Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Mar 1.
Published in final edited form as: J Res Adolesc. 2013 Feb 14;23(1):128–137. doi: 10.1111/j.1532-7795.2012.00801.x

Linking Life Skills and Norms With Adolescent Substance Use and Delinquency in South Africa

Mary H Lai 1, John W Graham 1, Edward A Smith 1, Linda L Caldwell 1, Stephanie A Bradley 1, Tania Vergnani 1, Cathy Mathews 1, Lisa Wegner 1
PMCID: PMC3613155  NIHMSID: NIHMS366566  PMID: 23559844

Abstract

We examined factors targeted in two popular prevention approaches with adolescent drug use and delinquency in South Africa. We hypothesized adolescent life skills to be inversely related, and perceived norms to be directly related to later drug use and delinquency. Multiple regression and a relative weights approach were conducted for each outcome using a sample of 714 South African adolescents ages 15 to 19 years (M = 15.8 years, 57% female). Perceived norms predicted gateway drug use. Conflict resolution skills (inversely) and perceived peer acceptability (directly) predicted harder drug use and delinquency. The “culture of violence” within some South African schools may make conflict resolution skills more salient for preventing harder drug use and delinquency.

According to the 2001 South Africa Census, young people ages 10–19 comprise 22% of the country’s population (Statistics South Africa, 2003). South African adolescents are developing within the post-Apartheid context of major political, economic, and social changes (Naude, 2001; Parry, Plüddemann, Louw, & Leggett, 2004), including racial discrimination and increasing violence (Brook, Morojele, Pahl, & Brook, 2006). This context promotes a constellation of risks for the development of adolescent substance use and delinquency and threatens the public health of South Africa.

Survey results tell a sobering tale. According to a national survey of youth and young adults ages 15–24 in South Africa, over 50% have tried alcohol and over 10% have tried drugs in their lifetime (Pettifor et al., 2004). Results from the 2008 National Youth Risk Behaviour Survey (NYRBS) conducted among 8th- to 11th-grade students showed lifetime substance use rates of 49.6% for alcohol, 29.5% for cigarettes, 12.7% for marijuana, 12.2% for inhalants, and 6.6% for methamphetamines (Reddy et al., 2010). Adolescent marijuana and alcohol use is more prevalent in the U.S. than in South Africa (Reddy, Resnicow, Omardien, & Kambaran, 2007). However, rates of lifetime “hard” drug use (cocaine, heroin, injectable drugs, and methamphetamines) are higher in South Africa. The increasing prevalence of methamphetamine use among South African adolescents and adults in recent years is of particular concern to public health (Plüddemann, Myers, & Parry, 2008).

Few empirical research or surveillance data exist on delinquent behaviors among South African adolescents. The results of the 2008 NYRBS indicate that 21.0% of South African adolescents reported getting into a physical fight at school in the past 6 months (Reddy et al., 2010). Significantly more South African males than females reported carrying weapons (such as gun, knife, “panga” or “kierrie”—South African terms for long knives or sticks, respectively) and being involved in a physical fight in the past month (Reddy et al., 2010). Some research also suggests that substance use co-occurs with delinquency among South African adolescents (Flisher, Ziervogel, Chalton, Leger, & Roberts, 1996). Because adolescent substance use and delinquency may share common risk factors (Hawkins, Jenson, Catalano, & Lishner, 1988), it is important to further the research in this area.

In Western samples, two successful approaches for preventing adolescent substance use and delinquency are the promotion of life skills and changing youths’ perception of social norms. The life skills promotion approach builds adolescents’ competencies for interpersonal relations, decision-making, critical thinking, and emotional coping skills (Mangrulkar, Whitman, & Posner, 2001). The Life Skills Training (LST) program developed by Botvin and colleagues in the U.S. is considered a successful intervention for reducing adolescent substance use (Botvin & Griffin, 2004) and delinquency (Botvin, Griffin, & Nichols, 2006). LST is theoretically based in social learning theory (Bandura, 1977) and problem behavior theory (Jessor & Jessor, 1977). From these perspectives, adolescent substance use and delinquency are behaviors learned from the interplay of adolescents’ contextual and personal risk factors.

Social norms approaches are based on the idea that perceptions of normal peer behaviors are important predictors of those behaviors (Fishbein & Ajzen, 1975). Hansen and colleagues showed that perceptions of peer norms regarding substance use are powerful predictors of young people’s own use and cessation of use (Hansen & McNeal, 2001). They also showed that a program designed to correct (lower) misperceptions of the prevalence and acceptability of peer substance use reduces young people’s own use (Hansen & Graham, 1991). Further, youth beliefs about substance use prevalence and acceptability mediated the beneficial effects of that substance abuse prevention program (Donaldson, Graham, & Hansen, 1994).

In South Africa, the life skills approach is part of a compulsory school subject called Life Orientation which prepares youth to “live meaningfully and successfully in a rapidly changing society” (Department of Basic Education, 2011). In their survey of youth-serving, substance-use prevention programs, Harker and colleagues (Harker, Myers, & Perry, 2008) noted that although 69% of school-based programs surveyed provided life skills training, this type of training is underrepresented among community-based programs in South Africa and may be a productive focus for future prevention efforts.

