Process Screening Device–Self-Report; CS = cutoff score.

Null hypothesis: True area = .5.

Using a binary logistic regression model, we analyzed the importance of some predictive variables—psychopathic traits, behavioral problems, self-reported delinquent behaviors, social desirability, and sex—in distinguishing between members of the forensic group and the community group. This was done as a part of the Portuguese validation of the instruments that measure these constructs and to analyze their potential value to the current psychometric assessment being done at the Portuguese youth detention centers. The dependent variable group was coded 0 (community group) or 1 (forensic group). The predictive variable gender was coded 0 (female) or 1 (male). Table 7 shows the independent variables that were statistically significant in the predictive model. Multicollinearity was checked with VIF. Only the SDQ H and the MCSDS-SF (social desirability) failed to reach statistical significance. The model was also used to classify the participants, and an overall correct classification rate of 90.6% was reached. The model had 80% sensitivity and 95.8% specificity. Sensitivity and specificity were calculated as a part of the binary logistic regression procedures.

Table 7.

Binary Logistic Regression Coefficients for the Forensic and Community Groups

Variables	B	SE	Wald	Exp(B)	p value
APSD N-I	−0.195	0.045	18.374	0.822	≤.001
APSD CU	0.372	0.073	25.838	1.451	≤.001
SDQ ES	0.230	0.089	6.628	1.259	≤.01
SDQ CP	0.441	0.092	22.726	1.555	≤.001
SDQ H	0.001	0.068	0.000	1.001	=.983
SDQ PP	0.237	0.086	7.512	1.268	≤.01
SDQ P	0.355	0.092	14.834	1.426	≤.001
ASDS	0.181	0.018	101.297	1.199	≤.001
MCSDS-SF	0.107	0.061	3.082	1.113	=.079
Gender	0.784	0.339	5.352	2.190	≤.05
Constant	−11.646	1.591	53.536	0.000	≤.001

Note: APSD = Antisocial Process Screening Device; N-I = narcissism–impulsivity; CU = callous– unemotional; SDQ = Strengths and Difficulties Questionnaire; ES = emotional symptoms; CP = conduct problems; H = hyperactivity; PP = peer problems; P = prosocial behavior; ASDS = Adapted Self-Report Delinquency Scale; MCSDS-SF = Marlowe–Crowne Social Desirability Scale–Short Form.

Due to the predictive variable Gender being statistically significant, it was decided to perform an additional binary logistic regression for males only, given the small sample size of the female forensic group (n = 29) that puts into question the accuracy of the results (Leech, Barrett, & Morgan, 2008). Table 8 shows the independent variables that were statistically significant in the predictive model. Only the SDQ H, the SDQ PP, and the MCSDS-SF (social desirability) failed to reach statistical significance. The model had an overall correct classification rate of 90.2%, a sensitivity of 83.7%, and a specificity of 94.7%.

Table 8.

Binary Logistic Regression Coefficients for the Male Forensic and Community Groups

Variables	B	SE	Wald	Exp(B)	p value
APSD N-I	−0.239	0.051	21.777	0.786	≤.001
APSD CU	0.424	0.083	26.088	1.529	≤.001
SDQ ES	0.397	0.114	11.998	1.487	≤.001
SDQ CP	0.468	0.108	18.632	1.597	≤.001
SDQ H	0.007	0.076	0.010	1.007	=.919
SDQ PP	0.160	0.098	2.653	1.174	=.103
SDQ P	0.315	0.101	9.757	1.371	≤.01
ASDS	0.168	0.018	83.846	1.183	≤.001
MCSDS-SF	0.095	0.070	1.830	1.099	=.176
Constant	−10.393	1.788	33.767	0.000	≤.001

Discussion

The present study had as its first purpose analysis of the psychometric properties of the APSD-SR with a sample of Portuguese youths, a measure that has been used in several past studies to assess traits associated with psychopathy (e.g., Caputo et al., 1999; Muñoz & Frick, 2007). PCA revealed a two-factor structure that was somewhat similar, albeit not identical (all the items in our sample loaded on at least one of the two factors), to the one found by Frick et al. (1994). The issue of factor structure is important due to the ongoing process of how best to define psychopathy (e.g., categorically vs. dimensionally) and measure it. Evidence obtained in our study shows that the two-factor model seems to be the most acceptable one for the Portuguese version of the APSD-SR.

