Abstract
Objective:
While peer influence is a well-documented risk factor for adolescent substance use, it remains unclear whether peer or parental attitudes have greater impact, and if this relationship is moderated by having a confidant and the relationship between adolescents and their confidant.
Method:
Pooled (2015–2018) National Survey on Drug Use and Health (NSDUH) data on adolescents (12–17 years) were used. Perceived peer and parental disapproval of tobacco, alcohol and marijuana use were dichotomized. We assessed associations between disapproval and past-month tobacco (N=51352), alcohol (N=51407), and marijuana use (N=51355) using separate multivariable logistic regression models. We explored effect modification by presence of a confidant, parental vs. non-parental, and peer vs. non-peer confidant relationship.
Results:
Peer and parental disapproval, presence of any confidant, and identifying a parental confidant were consistently protective against substance use; identifying a peer confidant increased odds of use across substances. For marijuana use, peer disapproval (adjusted Odds Ratio [aOR]: 0.07, 95% Confidence Interval [CI]: 0.06, 0.08) was more protective than parental disapproval (aOR: 0.13, 95%CI: 0.12, 0.15). The joint presence of peer/parental disapproval and any confidant decreased the odds of substance use beyond the individual effects of peer/parental disapproval and having a confidant. However, having a peer confidant attenuated the protective association between peer/parental disapproval and tobacco, alcohol and marijuana use.
Conclusions:
Both peer and parental relationships are salient when considering the social context of adolescent substance use and should be considered when studying the effects of perceived disapproval.
1. INTRODUCTION
Peer substance use is a well-established risk factor for adolescent substance use, particularly in reference to alcohol,1–4 marijuana,4,5 and cigarette use.4 Through the lens of social cognitive theory, previous literature has posited that using substances is a learned social behavior.6,7 This theory explains that individuals engage in certain behaviors, such as substance use, in order to emulate the behaviors of their peers and others.8 Similarly, peer selection can contribute to substance use initiation, as adolescents seek out peers with similar attitudes towards substances.9 Previous research has identified trends wherein adolescents who use substances grow their social networks of peers who also use substances over time,10 lending support to the peer selection theory. As both peer influence and selection may be acting in tandem, it is evident that peers are influential within the landscape of substance use in adolescence.
As adolescents age, social networks begin to change and friend support can take precedence over familial support.11 This could result from shifting social contexts wherein adolescents, as a means of gaining autonomy and independence, begin spending a greater amount of time with their peers.12 This could work to create a context wherein adolescents and young adults seek greater support from peers and work on developing stronger friendships. However, previous work notes that the parent-child relationship remains salient throughout adolescence and into adulthood, though this is dependent on the nature of the parental relationship.13 While several studies have explored the effects of various parenting methods and strategies on adolescent substance use behavior,14,15 the influence of the parental and adolescent relationship as it relates to peer relationships in the context of substance use remains understudied.
When considering social networks of adolescents and their influence on substance use, the impact of social isolation should be considered. Social isolation, which can be described as an objective lack of social relationships,16 and social exclusion, which is related to an absence of peer-based friendships,17 can inhibit the formation of social networks and affect social support. Depending on the degree and type of isolation (i.e., isolation by choice or via social exclusion), an association has been identified among those with lower perceptions of social cohesion and increased use of cigarettes, alcohol and marijuana,17 highlighting the importance of peer relationships as they relate to substance use. It has also been suggested that a lack of social embeddedness, the number of connections in a social network,18 or having fewer friends who use substances is protective due to diminished exposure to peers using substances. Similar findings have demonstrated that social embeddedness results in either more or less susceptibility to using substances;19 the most isolated adolescents smoke more cigarettes, whereas the most socially embedded drink more alcohol.20 Among adolescents with restricted peer networks, such as those without social support, it remains unclear whether greater emphasis is placed on parental relationships as opposed to peer relationships.
This study aimed to assess the influence of adolescents’ perceptions of peer and parental attitudes regarding substance use and actual adolescent substance use, in order to determine the magnitude of influence of parents and peers. Further, the present study explores how adolescents’ sources of support may modify the influence of social and familial attitudes towards substance use. To achieve this, we analyze whether the effect of peer or parental attitudes towards substance use on adolescent substance use varied based on (1) having a source of social support and (2) the relationship between adolescents and their source of support. We hypothesize that: perceptions of parental and peer disapproval will be protective against adolescent substance use, with peer disapproval showing a more protective effect; and, that this association will be moderated by the relationship between the adolescent and their confidant.
