- Email: [email protected]
Delinquency in male adolescents: The role of alexithymia and family structure
ARTICLE IN PRESS
Journal of Adolescence Journal of Adolescence 29 (2006) 321–332 www.elsevier.com/locate/jado
Delinquency in male adolescents: The role of alexithymia and family structure Gre´goire Zimmermann Department of Psychology and Institute for Psychotherapy, University of Lausanne, BFSH 2—Dorigny, CH-1015 Lausanne, Switzerland
Abstract The purpose of this study was to examine the linkages between alexithymia and delinquency in male adolescents (age ranging from 14 to 18 years), and to investigate whether alexithymia was a good discriminatory factor for juvenile delinquency. Thirty-six offender adolescents and 46 non-offender control adolescents participated in the study and completed the 20-item Toronto Alexithymia Scale (TAS-20) (alexithymia), the Revised Children’s Manifest Anxiety Scale (R-CMAS) (anxiety), the Liste d’Adjectifs Bipolaires et en Echelles de Likert (LABEL) (personality—Big Five) and demographic data. Findings revealed that the adolescents of the offender group scored high on alexithymia and that proportion of disrupted family structure in the offender group is higher than in the control group. Logistic regressions confirmed that alexithymia and family structure are the strongest discriminatory factors for juvenile delinquency. Limitations and clinical implications are discussed, and recommendations for future research are provided. r 2005 The Association for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved.
Introduction In the psychoanalytic literature, acting-out and other means of enacted repetitions, typical of juvenile offenders, are often conceptualized as the result of a deficit in mental elaboration Corresponding author. Tel.: +41 21 692 32 60; fax: +41 21 692 32 65.
E-mail address: [email protected]. 0140-1971/$30.00 r 2005 The Association for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.adolescence.2005.08.001
ARTICLE IN PRESS 322
G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
(Cahn, 1987). Currently, several authors consider the concept of alexithymia to be related to this deficit in mental elaboration (McDougall, 1982; Quartier & Zimmermann, 2004; Tardif, 1998). The construct of alexithymia, literally without word for feelings, is based on observations of psychosomatic patients and was first introduced in the early 1970s by Sifneos (1973). The alexithymia construct covers a cluster of cognitive and affective features, mainly characterized by a difficulty in identifying and expressing feelings, a striking paucity of fantasies, and an externally oriented cognitive style (Nemiah, Freyberger, & Sifneos, 1976). These features are thought to reflect a deficit in the cognitive processing and regulation of emotional states (Taylor, Bagby, & Parker, 1997). Alexithymia was first studied in classical psychosomatic or somatic disorders. However, during the last decade, it has been investigated from other perspectives (Taylor & Bagby, 2004), and has been associated with a great variety of medical disorders (e.g. hypertension and inflammatory bowel diseases) and psychiatric disorders (e.g. substance use and eating disorders) (Taylor, 2000; Taylor et al., 1997). Theoretical and empirical works regarding the aetiology of juvenile delinquency that include individual, familial and social factors are abundant. However, while clinicians and researchers can define groups of adolescents who might be at risk for delinquency, few authors have tried to predict why some of these at-risk youths develop delinquent behaviours, while others do not. The alexithymia construct may have implication for understanding this evolution and the development of repetitive and disruptive behaviours in juvenile offenders. Thirty years ago, Nemiah et al. (1976) indicated that alexithymic patients ‘‘may on occasion manifest brief but violent outbursts of affective behaviours’’ (p. 432), but are unaware of underlying feelings expressed. Since then, several authors have published clinical and theoretical works suggesting the presence of alexithymia among juvenile offenders (Jeammet, 2001; Marohn, 1990, 1992; Schreiber, 1992). However, empirical works studying alexithymic characteristics in this population are rare. In fact, besides two published studies (Langevin & Hare, 2001; Moriarty, Stough, Tidmarsh, Eger, & Dennison, 2001) there are, to our knowledge, no data on alexithymia in samples of juvenile offenders. Moriarty et al. (2001) investigated emotional intelligence in 15 male adolescent sex offenders and 49 male non-offenders and showed that sex offenders presented difficulty in identifying feelings and were more aggressive than control adolescents. However, even if the results indicated high scores on the alexithymia scale (20-item Toronto Alexithymia Scale (TAS-20)) (Bagby, Parker, & Taylor, 1994; Bagby, Taylor, & Parker, 1994), there were no significant differences between the two groups. In their study, Langevin and Hare (2001) examined the link between psychopathy, alexithymia and ability for symbolization in a sample of 60 male adolescents (40 offenders and 20 controls). As expected, they reported that alexithymia correlated positively with psychopathy and negatively with the capacity for symbolization. Moreover, descriptive statistics showed that juvenile offenders recorded higher score on TAS-20 than the control group. Finally, there is some evidence suggesting that juvenile offenders present an impairment of their facial affect recognition ability (McCown, Johnson, & Austin, 1986, 1988). This ability is commonly associated with alexithymia (Lane, Lee, Riedel, Shapiro, & Kaszniak, 2000; Lane et al., 1996). The aim of the present study is to examine the relative contribution of alexithymia to explain juvenile delinquency. Specifically, it is hypothesized that delinquents are more alexithymic than control group subjects and that alexthymia could better differentiate juvenile offenders from juvenile non-offenders than other personality, mood and demographic variables.
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
323
Method Sample Effect size (ES) for group differences on alexithymia was estimated, on the basis of previous research, as between medium and large (.50odo.80; see Cohen, 1988). With this expected ES (fixed at .65), the sample size required—for a two-tailed independent t-test—to detect the effect at a significance level of .05 with a power of .80 is 78 (Cohen, 1988). The study sample consisted of 82 male adolescents (36 offenders and 46 controls) ranging in age from 14 to 18 years (M ¼ 16:10, S.D.¼ 1:02). The non-offender adolescents were recruited in four secondary schools of the French-speaking part of Switzerland. Their mean age was 15.95 years (S.D.¼ :80). Juvenile offenders were inpatients in specialized residential facilities in the same area. Their mean age was 16.30 years (S.D.¼ 1:22). Thirty-one percent (11/36) of offenders were on remand (i.e. not currently conviced or unsentenced). Educative residential treatment varied from a few months to a few years and was ordered by the juvenile justice system. At the time of interview, the majority of the offenders (29/36–80.6%) had been treated in a residential centre for 12 months or longer. The arrest histories of the referred adolescents attest to the seriousness of their criminal involvement. The youths averaged 7.28 previous arrests (S.D.¼ 7:00) and they were quite young at the time of their first arrest (Age: M ¼ 12:61, S.D.¼ 2:41). The nature of the crimes committed by these adolescents included theft, drug trafficking, burglary and violent offences (i.e. aggravated assault, robbery, attempted rape). All adolescents included in this study were able to understand and write the French language. Our study is in compliance with the ethical code of the Swiss Society of Psychology. Measures For the purpose of the present study, we used the following questionnaires: (i) a demographic questionnaire, (ii) the TAS-20, (iii) the Revised Children’s Manifest Anxiety Scale (R-CMAS), and (iv) the Liste d’Adjectifs Bipolaires et en Echelles de Likert (LABEL). Demographic questionnaire We used a demographic questionnaire for registration of age, family structure, parent’s socioeconomic status (SES1) and nationality (Swiss citizen versus non-Swiss citizen). Concerning family structure, we differentiated intact and disrupted family structures, indicating the absence of a least one parent (i.e. children whose parents have divorced, separated or deceased). The TAS-20 Alexithymia was measured with one of the most reliable and validated self-report questionnaire: the TAS-20 (Bagby, Parker et al., 1994; Bagby, Taylor et al., 1994), in which participants indicate the extent of their agreement or disagreement with statements on a five-point Likert scale. A French-language version was applied in this study (Loas, Fremaux & Marchand, 1 SES is the International Socio-Economic Index of Occupational Status calculated according to Ganzeboom and Treiman (1996).