Current Study

Our overall aim is to conduct etiological research that contributes to the development of preventive interventions targeting adolescent substance use and delinquency in South Africa. We examine factors that are targeted in two popular prevention approaches (promotion of life skills and changing social norms around substance use). We hypothesize that an adolescent’s positive skills in anger and anxiety management, decision-making and risk management, and conflict resolution will be negatively associated with later substance use and delinquency. We also hypothesize that perceived acceptability and perceived prevalence of alcohol, cigarette, and marijuana use among peers will be positively associated with later substance use and delinquency. Substance use in South Africa consistently differs by gender, with males more likely to have used any substances and reporting a higher prevalence of ever using substances than females (Madu & Matla, 2003; Reddy et al., 2010), therefore we also include gender in the model.

Method

Procedures

Data for the current study came from the control group sample of an efficacy trial of HealthWise: South Africa (HWSA), a school-based prevention program for 8th and 9th graders conducted from 2004–2008 (Caldwell et al., 2004; Wegner et al., 2008). The goals of HWSA are to reduce adolescent substance use and risky sex and to promote positive leisure use. HWSA was adapted specifically to the South African context although the program model integrates aspects of prevention programs developed in the U.S. The theoretical framework and conceptualization of HWSA, and the process to culturally adapt the program to South Africa are described elsewhere (Caldwell et al., 2004; Wegner, Flisher, Caldwell, Vergnani, & Smith, 2008). HWSA is the result of an international collaboration between faculty at The Pennsylvania State University (PSU) in the United States, and in South Africa, The University of the Western Cape (UWC), The University of Cape Town (UCT), and the Western Cape Education Department. PSU has been the primary recipient of grant funds for this project from the National Institute on Drug Abuse (NIDA); UWC faculty and staff have assumed the primary role of program coordination and maintaining relationships with the schools, teachers, and students involved in the HWSA trial. Members of all three academic institutions have participated in data analysis and the dissemination of findings.

Parent consent and youth assent to participate in the study were collected prior to data collection. Surveys were administered using handheld computers (Palm Pilots) at the beginning and end of each school year of the study. Study protocols and data collection instruments were approved by the human subjects research review boards at both The Pennsylvania State University in the U.S. and Stellenbosch University in South Africa, which at the time served as the IRB body for research conducted through UWC.

Participants

Four treatment schools were randomly selected from secondary schools in the Mitchell’s Plain area that were deemed by local project collaborators as being functionally able to participate (6 of 25 secondary schools were excluded from the selection pool due to concerns regarding severe overcrowding, safety, and high levels of school disorganization). Five comparison schools were then matched with treatment schools based on demographic and other socioeconomic characteristics. The present study sample consisted of 714 adolescents from Cohort 3 (M= 15.8 years at the end of 9th grade, 57% female). The majority (94%) of youth in the current study sample are Coloured, a racial category used in South Africa for persons of mixed European, African, and Asian ancestry. Students reside in the Mitchell’s Plain township of South Africa where 48% of households earn less than the household subsistence level of R19, 200 ($2743 US) per year (Statistics South Africa, 2003). The most commonly spoken language at home is English; nearly 60% speak both English and Afrikaans at home.

Measures

All alpha values reported in this article were calculated in our data for the scales described below.

Life skills

Students reported their level of confidence in their ability to use, as well as their use of, personal and social skills. Ten items were adapted from the survey used to evaluate the Life Skills Training program (Botvin, Baker, Dusenbury, Tortu, & Botvin, 1990), two items from each of five skill areas: Anxiety Management (“When I feel anxious or nervous, I imagine myself being calm and relaxed”); Anger Management (“I am confident I can control my anger”); Decision Making (“I am confident I make good decisions”); Risk Avoidance (“I am confident I can avoid risky situations”); and Conflict Resolution (“I think it is important to resolve conflicts (fights or arguments) peacefully”). Students rated these questions on a scale of 0 (strongly disagree) to 4 (strongly agree). Preliminary analysis showed that Anxiety and Anger Management items were highly correlated and did not represent distinct constructs in our data. Similarly, Decision Making and Risk Avoidance items were highly correlated and did not represent distinct constructs in our data. Factor analysis of the 10 items suggested that the 10 items loaded on just three factors; thus we standardized and averaged the items to form three scales: (1) Anxiety/and Anger Management (4 items, α = .74); (2) Decision Making/Risk Management (4 items, α = .83); and (3) Conflict Resolution (2 items, α = .67).

Perceived acceptability and perceived prevalence

Perceived peer acceptability of substance use was measured through three items (adapted from Hansen & Graham, 1991) that asked whether most of each participating student’s friends think it is okay for someone their age to smoke cigarettes, drink alcohol, and use marijuana (α = .77). Response options were dichotomous such that a higher value represented agreement with the statement. Perceived prevalence of school-wide substance use was measured through three items (also adapted from Hansen & Graham, 1991) that asked the student how many learners their age at their school smoke cigarettes, drink alcohol, and use marijuana at least once a month (α = .78). Students rated these questions on a scale of 0 (none of them) to 3 (most of them).

