A Preliminary Study on Cognitive–Behavior Interventions for Hong Kong Chinese Delinquent Youth

Abstract

Objectives:

This study evaluated a culturally attuned cognitive–behavior intervention for reducing impulsivity and delinquency-related attitudes and behaviors among Chinese delinquent youth in Hong Kong.

Methods:

A quasi-experimental design was adopted. Twenty participants received CBI, and 20 received routine counseling for a period of 12 months. All participants filled out a standardized questionnaire before the intervention and at 6 and 12 months of intervention.

Results:

A series of 2 × 3 analyses of variance were used to examine the Time × Group Effects of the two groups of participants. Results suggested, when compared to the control group, the participants in the experimental group showed greater and continuous positive changes in overall impulsivity, nonplanning impulsivity, cognitive impulsivity, and overall delinquency, with effect sizes showing moderate to large magnitude of change.

Conclusion:

This culturally attuned CBI program may be a useful inclusion in the outreaching social services for Chinese delinquent youth in Hong Kong.

Keywords

delinquent youth cognitive–behavior interventions impulsivity gang affiliations

Juvenile delinquency is a serious concern in many societies, including Hong Kong. By definition, juvenile delinquency refers to “behaviors performed by young persons of not more than a specified age, usually 18 years, that is antisocial or in violation of the law” (Webster’s New World College Dictionary, 2014). According to various studies conducted in Hong Kong, delinquent youth are gang/triad members (Che, 1998); school dropouts who terminated their study before Form 3 (i.e., Grade 8 or 9); and who exhibit delinquent behaviors, such as sleeping out without family/parental consent, drug abuse, petty crimes, shoplifting, serious crimes, and drug trafficking (Che, 1998). The Hong Kong police records also indicate that juvenile offenders (aged 10–15 years) and young persons (aged 16–20 years) arrested by the police constituted approximately 21% of the overall arrests made in the year 2010 (Hong Kong Police Force, 2011). While some of them were incarcerated in closed detention facilities, a proportion of these juvenile offenders who committed less severe offenses were adjudicated to serve on a probation order (Chui & Chan, 2012).

Outreaching social work services (OSWS) have been implemented in Hong Kong since 1979 and seek to reach out to young people who are unattached to normal social systems but are inclined to be or are attached to some undesirable systems or elements. Contact is usually made in the places that these young people are known to frequently show up (e.g., playgrounds, parks, housing blocks, video game centers, etc.; Social Welfare Department, 2015). The purpose of OSWS is to identify and engage target clients to help them prevent themselves from further behavioral deterioration, to direct them to more positive lifestyles, and to foster the development of positive social values and attitudes (Social Welfare Department, 2015). In the literature, studies have shown that outreaching social services appear to be effective in treating delinquent youth generally (Cheung & Ngai, 2007; Chui, 2001). However, other studies also highlighted certain problems in the service, including the lack of skills of outreaching social workers in engaging and facilitating changes in youth-at-risk/juvenile delinquents (Lee, 2005). Moreover, there is a lack of a specification of the intervention approach and strategies that are useful to facilitate changes in this group of youth. Indeed, a search of the literature finds very few clinical outcome studies that had targeted the youth-at-risk/juvenile delinquents in Hong Kong, except for one study on young drug abusers (Cheung & Ngai, 2013).

Cognitive–Behavior Intervention for Juvenile Delinquents

Cognitive–behavior intervention (CBI) is based on the assumption that cognitive deficits and distortions found in juvenile delinquents are learned rather than inherent (Lipsey, Landenberger, & Wilson, 2007). Programs for delinquent youth attempt to teach delinquent youth to understand their thinking processes and choices that immediately preceded their delinquent behaviors, to develop strategies to self-monitor their thinking, and to identify and correct biased and risky thinking patterns. All cognitive–behavior interventions employ a set of structured techniques, aiming to build cognitive and behavioral skills in areas where delinquent youth show deficits (Lipsey et al., 2007). CBI techniques for delinquent youth typically involve cognitive skills training, anger management, and other supplementary components.

Emotion management

Emotion management training focuses on teaching delinquent youth to self-monitor their patterns of negative automatic thoughts in situations where they tend to react impulsively with negative emotions (e.g., aggression or violence). Various strategies, which include recognizing the “hot” or “trigger” cognitions and understanding how these negative thoughts and consequential maladaptive behaviors run in a vicious cycle, are taught to delinquent youth. In our project, we also introduced the five-steps strategies to delinquent youth, which include recognition of one’s physiological responses, thought stopping, cognitive restructuring, distraction, and positive self-talk. This set of skills was thought to be useful for dealing with delinquent youths’ impulsivities and anger outbursts.

Cognitive skills training

Cognitive skills training involves teaching “thinking skills,” which are related to interpersonal problem-solving, goal setting, long-term planning, perspective taking, and so on (Lipsey et al., 2007). While delinquent youth are guided to understand how a lack of any of these skills may be related to the development of their maladaptive habits and aggressive or criminal behaviors, they are also encouraged to role-play or practice in vivo the adaptive coping skills in and outside of the counseling sessions.

Supplementary components

CBI programs often include other components, such as social skills training or relapse prevention. Relapse prevention is increasingly popular and is aimed at developing cognitive risk management strategies, along with a set of behavioral contracts for avoiding or deescalating the precursors to offending behavior (e.g., high-risk situations, places, associates, or maladaptive coping responses; Lipsey et al., 2007).

Effectiveness of CBIs for Juvenile Delinquents

Non-Chinese subjects

Several meta-analyses have identified cognitive–behavior intervention as an effective intervention for reducing recidivism in juvenile offenders (Lipsey & Landenberger, 2006; Pearson, Lipton, Cleland, &Yee, 2002). Generally speaking, many of these studies showed that CBI had positive effects in reducing recidivism by 20–30%, compared to untreated control groups (Lipsey & Landenberger, 2006). However, variations in the rate of recidivism were found among different studies and might be related to factors such as the formats of the service (i.e., case or group), contents of the service (i.e., different programs with different emphases), types of clients served (e.g., sex offenders, drug offenders, or batterers), and treatment settings (e.g., prison or community; Lipsey & Landenberger, 2006). In addition, a recent study conducted by Pardini (2016) affirmed that cognitive–behavior treatment strategies produced larger effects for youth with more severe delinquency behaviors than those with less severe delinquency behaviors (Pardini, 2016).

Lipsey, Landenberger, and Wilson (2007) further examined the specific factors that might be related to positive outcomes in CBI for juvenile delinquents. They concluded that (1) the inclusion of anger control and interpersonal problem-solving components in the treatment program was associated with a larger effect size in recidivism; (2) none of the major CBI brand-name programs (e.g., The Reasoning and Rehabilitation Program, Moral Reconation Therapy, and Aggression Replacement Training) produced effects on recidivism that were significantly larger than the average effects of the other programs. They suggested that generic CBI approach might be responsible for the overall positive effects on recidivism; (3) high-quality implementation might be related to low proportions of treatment dropouts; (4) interventions with specific well-defined protocols (Pardini, 2016) and close monitoring of the quality and fidelity of treatment were needed to ensure good outcomes; and (5) adequate CBI training for providers was necessary for achieving good outcomes. In addition, a meta-analysis revealed that helping offenders to identify the precursors to offending behavior (i.e., the cognitive and behavioral chain) and role-playing high-risk situation (i.e., relapse prevention) were found to be effective in reducing future criminal behaviors (Dowden, Antonowicz, & Andrews, 2003). This information had informed the design of our culturally attuned CBI program for Chinese delinquent youth in Hong Kong.