ARTICLE IN PRESS 324
G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
1995). We used the total score, as well as the score of the three factors: (F1) difficulty-identifying feelings, (F2) difficulty describing feelings and (F3) externally oriented thinking. Even if alexithymia is conceptualized as a dimensional construct, an indicative cutoff score has been established empirically (X61 alexithymia) to facilitate comparisons of rates of alexithymia across studies (Taylor et al., 1997). In previous research, the French version of the TAS-20 has been found to have good internal consistency, and a three-factor structure consistent with the theoretical construct (Loas et al., 2001; Zimmermann, Rossier, Meyer de Stadelhofen, & Gaillard, 2005). The R-CMAS The R-CMAS subtitled ‘‘What I think and feel’’ (Reynolds & Richmond, 1985) is a widely used 37-item instrument to capture the degree and quality of anxiety experienced by children and adolescents. The R-CMAS consists of 37 dichotomous items (yes/no) divided into four subscales: Physiological Anxiety, Worry/Oversensitivity, Social Concerns/Concentration and the Lie Scale. A Total Anxiety score can be computed summing the 28 anxiety items. The R-CMAS has been found to be a reliable, valid and useful indicator of general anxiety (Gerard & Reynolds, 1999). In this study, we used the French-language validated version of the R-CMAS (Castro, 1999). The LABEL The Five Factor Model (FFM) of personality was measured with the 87-item LABEL developed and validated by Gendre and colleagues (Gendre & Capel, 2003; Gendre, Capel, & Monod, 2002). This 87-item semantic differential scale was designed to measure, among others dimensions, the five domains of FFM. The first factor (i.e. Neuroticism) assesses affective emotional instability versus adjustment. The second factor (i.e. Extraversion) assesses the quantity and intensity of interpersonal interactions, the need for stimulation and the capacity to experience joy. The third one, Openness to experience, is defined as the proactive seeking and appreciation of various experiences for their own sake, and as toleration for and exploration of the unfamiliar. As Extraversion evaluates the degree to which an individual enjoys being in the presence of others, the fourth factor (i.e. Agreeableness) examines the attitude a person holds towards other people, describing quality of interaction preference ranging from empathy to opposition. Finally, the fifth one (i.e. Conscientiousness) assesses the individual’s degree of organization, persistence, and motivation in goal-directed behaviour. Item consists in two bi-polar adjectives rated on a fivepoint scale. Respondents are required to choose, to a certain extent, one or the other adjective depending on what is most true for them. Procedure All the questionnaires were administered (in the order described above) in a private individual structured interview lasting approximately 60 min. Prior to the interview, adolescents were given explanations emphasizing the significance of their participation, the confidentiality of all data, and the opportunity of withdrawing from research interview at any moment. Moreover, participants have been encouraged to disclose their own opinion and reassured that there were no right or wrong answers.
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
325
Analyses Comparisons between juvenile offenders and non-offenders were performed using a w2 or a Fisher’s exact test for dichotomous variables and independent Student’s t-test for continuous variables. Moreover, ES were computed following Cohen’s procedures (1988). Pearson correlations were used to assess the relations between alexithymia, anxiety and personality. Last, a series of logistic regressions were performed to determine if alexithymia could predict juvenile delinquency. For the logistic regressions, group membership (delinquent/control) served as dependent variable (criterion) and demographic, personality, and mood variables as the independent variables (predictor). We set alpha level at .05 and confidence intervals of 95%. All analyses were computed using the statistical software SPSS 11.0.
Results First, we examined the potential differences in alexithymia, anxiety, FFM, and demographic variables between adolescent offenders and non-offenders. Demographic comparisons (see Table 1) between the offender and control group showed no differences for age, nationality and parent’s SES. The adolescents of the delinquent group significantly more often came from disrupted family structure (22/36–61.1%) than those in the control group (10/46–21.7%). Table 2 displays the means and standard deviations for the TAS-20, R-CMAS, and LABEL for the delinquent and control group. Compared to the control subjects, the adolescent offenders scored significantly higher on TAS-20 total score (upper medium ES) and on TAS-20 first factor (F1: difficulty-identifying feelings (upper medium ES)). Based on the cutoff score of X61 on the TAS-20, the proportion of alexithymic individuals was significantly higher (w2 ð1Þ ¼ 5:94, po:05) Table 1 Demographic characteristics Group
Statistical test
Effect size
Control (n ¼ 46)
Delinquent (n ¼ 36)
Age (S.D.) SES (S.D.)
15.95 (.80) 38.96 (13.78)
16.30 (1.22) 39.19 (16.82)
tð80Þ ¼ 1:49 tð80Þ ¼ :07
d ¼ :35 d ¼ :01
Nationality (%) Swiss Non-Swiss
35 (76.1) 11 (23.9)
22 (61.1) 14 (38.9)
w2 ð1Þ ¼ 2:14
w ¼ :17
Family structure (%) Intact Disrupted
36 (78.3) 10 (21.7)
14 (38.9) 22 (61.1)
w2 ð1Þ ¼ 13:16
w ¼ :41
Note: SES ¼ parent’s socioeconomic status. Cohen (1988) considered: d ¼ :20 as a ‘‘small’’ effect size, d ¼ :50 as a ‘‘medium’’ effect size and d ¼ :80 as a ‘‘large’’ effect size. For w effect size index values of w for ‘‘small’’, ‘‘medium’’ and ‘‘large’’ are, respectively, .10, .30 and .50. po:05.
3.24 5.00 1.63 9.87 52.38 47.45 48.79 54.24 51.10
R-CMAS (S.D.) Physiological anxiety Worry/oversensitivity Social concerns/concentration Total
LABEL (S.D.) Neuroticismb Extraversion Openness Agreeableness Conscientiousness (20.20) (20.28) (19.48) (15.54) (20.64)
(1.98) (2.62) (1.47) (4.66)
(5.12) (4.72) (4.06) (10.23)
46.95 53.26 51.54 44.58 48.60
3.64 4.61 2.19 10.44
20.69 16.08 22.47 59.25
(19.60) (19.43) (20.82) (23.70) (19.35)
(2.17) (2.75) (1.41) (4.83)
(6.03) (3.83) (5.25) (11.56)
5.43 +5.81 +2.75 9.66 2.50
+.40 .39 +.56 +.57
+3.56 +2.02 +1.99 +7.58
(14.26 to+3.40) (3.00 to +14.63) (6.14 to +11.64) (18.78 to –.51) (11.39 to +6.40)
(.51 to +1.31) (1.58 to +.80) (.07 to +1.20) (1.52 to +2.67)
(+1.11 to +6.02) (+.09 to +3.94) (.05 to +4.04) (+2.78 to +12.37)
tð80Þ ¼ 1:22 tð80Þ ¼ 1:31 tð80Þ ¼ :62 tð80Þ ¼ 2:12 tð80Þ ¼ :55
tð80Þ ¼ :87 tð80Þ ¼ :65 tð80Þ ¼ 1:76 tð80Þ ¼ :55
tð80Þ ¼ 2:89 tð80Þ ¼ 2:09 tð80Þ ¼ 1:94 tð80Þ ¼ 3:14
Statistical testa
d d d d d
d d d d
d d d d
¼ :27 ¼ :29 ¼ :14 ¼ :48 ¼ :12
¼ :14 ¼ :15 ¼ :38 ¼ :12
¼ :64 ¼ :45 ¼ :42 ¼ :66
Effect size
Note: F1 ¼ difficulty identifying feelings; F2 ¼ difficulty describing feelings; F3 ¼ externally oriented thinking. CI ¼ confidence interval. The LABEL’s dimensions were normed by computing T-scores, which have a mean of 50 and a standard deviation of 20. po:05=4 tests ¼ .0125 for TAS-20 and R-CMAS, po:05=5 tests ¼ .01 for LABEL. a Independent t-test were performed with a Bonferroni correction for multiple comparisons. b Compared to classical measure of Big Five dimensions, Neuroticism’s scale of LABEL is in reverse order.
17.13 14.07 20.48 51.67
TAS-20 (S.D.) F1 F2 F3 Total
Delinquent (n ¼ 36)
Mean differences (95% CI)
326
Control (n ¼ 46)
Group
Table 2 Means, standard deviations, mean differences, t-tests and effect sizes for the TAS-20, R-CMAS, LABEL for delinquent and control group
ARTICLE IN PRESS
G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
327
Table 3 Correlations between alexithymia, anxiety and personality TAS-20
TAS-20 R-CMAS LABEL N E O A C
—
R-CMAS
.41* —
LABEL N
E
O
A
C
.41* .48*
.06 .13
.08 .27*
.45* .19
.27* .21
—
.39* —
.32* .68* —
.67* .01 .02 —
.72* .09 .28* .48* —
po:05.