Adolescent substance use

For each substance, a new variable was created from two items adapted from Hansen & Graham (1991): a dichotomous item asking whether student used a particular substance in the past month; and the follow-up item for “yes” responses asking frequency of past month use. Responses for the alcohol item, for example, included 0 (did not use in the past month), 1 (1 or less drinks), 2 (2–3 drinks), and 3 (4 or more drinks). Then, composite scales were created representing past month frequency of alcohol, cigarette, and marijuana use (also known as gateway drugs since the use of these drugs commonly precede use of “hard” drugs, such as cocaine or heroin [Kandel, Yamaguchi, & Chen, 1992]; α = .64) and past month frequency of methamphetamine and inhalant (“sniff glue, paint, or petrol”) use (α = .48).

Delinquency

A delinquency scale was created using five items adapted from various large-scale youth surveys (Centers for Disease Control, 2007; Developmental Studies Center, 2005; Elliott, 1995; Liang, Flisher, & Lombard, 2007). Students reported their past month occurrence of bullying others at school, physically fighting (“hit, slapped, or physically hurt someone”), trespassing (“broken into a house, school, shop, or other building without permission”), vandalism (“caused serious damage to property that did not belong to you”), and stealing something worth more than 100 Rand (approximately $15 US; α = .73). Response options were dichotomous such that a higher value represented past month occurrence of the behavior.

Covariates

Gender (coded as 0=male, 1=female) and students’ prior substance use and delinquency were added as controls in the statistical models. Race and age were not included as covariates because the majority of our sample is of Coloured race and are from the same grade cohort in the study.

Analysis Plan

Our study goals were to examine the effects of life skills and perceived substance use norms (perceived acceptability and prevalence) on substance use (gateway drugs and harder drugs) and delinquency outcomes. Independent variables from Wave 4 (end of 9th grade) and were used to predict outcomes at Wave 5 (start of 10th grade). Multiple regression analyses were conducted for each dependent variable (DV) using SAS statistical software, version 9.2 (SAS Institute Inc., Cary, NC), controlling for gender, and using multiple imputation to deal with missing data.

Regression models were conducted without and with prior problem behaviors as covariates, respectively (i.e., an unconditional compared with a conditional model; Dwyer et al., 1989). The unconditional model assumes that all of the shared prediction with pretest measures of the DV belongs to the predictors; thus, prior problem behaviors are not included as covariates. The conditional model, which includes the pretest measure of the DV as a covariate, assumes that none of the shared prediction belongs to the predictors. However, it is more likely that some of the shared prediction does belong to the predictors. Johnson (2000) developed a statistical procedure for calculating the relative regression weight of independent variables in the presence of highly correlated predictors by dividing the shared prediction among predictors. We applied Johnson’s (2000) procedure for calculating relative weights and significance values to draw attention to the contribution of our independent variables of interest while controlling for prior problem behavior and gender.

Testing the significance of the relative weights has not been an easy matter; however, when there are no missing data, a straightforward solution is possible. The solution we used in this article stems from certain symmetries in the special case of multiple regression with orthogonal predictors, in which all of the prediction is carried in the regression coefficients. In this special case, the squared t-value for each predictor is proportional to the R2 attributable to each orthogonal predictor. Further, the sum of squared t-values is a constant regardless of the orthogonalization approach used. We obtained orthogonalized predictors using principal components and varimax rotated factor scores for our predictors. Then, we used these orthogonalized variables to predict our dependent variables and to calculate the sum of squared t-values. As a final step, we recalculate the squared t-values to be proportional to the relative weights (proportion of R2) as calculated by the Johnson (2000) procedure. These recalculated t-values with p-values, based on complete cases analysis, appear in the results tables described below along with results based on multiple imputation for the regular regression analyses.

Results

Regression Analyses

Tables 1, 2, and 3 present all of the regression results, and Table 4 presents a summary of main results. The top panel of each table presents the unconditional model for both the regular regression analysis and for the relative weights analysis. The pretest measure of the DV was significant in all of the conditional models; these effects are not presented below. Predictors with at least one statistically significant effect are described here.

Table 1.

Multiple Imputation Parameter Estimates for Alcohol, Cigarette, and Marijuana Use

Unconditional Model
Variable Regular Multiple Regression (with Multiple Imputation) Relative Weights (with Complete Cases)
b SE df FMI t p RW t p
Gender −0.200 0.065 797.73 0.35 −3.05 0.002 0.111 2.439 0.0151
Anger/anxiety mgmt 0.000 0.052 1051.80 0.31 0.00 NS 0.006 0.567 NS
Decision/risk mgmt 0.024 0.051 749.24 0.37 0.47 NS 0.031 1.289 NS
Conflict resolution −0.029 0.037 990.28 0.32 −0.79 NS 0.018 0.982 NS
Perceived acceptability 0.203 0.040 778.14 0.36 5.06 <.0001 0.594 5.643 <.0001
Perceived prevalence 0.138 0.042 644.90 0.39 3.28 0.001 0.24 3.587 0.0004
Conditional Model
Variable Regular Multiple Regression (with Multiple Imputation) Relative Weights (with Complete Cases)
b SE df FMI t p RW t p
Gender −0.075 0.055 365.14 0.52 −1.36 NS 0.012 2.438 0.0152
Anger/anxiety mgmt −0.006 0.041 492.83 0.45 −0.15 NS 0.002 0.995 NS
Decision/risk mgmt 0.072 0.043 346.62 0.54 1.68 0.090 0.011 2.334 0.020
Conflict resolution 0.004 0.034 471.93 0.46 −0.14 NS 0.002 0.995 NS
Perceived acceptability −0.001 0.034 392.51 0.50 −0.02 NS 0.063 5.586 <.0001
Perceived prevalence 0.026 0.035 339.55 0.54 0.75 NS 0.027 3.657 0.0003
Prior substance use 0.745 0.042 265.48 0.61 17.61 <.0001 0.884 20.926 <.0001
Open in a new tab

FMI = fraction of missing information. RW = relative weight as proportion of R2.