Chinese subjects

A search of the literature found only three articles evaluating the effectiveness of CBI for working with Chinese youth with delinquent behaviors. Cheung and Ngai (2013) examined the effects of CBI for increasing drug-free days of young service users in Hong Kong. Results indicated that CBI reduced playfulness (i.e., a belief in fun seeking and experimentation) and lengthened the drug-free days. However, this study had several limitations. First, the target group was youth with drug abuse and the findings might not be applicable to delinquent youth in Hong Kong. Second, there was an absence of a comparative analysis of the findings between the experimental and control subjects. Third, there was no continuous or follow-up assessment to examine the longer term effect of the intervention.

In China, there were two clinical trials evaluating cognitive–behavior therapy (CBT) for Chinese male violent juvenile offenders. Chen and colleagues (2014) evaluated the effectiveness of CBI in reducing overt aggressive behaviors in Chinese young male violent offenders in a reform school in Changsha, China. The experimental group received 8 weeks of CBI treatment, while the control group received treatment as usual. The CBI program followed the Williams Life Skills Training Program (Williams & Williams, 1997), which leaned toward a more behavioral approach with skills training centering on assertiveness, problem-solving, and communication skills. Results revealed a significantly greater reduction in aggressive behaviors for those offenders in the experimental group than those who were in the control group. Similarly, Zhang, Wang, Chen, Zhou, and Wang (2015) found that the CBI program could improve coping styles and interpersonal support compared to the control group after 9 weeks of the program. However, the CBI approach adopted in these studies was more behaviorally oriented, with little emphasis on treatments for cognitive dysfunctions among delinquent youth.

It would be of interest to find out whether a more cognitively oriented CBI approach would work well for Chinese juvenile youth in Hong Kong. Second, the target population was institutionalized young male offenders. It is necessary to examine whether CBI would be effective for juvenile youth in the community setting.

Cultural Consideration

According to Ngai and Cheung (2008), Hong Kong delinquent youth are characterized by weak moral beliefs and have higher endorsement of beating, gang activity, bullying, and use of illicit substances. They also appear to be more susceptible to friends’ approval of illegal activities/delinquent behaviors, such as frequent vandalism, beating, gang activity, bullying, and use of illicit substances (Ngai & Cheung, 2008). From these descriptions, it is not difficult to gauge that juvenile delinquents in Hong Kong may hold strongly to certain values and beliefs that reinforce their delinquent behaviors. For example, they believe that “if my gangster friends ask me for help, I must do so because it is a sign of comradeship,” “I must act tough and strong, otherwise, I would be looked down by friends/gang members,” and “others are responsible for making me angry and acting violently.” When planning our CBI program, our team took into consideration these characteristics and aimed to help delinquent youth to examine their antagonistic values and beliefs in gang activities and to develop strategies to modify these values and beliefs. Second, it is necessary to help them develop cognitive and behavioral self-control strategies to minimize their easily provoked and aggressive behaviors.

Traditional family values and beliefs may also constitute sources of stress and conflict for juvenile delinquents in Hong Kong and may further aggravate the intensity and frequency of their delinquent behaviors. Under the Confucian ethics, each member in a family must follow the prescribed standards of behavior and performance that are inherent in their respective roles. A failure to perform such behaviors can result in shame and guilt borne by individual members (Goodwin & Tang, 1996). Thus, it is not uncommon for delinquent youth to feel strongly that they have brought disgrace to their families and have let their families down. For parents, they may also experience strong shame and guilt, because they have failed to raise their children according to the socially expected standards. Indeed, families with delinquent youth are characterized by high familial conflicts, and such conflicts can be a source of further delinquency among delinquent youth in Chinese societies (Deng & Roosa, 2007). Indeed, the impacts of poor parent–child relationships on juvenile delinquency among Chinese youth have been highlighted in a recent meta-analytic study conducted by Weng, Ran, and Chui (2016). However, while these phenomena are by no means exclusive to Chinese families with delinquent youth, Chinese cultural values and beliefs are pervasive and can intensify the familial conflicts among those who hold strongly to these traditional beliefs. Consequently, family members who do not know how to face and handle the constant rule-breaking behaviors manifested by their delinquent youth, and who have strong internalized stigma regarding having a child with delinquent behaviors, may withhold from seeking outside help. In response to these cultural issues, our intervention intended to facilitate Chinese delinquent youth to examine their dysfunctional values and beliefs behind their familial conflicts and to help them learn to engage in perspective taking and other strategies to modify their rigidly held beliefs.

Individual and/or Group Treatments

Generally speaking, overseas programs for juvenile delinquents embrace a holistic approach, which provides intensive individual support and nurturing, as well as small group counseling, for juvenile delinquents (e.g., Esbensen & Maxon, 2012). This holistic approach does not only provide personalized attention, which is much needed for building a close and trustful relationship with the delinquent youth, but small group counseling also encourages young people to learn to build relationships and interact effectively with others—opportunities to talk about issues and concerns, mutual respect, acceptance, and to learn to negotiate the challenges and barriers in their circumstances.

In Hong Kong, a combined individual and group approach in working with juvenile delinquents is also commonly adopted in outreaching social services. Academics and social work practitioners advocate a “case in group work in natural environment” approach to work with delinquent youth (Lee, 2005). Essentially, social workers approach Chinese delinquent youth in places where they frequently go to and work with them individually in a nonstructured group setting. Previous studies suggested that this approach is effective in facilitating changes in delinquent youth in Hong Kong (e.g., Cheung & Ngai, 2007). In our program, CBI in individual and group counseling was incorporated into the generic outreaching social services for juvenile delinquents in Hong Kong. While a CBI approach, which included emotion management, cognitive skills training, social skills training, and relapse prevention, was used in individual counseling for delinquent youth, CBI in group counseling focused on deepening delinquent youths’ understanding and application of strategies in handling their negative emotions.

Based on the work of Lipsey et al. (2007), our project set out to develop a culturally attuned CBI program that is incorporated into the outreaching social services for Chinese delinquent youth in Hong Kong and to examine whether this CBI program would be effective for these delinquent youth. A time series design is common in intervention studies to examine intervention effects. This study set out to investigate the effects of the intervention by comparing scores at baseline, 6 months, and 12 months to assess the process of change (Gliner & Morgan, 2000). This study had the following hypothesis:
The participants in the experimental condition would show more improvements in delinquency-related attitudes and behaviors (i.e., fewer major, minor, and triad delinquent behaviors and less peer influence and fewer gang affiliations) than the participants in the control condition at the end of the 12-month intervention.

The participants in the experimental condition would show a lower level of impulsivity than the participants in the control condition at the end of the 12-month intervention.

Method

Participants

A quasi-experimental matched-pairs comparison design was employed in this study. The participants in the experimental and control groups were recruited by the outreaching social workers of The Hong Kong Federation of Youth Groups (HKFYG). The selection criteria included (a) aged between 14 and 24, (b) Cantonese-speaking Chinese, (c) met the screening requirement in the Hong Kong Delinquent Behavior Scale, and (d) had a minimum level of motivation for change. Those who had psychiatric illness and suicidal attempt in the past 3 months were excluded. The matched-pairs criteria included (a) gender, (b) similar age (±3 years), (c) education level (±2 years), and (d) self-proclaimed gang membership.