in the delinquent group (17/36–47.2%) than in the control group (10/46–21.7%). For the RCMAS and the LABEL, there were no significant differences between the two groups. As evident from Table 3, however, the TAS-20 is correlated with these other independent variables. Therefore, a set of hierarchical logistic regression analyses was performed to determine the power of alexithymia in predicting group membership. First, predictor variables were entered into the regression analyses in four blocks: alexithymia (TAS-20 total score) was entered first in the model as a block (step 1), anxiety (R-CMAS total score) was entered on the second block (step 2), followed by five-factor model personality dimensions (step 3) and demographic variables (SES, nationality and family structure) (step 4). Alexithymia was entered first to determine how well TAS-20 total score alone was associated with juvenile delinquency. The results from this analysis are presented in Table 4. The model after step 1 was significant, w2 ð1Þ ¼ 9:48, po:05, and provided correct predictions for 55.6% of the offender adolescents and 73.9% of the control adolescents (overall correct classification rate ¼ 65.9%). The addition of anxiety (step 2) and FFM of personality variables (step 3) did not significantly increase the overall fit of the model (see Table 4). However, the introduction in the fourth block (step 4) of demographic variables did significantly increase the overall fit of the model (w2chg ð3Þ ¼ 16:56, po:05). Of the variables entered in the fourth block, only the family structure made a significant contribution to the model. The full model was significant, w2 ð10Þ ¼ 32:03, po:05, and provided correct predictions for 72.2% of the offender adolescents and 84.8% of the control adolescents (overall correct classification rate ¼ 79.3%). In a second hierarchical logistic regression, only significant discriminatory factors of the previous model were used. The TAS-20 total score was entered first into the model, followed by the family structure. This model (see Table 5) was compared to the previous full model. Goodness-of-fit statistic showed no reliable differences between the models (w2 ð8Þ ¼ 10:29, p4:05), indicating no significant enhancement to prediction of juvenile delinquency by anxiety and FFM personality variables. This parsimonious model including only alexithymia and family structure correctly classified 63.9% of the offender adolescents and 78.3% of the control adolescents (overall correct classification rate ¼ 72%). In this model, the likelihood of being in the delinquent group increased by approximately 40% for each five points increase in the TAS-20
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
328
Table 4 Results from hierarchical logistic regression predicting group membership Step
Predictor
R2
w2
df
w2chg
df
Final model B
1 2 3
4
TAS-20 total R-CMAS total Block N—LABEL E—LABEL O—LABEL A—LABEL C—LABEL
.11 .12 .17
Block SES Nationality Family structure
.32
9.48* 10.09* 15.47*
32.03*
1 2 7
10
.61 5.37
16.56*
1 5
Wald
OR
OR 95% CI
.08 .07
6.12* .69
1.09 .93
(1.02–1.16) (.78–1.10)
.06 .04 .02 .00 .04
1.54 2.32 .42 .03 1.40
.94 1.04 1.02 1.00 1.04
(.86–1.04) (.99–1.08) (.96–1.09) (.96–1.05) (.97–1.12)
.01 1.24 2.03
.13 3.13 10.14*
1.01 3.45 7.59
(.97–1.05) (.87–13.65) (2.18–26.44)
3
Note: SES ¼ parent’s socioeconomic status, CI ¼ confidence interval, OR ¼ Odds ratio, R2 ¼ Cox & Snell R2. *po:05.
Table 5 Results from parsimonious hierarchical logistic regression predicting group membership Step
1 2
Predictor
TAS-20 total Family structure
R2
.11 .23
w2
9.48* 21.74*
df
1 2
w2chg
df
*
12.26
1
Final model B
Wald
OR
OR 95% CI
.07 1.75
7.21* 11.12*
1.07 5.78
(1.02–1.12) (2.06–16.19)
Note: SES ¼ parent’s socioeconomic status, CI ¼ confidence interval, OR ¼ Odds ratio, R2 ¼ Cox & Snell R2. *po:05:
total score. Moreover, adolescents from disrupted family structure were 5.8 times more likely to belong to delinquent group than those stemmed from intact home. These two discriminatory factors were independent (w2 ð1Þ ¼ :50, p4:05). In fact, the prevalence of alexithymia was similar for these two groups (16/50–32% of adolescents stemmed from intact family structure and 11/32–34.4% of adolescents stemmed from disrupted family structure).
Discussion The juvenile delinquency is a very complex issue and there are almost as many supposed causes of delinquency as there are specialists on this topic. The purpose of the present study was to investigate alexithymia in offender and non-offender adolescents.
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
329
The descriptive results indicate that disrupted family structure and high level of alexithymic characteristics (particularly difficulty in identifying feelings) are associated with juvenile delinquency. Furthermore, the prevalence rate of alexithymic individuals in the delinquent group (47.2%) is similar to rates that have been reported in studies of patients with psychiatric disorders (for reviews, see Taylor, 2000; Taylor et al., 1997), and the rate in the control group (21.7%) is in agreement with previous studies which showed that approximately 23.5% of ‘‘normal adolescents’’ are alexithymic (Horton, Gewirtz, & Kreutter, 1992). These results are consistent with previous reports indicating: (1) that family disruption has an impact on delinquency (Bailey, Thornton, & Weaver, 1994; Wells & Rankin, 1991) and (2) that adolescent offenders show deficits in emotional understanding and communication (Marohn, 1990, 1992). Furthermore, our results show that family structure and alexithymia clearly emerge as the strongest discriminatory factors for juvenile delinquency, indicating that these variables are significantly associated with group membership. The present findings indicate that two distinct factors (disrupted family structure and alexithymia) confer on adolescents a risk for juvenile delinquency. The topic of disrupted family structure as a central part of delinquency theory has a long research tradition (Wells & Rankin, 1991). In a detailed review, Pryor and Rodgers (2001) indicate that the risk of delinquency is more or less doubled for children from disrupted family structure compares to those from intact homes, and that this result is remarkably stable over time and place. However, the use of a dichotomous structural family variable (disrupted versus intact) can often be often disappointing because it ignores more relevant complex aspects of family functioning, and may consequently classify families with very different experiences in the same category (Juby & Farrington, 2001). Moreover, previous research has found that the family process variables, such as poor parental monitoring, family conflicts, weak attachments to parents or inconsistent parental discipline, are more important than structural variables in predicting antisocial behaviour and juvenile delinquency (Demo & Acock, 1988; Mucchielli, 2001). As Van Voorhis, Cullen, Mathers, and Garner (1998, p. 258) concluded, ‘‘bad homes not broken homes, it seems, place youths at risk.’’ In conformity with the hypotheses, alexithymia also predicts juvenile delinquency. Despite the exam of emotional intelligence (including alexithymia) in juvenile sexual offenders reported by Moriarty et al. (2001) and the study of the associations between alexithymia and psychopathy in juvenile offenders reported by Langevin and Hare (2001), no study, to our knowledge, has examined the role of alexithymia in predicting juvenile delinquency. According to the present findings, it is possible, then, that alexithymia may be considered as a vulnerability factor for delinquent behaviour. This is consistent with Krystal’s view (1979) indicating that unawareness of feelings may contribute to engage in uncontrollable behaviour, possibly violent. Several limitations of the current study should be noted. First, the offender adolescents were recruited from specialized residential facilities and may therefore not be representative of Swiss juvenile offenders. In Switzerland, residential sanctions are usually ordered by the juvenile justice system when minors have particular educative needs that parents are unable to fulfil (OFS, 2002). It is therefore possible that family disruption is overrepresented in the offender group. This might be part of the reason why a family structural variable was found to predict juvenile delinquency. As Mucchielli (2001) stated, this result could be a consequence of a social prejudice according to which single-parent families are less able to control and monitor their children. Consistent with this statement, Van Voorhis et al. (1988) showed that relations between family structure and
ARTICLE IN PRESS 330
G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
juvenile delinquency are stronger when the definition of adolescent offenders is based on official sanctions from the justice system (as was the case in this study), rather than when they are based on self-report measures of delinquency. Second, we did not differentiate adolescent offenders on the basis of conviction (such as violent versus non-violent offenders). Such a classification would have required a larger sample and could have obscured differences between the offender group and the control group. Third, alexithymia assessment relied exclusively on a self-report scale. Although, some authors have criticized the use of self-report measures to evaluate alexithymia (Lane, Ahern, Schwartz, & Kaszniak, 1997), Taylor and colleagues suggest that TAS-20, as a sole measure, can assess the concept satisfactorily (Taylor, Bagby, & Luminet, 2000). Nevertheless, we agree with the authors of the TAS-20 that empirical research is improved by the use of other types of assessment, like the Modified Version of the Beth Israel Psychosomatic Questionnaire (Taylor et al., 1997, 2000). Fourth, the control group was not matched on demographic variables that may have increased the risk of engaging in criminal activities. Finally, the present study used a cross-sectional design that precludes conclusions about the causality of juvenile delinquency. However, we hope that the results produced by this study will promote future longitudinal research, which is required to clarify causal risk factors that could be targeted in therapeutic and prevention programs. The present findings may have important implications regarding the aims of therapeutic interventions for such youths. Once in the system, especially in educative institutions, there is an important need for treatment in juvenile offenders to reduce risk of re-offending and to increase their chances to become ‘‘healthy’’ adults. Whereas subjects with high alexithymia respond poorly to traditional psychotherapies, group therapy is considered a useful alternative to help these youths to become more emotionally aware and expressive in a safe, supportive and empathic setting (Swiller, 1988). One could say that therapists should try to help these adolescents to mentalize (Fonagy, Gergely, Jurist, & Target, 2002), in order ‘‘to facilitate certain maturational steps, converting motor behaviour to verbal behaviour’’ (Marohn, 1990, p. 426). Whatever the limitations of our study, we are sure that the preliminary detection of some deficits in the mental elaboration of juvenile offenders (i.e. alexithymia) is beneficial and we hope that further empirical and clinical work may provide creative and practical ways to address this issue in offender adolescents.