Table 2.

Multiple Imputation Parameter Estimates for Methamphetamine and Inhalant Use

Unconditional Model
Variable Regular Multiple Regression (with Multiple Imputation) Relative Weights (with Complete Cases)
b SE df FMI t p RW t p
Gender 0.044 0.078 490.2 0.452 0.56 NS 0.108 1.516 NS
Anger/anxiety mgmt 0.024 0.060 695.4 0.379 0.40 NS 0.011 0.484 NS
Decision/risk mgmt −0.014 0.060 469.02 0.462 −0.23 NS 0.010 0.461 NS
Conflict resolution −0.132 0.047 421.11 0.487 −2.83 0.005 0.293 2.496 0.013
Perceived acceptability 0.183 0.055 295.4 0.582 3.36 0.001 0.478 3.188 0.002
Perceived prevalence 0.001 0.048 513.63 0.441 0.02 NS 0.100 1.458 NS
Conditional Model
Variable Regular Multiple Regression (with Multiple Imputation) Relative Weights (with Complete Cases)
b SE df FMI t p RW t p
Gender 0.142 0.067 276.2 0.472 1.95 0.052 0.073 1.599 NS
Anger/anxiety mgmt −0.015 0.048 372.57 0.433 −0.26 NS 0.003 0.324 NS
Decision/risk mgmt 0.030 0.048 330.18 0.511 0.52 NS 0.009 0.562 NS
Conflict resolution −0.100 0.034 405.71 0.527 −2.27 0.023 0.098 1.853 0.065
Perceived acceptability 0.091 0.042 256.85 0.569 1.83 0.068 0.133 2.159 0.031
Perceived prevalence 0.023 0.039 309.59 0.484 0.52 NS 0.048 1.297 NS
Prior substance use 0.435 0.054 215.54 0.819 5.84 <.0001 0.635 4.717 <.0001
Open in a new tab

FMI = fraction of missing information. RW = relative weight as proportion of R2.

Table 3.

Multiple Imputation Parameter Estimates for Delinquency

Unconditional Model
Variable Regular Multiple Regression (with Multiple Imputation) Relative Weights (with Complete Cases)
b SE df FMI t p RW t p
Gender −0.235 0.069 297.67 0.58 −3.41 0.001 0.415 3.651 <.0001
Anger/anxiety mgmt −0.017 0.051 429.77 0.48 −0.34 NS 0.002 0.253 NS
Decision/risk mgmt 0.036 0.05 356.03 0.53 0.72 NS 0.012 0.621 NS
Conflict resolution −0.066 0.035 472.15 0.46 −1.87 0.060 0.148 2.18 0.030
Perceived acceptability 0.118 0.043 276.89 0.501 2.74 0.007 0.335 3.281 0.001
Perceived prevalence 0.026 0.041 344.53 0.539 0.64 NS 0.088 1.681 0.093
Conditional Model
Variable Regular Multiple Regression (with Multiple Imputation) Relative Weights (with Complete Cases)
b SE df FMI t p RW t p
Gender −0.119 0.067 276.2 0.602 −1.78 0.076 0.092 2.976 0.003
Anger/anxiety mgmt −0.023 0.048 372.57 0.518 −0.47 NS 0.002 0.439 NS
Decision/risk mgmt 0.072 0.048 330.18 0.55 1.51 NS 0.014 1.161 NS
Conflict resolution −0.046 0.034 405.71 0.496 −1.38 NS 0.04 1.962 0.050
Perceived acceptability 0.057 0.042 256.85 0.624 1.38 NS 0.075 2.687 0.007
Perceived prevalence 0.032 0.039 309.59 0.568 0.83 NS 0.033 1.782 0.075
Prior delinquency 0.371 0.054 215.54 0.681 6.88 <.0001 0.744 8.462 <.0001
Open in a new tab

FMI = fraction of missing information. RW = relative weight as proportion of R2.

Table 4.

Summary of Main Results (t-Values Displayed)

Gateway Drugs Regular Regression Relative Weights
UC Cond. UC Cond.
 Perceived acceptability +5.06 −0.02 5.64 5.59
 Perceived prevalence +3.28 +0.75 3.59 3.66
 Decision/risk management +0.47 +1.68 1.29 2.33
Hard Drugs
 Conflict resolution −2.83 −2.27 2.50 1.85
 Perceived prevalence +3.36 +1.83 3.19 2.16
Delinquency
 Conflict resolution −1.87 −1.38 2.18 1.96
 Perceived prevalence +2.74 +1.38 3.28 2.69
Open in a new tab

Note. Effects (t-values) in bold are significant (p < .05) in the predicted direction. Effects in italics and underscored are significant or marginal in the direction opposite from what was predicted. All other effects were non-significant. Relative weights, as percents of R2, are all positive. UC = Unconditional Model (omitted pretest of DV); Cond. = Conditional Model (included pretest of DV as covariate).