2. METHODS
2.1. Participants and recruitment
This cross-sectional study includes pooled data from the 2015–2018 National Survey on Drug Use and Health (NSDUH), which is an annual survey of United States (U.S.) households sponsored by the Substance Abuse and Mental Health Services Administration21. As past-month substance use prevalence was <10% within this study sample of adolescents, data were pooled in order to achieve a more stable estimate over the course of four years. NSDUH employs a multistage area probability sample design to generate U.S. nationally representative prevalence estimates, and includes non-institutionalized individuals aged 12 and above. Households were randomly selected, and a maximum of two individuals, though potentially none, were asked to complete the in-person interview21. Each respondent who completes the survey is given 30 USD in cash. The NSDUH has been registered and approved by the review board at RTI International, a non-profit organization through which the survey is administered. The yearly weighted interview response rates for the NSDUH surveys, from year 2015 to 2018, ranged from 66.56% to 69.66%21–24. Though the NSDUH collects information on both adolescents and adults, this analysis is restricted to adolescents aged 12–17 as prior work within NSDUH operationalizes adolescence as between the ages of 12–17, and those aged 18–25 as young adults.25–27 IRB approval was not required, as data are publicly available.
2.2. Measure of perceived parental and peer attitudes
Main exposures in this study were divided into two groups: perceived parental attitudes and perceived peer attitudes. Survey questions captured attitudes towards various substances, including: “How do you think your parents would feel about you smoking one or more packs of cigarettes per day?” (neither approve nor disapprove, somewhat disapprove, strongly disapprove); “How do you think your parents would feel about you having one or two drinks of an alcoholic beverage nearly every day?”; and “How do you think your parents would feel about you using marijuana or hashish once a month or more?”. Similar questions were posed to respondents regarding peer behaviors (“How do you think your close friends would feel about you smoking one or more packs of cigarettes a day?”). The parental disapproval variable (yes vs. no) and the peer disapproval variable (yes vs. no) were dichotomized from these survey questions as per previous research,28 where any degree of perceived disapproval (somewhat disapprove and strongly disapprove) was classified as “yes” and responding “neither approve nor disapprove” was categorized as “no”. Missing responses (N=1584, 2.9%) were excluded from analysis.
2.3. Substance use outcomes
This analysis focused on dichotomous (yes vs. no) measures of past-month substance use, including tobacco, alcohol, and marijuana use. This analysis does not assess whether adolescents used electronic cigarettes as this information was not available in the NSDUH questionnaire from 2015–2018. Past-month substance use measures were employed in lieu of past-year measures to assess responses as close in time as possible to when adolescents reported perceived disapproval and sources of support. Polysubstance use was not considered in this analysis. Imputed measures for substance use, available directly from the NSDUH dataset, were used to address missingness.
2.4. Social support
Social support was measured by the presence of a confidant.29 Three dichotomous variables were constructed to assess whether adolescents identified peers and/or parental figures as their confidants. In response to “If you wanted to talk to someone about a serious problem, which of the following people would you turn to?”, participants could select nobody, parent or guardian, boyfriend or girlfriend, some other adult, or some other person. Adolescents who reported any person were classified as having a confidant. Those who identified a parent, guardian or other adult as a trusted source of support were coded as having a parental-confidant relationship and those who indicated that they would turn to a boyfriend, girlfriend, or some other person were coded as having a peer-confidant relationship. Respondents could identify multiple sources of support; parental- and peer-confidant relationships are not mutually exclusive. Missing responses were excluded from analysis (N=4145, 7.6%). Social support has been operationalized similarly in previous studies.29,30
2.5. Confounder measures
Potential confounders were identified a priori as causes of both perceived peer and parental attitudes and substance use outcomes. Confounders included: age (12–14 vs. 15–17), sex (male vs. female) and past-year major depressive episode (MDE) (yes vs. no). MDE was operationalized as meeting five out of nine criteria for major depression.21 Demographic measures were included as older age31,32 and gender32 can impact substance use. Additionally, social network composition may change as adolescents age and develop;12 composition of networks by gender can impact substance use,33 which could in turn influence perceptions of peer and parental disapproval. Further, this analysis adjusts for MDE given the association between depression, substance use34 and withdrawal from social networks.35 We did not adjust for income or race and ethnicity, as we did not hypothesize either of these factors to influence adolescent perceptions of peer or parental disapproval. As the analytic sample includes respondents pooled from multiple surveys (2015–2018), we adjusted for survey year. Imputed variables, obtained directly from the NSDUH dataset, were used for age and sex; missing responses (N=1581, 2.9%) were removed from the depression variable and thus excluded from the analysis. All a priori hypothesized confounders were included in the final models.
