Child Maltreatment, Relationship with Father, Peer Substance Use, and Adolescent Marijuana Use

Abstract

This longitudinal prospective study examined the relationship between child maltreatment as per reports to child protective services (CPS) and adolescent self-reported marijuana use, and the association between relationships with mothers and fathers and use of marijuana. The association between relationships with parents early in childhood (ages 6-8 years) and during adolescence with adolescent marijuana use were also probed. Another aim examined whether relationships with parents moderated the link between child maltreatment and youth marijuana use. The sample included 702 high risk adolescents from the Longitudinal Studies of Child Abuse and Neglect (LONGSCAN), a consortium of 5 studies related to maltreatment. Children were recruited at age 4 or 6 years together with their primary caregiver. Some were recruited due to their risk for child maltreatment, others were already involved with CPS, and children in one site had been placed in foster care.

Logistic regression analysis was performed using youth self-report of marijuana use as the criterion variable and child maltreatment and the relationships with parents as predictor variables, controlling for youths’ perceptions of peer substance use and parental monitoring, parental substance use, race/ethnicity, sex and study site. Approximately half the youth had used marijuana. Most of them described quite positive relationships with their mothers and fathers. Participant marijuana Use was associated with a poorer quality of relationship with mother during adolescence, and with peer and parental substance use. A better relationship with father, but not mother, during adolescence attenuated the connection between Child Maltreatment and youth Marijuana Use.

Marijuana is the most commonly used illicit substance by adolescents and young adults (Johnston, O’Malley, Bachman, Schulenberg, & Miech, 2014). Clear associations have been observed between marijuana use and negative health outcomes, mental health problems, cognitive problems, impaired academic performance, and psychosocial problems including difficult relationships with partners (Brook, Stimmel, Zhang, & Brook, 2008; Degenhardt et al., 2013; Hall, 2015; Homel, Thompson, & Leadbeater, 2014; Volkow, Baler, Compton, & Weiss, 2014). Effective prevention strategies require a fuller understanding of the risk and protective factors that influence marijuana use, especially in high-risk populations such as maltreated children.

The present study was guided by the ecological framework developed by Bronfenbrenner and colleagues (Bronfenbrenner & Morris, 1998) and examined by other researchers (Tudge, Mokrova, Hatfield & Karnik, 2009). The full ecological theory considers four salient interacting areas in influencing outcomes such as substance use. First, “process” refers to interactions that occur on a regular basis, such as parent-child relationships. Second, “person” includes different characteristics such as those related to emotional resources involving past experiences; having been abused or neglected is an example. Third, “context” refers to any environment - home, school or peer group. Peer use of substances is considered in the present study. Fourth, “time” includes “meso-time” when interactions occur consistently in the developing person’s environment. The adolescents’ relationships with their parents over time were probed in the present paper.

With regard to the ecology influencing adolescent development and behavior, many risk factors for marijuana use have been identified. First, several aspects of the family environment can be influential. Family interactional theory posits that the stronger the relationship between parent and child, the greater the influence and the lower the likelihood of deviant behavior (Norem-Hebeisen, Johnson, Anderson, & Johnson, 1984).While some research suggests that the absence of a mother or father increases the risk of marijuana use, other studies have shown no or less of a relationship after controlling for parenting characteristics (Barrett & Turner, 2006; Crawford & Novak, 2008). This suggests that the quality of the relationship between children and parents and parenting practices might have more impact than simply whether the parent lives in the home or not. Several parenting characteristics appear to increase the risk of marijuana use. In particular, low levels of monitoring of the adolescent’s whereabouts, activities and friends are associated with substance use (Cottrell et al., 2003; Wood, Read, Mitchell, & Brand, 2004). In addition, adolescents who report strong attachment to their parents are less likely to use marijuana (Kostelecky, 2005). Some studies have concluded that a close relationship, characterized by warmth, support and closeness can reduce the risk of marijuana use, but it is difficult to disentangle the inter-relationships between different aspects of parenting and the risk for substance use (Stice, Barrera, & Chassin, 1993). Ralston and colleagues observed that positive parent-child affective quality lowered the risk of substance use during the two years following high school (Ralston, Trudeau, & Spoth, 2012). In addition to adolescents’ current relationships with their parents, it is also possible that earlier relationships play a role and that many of the above characteristics influence later substance use.