Alcohol, cigarette, and marijuana use

Table 1 presents the regression results for gateway drugs as the DV. Gender was associated with gateway drug use (males used more than females). The effect reached statistical significance in both relative weights models and in the unconditional model with regular regression. Perceived acceptability was positively related to later drug use, and was statistically significant in both unconditional models and in the relative weights version of the conditional model. Perceived prevalence was positively related to later drug use and was statistically significant in both unconditional models and in the relative weights version of the conditional model. The decision and risk management predictor was positive (better management, more drug use) in all four models. However, the effect was marginally significant in the regular regression version of the conditional model and was significant in the relative weights version of the conditional model.

Methamphetamine and inhalant use

Table 2 presents the results for methamphetamine and inhalant (hard drug) use as the DV. Better conflict resolution was associated with less drug use; the effect was significant for both versions of the unconditional model and for the regular regression version of the conditional model. The effect was marginally significant for the relative weights version of the conditional model. Perceived acceptability was positively related to use of these drugs (higher perceived acceptability was associated with more use). The effect was significant for both versions of the unconditional model and for the relative weights version of the conditional model. The effect was marginally significant for the regular regression version of the conditional model.

Delinquency

Table 3 presents the results for delinquency as the DV. Gender was negatively related to delinquency (males reported more delinquent behaviors than females). The effect was significant for both versions of the unconditional model and for the relative weights version of the conditional model. The effect was marginally significant for the regular regression version of the conditional model. Conflict resolution was also related to this outcome (better conflict resolution was associated with less delinquency). The effect was statistically significant for the relative weights version of both unconditional and conditional models. The effect was marginally significant for the regular regression version of the unconditional model. Perceived acceptability was positively related to delinquency (more perceived acceptability was associated with more delinquency). The effect was significant for both versions of the unconditional model and for the relative weights version of the conditional model.

Summary of main results

Table 4 presents the t-values for the significant findings (p < .05) for each outcome. Gender, perceived acceptability, and prevalence of gateway drug use were associated with later gateway drug use in the hypothesized direction. Decision and risk management skills were associated with later gateway drug use but not in the hypothesized direction. Conflict resolution skills and perceived acceptability were associated with later hard drug use in the hypothesized direction. Gender (being male), conflict resolution skills, and perceived acceptability were associated with later delinquency in the hypothesized direction.

Discussion

Overall, results suggest that different developmental processes connect gateway drug use, hard drug use, and delinquency. Perceived acceptability and perceived prevalence of peer (gateway) substance use were consistently associated with an adolescent’s own gateway drug use, based on the relative weights in both the unconditional and conditional models. Experimentation with gateway drugs is prevalent during adolescence such that some substance use can be considered normative during this developmental period (e.g., Shedler & Block, 1990). The influence of peer norms on adolescent substance use and development in general is well known, especially as peers increase in importance during adolescence (Brown & Larson, 2009). Further research on understanding the relation of peer norms to adolescent substance use could compare the relative effects of perceived peer substance use versus peers’ actual substance use.

Our results suggest a slightly different story about hard drug use and delinquency. The lack of conflict resolution skills and perceived peer acceptability of gateway substance use were significantly associated with hard drug use and delinquency. Not surprisingly, our results showed the same “predictors” for hard substance use and delinquency since these problem behaviors are known to co-occur, possibly through common risk factors (Hawkins et al., 1988; Jessor & Jessor, 1977); however, a contribution of our study is that conflict resolution skills may be one way in which these problem behaviors are linked.

In South Africa many young people are exposed and vulnerable to ongoing violence. A recent study posited that there is a “culture of violence” in South Africa that manifests itself in the wider society, communities, schools, and families as a result of family dysfunction, poverty, violence perpetrated by students and by teachers through the use of corporal punishment (Centre for the Study of Violence and Reconciliation, 2010; van der Westhuizen & Maree, 2009). Given these circumstances it is likely that conflict resolution skills play a crucial role in helping adolescents negotiate the ongoing pressures and tensions in the school and may be critical to shaping adolescents’ developmental opportunities and pathways.

Association with delinquent peers may be another common link between delinquency and hard drug use. While our measures addressed peer substance use norms, differential association (Matsueda, 1988) and social learning theories (Bandura, 1977), for example, would suggest that an adolescent who is more sensitive to the behavior of his or her peers would be susceptible to the influence of, and model, delinquent peers. Compared to the “normative” nature of gateway drug use during adolescence, the use of hard drugs, such as methamphetamines and inhalants, is non-normative. These substances are highly addictive and have severely negative implications for brain development and functioning, and physical health (Parry, Myers, & Plüddemann, 2004). It could be that associating with delinquent peers encourages further antisocial behavior and exposure to more severe substances, such as methamphetamines and inhalants. Future research should investigate what mediates the association between perceived peer acceptability of substance use and delinquency.