The first factor found was a mixed one similar to the poor impulse control–conduct problems. The mixed factor (Items 1, 2, 4, 5, 6, 8, 9, 10, 11, 13, 14, 15, 16, and 17) can be further separated into narcissism (Items 5, 8, 10, 11, 14, 15, and 16) and impulsivity (Items 1, 4, 9, 13, and 17). Items 2 and 6, which originally did not load on any of the APSD factors (Muñoz & Frick, 2007), did load on the mixed factor. The second factor (Items 3, 7, 12, 18, 19, and 20) was similar to the CU traits factor also reported by Frick et al. (1994). Other authors (e.g., Fite et al., 2009; Pardini et al., 2003) have found supporting evidences for the two-dimensional nature of psychopathy in youths as the best and more parsimonious option, and this seems to apply also to the Portuguese cultural/ethnic reality.

Some psychometric problems were found. Analysis of the internal consistency revealed exceedingly low values (Cortina, 1993) for the CU factor and the impulsivity factor, similar to those found by Muñoz and Frick (2007). Such low values put into question the fidelity of measurements for these two factors when taken separately. Regarding the mean interitem correlations, some problems were also found that reveal item heterogeneity. The APSD total score did not reach the recommended minimum value of .15 (Clark & Watson, 1995) in most of the samples, despite the fact that APSD factors did reach it. Regarding the corrected item-total correlation range, the APSD total score and its factors did not reach, in most of the samples, the minimum recommended value of .20 (Nunnally & Bernstein, 1994), which indicates some weak associations between the items.

There were also some more positive findings. The 3-month stability reached a statistically significant value of .80, which is considered a good result (Kline, 2000). This result is even better than the one obtained by Muñoz and Frick (2007) in their study of the stability of the APSD-SR. The convergent validity of the APSD and its factors with CATS revealed moderate correlations, always statistically significant; the strongest correlation was obtained with the APSD total score, demonstrating the expected construct overlap (DeVellis, 1991; Kline, 2000). The divergent validity with the RSES showed mostly nonexisting or weak associations, with the exception of the impulsivity factor.

The concurrent validity of the APSD and its factors with DSM-IV-TR’s conduct disorder diagnosis showed modest statistically significant correlations; the strongest correlation was also obtained with the APSD total score. These correlations fell somewhat short of those obtained by Frick, Barry, et al. (2000), which while higher than what we observed were nonetheless similar to the ones obtained by other authors (e.g., Fung et al., 2010). The discriminant validity revealed that the APSD and its factors could significantly discriminate between the forensic group and the community group, conceptualized as structurally different and mutually exclusive (Maroco, 2010). The appraisal of the most balanced cutoff score took into account three criteria: sensitivity, specificity, and ROC area; the chosen cutoff score was 12, due to the best balance of the three parameters.

Another purpose of this study was to evaluate the importance of some variables in discriminating between inmate delinquent youth and community youth. The first regression model (for both the male and female genders) offered support for the importance of research variables such as psychopathic traits (e.g., Kruh et al., 2005), delinquent behavior (e.g., Forth, 1995), and behavior problems (e.g., Frick et al., 1994) in predicting group membership of delinquent youths and community youths. The second regression model (for the male gender only) reached almost the same results. The fact that the social desirability predictive variable did not reach statistical significance indicates that it did not influence the participant’s responses in a significant manner.

Overall, these findings provide some additional support for the extension of the psychopathy construct to adolescents and its potential generalization across different cultures and ethnic groups. We must conclude that it was possible to demonstrate some appropriate psychometric properties that justify the future use of the APSD-SR with the Portuguese youth population. Some nuances in the factor structure were detected, but these are consistent with the argument of some authors (e.g., Fritz et al., 2008) that the factor structure of the APSD-SR may vary somewhat between cultures. To our knowledge, this is the first study that attempts to extend the validation of the APSD to the Portuguese language and society.