2.6. Statistical analysis
Multivariable weighted regression models were built by analyzing the impact of peer or parental attitudes of a substance (e.g. perceptions regarding cigarette smoking) and adolescent past-month use of the concordant substance (e.g. past-month tobacco use) (Supplemental Table 1). Annual survey weights provided by NSDUH were divided by four such that results are interpreted as average, nationally representative results from years 2015 through 2018. It should be noted that perceived parental and perceived peer attitudes were not included in the same model as both were considered main exposures, and multiple exposures cannot be interpreted simultaneously.36 Overall, the proportion of data excluded due to missingness across models considering the tobacco (6.4%), alcohol (6.3%) and marijuana (6.4%) outcomes were similar; a greater proportion of missing data was excluded when conducting moderation analyses due to missingness in the confidant variables across the tobacco (10.9%), alcohol (10.8%) and marijuana (10.8%) outcomes.
Effect modification analyses were carried out to assess whether the presence of a confidant and the nature of an adolescent’s relationship with their confidant moderates the impact of peer or parental disapproval on substance use (Figure 1). When assessing the moderating effect of the relationship between adolescents and their confidant, participants who were previously coded as having no confidant were dropped from analysis. Models were run separately using cross-product terms to assess the statistical significance of multiplicative interaction and with indicator variables for calculation of odds ratios for specific combinations of peer/parental disapproval and confidant measures. All analyses were conducted in SAS version 9.4, taking into account NSDUH’s complex survey design.21
3. RESULTS
3.1. Descriptive Statistics
The sample included a total of 54866 adolescent respondents. Overall, the majority of the sample was between the ages of 15–17 (N=28046, 51.5%), male (N=28032, 51%), non-Hispanic white (N=29109, 52.9%) with household income more than two times the poverty threshold (N=30572, 56.1%), as presented in Table 1. The majority of adolescents identified a confidant (N=50721, 92.4%); 44.2% (N=23910) identified a parental source, 14.7% (N=8059) identified a peer source of support, and 33.5% (N=18752) identified both a parental and peer source of support. As parental and peer confidants were not mutually exclusive, they are operationalized and presented as two distinct variables in Table 1.
Table 1.
Sample statistics of respondents (N=54866) from the NSDUH survey (2015–2018), with weighted percentages using survey weights.
| Variable | Overall (N=54866, %) |
Tobacco Usea
(N=2109, 3.8%) |
Alcohol Usea (N=5334, 9.7%) |
Marijuana Usea (N=3924, 7.2%) |
|---|---|---|---|---|
| Perceived parental disapproval | ||||
| No | 15.5%b | 7.8%c | 7.5%d | |
| Yes | 81.8%b | 91.1%c | 90.7%d | |
| Missing | 2.8%b | 1.1%c | 1.8%d | |
| Perceived peer disapproval | ||||
| No | 35.5%b | 28.3%c | 20.2%d | |
| Yes | 61.5%b | 70.2%c | 77.9%d | |
| Missing | 3.0%b | 1.4%c | 1.9%d | |
| Age | ||||
| 12–14 | 48.5% | 12.9% | 13.6% | 12.2% |
| 15–17 | 51.5% | 87.1% | 86.4% | 87.8% |
| Sex | ||||
| Male | 51.0% | 55.4% | 48.1% | 52.6% |
| Female | 49.1% | 44.6% | 51.9% | 47.4% |
| Race/ethnicity | ||||
| Non-Hispanic white | 52.9% | 71.0% | 62.8% | 56.2% |
| Non-Hispanic Black | 13.7% | 6.5% | 8.8% | 13.3% |
| Non-Hispanic othere | 9.7% | 8.2% | 7.3% | 8.5% |
| Hispanic | 23.7% | 14.2% | 21.1% | 22.0% |
| Poverty level | ||||
| Living in poverty | 22.0% | 23.9% | 15.8% | 21.0% |
| Income ≤2x poverty threshold | 22.0% | 26.3% | 21.3% | 23.1% |
| Income >2x poverty threshold | 56.1% | 50.0% | 62.8% | 56.0% |
| Past year MDE | ||||
| No | 86.7% | 72.3% | 76.8% | 74.3% |
| Yes | 13.3% | 27.7% | 23.2% | 25.7% |
| Missing | 2.9% | 4.4% | 2.4% | 2.9% |
| Presence of a confidant | ||||
| No | 4.8% | 11.1% | 6.9% | 9.1% |
| Yes | 92.4% | 84.2% | 90.1% | 86.8% |
| Missing | 2.7% | 4.8% | 3.0% | 4.1% |
| Parental confidant | ||||
| Nof | 48.2% | 59.8% | 62.9% | 61.2% |
| Yesg | 44.2% | 24.4% | 27.2% | 25.6% |
| Missing/No oneh | 7.6% | 15.8% | 9.9% | 13.2% |
| Peer confidant | ||||
| Nog | 77.7% | 55.8% | 64.6% | 59.0% |
| Yesf | 14.7% | 28.4% | 25.5% | 27.8% |
| Missing/No oneh | 7.6% | 15.8% | 9.9% | 13.2% |
| Year | ||||
| 2015 | 25.0% | 32.2% | 25.9% | 26.6% |
| 2016 | 25.0% | 25.2% | 23.8% | 23.8% |
| 2017 | 25.0% | 23.2% | 26.3% | 24.7% |
| 2018 | 25.0% | 19.5% | 24.0% | 24.8% |
MDE: Major depressive episode.