Acknowledgements I would like to thank the adolescents who kindly volunteered to participate in the study. I thank the principals of the schools and the specialized residential facilities. I am also grateful to Sarah Blanc, Vanessa Braunschweig, Olivia Lempen and Laurence Maret for assistance with data collection. I also kindly acknowledge Franz Meyer de Stadelhofen for his help in editing the English.
References Bagby, R. M., Parker, J. D. A., & Taylor, G. J. (1994). The twenty-item toronto alexithymia scale—I. Item selection and cross validation of the factor structure. Journal of Psychosomatic Research, 38(1), 23–32.
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
331
Bagby, R. M., Taylor, G. J., & Parker, J. D. A. (1994). The twenty-item toronto alexithymia scale—II. Convergent, discriminant and concurrent validity. Journal of Psychosomatic Research, 38(1), 33–40. Bailey, S. M., Thornton, L., & Weaver, A. B. (1994). The first 100 admissions to an adolescent secure unit. Journal of Adolescence, 17(3), 207–220. Cahn, R. (1987). L’agir dans le fonctionnement mental de l’adolescent. Revue Franc- aise de Psychanalyse, 4, 1147–1154. Castro, D. (1999). Manuel de l’Echelle Re´vise´e d’Anxie´te´ Manifeste pour Enfant (R-CMAS). Paris: Les Editions du Centre de Psychologie Applique´e. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New York: Academic Press. Demo, D. H., & Acock, A. C. (1988). The impact of divorce on children. Journal of Marriage and the Family, 50(3), 619–648. Fonagy, P., Gergely, G., Jurist, E. L., & Target, M. (2002). Developmental issues in normal adolescence and adolescent breakdown. In P. Fonagy, G. Gergely, E. L. Jurist, & M. Target (Eds.), Affect regulation, mentalization, and the development of the self (pp. 317–340). New York: Other Press. Ganzeboom, H. B. G., & Treiman, D. G. (1996). Internationaly comparable measures of occupational status for the 1988 international standard classification of occupations. Social Science Research, 25, 201–239. Gendre, F., & Capel, R. (2003). L.A.B.E.L (Liste d’Ajectifs Bipolaires et en Echelles de Likert). Manuel pratique (2e ed.). Lausanne: Institut de Psychlogie, Universite´ de Lausanne. Gendre, F., Capel, R., & Monod, D. (2002). L.A.B.E.L (Liste d’Ajectifs Bipolaires et en Echelles de Likert). Un mode`le d’e´valuation de la personnalite´ a` vise´e universelle. Psychologie et Psychome´trie, 23(1/2), 101–133. Gerard, A. B., & Reynolds, C. R. (1999). Characteristics and applications of the Revised children’s manifest anxiety scale (RCMAS). In M. E. Maruish (Ed.), The use of psychological testing for treatment planning and outcomes assessment (pp. 323–340). Mahwah, NJ: Lawrence Erlbaum. Horton, P. C., Gewirtz, H., & Kreutter, K. J. (1992). Alexithymia—State and trait. Psychotherapy and Psychosomatics, 58, 91–96. Jeammet, P. (2001). Approche psychodynamique de la violence chez les adolescents. Perspectives in Psychiatry, 40(3), 182–188. Juby, H., & Farrington, D. P. (2001). Disentangling the link between disrupted families and delinquency. British Journal of Criminology, 41, 22–40. Krystal, H. (1979). Alexithymia and psychotherapy. American Journal of Psychotherapy, 33(1), 17–31. Lane, R. D., Ahern, G. L., Schwartz, G. E., & Kaszniak, A. W. (1997). Is alexithymia the emotional equivalent of blindsight? Biological Psychiatry, 42, 834–844. Lane, R. D., Lee, S., Riedel, R., Shapiro, D. E., & Kaszniak, A. W. (2000). Pervasive emotion recognition deficit common to alexithymia and the repressive coping style. Psychosomatic Medicine, 62, 492–501. Lane, R. D., Lee, S., Reidel, R., Weldon, V., Kaszniak, A. W., & Schwartz, G. E. (1996). Impaired verbal and nonverbal emotion recognition in alexithymia. Psychosomatic Medicine, 58(3), 203–210. Langevin, R., & Hare, R. D. (2001). Psychopathie et alexithymie chez un groupe de jeunes contrevenants. Revue de Psychoe´ducation et d’Orientation, 30(2), 227–236. Loas, G., Corcos, M., Stephan, P., Pellet, J., Bizouard, P., Venisse, J. L., et al. (2001). Factorial structure of the 20-item Toronto alexithymia scale. Confirmatory factorial analyses in nonclinical and clinical samples. Journal of Psychosomatic Research, 50, 255–261. Loas, G., Fremaux, D., & Marchand, M. P. (1995). Etude de la structure factorielle et de la cohe´rence interne de la version franc- aise de l’e´chelle d’alexithymie de Toronto a` 20 items (TAS-20) chez un groupe de 183 sujets sains. L’Ence´phale, 21, 117–122. Marohn, R. C. (1990). Violence and unrestrained behavior in adolescents. Adolescent Psychiatry, 17, 419–432. Marohn, R. C. (1992). Management of the assaultive adolescent. Hospital and Community Psychiatry, 43(6), 622–624. McCown, W., Johnson, J., & Austin, S. (1986). Inability of delinquents to recognize facial affects. Journal of Social Behavior and Personality, 1(4), 489–496. McCown, W. G., Johnson, J. L., & Austin, S. H. (1988). Patterns of facial affect recognition errors in delinquent adolescent males. Journal of Social Behavior and Personality, 3(3), 215–224. McDougall, J. (1982). Alexithymia: A psychoanalytic viewpoint. Psychotherapy and Psychosomatics, 38, 81–90.
ARTICLE IN PRESS 332
G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
Moriarty, N., Stough, C., Tidmarsh, P., Eger, D., & Dennison, S. (2001). Deficits in emotional intelligence underlying adolescent sex offending. Journal of Adolescence, 24, 743–751. Mucchielli, L. (2001). Monoparentalite´, divorce et de´linquance juve´nile: une liaison empiriquement contestable. De´viance et Socie´te´, 25(2), 209–228. Nemiah, J. C., Freyberger, H., & Sifneos, P. E. (1976). Alexthymia: A view of the psychosomatic process. In O. W. Hill (Ed.), Modern trends in psychosomatic medicine, Vol. 3 (pp. 430–439). London: Butterworths. Office fe´de´ral de la statistique. (2002). Statistiques des jugements pe´naux de mineurs, 2000. Pryor, J., & Rodgers, B. (2001). Children in changing families: Life after parental separation. Oxford: Blackwell. Quartier, V., & Zimmermann, G. (2004). Agir et mentalisation chez l’enfant et l’adolescent. Actualite´s Psychologiques, 17, 69–85. Reynolds, C. R., & Richmond, B. O. (1985). Revised children’s manifest anxiety scales (RCMAS). Los Angeles: Western Psychological Services. Schreiber, K. (1992). The adolescent crack dealer: A failure in the development of empathy. Journal of the American Academy of Psychoanalysis, 20(2), 241–249. Sifneos, P. E. (1973). The prevalence of ‘‘alexithymic’’ charcteristics in psychosomatic patients. Psychotherapy and Psychosomatics, 22, 255–262. Swiller, H. I. (1988). Alexithymia: Treatment utilizing combined individual and group psychotherapy. International Journal of Group Psychotherapy, 38(1), 47–61. Tardif, M. (1998). Le de´terminisme de la carence d’e´laboration psychique dans le passage a` l’acte. In F. Millaud (Ed.), Le passage a` l’acte: Aspects cliniques et psychodynamiques (pp. 25–40). Paris: Masson. Taylor, G. J. (2000). Recent developments in alexithymia. Theory and research. Canadian journal of Psychiatry, 45, 134–142. Taylor, G. J., & Bagby, M. R. (2004). New trends in alexithymia research. Psychotherapy and Psychosomatics, 73, 68–77. Taylor, G. J., Bagby, R. M., & Luminet, O. (2000). Assessement of alexithymia. Self-report and observer-rated measures. In J. D. A. Parker, & R. Bar-On (Eds.), The handbook of emotional intelligence (pp. 301–319). San Fransisco, CA: Jossey Bass. Taylor, G. J., Bagby, R. M., & Parker, J. D. A. (1997). Disorders of affect regulation: Alexithymia in medical and psychiatric illness. Cambridge: Cambridge University Press. Van Voorhis, P., Cullen, F., Mathers, R., & Garner, C. (1988). The impact of family structure and quality on delinquency: A comparative assessment of structural and functional factors. Criminology, 26, 235–261. Wells, L. E., & Rankin, J. H. (1991). Families and delinquency: A meta-analysis of the impact of broken homes. Social Problems, 38(1), 71–93. Zimmermann, G., Rossier, J., Meyer de Stadelhofen, F., & Gaillard, F. (2005). Alexithymia assessment and relations with dimensions of personality. European Journal of Psychological Assessment, 21(1), 23–33.