Since Wilson, Bouffard, and MacKenzie (2005) found an overall small-to-medium mean effect size for CBT programs, sample size was calculated based on an effect size of .40 for clinical outcome research. The minimum sample size was calculated by GPower 3.1.3. For 80% power, an α error of .05 and a test of two independent groups, the total required number was 36. The participants were recruited by HKFYG, a social service organization in Hong Kong. A total of 44 participants were recruited to join the groups, 22 adolescents participated in the experimental group and another 22 matched-pairs cases in the control group. However, two participants from each group did not complete the project. Three of them withdrew from the services, due to a lack of interest (two from experimental and one from control groups), and one from the control group was incarcerated and could not continue to join.

There were 20 participants in the experimental group and 20 in the control group, with disproportionately more males than females (i.e., 90% and 10%). Demographic information between the experimental and control groups is displayed in Table 1. In the experimental group, the mean age was about 16.65 years, and more than half of them were students. About 65% had achieved junior secondary school (Grades 7–9) levels of schooling. Most participants lived in public housing estates and were living with their parents and have a family income between HK$10,000 and HK$20,000 (Table 1). For participants of the control group, the mean age was about 16.60, and the majority of participants were employed. The majority were junior secondary school students (Grades 7–9) or had achieved such levels of schooling. Most participants lived in public housing estates and were living with their parents and have a family income between HK$10,000 and HK$20,000 (Table 1). χ² tests and analyses of variance (ANOVAs) were performed and revealed no statistical significance in demographic characteristics between the participants of the experimental and control groups. This profile suggested that the participants belonged to lower socioeconomic backgrounds.

Table 1.
Demographics of Participants.

Outcome Variables Control Group (N = 20) Experimental Group (N = 20)

n (%) n (%)

Gender Male 18 (90.00) 18 (90.00)

Female 2 (10.00) 2 (10.00)

Birth place HK 17 (85.00) 16 (80.00)

China (mean years of stay in HK) 3 (8.00) 4 (11.00)

Age Mean (SD) 16.60 (1.67) 16.65 (1.93)

Employment

Full time 5 (25.00) 5 (25.00)

Part time 5 (25.00) 2 (10.00)

Unemployed 2 (10.00) 2 (10.00)

Student 8 (40.00) 11 (55.00)

Education Primary and below 1 (5.00) —

S.1–S.3 17 (85.00) 13 (65.00)

S.4–S.6 2 (10.00) 7 (35.00)

Housing type Public housing 12 (60.00) 12 (60.00)

Village housing 1 (5.00) 3 (15.00)

Private 3 (15.00) 3 (15.00)

Subsidized 2 (10.00) 2 (10.00)

Others 2 (10.00) —

Living in the same household^a With mum 15 (75.00) 13 (65.00)

With dad 15 (75.00) 18 (90.00)

With siblings 13 (65.00) 4 (20.00)

With others 3 (15.00) 3 (15.00)

Monthly household income HK$5,000–9,999 5 (25.00) 3 (15.00)

HK$10,000–19,999 8 (40.00) 10 (50.00)

HK$20,000+ 7 (35.00) 7 (35.00)

^aVariable that allows more than one primary answer.

Measures

The research team developed a self-administered questionnaire containing the following instruments.

Hong Kong Delinquency Behaviors (HKDBS) Scale

HKDBS was modified from the 17-item Delinquency Scale developed by Miller, Melnick, Barnes, Sabo, and Farrell (2007). The scale was used to measure juvenile delinquency behaviors of the respondents. The scale is composed of three subscales: Minor Delinquency subscale, Major Delinquency subscale, and Triad Delinquency subscale. The scale asks about the frequency in which the participants have engaged in a series of delinquency behaviors in the past few months. The scale ranges from 0 to 4 (0 = never, 1 = seldom, 2 = sometimes, 3 = often, and 4 = always). A pilot test of this scale was done with around 50 delinquent youth who were invited to rate and comment on the list of delinquent behaviors. In this study, the Cronbach’s α for the Major Delinquency subscale (pre-intervention = .77, 6 months = .72, 12 months = .73), Minor Delinquency subscale (pre-intervention = .64, 6 months = .61, 12 months = .68), and Triad Delinquency subscale (pre-intervention = .84, 6 months = .88, 12 months = .89) were acceptable. The reliability of the whole scale was good (pre-intervention = .88, 6 months = .87, 12 months = .89).

Resistance to Peer Influence Scale

This scale was developed to assess the degree to which adolescents act autonomously in interactions with their peer groups (Steinberg & Monahan, 2007). The Chinese version has been used in previous studies to assess risky behaviors in Chinese adolescents, with an acceptable Cronbach’s α of .66 (An et al., 2013). The scale consists of 10 items. Each item presents the respondent with two statements and two options of response. An example of such item is “some people accompany their friends to make them happy” (i.e., “very true to me” or “partly true to me”) or “some people refuse to accompany their friends even though they know their friends would be unhappy” (i.e., “very true to me” or “partly true to me”). The respondents have to indicate which group and which option they prefer. Similar to the results of An et al.’s (2013) study, the Cronbach’s α in the present study was considered marginally acceptable (pre-intervention = .49, 6 months = .68, 12 months = .63).

Attitudes toward gangs questionnaire

This questionnaire was developed by Nadel, Spellmann, Alvarez-Canino, Lausell-Bryant, and Landsbero (1996). This scale has been used in youth gang research in other Asian contexts (Chu, Daffern, Thomas, Ang, & Long, 2014). It is comprised of 9 items with true or false responses. The questions intend to examine respondents’ attitudes and bonding to gang activities and membership. The internal consistency of this measure was .74, with high school-age students. In the present study, this scale achieved an acceptable reliability (pre-intervention = .65, 6 months = .71, 12 months = .65).

Barratt Impulsiveness Scale (BIS-11)

The 11th version of BIS-11 is used in this study. This is a self-rated questionnaire comprised of 30 statements that assess the level of impulsivity of the respondent (Patton, Stanford, & Barratt, 1995). The scale measures three aspects: cognitive impulsiveness, motor impulsiveness, and nonplanning impulsiveness. It uses a 4-point scale, which ranges from 1 (never/rarely) to 4 (almost always). Higher summative score denotes higher level of impulsivity. The validated Chinese translation of BIS-11 was adopted for use in this study (Chan et al., 2003). This scale had acceptable reliability across all three subscales with cognitive impulsiveness subscale reaching (pre-intervention = .66, 6 months = .61, 12 months = .60; motor impulsiveness: pre-intervention = .55, 6 months = .64, 12 months = .63; and nonplanning impulsiveness: pre-intervention = .68, 6 months = .60, 12 months = .75). The reliability of the full scale was good (pre-intervention = .76, 6 months = .68, 12 months = .76).

Procedures

Figure 1 presents a flow diagram of participants at each stage of the study. Informed consent was obtained from either the parents/guardians of the participants or the participants themselves. All participants were provided with a battery of self-administered instruments, including basic demographics and the outcome measurements. The social workers collected the questionnaires at three time points: pre-intervention, at the end of the 6-month, and at the end of the 12-month intervention. Ethical approval had been sought from City University of Hong Kong’s Ethics Committee.

Figure 1.
Flow diagram of intervention through the stages of the study.

Interventions

Routine counseling

Routine individual counseling services were given for 1 year to participants in the control conditions by the outreaching social workers of HKFYG. Generally speaking, the key elements of the individual counseling services included practical support and advice, problem-solving skills and social skills training, crisis intervention, and family counseling. Nothing about CBI materials was shared in the counseling process. The counseling services were provided every other week, and each session lasted for 45–90 min, with an average of 16 sessions, over a 1-year period.