Anger and anxiety management skills were not associated with any of our outcomes. This was an unexpected result given that anger and poor self-regulation skills are known to be related to early adolescent substance use (Swaim, Oetting, Edwards, & Beauvais 1989; Wills, 1986). There are two possible explanations for this finding. First, it is possible that these important intrapersonal skills are overshadowed by the need for social skills, such as conflict resolution, in high risk contexts. Second, some research suggests that intrapersonal factors, such as anger and anxiety, may be less relevant for predicting substance use during adolescence as peers play a more prominent role (Swaim et al., 1989). Another unexpected result of our study was that higher decision and risk management skills were associated with gateway drug use. However, this was a limited finding as it was a significant association in only one of four models and may reflect the “normative” experimentation in gateway drug use during adolescence. Overall, more research is needed to explore the roles of life skills in high risk contexts such as South Africa.

In terms of gender, our results showed that males were more likely to engage in gateway substance use and delinquency, corresponding with known trends (Reddy et al., 2010). However, we did not find any gender differences for methamphetamine and inhalant use in our study while other research has shown that males are more likely to try these hard drugs (Reddy et al., 2010).

Limitations

One limitation of this study is that all measures were based on students’ self-report. Thus, the life skills measures represent an individual’s perception of their ability to manage their anger, anxiety, decisions, risks, and conflict resolution values. Another limitation is that the study sample may not be representative of all South African adolescents because the majority of our sample were Coloured. It is also important to note that our study focused on individual-level characteristics and their associations with negative outcomes. From a social ecological theory perspective (Bronfenbrenner, 1995) there are clearly many other social and cultural factors that influence risk behaviors. As South Africa is still in the process of recovering from the lasting effects of apartheid, these issues are no doubt strong and further research should consider them. Despite these important limitations, however, our study highlighted the potential need for different prevention approaches to address adolescent gateway drug use rather than for hard drug use and delinquency.

Implications for Prevention

Study findings suggest that different prevention approaches may be needed depending on the type of substance targeted. A preventive and universal approach for decreasing gateway substance use should focus on changing adolescent perceptions about the acceptability and prevalence of substance use. As our study found that males were more likely than females to engage in gateway substance use as well as delinquency, further research is needed on the utility of gender-specific approaches to substance use and delinquency prevention.

The present study highlights that conflict resolution training may be a critical component in developing a targeted intervention focusing on higher-risk problem behaviors, such as methamphetamine and inhalant use and delinquency. Helping youth build a robust skill set for addressing multiple social and cultural pressures may better enable them to successfully avoid engaging in deleterious behaviors during adolescence and beyond. Without such skills, youth may resort to delinquency and hard drug use such as those in this study.

Due to the link between hard drug use and delinquency, prevention programs may need to address both of these behaviors in tandem (Hawkins et al., 1988). Program developers should note, however, that addressing co-occurring adolescent substance use and delinquency in the same prevention program may be not be as simple as combining what is known separately about substance use and delinquency. Research in the U.S. has shown an “intensification effect” of co-occurring substance use and delinquent behaviors in which the impact of co-occurring substance use and delinquency is more than simply double the effect of each problem behavior alone. For example, Tubman, Gil, and Wagner (2004) found that rates of adolescent substance use and delinquency were four times higher than those of adolescents who reported engaging in only substance use or delinquency. Prevention efforts attempting to reduce these co-occurring behaviors may need to be especially strengthened and warrant further research.

Acknowledgments

This study was supported by the National Institute on Drug Abuse award numbers R01 DA017491 (Smith), T32 DA017629 (Greenberg), and F31 DA028155 (Lai). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute on Drug Abuse or the National Institutes of Health.