Journal of Adolescence Journal of Adolescence 29 (2006) 321–332 www.elsevier.com/locate/jado
Delinquency in male adolescents: The role of alexithymia and family structure Gre´goire Zimmermann Department of Psychology and Institute for Psychotherapy, University of Lausanne, BFSH 2—Dorigny, CH-1015 Lausanne, Switzerland
Abstract The purpose of this study was to examine the linkages between alexithymia and delinquency in male adolescents (age ranging from 14 to 18 years), and to investigate whether alexithymia was a good discriminatory factor for juvenile delinquency. Thirty-six offender adolescents and 46 non-offender control adolescents participated in the study and completed the 20-item Toronto Alexithymia Scale (TAS-20) (alexithymia), the Revised Children’s Manifest Anxiety Scale (R-CMAS) (anxiety), the Liste d’Adjectifs Bipolaires et en Echelles de Likert (LABEL) (personality—Big Five) and demographic data. Findings revealed that the adolescents of the offender group scored high on alexithymia and that proportion of disrupted family structure in the offender group is higher than in the control group. Logistic regressions confirmed that alexithymia and family structure are the strongest discriminatory factors for juvenile delinquency. Limitations and clinical implications are discussed, and recommendations for future research are provided. r 2005 The Association for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved.
Introduction In the psychoanalytic literature, acting-out and other means of enacted repetitions, typical of juvenile offenders, are often conceptualized as the result of a deficit in mental elaboration Corresponding author. Tel.: +41 21 692 32 60; fax: +41 21 692 32 65.
E-mail address: [email protected]. 0140-1971/$30.00 r 2005 The Association for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.adolescence.2005.08.001
ARTICLE IN PRESS 322
G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
(Cahn, 1987). Currently, several authors consider the concept of alexithymia to be related to this deficit in mental elaboration (McDougall, 1982; Quartier & Zimmermann, 2004; Tardif, 1998). The construct of alexithymia, literally without word for feelings, is based on observations of psychosomatic patients and was first introduced in the early 1970s by Sifneos (1973). The alexithymia construct covers a cluster of cognitive and affective features, mainly characterized by a difficulty in identifying and expressing feelings, a striking paucity of fantasies, and an externally oriented cognitive style (Nemiah, Freyberger, & Sifneos, 1976). These features are thought to reflect a deficit in the cognitive processing and regulation of emotional states (Taylor, Bagby, & Parker, 1997). Alexithymia was first studied in classical psychosomatic or somatic disorders. However, during the last decade, it has been investigated from other perspectives (Taylor & Bagby, 2004), and has been associated with a great variety of medical disorders (e.g. hypertension and inflammatory bowel diseases) and psychiatric disorders (e.g. substance use and eating disorders) (Taylor, 2000; Taylor et al., 1997). Theoretical and empirical works regarding the aetiology of juvenile delinquency that include individual, familial and social factors are abundant. However, while clinicians and researchers can define groups of adolescents who might be at risk for delinquency, few authors have tried to predict why some of these at-risk youths develop delinquent behaviours, while others do not. The alexithymia construct may have implication for understanding this evolution and the development of repetitive and disruptive behaviours in juvenile offenders. Thirty years ago, Nemiah et al. (1976) indicated that alexithymic patients ‘‘may on occasion manifest brief but violent outbursts of affective behaviours’’ (p. 432), but are unaware of underlying feelings expressed. Since then, several authors have published clinical and theoretical works suggesting the presence of alexithymia among juvenile offenders (Jeammet, 2001; Marohn, 1990, 1992; Schreiber, 1992). However, empirical works studying alexithymic characteristics in this population are rare. In fact, besides two published studies (Langevin & Hare, 2001; Moriarty, Stough, Tidmarsh, Eger, & Dennison, 2001) there are, to our knowledge, no data on alexithymia in samples of juvenile offenders. Moriarty et al. (2001) investigated emotional intelligence in 15 male adolescent sex offenders and 49 male non-offenders and showed that sex offenders presented difficulty in identifying feelings and were more aggressive than control adolescents. However, even if the results indicated high scores on the alexithymia scale (20-item Toronto Alexithymia Scale (TAS-20)) (Bagby, Parker, & Taylor, 1994; Bagby, Taylor, & Parker, 1994), there were no significant differences between the two groups. In their study, Langevin and Hare (2001) examined the link between psychopathy, alexithymia and ability for symbolization in a sample of 60 male adolescents (40 offenders and 20 controls). As expected, they reported that alexithymia correlated positively with psychopathy and negatively with the capacity for symbolization. Moreover, descriptive statistics showed that juvenile offenders recorded higher score on TAS-20 than the control group. Finally, there is some evidence suggesting that juvenile offenders present an impairment of their facial affect recognition ability (McCown, Johnson, & Austin, 1986, 1988). This ability is commonly associated with alexithymia (Lane, Lee, Riedel, Shapiro, & Kaszniak, 2000; Lane et al., 1996). The aim of the present study is to examine the relative contribution of alexithymia to explain juvenile delinquency. Specifically, it is hypothesized that delinquents are more alexithymic than control group subjects and that alexthymia could better differentiate juvenile offenders from juvenile non-offenders than other personality, mood and demographic variables.
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
323
Method Sample Effect size (ES) for group differences on alexithymia was estimated, on the basis of previous research, as between medium and large (.50odo.80; see Cohen, 1988). With this expected ES (fixed at .65), the sample size required—for a two-tailed independent t-test—to detect the effect at a significance level of .05 with a power of .80 is 78 (Cohen, 1988). The study sample consisted of 82 male adolescents (36 offenders and 46 controls) ranging in age from 14 to 18 years (M ¼ 16:10, S.D.¼ 1:02). The non-offender adolescents were recruited in four secondary schools of the French-speaking part of Switzerland. Their mean age was 15.95 years (S.D.¼ :80). Juvenile offenders were inpatients in specialized residential facilities in the same area. Their mean age was 16.30 years (S.D.¼ 1:22). Thirty-one percent (11/36) of offenders were on remand (i.e. not currently conviced or unsentenced). Educative residential treatment varied from a few months to a few years and was ordered by the juvenile justice system. At the time of interview, the majority of the offenders (29/36–80.6%) had been treated in a residential centre for 12 months or longer. The arrest histories of the referred adolescents attest to the seriousness of their criminal involvement. The youths averaged 7.28 previous arrests (S.D.¼ 7:00) and they were quite young at the time of their first arrest (Age: M ¼ 12:61, S.D.¼ 2:41). The nature of the crimes committed by these adolescents included theft, drug trafficking, burglary and violent offences (i.e. aggravated assault, robbery, attempted rape). All adolescents included in this study were able to understand and write the French language. Our study is in compliance with the ethical code of the Swiss Society of Psychology. Measures For the purpose of the present study, we used the following questionnaires: (i) a demographic questionnaire, (ii) the TAS-20, (iii) the Revised Children’s Manifest Anxiety Scale (R-CMAS), and (iv) the Liste d’Adjectifs Bipolaires et en Echelles de Likert (LABEL). Demographic questionnaire We used a demographic questionnaire for registration of age, family structure, parent’s socioeconomic status (SES1) and nationality (Swiss citizen versus non-Swiss citizen). Concerning family structure, we differentiated intact and disrupted family structures, indicating the absence of a least one parent (i.e. children whose parents have divorced, separated or deceased). The TAS-20 Alexithymia was measured with one of the most reliable and validated self-report questionnaire: the TAS-20 (Bagby, Parker et al., 1994; Bagby, Taylor et al., 1994), in which participants indicate the extent of their agreement or disagreement with statements on a five-point Likert scale. A French-language version was applied in this study (Loas, Fremaux & Marchand, 1 SES is the International Socio-Economic Index of Occupational Status calculated according to Ganzeboom and Treiman (1996).