CBI individual and group counseling

The participants in the experimental condition were provided with CBI individual counseling for an average of 16 sessions, over a 1-year period, by the outreaching social workers. Since this is a difficult-to-engage target group, it is understandable that a much longer duration of counseling sessions was expected. Moreover, a few of these sessions were devoted to crisis management and little CBI work could be done. Individual counseling in CBI followed the generic model, which included emotion management training, cognitive skills training (such as interpersonal problem-solving, perspective taking, and long-term planning), and other skills training (such as social skills training and relapse prevention) (Lipsey et al., 2007). In the initial phase of the intervention, the participants were facilitated by the social workers to explore their motivations for change and to introduce the CBI model to the participants. They were also helped to learn to self-monitor their patterns of dysfunctional responses to stressful and unpleasant situations, to recognize their “hot” or “trigger” cognitions, and to understand how these negative thoughts and consequential maladaptive behaviors ran in a vicious cycle. In the middle phase, the participants had the opportunity to learn and practice the various adaptive cognitive and behavioral coping strategies, such as thought stopping, cognitive restructuring, and positive self-talk. In addition, they were assisted to examine how their personal values and beliefs, albeit socially and culturally constructed, might be affecting their interpersonal relationships with families and peers and to learn CBI strategies to relax their rigid beliefs and values. Each participant was seen every 2 weeks by an outreaching social worker, and each session lasted for 45–90 min.

CBI group counseling was also run for the participants who were interested in joining the group. Each group had six sessions, 2 hr each, with each group having four to five participants over 1½ months. The group contents revolved around teaching emotion management skills—self-monitoring of their patterns of dysfunctional responses to stressful and unpleasant situations, strategies for recognizing their “hot” or “trigger” cognitions, understanding how these negative thoughts and consequential maladaptive behaviors ran in a vicious cycle, and, lastly, learning adaptive cognitive and behavioral coping strategies, such as thought stopping, cognitive restructuring, and positive self-talk.

Treatment Fidelity

In order to ensure fidelity to CBI treatments, all involved social workers had to achieve an undergraduate level of social work qualification with a minimum of 3 years of outreaching social work experience. They had to undergo a 1-year CBT training and supervision provided by a social work academic, who was a cognitive therapist, trained at the Beck Institute, and a clinical supervisor, who had received CBT training from a local institution. The training package included (1) a 3-day theoretical and practical skills training in individual and group CBT counseling, (2) an eight-session self-exploration and reflection group run by the clinical supervisor to facilitate the social workers to appreciate the therapeutic essences of Beck’s CBT framework, (3) another eight-session CBT skills training to help social workers learn and practice the major CBT skills, and lastly (4) monthly supervision for a period of 12 months.

Statistical Analyses

Data analyses were performed on an intention to treat principle, with the missing data for each outcome measure inputted, using “last observation carried forward” (Streiner & Geddes, 2001). Differences in each demographic variable between the experimental and control groups were examined using χ² analyses for categorical variables and ANOVAs for continuous variables. A series of 2 × 3 ANOVAs were performed to examine the main and interaction effects of time and group on the outcome measures at pre-intervention and at the end of 6-month and 12-month intervention (Howell, 2013). Assumption tests for ANOVAs such as independence of observations, normal distribution, and homogeneity of variance were met. Post hoc pairwise comparisons using Bonferroni test were used to examine the level of significance between pre-intervention and 6-month intervention and pre-intervention and 12-month intervention in the outcome measures in the experimental group (Howell, 2013). Cohen’s d was also used to measure the magnitude of change between pre-intervention and 6-month intervention and pre-intervention and 12-month intervention in both experimental and control groups (Cohen, 1988).

Results

The means, standard deviations, and interaction effects of the outcome measures are presented in Tables 2 and 3. The effect sizes for all outcome variables for the experimental and control group are presented in Table 4. Significant Time × Group Effects were found for overall impulsivity (F = 4.27, p = .03), nonplanning impulsivity (F = 9.42, p < .001), cognitive impulsivity (F = 3.27, p = .05), and total delinquency (F = 3.71 p = .03). This reveals that there were patterns of difference between the experimental and control groups after the implementation of intervention in the specified variables. Since interaction effects were significant, simple main effects and post hoc test using Bonferroni adjustment were conducted to identify where the differences lie (i.e., pre-intervention, 6-month intervention, 12-month intervention within each group). Simple main effects for overall delinquency were significant for experimental group (F = 13.38, p < .001) and control group (F = 3.61, p = .04). Further pairwise comparisons using Bonferroni adjustment revealed significant differences between pre-intervention and 6-month intervention (M _Diff = 12.95, p < .001, Cohen’s d = 0.68) and between pre-intervention and 12-month intervention (M _Diff = 13.89, p < .001 Cohen’s d = 0.62) in the experimental group, while significant differences were observed only between pre-intervention and 12-month intervention in the control group (M _Diff = 7.35, p = .04, 95% confidence interval [CI] = [0.18, 14.52], Cohen’s d = 0.60). The results cast favors toward the experimental group, showing greater and continuous positive changes in overall delinquency.

Table 2.
Means and Standard Deviations for Pre-intervention and at 6 Months and 12 Months Interventions by Outcome Measures.

Experimental Group Pre-intervention 6 Months 12 Months

M (SD) M (SD) M (SD)

Gang affiliations 4.35 (2.21) 4.45 (2.19) 2.91 (1.58)

Resistance to peer influence 2.38 (0.48) 2.51 (0.53) 2.69 (0.52)

Impulsivity (total) 76.27 (11.86) 70.52 (8.62) 68.88 (9.11)

Cognitive impulsivity 19.75 (3.75) 17.30 (2.74) 17.10 (2.59)

Motor impulsivity 24.74 (3.91) 24.82 (4.42) 24.63 (4.55)

Nonplanning impulsivity 30.20 (5.93) 28.40 (4.36) 27.15 (5.83)

Delinquency (total) 25.75 (15.98) 12.80 (8.35) 11.86 (12.57)

Major delinquency 7.25 (5.95) 2.80 (3.55) 2.45 (3.68)

Minor delinquency 10.85 (5.34) 6.80 (3.30) 6.46 (3.17)

Triad delinquency 7.65 (8.39) 3.20 (4.41) 3.20 (4.41)

Control group

Gang affiliations 4.05 (1.99) 3.80 (3.42) 2.73 (2.22)

Resistance to peer influence 2.49 (0.33) 2.55 (0.05) 2.63 (0.43)

Impulsivity (total) 73.54 (8.60) 74.49 (8.35) 72.82 (9.65)

Cognitive impulsivity 19.75 (4.62) 20.00 (4.36) 18.50 (4.11)

Motor impulsivity 26.32 (4.23) 25.61 (4.49) 25.79 (4.90)

Nonplanning impulsivity 24.84 (4.08) 30.05 (4.39) 28.63 (4.47)

Delinquency (total) 18.30 (10.65) 16.05 (12.14) 10.95 (8.18)

Major delinquency 4.95 (4.17) 7.25 (5.95) 3.45 (3.25)

Minor delinquency 8.40 (3.98) 7.25 (4.46) 6.25 (4.76)

Triad delinquency 4.95 (4.77) 5.35 (6.60) 2.45 (3.75)

Table 3.
Interaction Effects, Pairwise Comparison of Outcome Measures.