Historically, research on family relationships has largely focused on the mother–child relationship, although research has generally suggested that fathers can influence children’s wellbeing (Goodvin, Meyer, Thompson, & Hayes, 2008). Nurturing fathers have been found to contribute to young children’s wellbeing and development (Black, Dubowitz, & Starr Jr, 1999). Positive relationships between adolescents and fathers have been found to decrease their use of substances and other risky behaviors, particularly among boys (Bronte-Tinkew, Moore, & Carrano, 2006). In another study, the relationship between peer substance use and adolescent marijuana use was attenuated by closeness to father, supporting the notion that authoritative parenting may counter peer pressure on adolescents to use substances (Dorius, Bahr, Hoffmann, & Harmon, 2004). Among African American families, father absence was found to increase use of substances, but only in boys (Mandara & Murray, 2006). In another study, fathers’ knowledge of adolescents’ activities was the sole protective factor associated with marijuana use, across genders (Farhat, Simons-Morton, & Luk, 2011). Father-youth connectedness has been generally associated with reductions in problem behavior (Fosco, Stormshak, Dishion, & Winter, 2012). These findings did not differ for boys and girls, or for families with resident or nonresident fathers. Alternatively, in one study, harsh parenting by fathers was the only family variable associated with marijuana-related problems youth experienced (Bares, Delva, Grogan-Kaylor, & Andrade, 2011). In sum, while fathers appear to play a role in adolescents’ substance use, few of these studies involved maltreated and high risk youth. Neglect is by far the most common form of child maltreatment (also referred to as ‘maltreatment’) and is usually attributed to omissions in care by mothers, the primary caregivers. We were interested to probe whether positive relationships between adolescents and their parents, particularly fathers, would attenuate the link between maltreatment and marijuana use.

The role of peers is clearly important in understanding adolescent substance use. Affiliation with peers engaged in alcohol and cigarette use has been strongly associated with adolescents’ use of these substances (Brechwald & Prinstein, 2011). There have also been a few such studies linking peer substance use to adolescent marijuana use (Andreas, Pape, & Bretteville-Jensen, 2016; Su & Supple, 2016; Tucker, De La Haye, Kennedy, Green, & Pollard, 2014).

Importantly, much of the research concerning parenting and peer factors with marijuana use utilized samples from either the general population, specific community-based samples (e.g., adolescents growing up in an urban or rural setting), or school-based samples. We are not aware of studies that focused on examining these associations among maltreated and high risk youth, where relationships may be quite different.

Child maltreatment, including abuse and neglect, is a major public health problem. In 2017, 7.5 million children were reported for abuse or neglect (U.S. Department of Health & Human Services, 2019). Exposure to maltreatment and other trauma in childhood and adolescence have been predictive of later substance use including marijuana (Dubowitz et al., 2016; Wright, Fagan, & Pinchevsky, 2013). The mechanisms linking such trauma to substance use are not well understood, but most likely involve a complex interplay between the characteristics of the trauma, individual coping and the presence of other trusted adults that might buffer the impact of stressful experiences. In this respect, the degree of caring and supportive involvement by parents might be very influential in attenuating the relationship between maltreatment and marijuana use. In addition, fathers are often not that involved in the lives of maltreated children (Dubowitz, Black, Kerr, Starr, & Harrington, 2000); their lack of involvement might exacerbate the risk of marijuana use via less parental monitoring.

The current study utilized data from a longitudinal, prospective study to understand associations between child maltreatment, the quality of the child-parent and adolescent-parent relationships, peer substance use, parental monitoring and adolescent marijuana use. The first aim was to evaluate the strength of the association between child maltreatment and marijuana use, controlling for sex, race/ethnicity, study site, peer substance use and parental substance use. We controlled for the two last variables as they may influence adolescent marijuana use, independent of child maltreatment. Second, we evaluated the effect of the overall relationships with mothers and fathers on marijuana use after adjusting for demographic variables and parental and peer substance use. The last two aims focused on understanding the distinct impacts of the quality of parent-child relationships (i.e., mother and father-child separately) early in childhood vs. later in adolescence on marijuana use. Finally, we tested the possible moderating impact of the quality of parent-child relationships on the link between child maltreatment and marijuana use.