ARTICLE IN PRESS 324
G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
1995). We used the total score, as well as the score of the three factors: (F1) difficulty-identifying feelings, (F2) difficulty describing feelings and (F3) externally oriented thinking. Even if alexithymia is conceptualized as a dimensional construct, an indicative cutoff score has been established empirically (X61 alexithymia) to facilitate comparisons of rates of alexithymia across studies (Taylor et al., 1997). In previous research, the French version of the TAS-20 has been found to have good internal consistency, and a three-factor structure consistent with the theoretical construct (Loas et al., 2001; Zimmermann, Rossier, Meyer de Stadelhofen, & Gaillard, 2005). The R-CMAS The R-CMAS subtitled ‘‘What I think and feel’’ (Reynolds & Richmond, 1985) is a widely used 37-item instrument to capture the degree and quality of anxiety experienced by children and adolescents. The R-CMAS consists of 37 dichotomous items (yes/no) divided into four subscales: Physiological Anxiety, Worry/Oversensitivity, Social Concerns/Concentration and the Lie Scale. A Total Anxiety score can be computed summing the 28 anxiety items. The R-CMAS has been found to be a reliable, valid and useful indicator of general anxiety (Gerard & Reynolds, 1999). In this study, we used the French-language validated version of the R-CMAS (Castro, 1999). The LABEL The Five Factor Model (FFM) of personality was measured with the 87-item LABEL developed and validated by Gendre and colleagues (Gendre & Capel, 2003; Gendre, Capel, & Monod, 2002). This 87-item semantic differential scale was designed to measure, among others dimensions, the five domains of FFM. The first factor (i.e. Neuroticism) assesses affective emotional instability versus adjustment. The second factor (i.e. Extraversion) assesses the quantity and intensity of interpersonal interactions, the need for stimulation and the capacity to experience joy. The third one, Openness to experience, is defined as the proactive seeking and appreciation of various experiences for their own sake, and as toleration for and exploration of the unfamiliar. As Extraversion evaluates the degree to which an individual enjoys being in the presence of others, the fourth factor (i.e. Agreeableness) examines the attitude a person holds towards other people, describing quality of interaction preference ranging from empathy to opposition. Finally, the fifth one (i.e. Conscientiousness) assesses the individual’s degree of organization, persistence, and motivation in goal-directed behaviour. Item consists in two bi-polar adjectives rated on a fivepoint scale. Respondents are required to choose, to a certain extent, one or the other adjective depending on what is most true for them. Procedure All the questionnaires were administered (in the order described above) in a private individual structured interview lasting approximately 60 min. Prior to the interview, adolescents were given explanations emphasizing the significance of their participation, the confidentiality of all data, and the opportunity of withdrawing from research interview at any moment. Moreover, participants have been encouraged to disclose their own opinion and reassured that there were no right or wrong answers.
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
325
Analyses Comparisons between juvenile offenders and non-offenders were performed using a w2 or a Fisher’s exact test for dichotomous variables and independent Student’s t-test for continuous variables. Moreover, ES were computed following Cohen’s procedures (1988). Pearson correlations were used to assess the relations between alexithymia, anxiety and personality. Last, a series of logistic regressions were performed to determine if alexithymia could predict juvenile delinquency. For the logistic regressions, group membership (delinquent/control) served as dependent variable (criterion) and demographic, personality, and mood variables as the independent variables (predictor). We set alpha level at .05 and confidence intervals of 95%. All analyses were computed using the statistical software SPSS 11.0.
Results First, we examined the potential differences in alexithymia, anxiety, FFM, and demographic variables between adolescent offenders and non-offenders. Demographic comparisons (see Table 1) between the offender and control group showed no differences for age, nationality and parent’s SES. The adolescents of the delinquent group significantly more often came from disrupted family structure (22/36–61.1%) than those in the control group (10/46–21.7%). Table 2 displays the means and standard deviations for the TAS-20, R-CMAS, and LABEL for the delinquent and control group. Compared to the control subjects, the adolescent offenders scored significantly higher on TAS-20 total score (upper medium ES) and on TAS-20 first factor (F1: difficulty-identifying feelings (upper medium ES)). Based on the cutoff score of X61 on the TAS-20, the proportion of alexithymic individuals was significantly higher (w2 ð1Þ ¼ 5:94, po:05) Table 1 Demographic characteristics Group
Statistical test
Effect size
Control (n ¼ 46)
Delinquent (n ¼ 36)
Age (S.D.) SES (S.D.)
15.95 (.80) 38.96 (13.78)
16.30 (1.22) 39.19 (16.82)
tð80Þ ¼ 1:49 tð80Þ ¼ :07
d ¼ :35 d ¼ :01
Nationality (%) Swiss Non-Swiss
35 (76.1) 11 (23.9)
22 (61.1) 14 (38.9)
w2 ð1Þ ¼ 2:14
w ¼ :17
Family structure (%) Intact Disrupted
36 (78.3) 10 (21.7)
14 (38.9) 22 (61.1)
w2 ð1Þ ¼ 13:16
w ¼ :41
Note: SES ¼ parent’s socioeconomic status. Cohen (1988) considered: d ¼ :20 as a ‘‘small’’ effect size, d ¼ :50 as a ‘‘medium’’ effect size and d ¼ :80 as a ‘‘large’’ effect size. For w effect size index values of w for ‘‘small’’, ‘‘medium’’ and ‘‘large’’ are, respectively, .10, .30 and .50. po:05.
3.24 5.00 1.63 9.87 52.38 47.45 48.79 54.24 51.10
R-CMAS (S.D.) Physiological anxiety Worry/oversensitivity Social concerns/concentration Total
LABEL (S.D.) Neuroticismb Extraversion Openness Agreeableness Conscientiousness (20.20) (20.28) (19.48) (15.54) (20.64)
(1.98) (2.62) (1.47) (4.66)
(5.12) (4.72) (4.06) (10.23)
46.95 53.26 51.54 44.58 48.60
3.64 4.61 2.19 10.44
20.69 16.08 22.47 59.25
(19.60) (19.43) (20.82) (23.70) (19.35)
(2.17) (2.75) (1.41) (4.83)
(6.03) (3.83) (5.25) (11.56)
5.43 +5.81 +2.75 9.66 2.50
+.40 .39 +.56 +.57
+3.56 +2.02 +1.99 +7.58
(14.26 to+3.40) (3.00 to +14.63) (6.14 to +11.64) (18.78 to –.51) (11.39 to +6.40)
(.51 to +1.31) (1.58 to +.80) (.07 to +1.20) (1.52 to +2.67)
(+1.11 to +6.02) (+.09 to +3.94) (.05 to +4.04) (+2.78 to +12.37)
tð80Þ ¼ 1:22 tð80Þ ¼ 1:31 tð80Þ ¼ :62 tð80Þ ¼ 2:12 tð80Þ ¼ :55
tð80Þ ¼ :87 tð80Þ ¼ :65 tð80Þ ¼ 1:76 tð80Þ ¼ :55
tð80Þ ¼ 2:89 tð80Þ ¼ 2:09 tð80Þ ¼ 1:94 tð80Þ ¼ 3:14
Statistical testa
d d d d d
d d d d
d d d d
¼ :27 ¼ :29 ¼ :14 ¼ :48 ¼ :12
¼ :14 ¼ :15 ¼ :38 ¼ :12
¼ :64 ¼ :45 ¼ :42 ¼ :66
Effect size
Note: F1 ¼ difficulty identifying feelings; F2 ¼ difficulty describing feelings; F3 ¼ externally oriented thinking. CI ¼ confidence interval. The LABEL’s dimensions were normed by computing T-scores, which have a mean of 50 and a standard deviation of 20. po:05=4 tests ¼ .0125 for TAS-20 and R-CMAS, po:05=5 tests ¼ .01 for LABEL. a Independent t-test were performed with a Bonferroni correction for multiple comparisons. b Compared to classical measure of Big Five dimensions, Neuroticism’s scale of LABEL is in reverse order.
17.13 14.07 20.48 51.67
TAS-20 (S.D.) F1 F2 F3 Total
Delinquent (n ¼ 36)
Mean differences (95% CI)
326
Control (n ¼ 46)
Group
Table 2 Means, standard deviations, mean differences, t-tests and effect sizes for the TAS-20, R-CMAS, LABEL for delinquent and control group
ARTICLE IN PRESS
G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
327
Table 3 Correlations between alexithymia, anxiety and personality TAS-20
TAS-20 R-CMAS LABEL N E O A C
—
R-CMAS
.41* —
LABEL N
E
O
A
C
.41* .48*
.06 .13
.08 .27*
.45* .19
.27* .21
—
.39* —
.32* .68* —
.67* .01 .02 —
.72* .09 .28* .48* —
po:05.