Outcome Variables Interaction Effects (Time × Group) Experimental Group Control Group

Bonferroni Pairwise Comparison Bonferroni Pairwise Comparison

M _Diff [95% CIs] M _Diff [95% CIs]

F Pre-6 Months Pre-12 Months Pre-6 Months Pre-12 Months

Gang affiliations 0.29 −0.10 [−0.91, 0.71] 1.44 [0.23, 2.66] 0.25 [−0.38, 0.25] 1.33* [0.11, 2.54]

Resistance to peer influence 0.62 −0.13 [−0.45, 0.19] −0.31 [−0.63, 0.01] −0.07 [−0.38, 0.25] −0.14 [−0.46, 0.18]

Impulsivity (total) 4.26* 5.75 [0.95, 10.55] 7.39 [1.49, 13.28] −0.95 [−5.75, 3.85] 0.72 [−5.18, 6.62]

Cognitive impulsivity 3.27* 2.45*** [1.00, 3.90] 2.65* [0.41, 4.89] −0.25 [−1.70, 1.20] 1.25 [−0.99, 3.49]

Motor impulsivity 0.11 −0.08 [−3.60, 3.45] 0.11 [−3.63, 3.85] 0.71 [−2.81, 4.24] 0.53 [−3.20, 4.27]

Nonplanning impulsivity 9.42* 1.80 [−1.39, 4.99] 3.05 [−0.77, 6.88] −5.21* [−8.40, −2.02] −3.79 [−7.62, 0.03]

Delinquency total 3.71* 12.95* [5.66, 20.24] 13.89* [6.72, 21.06] 2.25 [−5.04, 9.54] 7.35* [0.18, 14.52]

Major delinquency 1.94 4.45* [1.33, 7.57] 4.80* [2.12, 7.48] 1.50 [−1.62, 4.62] 2.95* [0.27, 5.63]

Minor delinquency 2.22 4.05* [1.62, 6.48] 4.39* [1.34, 7.45] 1.15 [−1.28, 3.58] 2.15 [−0.90, 5.21]

Triad delinquency 2.77 4.45* [0.55, 8.35] 4.70** [1.20, 8.20] −0.40 [−4.30, 3.50] 2.25 [−1.25, 5.75]

Note. CI = confidence interval.

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

Table 4.
Effect Size of Outcome Measures Between Experimental Group and Control Group.

Outcome Variables Experimental Group Control Group

Effect Size Effect Size Effect Size Effect Size

Cohen’s d Cohen’s d Cohen’s d Cohen’s d

Pre-6 Months Pre-12 Months Pre-6 Months Pre-12 Months

Gang affiliations −0.05 0.74 0.11 0.63

Resistance to peer influence 0.26 0.62 0.17 0.37

Impulsivity (total) 0.55 0.70 −0.11 0.08

Cognitive impulsivity 0.75 0.82 −0.06 0.29

Motor impulsivity −0.02 0.03 0.16 0.12

Nonplanning impulsivity 0.35 0.54 −1.23 −0.89

Delinquency total 0.93 0.97 0.20 0.77

Major delinquency 0.91 0.97 −0.40 0.89

Minor delinquency 0.91 0.83 −0.27 0.49

Triad delinquency 0.68 0.62 −0.07 0.60

Note. Cohen’s d with a negative value denotes change in the undesired direction (i.e., increased).

Simple main effects were also tested for overall impulsivity. The results showed significant changes in the experimental group only (F = 4.90, p = .01), and the changes in control group remained insignificant (F = 1.34, p = .27). Pairwise comparisons using Bonferroni adjustment revealed significant differences between pre-intervention and 6-month intervention (M _Diff = 5.75, p = .01, Cohen’s d= 0.55) and between pre-intervention and 12-month intervention (M _Diff =7.38, p = .01, Cohen’s d = 0.70) in the experimental group. Similarly, simple main effects were tested for cognitive impulsivity, and significant changes were observed in the experimental group only (F = 8.92, p < .001), while no significant changes were found in the control group (F = 2.07, p = .14). Pairwise comparisons using Bonferroni adjustment revealed significant differences in experimental group between pre-intervention and 6-month intervention (M _Diff = 2.45, p < .001, Cohen’s d = 0.75) and pre--intervention and 12-month intervention (M _Diff = 2.65, p = .02, Cohen’s d = 0.82). Lastly, simple main effects for nonplanning impulsivity were examined. The control group showed a significant increase in the undesirable direction (F = 8.39, p = .01), while the experimental group showed an insignificant decrease in the desirable direction (F = 1.94, p = .16). Pairwise comparisons using Bonferroni adjustment revealed significant increases in the control group between pre-intervention and 6-month intervention (M _Diff = −5.21, p ≤ .001, 95% CI [−8.40, −2.02], Cohen’s d = −1.23) as well as between pre-intervention and 12-month intervention (M _Diff = −3.79, p = .05, 95% CI [−7.62, 0.03], Cohen’s d = −0.89). Overall speaking, the results showed that CBI significantly improved overall impulsivity and cognitive impulsivity when compared to the control group. Effect size statistics measuring the magnitude of change in outcome measures in the participants of the experimental group revealed mainly large improvements in delinquency-related attitudes and behaviors (i.e., Cohen’s d ranging from 0.62 to 0.97) at the end of 12-month intervention. On the other hand, most outcome variables in the control group also achieved some improvements, with delinquency-related attitudes and behaviors achieving mainly moderate improvements (with Cohen’s d ranging from 0.37 to 0.89). As hypothesized, the participants in the experimental group appeared to show improvements in more outcome measures and had achieved greater magnitude of improvements in outcome measures than the participants in the control group at 12-month intervention.

Discussion and Applications to Practice

To begin with, our hypothesis was partially confirmed. When compared to the control group, the participants in the experimental group had shown significant improvements in a number of emotion-related (i.e., overall impulsivity, cognitive impulsivity) and delinquency-related attitude and behavioral outcomes (i.e., total delinquency), and the magnitudes of change in these areas were mostly moderate to large over a 12-month period. The results are encouraging, because these constitute initial empirical evidence that supports the effectiveness of CBI for Chinese delinquent youth in a community setting. These positive changes echo previous studies found overseas (e.g., Lipsey & Landenberger, 2006; Pearson et al., 2002) and may be ascribed to the fact that our CBI program was able to modify participants’ impulsivities (i.e., cognitive and overall impulsivities), which in turn led to a reduction in their frequencies of delinquent behavior. However, due to a small sample size, our study could not perform correlational or regression analyses to ascertain such relationships. Future studies with a larger sample size should aim to identify these possible process variables that may have contributed to the outcomes in a CBI program for delinquent youth. This information is of paramount importance, because previous studies had found that impulsivity and negative emotions were related to the development and maintenance of delinquent behaviors (Calvete & Cardenoso, 2005; Chen et al., 2014). Indeed, if proven to be true, such findings can help practitioners develop CBI programs that target facilitating delinquent youth to understand the cognitive and behavioral processes underlying their impulsive acts and negative emotions and to learn adaptive cognitive and behavioral coping strategies to deal with these issues.

Our findings also revealed that participants in the control group had achieved significant changes in some outcome measures. The findings support previous studies that suggested outreaching social services in Hong Kong are generally effective in helping Chinese delinquent youth (Cheung & Ngai 2007; Chui, 2001). These results should not be surprising, because the participants in the control group also received 12 months of counseling from qualified outreaching social workers who used conventional generic counseling approach to work with juvenile delinquents. However, what is noteworthy is that the inclusion of CBI in outreaching social services appears to have yielded greater scope and magnitude of therapeutic changes in the outcome measures for juvenile delinquents. It is therefore possible that the incorporation of a specific therapeutic approach in the generic model of outreaching social services, such as CBI, may be able to create greater therapeutic benefits for juvenile delinquents in Hong Kong. On a different note, this culturally attuned CBI program for Chinese adolescents with delinquent behaviors, if and with further adaptations in specific sociocultural contexts, may be useful for Chinese adolescents with similar concerns in other Chinese communities, such as Taiwan and China. However, given the preliminary nature of this study, future studies should gather more empirical evidence to ascertain the benefits of CBI for Chinese adolescents with delinquent behaviors in Hong Kong and elsewhere.