Conclusions

We were able to demonstrate the utility of the APSD-SR as a screening measure generalizable to the Portuguese cultural/ethnic reality. It can serve as a means of early identification of children and adolescents with high psychopathic traits and thus promote a basis for effective interventions that can save expenses on future rehabilitation. Still some further validation procedures are recommended and should be done in the future (e.g., cross-validation using other samples, concurrent validity with the PCL-YV).

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author.

Ballard

(1992). Short forms of the Marlowe-Crowne Social Desirability Scale. Psychological Reports, 71, 1155-1160.

Baron

(1995). Serious offenders. In Creechman

Silverman

(Eds.), Canadian delinquency (pp. 135–147). Scarborough, Canada: Prentice Hall.

Barry

Grafeman

Adler

Pickard

(2007). The relations among narcissism, self-esteem, and delinquency in a sample of at-risk adolescents. Journal of Adolescence, 30, 933-942.

Bentler

(2004). EQS 6 Structural Equations Program Manual. Encino, CA: Multivariate Software.

Caputo

Frick

Brodsky

(1999). Family violence and juvenile sex offending. Criminal Justice and Behavior, 26, 338-356.

Carroll

Durkin

Houghton

Hattie

(1996). An adaptation of Mak’s self-reported delinquency scale for Western Australian adolescents. Australian Journal of Psychology, 48, 1-7.

Clark

Watson

(1995). Constructing validity: Basic issues in objective scale development. Psychological Assessment, 7, 309-319.

Corcoran

Fischer

(2000). Measures for clinical practice: A sourcebook (3rd ed., Vol. 1). New York, NY: Free Press.

10.

Cortina

(1993). What is coefficient alpha? An examination of theory and applications. Journal of Applied Psychology, 78, 98-104.

11.

Crowne

Marlowe

(1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology, 24, 349-354.

12.

Dadds

M. R.

Fraser

Frost

Hawes

D. J.

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

13.

DeVellis

(1991). Scale development: Theory and applications. London, England: SAGE.

14.

Elliott

Ageton

(1980). Reconciling race and class differences in self-reported and official estimates of delinquency. American Sociological Review, 45, 95-110.

15.

Farrington

Loeber

Kalb

(2001). Key research and policy issues. In Loeber

Farrington

(Eds.), Child delinquents: Development, intervention, and service needs (pp. 385-394). Thousand Oaks, CA: SAGE.

16.

Fite

Greening

Stoppelbein

Fabiano

(2009). Confirmatory factor analysis of the antisocial process screening device with a clinical inpatient population. Assessment, 16, 103-114.

17.

Forth

(1995). Psychopathy and young offenders. Ottawa, Ontario: Ministry of the Solicitor General of Canada.

18.

Frick

Barry

Bodin

(2000). Applying the concept of psychopathy to children: Implications for the assessment of antisocial youth. In Gacono

(Ed.), The clinical and forensic assessment of psychopathy: A practitioner’s guide (pp. 1-24). Mahwah, NJ: Lawrence Erlbaum.

19.

Frick

Bodin

Barry

(2000). Psychopathic traits and conduct problems in community and clinic-referred samples of children: Further development of the psychopathy screening device. Psychological Assessment, 12, 382-393.

20.

Frick

Hare

(2001). The antisocial process screening device. Toronto, Ontario, Canada: Multi-Health System.

21.

Frick

O’Brien

Wootton

McBurnett

(1994). Psychopathy and conduct problems in children. Journal of Abnormal Psychology, 103, 700-707.

22.

Fritz

Ruchkin

Kaposov

Klinteberg

(2008). Antisocial process screening device: Validation on a Russian sample of juvenile delinquents with the emphasis on the role of personality and parental rearing. International Journal of Law and Psychiatry, 31, 438-446.

23.