METHODS

Sample

The sample was derived from the Longitudinal Studies of Child Abuse and Neglect (LONGSCAN), a consortium of 5 studies of antecedents and consequences of child maltreatment (Runyan et al., 1998). The subsamples were deliberately varied to represent those at risk for maltreatment, according to specified criteria such as having a mother at risk for Human Immuno Virus infection, or already identified as maltreated and involved with CPS, and one subsample had been placed in foster care. By age 12, 86% of the children had been reported to CPS. Eighty six percent involved neglect, 48% emotional abuse, 47% physical abuse, and 23% involved sexual abuse. At age 12 years, 2.5% reported having used marijuana. This steadily increased to 32.6% by age 18. Inclusion criteria for the analyses were not missing data on: 1) marijuana use during adolescence and 2) assessment of relationships with both parents. Of the original LONGSCAN sample of 1354 children, 702 (52%) met the inclusion criteria. Chi-square tests revealed no significant differences between the included and excluded participants in terms of sex, race/ethnicity and a CPS report.

Procedure

The consortium began the common study protocols when the children were 4 or 6 years old, and collected data from the primary caregivers and children/youth every 2 years between 4 and 18, mostly through face-to-face interviews and participants completing computerized self-administered protocols (Runyan et al., 1998). CPS data were obtained by each site from the local agency approximately every 2 years. All procedures were approved by local Institutional Review Boards.

Measures

Predictors

Child Maltreatment.

A participant was defined as Maltreated if there had been a CPS report between birth and 12 years, based on regular review of child welfare records. This variable was represented by the number of reports regardless of substantiation. Research has found few differences regarding recidivism for children with substantiated vs. unsubstantiated reports (Drake, Jonson-Reid, Way, & Chung, 2003).

The Child-Father Quality of Relationship (QoR).

(ages 6-8 years) was measured by a LONGSCAN four-item scale where primary caregivers, mostly mothers, rated father’s emotional and financial support and physical care. One example of a question is: “how much does he show that he cares about CHILD?” Each item was rated from 1 (none) to 4 (a lot); mean scores were computed for ages 6 and 8 years. The Child-Father QoR was based on the average of the two scores, or the single score if only one was available. Cronbach’s alpha was .77 and .76 for ages 6 and 8 years, respectively.

The Teen-Father QoR

(ages 12 – 18 years) was measured youths’ perceptions of the quality of relationship with his or her father or father-figure with a 9-item scale from the National Longitudinal Study on Adolescent Health completed by the youth at 12, 14, 16 and 18 years (Resnick et al, 1997). A sample question is: “Have you talked about your friends or things you were doing with friends with him?” Each item was rated on a 5-point scale; a mean was computed at each age; correlations across ages ranged from 0.3-0.5. The Teen-Father QoR score was the average of the means; if only one score was available, that score was used. Cronbach’s alphas were .83, .87, .89 and .89 for ages 12, 14, 16 and18, respectively. Validity of the measure was supported by moderate correlations with the youth report of relationship with mother.

The Child-Mother QoR

(Early, age 6 years) was measured by the child’s report at age 6 on the Mother Acceptance Subscale of the Pictorial Scale of Perceived Competence and Social Acceptance (Harter & Pike, 1984). Mean scores of the 6 items comprising the Subscale were used. Cronbach’s alpha was .85.

The Teen-Mother QoR

(Late, ages 12 – 18 years) consisted of the same 9-item scale used regarding fathers; the score was computed as the average of the means; if only one score was available, that score was used. Cronbach’s alphas for Teen-Mother QoR were .81, .84, .84 and .88 for ages 12, 14, 16 and18 years respectively.

Parental Monitoring.

Youths’ perception of caregivers’ knowledge of their activities, friends, whereabouts and money use was assessed by a 5-item scale adapted from the Patterson and Stouthamer-Loeber measure at ages 12, 14 and 16 years (Patterson & Stouthamer-Loeber, 1984). Response were: 0 (don’t know), 1 (know a little), and 2 (know a lot). The Parental Monitoring score was computed by averaging scores across the three interviews, and ranged from 0-2. For inclusion in the analyses, data from a minimum of two interviews were required.Cronbach alphas were .68 at age 12 and 14 years. Modest correlations −.11 to −.25 with child behaviors at age 12 supported validity.

Parental Substance Use.

The total number of legal and illegal substances, including alcohol, marijuana, cocain, hallucinogents, heroin, stimulants and tranquillizers, reported by the parent, usually mother, when the child was 8 years old was used as a measure of parental substance use.

Peer Marijuana Use.