There are several limitations to this study. First, since this study had a rather small sample size and was the first of its kind to explore the effectiveness of this CBI intervention for delinquent youth in Hong Kong, future studies should replicate this study and increase its sample, so that the results can be more robust. Second, it is necessary to conduct a follow-up test, say 6–12 months after the cessation of intervention, to ascertain the longer term effects of CBI for Chinese juvenile delinquents in Hong Kong. Third, due to the quasi-experimental nature of this study design, there might be confounding factors that had affected the outcomes of this study (e.g., John Henry effect and diffusion/imitation of treatments, etc.). Future studies should attempt to increase the internal validity of this kind of study through randomization, single- or double-blinded trial, increase in the sample size, and the choice of comparison group. Finally, all Cronbach’s α statistics for the outcome variables were found to be within the marginally acceptable level (.60–.70) to acceptable (.70–.80; DeVellis, 2012) except for Resistance to Peer Influence Scale at pre-intervention. This might be due to the small sample size of this study that could have increased the variability of the scores.

To conclude, this study is the first of its kind to systematically study the effectiveness of a culturally attuned CBI program for juvenile delinquents in Hong Kong. Preliminary results appear to lend support toward the effectiveness of CBIs in reducing some of the delinquency-related attitudes and behaviors and emotional outcomes measures. From a research point of view, it is necessary to conduct a larger scale study with a more vigorous research design so as to provide more solid evidence to support the effectiveness of CBI for Chinese delinquent youth. As a practitioner, it is advisable for the government to provide funding for social service agencies to train outreaching social workers to be equipped with a clinical approach, such as CBI in working with juvenile delinquents. Many social workers have basic training in counseling skills in Hong Kong and elsewhere but may not be adequately trained to practice a specific intervention approach in working with delinquent youth.

Outcome Variables		Control Group (N = 20)	Experimental Group (N = 20)
Gender	Male	18 (90.00)	18 (90.00)
Female	2 (10.00)	2 (10.00)
Birth place	HK	17 (85.00)	16 (80.00)
China (mean years of stay in HK)	3 (8.00)	4 (11.00)
Age	Mean (SD)	16.60 (1.67)	16.65 (1.93)
Employment
	Full time	5 (25.00)	5 (25.00)
Part time	5 (25.00)	2 (10.00)
Unemployed	2 (10.00)	2 (10.00)
Student	8 (40.00)	11 (55.00)
Education	Primary and below	1 (5.00)	—
S.1–S.3	17 (85.00)	13 (65.00)
S.4–S.6	2 (10.00)	7 (35.00)
Housing type	Public housing	12 (60.00)	12 (60.00)
Village housing	1 (5.00)	3 (15.00)
Private	3 (15.00)	3 (15.00)
Subsidized	2 (10.00)	2 (10.00)
Others	2 (10.00)	—
Living in the same household^a	With mum	15 (75.00)	13 (65.00)
With dad	15 (75.00)	18 (90.00)
With siblings	13 (65.00)	4 (20.00)
With others	3 (15.00)	3 (15.00)
Monthly household income	HK$5,000–9,999	5 (25.00)	3 (15.00)
HK$10,000–19,999	8 (40.00)	10 (50.00)
HK$20,000+	7 (35.00)	7 (35.00)

Experimental Group	Pre-intervention	6 Months	12 Months
Gang affiliations	4.35 (2.21)	4.45 (2.19)	2.91 (1.58)
Resistance to peer influence	2.38 (0.48)	2.51 (0.53)	2.69 (0.52)
Impulsivity (total)	76.27 (11.86)	70.52 (8.62)	68.88 (9.11)
Cognitive impulsivity	19.75 (3.75)	17.30 (2.74)	17.10 (2.59)
Motor impulsivity	24.74 (3.91)	24.82 (4.42)	24.63 (4.55)
Nonplanning impulsivity	30.20 (5.93)	28.40 (4.36)	27.15 (5.83)
Delinquency (total)	25.75 (15.98)	12.80 (8.35)	11.86 (12.57)
Major delinquency	7.25 (5.95)	2.80 (3.55)	2.45 (3.68)
Minor delinquency	10.85 (5.34)	6.80 (3.30)	6.46 (3.17)
Triad delinquency	7.65 (8.39)	3.20 (4.41)	3.20 (4.41)
Control group
Gang affiliations	4.05 (1.99)	3.80 (3.42)	2.73 (2.22)
Resistance to peer influence	2.49 (0.33)	2.55 (0.05)	2.63 (0.43)
Impulsivity (total)	73.54 (8.60)	74.49 (8.35)	72.82 (9.65)
Cognitive impulsivity	19.75 (4.62)	20.00 (4.36)	18.50 (4.11)
Motor impulsivity	26.32 (4.23)	25.61 (4.49)	25.79 (4.90)
Nonplanning impulsivity	24.84 (4.08)	30.05 (4.39)	28.63 (4.47)
Delinquency (total)	18.30 (10.65)	16.05 (12.14)	10.95 (8.18)
Major delinquency	4.95 (4.17)	7.25 (5.95)	3.45 (3.25)
Minor delinquency	8.40 (3.98)	7.25 (4.46)	6.25 (4.76)
Triad delinquency	4.95 (4.77)	5.35 (6.60)	2.45 (3.75)

Outcome Variables	Interaction Effects (Time × Group)	Experimental Group	Control Group
Gang affiliations	0.29	−0.10 [−0.91, 0.71]	1.44* [0.23, 2.66]	0.25 [−0.38, 0.25]	1.33* [0.11, 2.54]
Resistance to peer influence	0.62	−0.13 [−0.45, 0.19]	−0.31 [−0.63, 0.01]	−0.07 [−0.38, 0.25]	−0.14 [−0.46, 0.18]
Impulsivity (total)	4.26*	5.75** [0.95, 10.55]	7.39** [1.49, 13.28]	−0.95 [−5.75, 3.85]	0.72 [−5.18, 6.62]
Cognitive impulsivity	3.27*	2.45*** [1.00, 3.90]	2.65* [0.41, 4.89]	−0.25 [−1.70, 1.20]	1.25 [−0.99, 3.49]
Motor impulsivity	0.11	−0.08 [−3.60, 3.45]	0.11 [−3.63, 3.85]	0.71 [−2.81, 4.24]	0.53 [−3.20, 4.27]
Nonplanning impulsivity	9.42***	1.80 [−1.39, 4.99]	3.05 [−0.77, 6.88]	−5.21*** [−8.40, −2.02]	−3.79 [−7.62, 0.03]
Delinquency total	3.71*	12.95*** [5.66, 20.24]	13.89*** [6.72, 21.06]	2.25 [−5.04, 9.54]	7.35* [0.18, 14.52]
Major delinquency	1.94	4.45*** [1.33, 7.57]	4.80*** [2.12, 7.48]	1.50 [−1.62, 4.62]	2.95* [0.27, 5.63]
Minor delinquency	2.22	4.05*** [1.62, 6.48]	4.39*** [1.34, 7.45]	1.15 [−1.28, 3.58]	2.15 [−0.90, 5.21]
Triad delinquency	2.77	4.45* [0.55, 8.35]	4.70** [1.20, 8.20]	−0.40 [−4.30, 3.50]	2.25 [−1.25, 5.75]

Outcome Variables	Experimental Group	Control Group
Gang affiliations	−0.05	0.74	0.11	0.63
Resistance to peer influence	0.26	0.62	0.17	0.37
Impulsivity (total)	0.55	0.70	−0.11	0.08
Cognitive impulsivity	0.75	0.82	−0.06	0.29
Motor impulsivity	−0.02	0.03	0.16	0.12
Nonplanning impulsivity	0.35	0.54	−1.23	−0.89
Delinquency total	0.93	0.97	0.20	0.77
Major delinquency	0.91	0.97	−0.40	0.89
Minor delinquency	0.91	0.83	−0.27	0.49
Triad delinquency	0.68	0.62	−0.07	0.60

Footnotes

Authors’ Note

The authors alone are responsible for the content and writing of the article.