Fung

Gao

Raine

(2010). The utility of the child and adolescent psychopathy construct in Hong Kong, China. Journal of Clinical Child and Adolescent Psychology, 39, 134-140.

24.

Goodman

Meltzer

Bailey

(1998). The strengths and difficulties questionnaire: A pilot study on the validity of the self-report version. European Child & Adolescent Psychiatry, 7, 125-130.

25.

Hambleton

(2001). The next generation of the ITC test translation and adaptation guidelines. European Journal of Psychological Assessment, 17, 164-172.

26.

Hare

(1991). The Hare Psychopathy Checklist–Revised. Toronto, Ontario, Canada: Multi-Health Systems.

27.

Hare

(2003). The Hare Psychopathy Checklist–Revised: Technical manual (2nd ed.). Toronto, Ontario, Canada: Multi-Health Systems.

28.

Harris

Rice

Quinsey

(1994). Psychopathy as a taxon: Evidence that psychopaths are a discrete class. Journal of Consulting and Clinical Psychology, 62, 387-397.

29.

IBM SPSS. (2010). IBM SPSS Statistics Base 19. Chicago, IL: SPSS Inc.

30.

Kline

(2000). The handbook of psychological testing. London, England: Routledge.

31.

Kotler

McMahon

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

32.

Kruh

Frick

Clements

(2005). Historical and personality correlates to the violence patterns of juveniles tried as adults. Criminal Justice and Behavior, 92, 69-96.

33.

Lee

Vincent

Hart

Corrado

(2003). The validity of the antisocial process screening device as a self-report measure of psychopathy in adolescent offenders. Behavioral Sciences & the Law, 21, 771-786.

34.

Leech

Barrett

Morgan

(2008). SPSS for intermediate statistics. New York, NY: LEA.

35.

Maroco

(2010). Statistical analysis with PASW Statistics (ex-SPSS). Pero Pinheiro: ReportNumber.

36.

Moffitt

(1993). Adolescent-limited and life-persistent antisocial behaviour: A developmental taxonomy. Psychological Review, 100 (4), 674-701.

37.

Muñoz

Frick

(2007). The reliability, stability, and predictive utility of the self-report version of the antisocial process screening device. Scandinavian Journal of Psychology, 48, 299-312.

38.

Nunnally

Bernstein

(1994). Psychometric theory. New York, NY: McGraw-Hill.

39.

Pardini

Lochman

Frick

(2003). Callous/unemotional traits and social-cognitive processes in adjudicated youths. Journal of American Academy of Child & Adolescent Psychiatry, 42, 364-371.

40.

Patrick

(2010). Conceptualizing the psychopathic personality: Disinhibited, bold, or just plain mean? In Salekin

Lynam

(Eds.), Handbook of child and adolescent psychopathy (pp. 15-48). New York, NY: Guilford.

41.

Quinsey

Harris

Rice

Cormier

(2006). Violent offenders: Appraising and managing risk (2nd ed.). Washington, DC: American Psychological Association.

42.

Rosenberg

(1979). Conceiving the self. New York, NY: Basic Books.

43.

Rosenberg

(1989). Society and the adolescent self-image (Rev. ed.). Middletown, CT: Wesleyan University Press.

44.

Salekin

(2010). Treatment of child and adolescent psychopathy: Focusing on change. In Salekin

Lynam

(Eds.), Handbook of child and adolescent psychopathy (pp. 343-373). New York, NY: Guilford.

45.

Salekin

Lynam

(2010). Child and adolescent psychopathy: An introduction. In Salekin

Lynam

(Eds.), Handbook of child and adolescent psychopathy (pp. 1-12). New York, NY: Guilford.

46.

Seagrave

Grisso

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

47.

Van

Vijver

Hambleton

(1996). Translating tests: Some practical guidelines. European Psychologist, 1, 89-99.

48.

Verona

Sadeh

Javdani

(2010). The influences of gender and culture on child and adolescent psychopathy. In Salekin

Lynam

(Eds.), Handbook of child and adolescent psychopathy (pp. 317-342). New York, NY: Guilford.