in the delinquent group (17/36–47.2%) than in the control group (10/46–21.7%). For the RCMAS and the LABEL, there were no significant differences between the two groups. As evident from Table 3, however, the TAS-20 is correlated with these other independent variables. Therefore, a set of hierarchical logistic regression analyses was performed to determine the power of alexithymia in predicting group membership. First, predictor variables were entered into the regression analyses in four blocks: alexithymia (TAS-20 total score) was entered first in the model as a block (step 1), anxiety (R-CMAS total score) was entered on the second block (step 2), followed by five-factor model personality dimensions (step 3) and demographic variables (SES, nationality and family structure) (step 4). Alexithymia was entered first to determine how well TAS-20 total score alone was associated with juvenile delinquency. The results from this analysis are presented in Table 4. The model after step 1 was significant, w2 ð1Þ ¼ 9:48, po:05, and provided correct predictions for 55.6% of the offender adolescents and 73.9% of the control adolescents (overall correct classification rate ¼ 65.9%). The addition of anxiety (step 2) and FFM of personality variables (step 3) did not significantly increase the overall fit of the model (see Table 4). However, the introduction in the fourth block (step 4) of demographic variables did significantly increase the overall fit of the model (w2chg ð3Þ ¼ 16:56, po:05). Of the variables entered in the fourth block, only the family structure made a significant contribution to the model. The full model was significant, w2 ð10Þ ¼ 32:03, po:05, and provided correct predictions for 72.2% of the offender adolescents and 84.8% of the control adolescents (overall correct classification rate ¼ 79.3%). In a second hierarchical logistic regression, only significant discriminatory factors of the previous model were used. The TAS-20 total score was entered first into the model, followed by the family structure. This model (see Table 5) was compared to the previous full model. Goodness-of-fit statistic showed no reliable differences between the models (w2 ð8Þ ¼ 10:29, p4:05), indicating no significant enhancement to prediction of juvenile delinquency by anxiety and FFM personality variables. This parsimonious model including only alexithymia and family structure correctly classified 63.9% of the offender adolescents and 78.3% of the control adolescents (overall correct classification rate ¼ 72%). In this model, the likelihood of being in the delinquent group increased by approximately 40% for each five points increase in the TAS-20
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
328
Table 4 Results from hierarchical logistic regression predicting group membership Step
Predictor
R2
w2
df
w2chg
df
Final model B
1 2 3
4
TAS-20 total R-CMAS total Block N—LABEL E—LABEL O—LABEL A—LABEL C—LABEL
.11 .12 .17
Block SES Nationality Family structure
.32
9.48* 10.09* 15.47*
32.03*
1 2 7
10
.61 5.37
16.56*
1 5
Wald
OR
OR 95% CI
.08 .07
6.12* .69
1.09 .93
(1.02–1.16) (.78–1.10)
.06 .04 .02 .00 .04
1.54 2.32 .42 .03 1.40
.94 1.04 1.02 1.00 1.04
(.86–1.04) (.99–1.08) (.96–1.09) (.96–1.05) (.97–1.12)
.01 1.24 2.03
.13 3.13 10.14*
1.01 3.45 7.59
(.97–1.05) (.87–13.65) (2.18–26.44)
3
Note: SES ¼ parent’s socioeconomic status, CI ¼ confidence interval, OR ¼ Odds ratio, R2 ¼ Cox & Snell R2. *po:05.
Table 5 Results from parsimonious hierarchical logistic regression predicting group membership Step
1 2
Predictor
TAS-20 total Family structure
R2
.11 .23
w2
9.48* 21.74*
df
1 2
w2chg
df
*
12.26
1
Final model B
Wald
OR
OR 95% CI
.07 1.75
7.21* 11.12*
1.07 5.78
(1.02–1.12) (2.06–16.19)
Note: SES ¼ parent’s socioeconomic status, CI ¼ confidence interval, OR ¼ Odds ratio, R2 ¼ Cox & Snell R2. *po:05:
total score. Moreover, adolescents from disrupted family structure were 5.8 times more likely to belong to delinquent group than those stemmed from intact home. These two discriminatory factors were independent (w2 ð1Þ ¼ :50, p4:05). In fact, the prevalence of alexithymia was similar for these two groups (16/50–32% of adolescents stemmed from intact family structure and 11/32–34.4% of adolescents stemmed from disrupted family structure).
Discussion The juvenile delinquency is a very complex issue and there are almost as many supposed causes of delinquency as there are specialists on this topic. The purpose of the present study was to investigate alexithymia in offender and non-offender adolescents.
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
329
The descriptive results indicate that disrupted family structure and high level of alexithymic characteristics (particularly difficulty in identifying feelings) are associated with juvenile delinquency. Furthermore, the prevalence rate of alexithymic individuals in the delinquent group (47.2%) is similar to rates that have been reported in studies of patients with psychiatric disorders (for reviews, see Taylor, 2000; Taylor et al., 1997), and the rate in the control group (21.7%) is in agreement with previous studies which showed that approximately 23.5% of ‘‘normal adolescents’’ are alexithymic (Horton, Gewirtz, & Kreutter, 1992). These results are consistent with previous reports indicating: (1) that family disruption has an impact on delinquency (Bailey, Thornton, & Weaver, 1994; Wells & Rankin, 1991) and (2) that adolescent offenders show deficits in emotional understanding and communication (Marohn, 1990, 1992). Furthermore, our results show that family structure and alexithymia clearly emerge as the strongest discriminatory factors for juvenile delinquency, indicating that these variables are significantly associated with group membership. The present findings indicate that two distinct factors (disrupted family structure and alexithymia) confer on adolescents a risk for juvenile delinquency. The topic of disrupted family structure as a central part of delinquency theory has a long research tradition (Wells & Rankin, 1991). In a detailed review, Pryor and Rodgers (2001) indicate that the risk of delinquency is more or less doubled for children from disrupted family structure compares to those from intact homes, and that this result is remarkably stable over time and place. However, the use of a dichotomous structural family variable (disrupted versus intact) can often be often disappointing because it ignores more relevant complex aspects of family functioning, and may consequently classify families with very different experiences in the same category (Juby & Farrington, 2001). Moreover, previous research has found that the family process variables, such as poor parental monitoring, family conflicts, weak attachments to parents or inconsistent parental discipline, are more important than structural variables in predicting antisocial behaviour and juvenile delinquency (Demo & Acock, 1988; Mucchielli, 2001). As Van Voorhis, Cullen, Mathers, and Garner (1998, p. 258) concluded, ‘‘bad homes not broken homes, it seems, place youths at risk.’’ In conformity with the hypotheses, alexithymia also predicts juvenile delinquency. Despite the exam of emotional intelligence (including alexithymia) in juvenile sexual offenders reported by Moriarty et al. (2001) and the study of the associations between alexithymia and psychopathy in juvenile offenders reported by Langevin and Hare (2001), no study, to our knowledge, has examined the role of alexithymia in predicting juvenile delinquency. According to the present findings, it is possible, then, that alexithymia may be considered as a vulnerability factor for delinquent behaviour. This is consistent with Krystal’s view (1979) indicating that unawareness of feelings may contribute to engage in uncontrollable behaviour, possibly violent. Several limitations of the current study should be noted. First, the offender adolescents were recruited from specialized residential facilities and may therefore not be representative of Swiss juvenile offenders. In Switzerland, residential sanctions are usually ordered by the juvenile justice system when minors have particular educative needs that parents are unable to fulfil (OFS, 2002). It is therefore possible that family disruption is overrepresented in the offender group. This might be part of the reason why a family structural variable was found to predict juvenile delinquency. As Mucchielli (2001) stated, this result could be a consequence of a social prejudice according to which single-parent families are less able to control and monitor their children. Consistent with this statement, Van Voorhis et al. (1988) showed that relations between family structure and
ARTICLE IN PRESS 330
G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
juvenile delinquency are stronger when the definition of adolescent offenders is based on official sanctions from the justice system (as was the case in this study), rather than when they are based on self-report measures of delinquency. Second, we did not differentiate adolescent offenders on the basis of conviction (such as violent versus non-violent offenders). Such a classification would have required a larger sample and could have obscured differences between the offender group and the control group. Third, alexithymia assessment relied exclusively on a self-report scale. Although, some authors have criticized the use of self-report measures to evaluate alexithymia (Lane, Ahern, Schwartz, & Kaszniak, 1997), Taylor and colleagues suggest that TAS-20, as a sole measure, can assess the concept satisfactorily (Taylor, Bagby, & Luminet, 2000). Nevertheless, we agree with the authors of the TAS-20 that empirical research is improved by the use of other types of assessment, like the Modified Version of the Beth Israel Psychosomatic Questionnaire (Taylor et al., 1997, 2000). Fourth, the control group was not matched on demographic variables that may have increased the risk of engaging in criminal activities. Finally, the present study used a cross-sectional design that precludes conclusions about the causality of juvenile delinquency. However, we hope that the results produced by this study will promote future longitudinal research, which is required to clarify causal risk factors that could be targeted in therapeutic and prevention programs. The present findings may have important implications regarding the aims of therapeutic interventions for such youths. Once in the system, especially in educative institutions, there is an important need for treatment in juvenile offenders to reduce risk of re-offending and to increase their chances to become ‘‘healthy’’ adults. Whereas subjects with high alexithymia respond poorly to traditional psychotherapies, group therapy is considered a useful alternative to help these youths to become more emotionally aware and expressive in a safe, supportive and empathic setting (Swiller, 1988). One could say that therapists should try to help these adolescents to mentalize (Fonagy, Gergely, Jurist, & Target, 2002), in order ‘‘to facilitate certain maturational steps, converting motor behaviour to verbal behaviour’’ (Marohn, 1990, p. 426). Whatever the limitations of our study, we are sure that the preliminary detection of some deficits in the mental elaboration of juvenile offenders (i.e. alexithymia) is beneficial and we hope that further empirical and clinical work may provide creative and practical ways to address this issue in offender adolescents.
Acknowledgements I would like to thank the adolescents who kindly volunteered to participate in the study. I thank the principals of the schools and the specialized residential facilities. I am also grateful to Sarah Blanc, Vanessa Braunschweig, Olivia Lempen and Laurence Maret for assistance with data collection. I also kindly acknowledge Franz Meyer de Stadelhofen for his help in editing the English.