Declaration of Conflicting Interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Daniel F. K. Wong, Xiao Yu Zhuang, and Priscilla S. Y. Ip declare no conflict of interest, while Man Ho Chan is employed under Hong Kong Federation of Youth Groups as a social work supervisor.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was funded by both the Hong Kong Federation of Youth Groups and The Social Welfare Development Fund of the HKSAR Government.

References

Sun

Wang

Mai

J. C.

Liang

J. P.

… Tao

F. B.

(2013). Correlation of resistance to peer pressure and risky decision-making with adolescent health risk behaviors. Chinese Journal of Preventive Medicine, 47, 238–244.

Calvete

Cardenosa

(2005). Gender differences in cognitive vulnerability to depression and behavior problems in adolescents. Journal of Abnormal Child Psychology, 33, 179–192.

Chan

Chen

E. Y. H.

Chan

R. C. K.

Hung

S. F.

Cheung

E. F. C.

Law

C. W.

Lam

M. L.

(2003). A comparison of psychological profiles between suicidal ideators and non ideators among psychiatric patients: A preliminary report. Schizophrenia Research, 60, 337.

Che

W. K.

(1998). The types and trends of juvenile delinquency in Hong Kong. Journal of Youth Studies, 1. Retrieved from http://yrc.hkfyg.org.hk/page.aspx?corpname=yrc&i=4494

Chen

Wang

J. J.

Zhou

J. S.

Wang

X. P.

(2014). Cognitive behavioral therapy to reduce overt aggression behavior in Chinese young male violent offenders. Aggressive Behavior, 40, 329–336.

Cheung

C. K.

Ngai

S. S. Y.

(2007). Effective group work with delinquents in Hong Kong. Adolescence, 42, 151–165.

Cheung

C. K.

Ngai

S. S. Y.

(2013). Reducing youth’s drug abuse through training social workers for cognitive–behavioral integrated treatment. Children and Youth Services Review, 35, 302–311.

Chu

C. M.

Daffern

Thomas

S. D. M.

Ang

Long

(2014). Criminal attitudes and psychopathic personality attributes of youth gang offenders in Singapore. Psychology, Crime & Law, 20, 284–301.

Chui

W. H.

(2001). Users’ views of an outreaching social work team in Hong Kong. International Social Work, 44, 455–469.

10.

Chui

W. H.

Chan

O. H. C.

(2012). Criminal recidivism among Hong Kong male juvenile probationers. Journal of Child Family Studies, 21, 857–868.

11.

Cohen

(1988). Statistical power analysis for the behavioral sciences. Mahwah, NJ: Lawrence Erlbaum.

12.

Deng

Roosa

M. W.

(2007). Family influences on adolescent delinquent behaviors: Applying the social development model to a Chinese sample. American Journal of Community Psychology, 40, 433–444.

13.

DeVellis

R. F.

(2012). Scale development: Theory and applications. Los Angeles, CA: Sage.

14.

Dowden

Antonowicz

Andrews

(2003). The effectiveness of relapse prevention with offenders: A meta-analysis. International Journal of Offender Therapy and Comparative Criminology, 47, 516–528.

15.

Esbensen

F. A.

Maxon

C. L.

(2012). Youth gangs in international perspective: Results from the Eurogang program of research. New York, NY: Springer.

16.

Gliner

J. A.

Morgan

G. A.

(2000). Research methods in applied settings. Mahwah, NJ: Lawrence Erlbaum.

17.

Goodwin

Tang

C. K. S.

(1996). Chinese personal relationships. In Bond

M. H.

(Ed.), The handbook of Chinese psychology (pp. 294–308). Hong Kong: Oxford University Press.

18.

Hong Kong Police Force. (2011). Hong Kong crime statistics. Retrieved from http://www.police.gov.hk/

19.

Howell

(2013). Statistical methods for psychology (8th ed.). Belmont, CA: Wadsworth.

20.

Lee

F. W. L.

(2005). Effectiveness of services to “youth-at risk”: The case of outreaching social work. In Lee

F. W. L.

(Ed.), Working with youth-at-risk in Hong Kong (pp. 87–98). Hong Kong: Hong Kong University Press.

21.

Lipsey

M. W.

Landenberger

N. A.

(2006). Cognitive behavioral programs for juvenile and adult offenders: A meta-analysis of controlled intervention. Retrieved from www.campbellcollaboration.org

22.

Lipsey

M. W.

Landenberger

N. A.

Wilson

S. J.

(2007). Effects of cognitive–behavioral programs for criminal offenders. Campbell Systematic Reviews, 6, 1–27.

23.

Miller

K. E.

Melnick

M. J.

Barnes

G. M.

Sabo

Farrell

M. P.

(2007). Athletic involvement and adolescent delinquency. Journal of Youth and Adolescence, 36, 711–723.

24.

Nadel

Spellmann

Alvarez-Canino

Lausell-Bryant

Landsbero

(1996). The cycle of violence and victimization: A study of the school based intervention of a multidisciplinary youth violence prevention program. American Journal of Preventive Medicine, 12, 109–119.

25.

Ngai

N. P.

Cheung

C. K.

(2008). Subcultural factors underlying Chinese marginal youth’s gang involvement: A comparative study. In Bendit

Hahn-Bleibtreu

(Eds.), Youth and transitions: Processes of social inclusion and patterns of vulnerability in a globalized world (pp.317–337). Germany: Barbara Budrich Leverkusen Opladen.

26.

Pardini

(2016). Empirically based strategies for preventing juvenile delinquency. Child and Adolescent Psychiatric Clinics of North America, 25, 257–268.

27.

Patton

J. H.

Stanford

Barratt

E. S.

(1995). Factor analysis of the Barratt Impulsiveness Scale. Journal of Clinical Psychology, 51, 768–774.

28.

Pearson

F. S.

Lipton

D. S.

Cleland

C. M.

Yee

D. S.

(2002). The effects of behavioral/cognitive–behavioral programs on recidivism. Crime & Delinquency, 48, 476–496.

29.

Social Welfare Department. (2015). Funding and services agreement. Retrieved from www.swd.gov.hk/doc/fsa_sd/088a.pdf

30.

Steinberg

Monahan

K. C.

(2007). Age differences in resistance to peer influence. Developmental Psychology, 43, 1531–1543.

31.

Streiner

Geddes

(2001). Intention to treat analysis in clinical trials when there are missing data. Evidenced Based Mental Health, 4, 70–71.

32.

Webster’s New World College Dictionary. (2014). Retrieved from http://www.yourdictionary.com/juvenile-delinquency#websters.

33.