At ages 12, 14 and 16 years, youth were asked “how many of your close friends smoke marijuana?” Response options were “none of my friends’ (0), ‘Some of my friends’ (1), and ‘Most of my friends’ (2). If at any interview they indicated ‘‘some’ or ‘most’, they were considered to be positive for Peer Use. If they did not indicate any peer use in at least two interviews, they were considered negative for Peer Use.

Outcome Variable

Adolescent Marijuana Use was assessed at 12, 14, 16 and 18 years. Participants had to have indicated marijuana use with standardized measures including the Diagnostic Interview Schedule for Children (DISC) at any of these interviews (Used Marijuana), or, to have indicated at 18 that they had never used marijuana (Never Used Marijuana). Youth who never indicated marijuana use and who did not complete the age 18 interview were considered as having missing data and were excluded. At age 12, participants completed the Adolescent Substance Involvement Measure, including “In the past year, did you ever use or try marijuana?” Those indicating a positive response were then asked about the frequency of use. At ages 14 and 18, respondents completed the Marijuana Use module of the DISC, indicating whether they had ever used marijuana. At age 16 and 18 years respondents completed the LONGSCAN developed Tobacco, Alcohol, and Drug Use measure regarding marijuana use in the past year.

Control Variables

Sex and race/ethnicity data were gathered from primary caregivers. A variable code for study site was constructed. We controlled for these variables due to their possible influence on marijuana use based on ours and other’s research. Site was included because sampling variation may too have been influential. For example, some sites were higher risk than others.

Data Analysis

The analyses examined the relationships among Maltreatment, Parent-Child/Adolescent Relationship Quality, Parental Monitoring, Parental Substance Use, Peer Marijuana Use and Adolescent Marijuana Use. Preliminary chi-square tests of independence were evaluated to assess the bivariate relationships among covariates and the outcome variable. Next, the bivariate relationships were probed between adolescent Marijuana Use and Child Maltreatment, Parental Monitoring, Parental Substance Use and Peer Marijuana. To evaluate Aims 1 and 2, logistic regression models were used to predict the primary outcome of adolescent Marijuana Use. For Aim 1, Child Maltreatment was the primary predictor of interest. The importance of this predictor variable, however, was evaluated while controlling for other substantive variables of interest (Parental Monitoring, Parental Substance Use, Peer Substance Use) and general covariates (site, race/ethnicity, sex). For Aim 2, the quality of the relationship (QoR) with father and mother during childhood and during adolescence, respectively, were the primary predictors of interest. These models, identical to Aim 1, evaluated the relationships between QoR and adolescent Marijuna Use in the context of the same substantive and general covariates. To determine if QoR moderated the Child Maltreatment-adolesent Marijuana Use relationship, each of the teen-parent QoR variables (both mothers and fathers) were tested in interaction terms added to the previous model. Simple slopes analyses were conducted as follow-up tests when statistically significant interactions were found (Cohen, Cohen, West, & Aiken, 2003). Separate regression models were tested for mothers and fathers. All analyses were conducted using SPSS version 24.

RESULTS

Table 1 shows that approximately half the sample had used marijuana at some point during adolescence, with substantial variation by study site, race/ethnicity, and maltreatment status. Black youth were less likely than all other race/ethnicity groups to have used marijuana. No sex differences in marijuana use were observed. Participants rated their relationships with their parents, both early and late, rather positively (i.e., means of approximately 4 on a 5-point scale). Similarly, they described considerable parental monitoring (mean = 1.5 on scale of 0-2 with 2 representing more monitoring).

Table 1.

Characteristics of the Sample and Adolescent Marijuana Use (N = 702)

	n	% Used Marijuana	pa
Site			.002
East	135 (19)	36
Midwest	104 (15)	39
South	127 (18)	50
Southwest	178 (25)	57
Northwest	158 (22)	51
Sex			.31
Male	339 (48)	50
Female	363 (52)	50
Race/ethnicity			.04
Black	384 (55)	43
White	183 (26)	51
Other (Native American, Asian/Pacific Islander, Mixed)	88 (12)	57
Hispanic	47 (7)	55

^aThe chi-squared probability of a 2 (Marijuana Use/No Use) by k −1 (Category) analysis.