References Bagby, R. M., Parker, J. D. A., & Taylor, G. J. (1994). The twenty-item toronto alexithymia scale—I. Item selection and cross validation of the factor structure. Journal of Psychosomatic Research, 38(1), 23–32.
ARTICLE IN PRESS G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
331
Bagby, R. M., Taylor, G. J., & Parker, J. D. A. (1994). The twenty-item toronto alexithymia scale—II. Convergent, discriminant and concurrent validity. Journal of Psychosomatic Research, 38(1), 33–40. Bailey, S. M., Thornton, L., & Weaver, A. B. (1994). The first 100 admissions to an adolescent secure unit. Journal of Adolescence, 17(3), 207–220. Cahn, R. (1987). L’agir dans le fonctionnement mental de l’adolescent. Revue Franc- aise de Psychanalyse, 4, 1147–1154. Castro, D. (1999). Manuel de l’Echelle Re´vise´e d’Anxie´te´ Manifeste pour Enfant (R-CMAS). Paris: Les Editions du Centre de Psychologie Applique´e. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New York: Academic Press. Demo, D. H., & Acock, A. C. (1988). The impact of divorce on children. Journal of Marriage and the Family, 50(3), 619–648. Fonagy, P., Gergely, G., Jurist, E. L., & Target, M. (2002). Developmental issues in normal adolescence and adolescent breakdown. In P. Fonagy, G. Gergely, E. L. Jurist, & M. Target (Eds.), Affect regulation, mentalization, and the development of the self (pp. 317–340). New York: Other Press. Ganzeboom, H. B. G., & Treiman, D. G. (1996). Internationaly comparable measures of occupational status for the 1988 international standard classification of occupations. Social Science Research, 25, 201–239. Gendre, F., & Capel, R. (2003). L.A.B.E.L (Liste d’Ajectifs Bipolaires et en Echelles de Likert). Manuel pratique (2e ed.). Lausanne: Institut de Psychlogie, Universite´ de Lausanne. Gendre, F., Capel, R., & Monod, D. (2002). L.A.B.E.L (Liste d’Ajectifs Bipolaires et en Echelles de Likert). Un mode`le d’e´valuation de la personnalite´ a` vise´e universelle. Psychologie et Psychome´trie, 23(1/2), 101–133. Gerard, A. B., & Reynolds, C. R. (1999). Characteristics and applications of the Revised children’s manifest anxiety scale (RCMAS). In M. E. Maruish (Ed.), The use of psychological testing for treatment planning and outcomes assessment (pp. 323–340). Mahwah, NJ: Lawrence Erlbaum. Horton, P. C., Gewirtz, H., & Kreutter, K. J. (1992). Alexithymia—State and trait. Psychotherapy and Psychosomatics, 58, 91–96. Jeammet, P. (2001). Approche psychodynamique de la violence chez les adolescents. Perspectives in Psychiatry, 40(3), 182–188. Juby, H., & Farrington, D. P. (2001). Disentangling the link between disrupted families and delinquency. British Journal of Criminology, 41, 22–40. Krystal, H. (1979). Alexithymia and psychotherapy. American Journal of Psychotherapy, 33(1), 17–31. Lane, R. D., Ahern, G. L., Schwartz, G. E., & Kaszniak, A. W. (1997). Is alexithymia the emotional equivalent of blindsight? Biological Psychiatry, 42, 834–844. Lane, R. D., Lee, S., Riedel, R., Shapiro, D. E., & Kaszniak, A. W. (2000). Pervasive emotion recognition deficit common to alexithymia and the repressive coping style. Psychosomatic Medicine, 62, 492–501. Lane, R. D., Lee, S., Reidel, R., Weldon, V., Kaszniak, A. W., & Schwartz, G. E. (1996). Impaired verbal and nonverbal emotion recognition in alexithymia. Psychosomatic Medicine, 58(3), 203–210. Langevin, R., & Hare, R. D. (2001). Psychopathie et alexithymie chez un groupe de jeunes contrevenants. Revue de Psychoe´ducation et d’Orientation, 30(2), 227–236. Loas, G., Corcos, M., Stephan, P., Pellet, J., Bizouard, P., Venisse, J. L., et al. (2001). Factorial structure of the 20-item Toronto alexithymia scale. Confirmatory factorial analyses in nonclinical and clinical samples. Journal of Psychosomatic Research, 50, 255–261. Loas, G., Fremaux, D., & Marchand, M. P. (1995). Etude de la structure factorielle et de la cohe´rence interne de la version franc- aise de l’e´chelle d’alexithymie de Toronto a` 20 items (TAS-20) chez un groupe de 183 sujets sains. L’Ence´phale, 21, 117–122. Marohn, R. C. (1990). Violence and unrestrained behavior in adolescents. Adolescent Psychiatry, 17, 419–432. Marohn, R. C. (1992). Management of the assaultive adolescent. Hospital and Community Psychiatry, 43(6), 622–624. McCown, W., Johnson, J., & Austin, S. (1986). Inability of delinquents to recognize facial affects. Journal of Social Behavior and Personality, 1(4), 489–496. McCown, W. G., Johnson, J. L., & Austin, S. H. (1988). Patterns of facial affect recognition errors in delinquent adolescent males. Journal of Social Behavior and Personality, 3(3), 215–224. McDougall, J. (1982). Alexithymia: A psychoanalytic viewpoint. Psychotherapy and Psychosomatics, 38, 81–90.
ARTICLE IN PRESS 332
G. Zimmermann / Journal of Adolescence 29 (2006) 321–332
Moriarty, N., Stough, C., Tidmarsh, P., Eger, D., & Dennison, S. (2001). Deficits in emotional intelligence underlying adolescent sex offending. Journal of Adolescence, 24, 743–751. Mucchielli, L. (2001). Monoparentalite´, divorce et de´linquance juve´nile: une liaison empiriquement contestable. De´viance et Socie´te´, 25(2), 209–228. Nemiah, J. C., Freyberger, H., & Sifneos, P. E. (1976). Alexthymia: A view of the psychosomatic process. In O. W. Hill (Ed.), Modern trends in psychosomatic medicine, Vol. 3 (pp. 430–439). London: Butterworths. Office fe´de´ral de la statistique. (2002). Statistiques des jugements pe´naux de mineurs, 2000. Pryor, J., & Rodgers, B. (2001). Children in changing families: Life after parental separation. Oxford: Blackwell. Quartier, V., & Zimmermann, G. (2004). Agir et mentalisation chez l’enfant et l’adolescent. Actualite´s Psychologiques, 17, 69–85. Reynolds, C. R., & Richmond, B. O. (1985). Revised children’s manifest anxiety scales (RCMAS). Los Angeles: Western Psychological Services. Schreiber, K. (1992). The adolescent crack dealer: A failure in the development of empathy. Journal of the American Academy of Psychoanalysis, 20(2), 241–249. Sifneos, P. E. (1973). The prevalence of ‘‘alexithymic’’ charcteristics in psychosomatic patients. Psychotherapy and Psychosomatics, 22, 255–262. Swiller, H. I. (1988). Alexithymia: Treatment utilizing combined individual and group psychotherapy. International Journal of Group Psychotherapy, 38(1), 47–61. Tardif, M. (1998). Le de´terminisme de la carence d’e´laboration psychique dans le passage a` l’acte. In F. Millaud (Ed.), Le passage a` l’acte: Aspects cliniques et psychodynamiques (pp. 25–40). Paris: Masson. Taylor, G. J. (2000). Recent developments in alexithymia. Theory and research. Canadian journal of Psychiatry, 45, 134–142. Taylor, G. J., & Bagby, M. R. (2004). New trends in alexithymia research. Psychotherapy and Psychosomatics, 73, 68–77. Taylor, G. J., Bagby, R. M., & Luminet, O. (2000). Assessement of alexithymia. Self-report and observer-rated measures. In J. D. A. Parker, & R. Bar-On (Eds.), The handbook of emotional intelligence (pp. 301–319). San Fransisco, CA: Jossey Bass. Taylor, G. J., Bagby, R. M., & Parker, J. D. A. (1997). Disorders of affect regulation: Alexithymia in medical and psychiatric illness. Cambridge: Cambridge University Press. Van Voorhis, P., Cullen, F., Mathers, R., & Garner, C. (1988). The impact of family structure and quality on delinquency: A comparative assessment of structural and functional factors. Criminology, 26, 235–261. Wells, L. E., & Rankin, J. H. (1991). Families and delinquency: A meta-analysis of the impact of broken homes. Social Problems, 38(1), 71–93. Zimmermann, G., Rossier, J., Meyer de Stadelhofen, F., & Gaillard, F. (2005). Alexithymia assessment and relations with dimensions of personality. European Journal of Psychological Assessment, 21(1), 23–33.