Weng

Ran

M. S.

Chui

W. H.

(2016). Juvenile delinquency in Chinese adolescents: An ecological review of the literature. Aggression and Violent Behavior. online first doi:10.1016/j.avb.2016.06.016

34.

Williams

R. B.

Williams

V. P.

(1997). Lifeskills. New York, NY: Random House.

35.

Wilson

D. B.

Bouffard

L. A.

MacKenzie

D. L.

(2005). A quantitative review of structured, group oriented, cognitive behavioral programs for offenders. Journal of Criminal Justice and Behavior, 32, 172–204.

36.

Zhang

Wang

Chen

Zhou

Wang

(2015). Efficacy of Williams Lifeskills Training in improving psychological health of Chinese male juvenile violent offenders: A randomized controlled study. Neuroscience Bulletin, 31, 53–60.

Outcome Variables		Control Group (N = 20)	Experimental Group (N = 20)
		n (%)	n (%)
Gender	Male	18 (90.00)	18 (90.00)
Gender	Female	2 (10.00)	2 (10.00)
Birth place	HK	17 (85.00)	16 (80.00)
Birth place	China (mean years of stay in HK)	3 (8.00)	4 (11.00)
Age	Mean (SD)	16.60 (1.67)	16.65 (1.93)
Employment
	Full time	5 (25.00)	5 (25.00)
	Part time	5 (25.00)	2 (10.00)
	Unemployed	2 (10.00)	2 (10.00)
	Student	8 (40.00)	11 (55.00)
Education	Primary and below	1 (5.00)	—
	S.1–S.3	17 (85.00)	13 (65.00)
	S.4–S.6	2 (10.00)	7 (35.00)
Housing type	Public housing	12 (60.00)	12 (60.00)
	Village housing	1 (5.00)	3 (15.00)
	Private	3 (15.00)	3 (15.00)
	Subsidized	2 (10.00)	2 (10.00)
	Others	2 (10.00)	—
Living in the same household^a	With mum	15 (75.00)	13 (65.00)
	With dad	15 (75.00)	18 (90.00)
	With siblings	13 (65.00)	4 (20.00)
	With others	3 (15.00)	3 (15.00)
Monthly household income	HK$5,000–9,999	5 (25.00)	3 (15.00)
	HK$10,000–19,999	8 (40.00)	10 (50.00)
	HK$20,000+	7 (35.00)	7 (35.00)

Experimental Group	Pre-intervention	6 Months	12 Months
Experimental Group	M (SD)	M (SD)	M (SD)
Gang affiliations	4.35 (2.21)	4.45 (2.19)	2.91 (1.58)
Resistance to peer influence	2.38 (0.48)	2.51 (0.53)	2.69 (0.52)
Impulsivity (total)	76.27 (11.86)	70.52 (8.62)	68.88 (9.11)
Cognitive impulsivity	19.75 (3.75)	17.30 (2.74)	17.10 (2.59)
Motor impulsivity	24.74 (3.91)	24.82 (4.42)	24.63 (4.55)
Nonplanning impulsivity	30.20 (5.93)	28.40 (4.36)	27.15 (5.83)
Delinquency (total)	25.75 (15.98)	12.80 (8.35)	11.86 (12.57)
Major delinquency	7.25 (5.95)	2.80 (3.55)	2.45 (3.68)
Minor delinquency	10.85 (5.34)	6.80 (3.30)	6.46 (3.17)
Triad delinquency	7.65 (8.39)	3.20 (4.41)	3.20 (4.41)
Control group
Gang affiliations	4.05 (1.99)	3.80 (3.42)	2.73 (2.22)
Resistance to peer influence	2.49 (0.33)	2.55 (0.05)	2.63 (0.43)
Impulsivity (total)	73.54 (8.60)	74.49 (8.35)	72.82 (9.65)
Cognitive impulsivity	19.75 (4.62)	20.00 (4.36)	18.50 (4.11)
Motor impulsivity	26.32 (4.23)	25.61 (4.49)	25.79 (4.90)
Nonplanning impulsivity	24.84 (4.08)	30.05 (4.39)	28.63 (4.47)
Delinquency (total)	18.30 (10.65)	16.05 (12.14)	10.95 (8.18)
Major delinquency	4.95 (4.17)	7.25 (5.95)	3.45 (3.25)
Minor delinquency	8.40 (3.98)	7.25 (4.46)	6.25 (4.76)
Triad delinquency	4.95 (4.77)	5.35 (6.60)	2.45 (3.75)

Outcome Variables	Interaction Effects (Time × Group)	Experimental Group		Control Group
		Bonferroni Pairwise Comparison		Bonferroni Pairwise Comparison
		M _Diff [95% CIs]		M _Diff [95% CIs]
	F	Pre-6 Months	Pre-12 Months	Pre-6 Months	Pre-12 Months
Gang affiliations	0.29	−0.10 [−0.91, 0.71]	1.44* [0.23, 2.66]	0.25 [−0.38, 0.25]	1.33* [0.11, 2.54]
Resistance to peer influence	0.62	−0.13 [−0.45, 0.19]	−0.31 [−0.63, 0.01]	−0.07 [−0.38, 0.25]	−0.14 [−0.46, 0.18]
Impulsivity (total)	4.26*	5.75** [0.95, 10.55]	7.39** [1.49, 13.28]	−0.95 [−5.75, 3.85]	0.72 [−5.18, 6.62]
Cognitive impulsivity	3.27*	2.45*** [1.00, 3.90]	2.65* [0.41, 4.89]	−0.25 [−1.70, 1.20]	1.25 [−0.99, 3.49]
Motor impulsivity	0.11	−0.08 [−3.60, 3.45]	0.11 [−3.63, 3.85]	0.71 [−2.81, 4.24]	0.53 [−3.20, 4.27]
Nonplanning impulsivity	9.42***	1.80 [−1.39, 4.99]	3.05 [−0.77, 6.88]	−5.21*** [−8.40, −2.02]	−3.79 [−7.62, 0.03]
Delinquency total	3.71*	12.95*** [5.66, 20.24]	13.89*** [6.72, 21.06]	2.25 [−5.04, 9.54]	7.35* [0.18, 14.52]
Major delinquency	1.94	4.45*** [1.33, 7.57]	4.80*** [2.12, 7.48]	1.50 [−1.62, 4.62]	2.95* [0.27, 5.63]
Minor delinquency	2.22	4.05*** [1.62, 6.48]	4.39*** [1.34, 7.45]	1.15 [−1.28, 3.58]	2.15 [−0.90, 5.21]
Triad delinquency	2.77	4.45* [0.55, 8.35]	4.70** [1.20, 8.20]	−0.40 [−4.30, 3.50]	2.25 [−1.25, 5.75]

Outcome Variables	Experimental Group		Control Group
	Effect Size	Effect Size	Effect Size	Effect Size
	Cohen’s d	Cohen’s d	Cohen’s d	Cohen’s d
	Pre-6 Months	Pre-12 Months	Pre-6 Months	Pre-12 Months
Gang affiliations	−0.05	0.74	0.11	0.63
Resistance to peer influence	0.26	0.62	0.17	0.37
Impulsivity (total)	0.55	0.70	−0.11	0.08
Cognitive impulsivity	0.75	0.82	−0.06	0.29
Motor impulsivity	−0.02	0.03	0.16	0.12
Nonplanning impulsivity	0.35	0.54	−1.23	−0.89
Delinquency total	0.93	0.97	0.20	0.77
Major delinquency	0.91	0.97	−0.40	0.89
Minor delinquency	0.91	0.83	−0.27	0.49
Triad delinquency	0.68	0.62	−0.07	0.60