Significant bivariate relationships were found between adolescent Marijuana Use and Peer Use (AOR = 5.10, p<.001; 95% CI 3.62-7.19), more Parental Substance Use (AoR = 1.16, p=.002; 95% CI: 1.06-1.27) and less Parental Monitoring (AoR = 0.32, p<.001; 95% CI: 0.19-0.53). Child Maltreatment was only marginally related to Marijuana Use during adolescence (AoR = 1.11, p=.054; 95% CI 0.99-1.24).

Table 2 presents the relationships predicting adolescent Marijuana Use with Child Maltreatment. As shown, Child Maltreatment was not significantly associated with Aaolescent Marijuana Use. However, greater Peer Use and Parental Substance Use were both significant predictors of more adolescent Marijuana Use by participants.

Table 2.

Regression Model Predicting Adolescent Marijuana Use, with Covariates.

	AOR	95% CI
Site
Midwest	1.39	0.62-3.09
South	1.35	0.64-2.83
Southwest	1.51	0.78-2.92
Northwest	1.38	0.69-2.76
East	Reference
Sex
Female	0.83	0.58-1.20
Male	Reference
Race/ethnicity
Black	0.82	0.51-1.30
Hispanic	1.34	0.61-2.95
Mixed/Other	1.15	0.64-2.04
White	Reference
CPS Report 0-12
Yes	0.74	0.26-2.10
No	Reference
Parental Monitoring	0.63	0.33-1.21
Peer Substance Use
Yes	4.76*	3.20-7.07
No	Reference
Parental Substance Use	1.16*	1.04-1.30

^*P < .05

Building on the previous model addressing Aim 1, the relationships with parents during childhood and adolescence were entered as predictors in logistic regressions models predicting adolescent Marijuana Use (see Tables 3 and 4 for models with Father’s and Mother’s QoR, respectively). The child-parent QoR variables were not associated with the adolescents’ Marijuana Use. In contrast, poorer relationships during adolescence with mothers, but not fathers, and greater Peer and Parental Use were associated with more Marijuana Use by youth participants.

Table 3.

Child Maltreatment and Child/Adolescent - Father Relationships Predicting Marijuana Use, with Covariates ^a

	AOR	95% CI
Child-Father Relationship	0.93	.0.67-1.28
Teen-Father Relationship	0.76	0.55-1.06
CPS (CM) Report 0-12
Yes	0.73	0.23-2.33
No	Reference
Parental Monitoring	0.60	0.28-1.27
Peer Substance Use
Yes	4.60*	2.98-7.11
No	Reference
Parental Substance Use	1.16*	1.02-1.32
Interaction Terms
CM X Child-Father Relationship	1.01	0.81-1.25
CM X Tenn-Father Relationship	0.09*	0.01-0.83

^aCovariates Sex, Race/Ethnicity and Site not shown; all NS.

^*P<.05

Table 4.

Child Maltreatment and Child/Adolescent - Mother Relationships Predicting Marijuana Use, with Covariates^a

	AOR	95% CI
Child-Mother Relationship	1.00	0.94-1.05
Teen-Mother Relationship	0.64*	0.42-0.97
CPS Reports 0-12
Yes	0.92	0.32-2.68
No	Reference
Parental Monitoring	0.76	0.37-1.58
Peer Substance Use
Yes	4.63*	3.05-7.02
No	Reference
Parental Substance Use	1.14*	1.01-1.28
Interaction Terms
CM X Child-Mother Relationship	0.97	0.93-1.01
CM X Teen-Mother Relationsip	0.97	0.75-1.25

^aCovariates Sex, Race/Ethnicity and Site not shown; all NS.

^*P<.05,

Finally, we tested the possible moderating impact of the quality of parent-child relationships on the link between Child Maltreatment and Marijuana Use via CM X parent QoR interactions. A statistically significant CM X Teen-Father’s QoR interaction was evident (AoR = 0.09, p=.033; 95% CI 0.01-0.83), indicating a protective effect of a positive relationship with father among teens who had been maltreated, but not for the earlier child-father relationship (see Table 3). Specifically, for those who had a CPS report from 0-12 years of age, a significant negative association was found between Marijuana Use and Teen-Father’s QoR (AoR = 0.70, p=.038; 95% CI 0.50-0.98); a better relationship with father was linked to less use of marijuana. This effect was not evident among those who had not been reported for maltreatment and also when replicating the analyses substituting QoR with mothers. All effects were powered at 80%.