References

  1. Bandura A. Social learning theory. Englewood Cliff, NJ: Prentice Hall; 1977. [Google Scholar]
  2. Botvin GJ. Preventing drug abuse in schools: Social and competence enhancement approaches targeting individual-level etiological factors. Addictive Behaviors. 2000;25(6):887–897. doi: 10.1016/S0306-4603(00)00119-2. [DOI] [PubMed] [Google Scholar]
  3. Botvin GJ, Griffin KW. Life Skills Training: Empirical findings and future directions. Journal of Primary Prevention. 2004;25(2):211–232. doi: 10.1023/B:JOPP.0000042391.58573.5b. [DOI] [Google Scholar]
  4. Botvin GJ, Baker E, Dusenbury L, Tortu S, Botvin EM. Preventing adolescent drug abuse through a multimodal cognitive-behavioral approach: Results of a three-year study. Journal of Consulting and Clinical Psychology. 1990;58:437–446. doi: 10.1037/0022-006X.58.4.437. [DOI] [PubMed] [Google Scholar]
  5. Botvin GJ, Griffin KW, Nichols TD. Preventing youth violence and delinquency through a universal school-based prevention approach. Prevention Science. 2006;7(4):403–408. doi: 10.1007/s11121-006-0057-y. [DOI] [PubMed] [Google Scholar]
  6. Bronfenbrenner U. Developmental ecology through space and time: A future perspective. In: Moen P, Elder GH Jr, Luscher K, editors. Examining lives in context: Perspectives on the ecology of human development. Washington, DC: American Psychological Association; 1995. pp. 619–647. [DOI] [Google Scholar]
  7. Brook JS, Morojele NK, Pahl K, Brook DW. Predictors of drug use among South African adolescents. Journal of Adolescent Health. 2006;38(1):26–34. doi: 10.1016/j.jadohealth.2004.08.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Brown B, Larson J. Peer relationships in adolescence. In: Lerner R, Steinberg L, editors. Handbook of adolescent psychology (Vol. 2): Contextual influences on adolescent development. 3. New York, NY: Wiley; 2009. pp. 74–103. [Google Scholar]
  9. Caldwell LL, Smith EA, Wegner L, Vergnani T, Mpofu E, Flisher A, Matthews C. HealthWise South Africa: Developing a Life Skills curriculum for young adults. World Leisure Journal. 2004;46(3):4–17. doi: 10.1080/04419057.2004.9674362. [DOI] [Google Scholar]
  10. Centers for Disease Control. Middle School Youth Risk Behavior Survey (YRBS) 2007 Retrieved from http://www.cdc.gov/HealthyYouth/yrbs/
  11. Centre for the Study of Violence and Reconciliation. Tackling armed violence: Key findings and recommendations of the Study on the Violent Nature of Crime in South Africa. Cape Town, South Africa: 2010. Retrieved from http://www.csvr.org.za/ [Google Scholar]
  12. Department of Basic Education. Curriculum and Assessment Policy Statement (CAPS): Life Orientation grades 7–9, Final Draft. Pretoria, South Africa: Author; 2011. [Google Scholar]
  13. Developmental Studies Center. Scales from student questionnaire, Child Development Project middle school student follow-up study (Grades 6–8) 2005 Retrieved from http://www.devstu.org/pdfs/cdp/DSC_MidSch_scales.pdf.
  14. Donaldson SI, Graham JW, Hansen WB. Testing the generalizability of intervening mechanism theories: Understanding the effects of adolescent drug use prevention interventions. Journal of Behavioral Medicine. 1994;17:195–216. doi: 10.1007/BF01858105. [DOI] [PubMed] [Google Scholar]
  15. Dwyer JH, MacKinnon DP, Pentz MA, Flay BR, Hansen WB, Wang EYI, Johnson CA. Estimating intervention effects in longitudinal studies. American Journal of Epidemiology. 1989;130:781–795. doi: 10.1093/oxfordjournals.aje.a115399. Retrieved from http://aje.oxfordjournals.org/ [DOI] [PubMed] [Google Scholar]
  16. Elliott D. National Youth Survey: Wave VII, 1987. Boulder, CO: Behavioral Research Institute; 1995. [Google Scholar]
  17. Fishbein M, Ajzen I. Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley; 1975. [Google Scholar]
  18. Flisher AJ, Ziervogel CF, Chalton DO, Leger PH, Roberts BA. Risk-taking behaviour of Cape Peninsula high-school students. Part IX. Evidence for a syndrome of adolescent risk behavior. South African Medical Journal. 1996;86(9):1090–1093. [PubMed] [Google Scholar]
  19. Hansen WB, Graham JW. Preventing alcohol, marijuana, and cigarette use among adolescents: Peer pressure resistance training versus establishing conservative norms. Preventive Medicine. 1991;20:414–430. doi: 10.1016/0091-7435(91)90039-7. [DOI] [PubMed] [Google Scholar]
  20. Hansen WB, McNeal RB. Self–initiated cessation from substance use: A longitudinal study of the relationship between postulated mediators and quitting. Journal of Drug Issues. 2001;31:957–976. Retrieved from http://www2.criminology.fsu.edu/~jdi/ [Google Scholar]
  21. Harker N, Myers B, Parry C. Audit of prevention programmes targeting substance use among young people in the greater Cape Town metropole: Technical report. Cape Town, South Africa: Medical Research Council; 2008. [Google Scholar]
  22. Harrison A, Newell ML, Imrie J, Hoddinott G. HIV prevention for South African youth: Which interventions work? A systematic review of current evidence. BMC Public Health. 2010;10 (102):1–12. doi: 10.1186/1471-2458-10-102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hawkins JD, Jenson JM, Catalano RF, Lishner DM. Delinquency and drug abuse: Implications for social services. Social Service Review. 1988;62(2):258–284. doi: 10.1086/644546. [DOI] [Google Scholar]
  24. Holborn L, Eddy G. First steps to healing the South African Family. Johannesburg, South Africa: South African Institute of Race Relations; 2011. [Google Scholar]
  25. Jessor R, Jessor SL. Problem behavior and psychosocial development: A longitudinal study of youth. New York, NY: Academic Press; 1977. [Google Scholar]
  26. Johnson JW. A heuristic method for estimating the relative weight of predictor variables in multiple regression. Multivariate Behavioral Research. 2000;35:1–19. doi: 10.1207/S15327906MBR3501_1. [DOI] [PubMed] [Google Scholar]
  27. Kandel DB, Yamaguchi K, Chen K. Stages of progression in drug involvement from adolescence to adulthood: Further evidence for the Gateway Theory. Journal of Studies of Alcohol. 1992;53:447–457. doi: 10.15288/jsa.1992.53.447. Retrieved from http://www.jsad.com/jsad/volumes. [DOI] [PubMed] [Google Scholar]
  28. Liang H, Flisher AJ, Lombard CJ. Bullying, violence, and risk behavior in South African school students. Child Abuse and Neglect. 2007;31:161–171. doi: 10.1016/j.chiabu.2006.08.007. [DOI] [PubMed] [Google Scholar]
  29. Madu SN, Matla MQP. Illicit drug use, cigarette smoking and alcohol drinking behaviour among a sample of high school adolescents in the Pietersburg area of the Northern Province, South Africa. Journal of Adolescence. 2003;26:121–136. doi: 10.1016/S0140-1971(02)00120-3. [DOI] [PubMed] [Google Scholar]
  30. Mangrulkar L, Whitman CV, Posner M. Life skills approach to child and adolescent healthy human development. Washington, DC: Pan American Health Organization; 2001. [Google Scholar]
  31. Matsueda RL. The current state of differential association theory. Crime and Delinquency. 1988;34:377–306. doi: 10.1177/0011128788034003005. [DOI] [Google Scholar]
  32. Moore G, Lemmer E. Towards cultural proficiency in South African secondary schools: Ethnodrama as educational tool. Education As Change. 2010;14(1):5–18. doi: 10.1080/16823206.2010.487356. [DOI] [Google Scholar]
  33. Naude B. South Africa. In: Hoffman AM, Summers RW, editors. Teen violence: A global view. Westport, CT: Greenwood Press; 2001. pp. 145–157. [Google Scholar]
  34. Parry CDH, Myers B, Plüddemann A. Editorial: Drug policy for methamphetamine use urgently needed. South African Medical Journal. 2004;94(12):964–965. Retrieved from http://www.samj.org.za/index.php/samj. [PubMed] [Google Scholar]
  35. Parry CDH, Plüddemann A, Louw A, Leggett T. The 3-metros study of drugs and crime in South Africa: Findings and policy implications. American Journal of Drug and Alcohol Abuse. 2004;30(1):167–185. doi: 10.1081/ADA-120029872. [DOI] [PubMed] [Google Scholar]
  36. Pettifor AE, Rees HV, Steffenson A, Hlongwa-Madikizela L, McPhail C, Vermaak K, Kleinschmidt I. HIV and sexual behavior among young South Africans: A national survey of 15–24 year olds. Johannesburg, South Africa: Reproductive Health Research Unit, University of the Witwatersrand; 2004. [Google Scholar]
  37. Plüddemann A, Myers BJ, Parry CDH. Surge in treatment admissions related to methamphetamine use in Cape Town, South Africa: Implications for public health. Drug and Alcohol Review. 2008;27:185–189. doi: 10.1080/09595230701829363. [DOI] [PubMed] [Google Scholar]
  38. Reddy SP, Panday S, Swart D, Jiabhai CC, Amosum SL, Monyeki JS, Van den Borne HW. Umthenthe uhlaba usamila: The South African Youth Risk Behaviour Survey 2002. Cape Town, South Africa: South African Medical Research Council; 2003. [Google Scholar]
  39. Reddy P, Resnicow K, Omardien R, Kambaran N. Prevalence and correlates of substance use among high school students in South Africa and the United States. American Journal of Public Health. 2007;97(10):1859–1864. doi: 10.2105/AJPH.2006.086330. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Reddy SP, James S, Sewpaul R, Koopman F, Funanl NI, Slfunda S, Omardien RG. Umthente uhlaba usamila – The South African Youth Risk Behavior Survey 2008. Cape Town, South Africa: South Africa Medical Research Council; 2010. [Google Scholar]
  41. Shedler J, Block J. Adolescent drug use and psychological health. American Psychologist. 1990;45:612–630. doi: 10.1037/0003-066X.45.5.612. [DOI] [PubMed] [Google Scholar]
  42. Snodgrass L, Blunt R. The value of play for conflict management: A case study. South African Journal of Education. 2009;29:53–67. doi: 10.1590/S0256-01002009000100004. [DOI] [Google Scholar]
  43. Statistics South Africa. Census 2001: Census in brief. Pretoria, South Africa: Statistics South Africa; 2003. [Google Scholar]
  44. Swaim RC, Oetting ER, Edwards RW, Beauvais F. Links from emotional distress to adolescent drug use: A path model. Journal of Consulting and Clinical Psychology. 1989;57(2):227–231. doi: 10.1037/0022-006X.57.2.227. [DOI] [PubMed] [Google Scholar]
  45. Tubman JG, Gil AG, Wagner EF. Co-occurring substance use and delinquent behavior during early adolescence: Emerging relations and implications for intervention strategies. Criminal Justice and Behavior. 2004;31(4):463–488. doi: 10.1177/0093854804265178. [DOI] [Google Scholar]
  46. Van der Westhuizen CN, Maree JG. The scope of violence in a number of Gauteng schools. Acta Criminologica. 2009;22(3):43–62. Retrieved from http://www.journals.co.za/crim/acta/index.html. [Google Scholar]
  47. Wegner L, Flisher AJ, Caldwell LL, Vergnani T, Smith EA. Healthwise South Africa: Cultural adaptation of a school-based risk prevention programme. Health Education Research. 2008;23(6):1085–1096. doi: 10.1093/her/cym064. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Wills TA. Stress and coping in early adolescence: Relationships to substance use in urban school samples. Health Psychology. 1986;5(6):503–529. doi: 10.1037/0278-6133.5.6.503. [DOI] [PubMed] [Google Scholar]

RESOURCES