DISCUSSION

Given the potentially negative outcomes of marijuana use in adolescence, it is important to understand factors that influence their use (Brook et al., 2008). As noted earlier, prior research suggests that parent-child relationships, parental monitoring, child maltreatment and exposure to other trauma and peer use might predict adolescent marijuana use (Brechwald & Prinstein, 2011; Cottrell et al., 2003; Hussey et al., 2005; Kostelecky, 2005; Ralston et al., 2012; Resnick et al., 1997). The current analyses extend these findings to a high risk sample of children who had been maltreated or were at high risk of maltreatment while probing separately and together the influence of relationships with mothers and fathers, both early in childhood and during adolescence. Child maltreatment generally occurs in the context of family dysfunction and troubled parent-child relationships. As stated earlier, clinicians have an interest in identifying and working with strengths; there is a need to better understand how fathers can play a potentially helpful role fathers in troubled families.

The prevalence of marijuana use among this sample was higher than a national estimates of 39.5 percent of youth (Center for Behavioral Health Statistics and Quality, 2016). The first noteworthy finding is that the quality of relationships during adolescence with mothers and particularly with fathers influenced Marijuana Use in these high risk youth. Other studies found similar associations in different kinds of samples of youth (Cottrell et al., 2003; Dishion & McMahon, 1998; Luk, Farhat, Iannotti, & Simons-Morton, 2010; Wood et al., 2004). Further, the finding that average or good relationships with just fathers lessened Marijuana Use in the maltreated subsample is important. This supports other research findings cited earlier (Bronte-Tinkew et al., 2006; Farhat et al., 2011). Also important, the youth described the quality of relationships with parents as quite good on average. This highlights that even youth who are “high risk” often have important resources that can be marshalled in addressing or buffering risks. Clinicians working with these families can help foster positive relationships between adolescents and fathers, and mothers, to reduce the likelihood of marijuana use by high risk youth.

The influence of Peer Marijuana Use on use by adolescents is consistent with prior research (Andreas et al., 2016; Su & Supple, 2016; Tucker et al., 2014). In addition to asking adolescents about their possible marijuana use, clinicians are encouraged to ask about use by their peers, and help them navigate possible pressure to use substances. This recommendation is consistent with physician guidelines for screening for alcohol use among youth with two questions, one about peer use and one about the patient’s own use (National Institute of Alcoholism and Alcohol Abuse, 2013).

The study also found that Maltreatment (0-12 years) was only marginally related to Marijuana Use, supporting findings in other studies (Dubowitz et al., 2016; Duncan et al., 2008; Wright et al., 2013). There are several possible explanations, including using marijuana as a maladaptive coping mechanism. Maltreatment is also an indicator of underlying family dysfunction that may contribute to substance use, and could influence peer selection that in turn increases risk for using substances. It is interesting that the link between maltreatment and Marijuana Use was not stronger, suggesting its modest role in the context of the multiple influences on adolescents’ risk taking behavior. More proximal variables such as Peer Substance Use appear to be more influential than earlier Maltreatment in influencing Marijuana Use. It should be noted, however, that most of the children had been reported to CPS and that the comparison of Maltreatment is with other high risk youth. Nevertheless, asking youth about possible maltreatment can help clinicians address their needs.

The study has several limitations. CPS reports are a crude proxy for possible maltreatment. Clearly, this problem is often not reported to CPS. The findings are only generalizable to similar high risk youth, many of whom had been maltreated. We were unable to disentangle the timing of the relationships with parents, Peer Use and Marijuana Use by the adolescents. Thus, the direction of associations is unclear. For example, youth using marijuana may have selected peers doing the same, or, they may have been encouraged to use the substance by their peers. It also needs to be noted that researchers have used varying measures of substance use and psychometric data on some measures was limited.

CONCLUSIONS

Despite increasing acceptance of marijuana use in the U.S. and prevalent use by adolescents, marijuana can impair health and development. The findings in this study suggest that clinicians be especially attentive to marijuana use among all youth, but especially those who have been maltreated, and inquire about peer use as a risk marker for further assessment and early intervention. The findings of the particularly positive role that fathers can play signals the need for clinicians to pay special attention to fostering such relationships. This may well lessen the likelihood of high risk adolescents using marijuana while benefiting them and their families in many ways. The study also suggests the need for more research focused on fathers, particularly the potentially constructive role they can play in troubled families. Fathers too are often an important influence on adolescent’s health and development.