Refers to past month use.
Refers to either perceived parental or perceived peer disapproval towards smoking one or more packs of cigarettes daily.
Refers to either perceived parental or perceived peer disapproval towards drinking daily.
Refers to either perceived parental or perceived peer disapproval towards using marijuana.
Includes people who are Native American, Alaskan Native, Native Hawaiian, other Pacific Islander, Asian, and more than one race
Includes adolescents who reported a boyfriend/girlfriend or other person as their confidant.
Includes adolescents who reported a parent, guardian or other adult as their confidant.
Includes missing responses and adolescents who responded not having a confidant.
3.2. Multivariable Models
Results from the multivariable logistic regression models are presented below, categorized by the substance use-related outcome of interest for each model. Estimates for main effects, and main effects with the addition of effect modification, are presented separately when the interaction term was determined to be significant. Complete tables with non-significant interaction terms and comparison groups are available in the supplemental files (Supplemental Tables 2, 3, 4).
3.2.1. Past-Month Tobacco Use
A total of 3.8% (N=2109) of adolescents used tobacco in the past month (Table 1). The impact of perceived parental and peer disapproval was similar. Adolescents who perceived parental disapproval (versus those who did not perceive parental disapproval) had an 83% decrease in the odds of using tobacco (adjusted Odds Ratio [aOR]: 0.17; 95% Confidence Interval [CI]: 0.14, 0.21) (Table 2). Those who perceived peer disapproval had an 84% decrease in the odds of using tobacco (aOR: 0.16; 95%CI: 0.14, 0.19). Reporting any confidant resulted in a 56% decrease in the odds of using tobacco (aOR: 0.44; 95%CI: 0.36, 0.54); identifying a parental confidant decreased odds of using by 39% (aOR: 0.61; 95%CI: 0.53, 0.71) and identifying a peer confidant increased odds of using by 117% (aOR: 2.17; 95%CI: 1.85, 2.54).
Table 2.
Multivariable logistic regression models quantifying the association between parental and peer perceptions regarding smoking and actual adolescent tobacco use with effect modification by presence of a confidant and relationship of adolescent with confidant.
| Exposure | Tobacco use (Yes vs. no) | ||
|---|---|---|---|
| N | aOR | 95% CI | |
| Main effects | |||
| Parental disapproval towards smoking one or more packs/day | |||
| Yes | 51352 | 0.17*** | (0.14, 0.20) |
| Peer disapproval towards smoking one or more packs/day | |||
| Yes | 51352 | 0.16*** | (0.14, 0.19) |
| Presence of a confidanta | 51352 | 0.44*** | (0.36, 0.54) |
| Parental confidantb | 48909 | 0.61*** | (0.53, 0.71) |
| Peer confidantc | 48909 | 2.17*** | (1.85, 2.54) |
| Interaction terms | |||
| Parental disapproval towards smoking one or more packs/day × Presence of a confidant* | 51352 | ||
| Parental disapproval, no confidant | 0.28*** | (0.18, 0.44) | |
| No parental disapproval, any confidant | 0.84 | (0.55, 1.30) | |
| Parental disapproval, any confidant | 0.14*** | (0.10, 0.21) | |
| Parental disapproval towards smoking one or more packs/day × Peer confidant* | 48909 | ||
| Parental disapproval, non-peer confidant | 0.15*** | (0.12, 0.19) | |
| No parental disapproval, peer confidant | 1.31 | (0.86, 1.98) | |
| Parental disapproval, peer confidant | 0.32*** | (0.25, 0.43) | |
| Peer disapproval towards smoking one or more packs/day × Presence of a confidant*** | 51352 | ||
| Peer disapproval, no confidant | 0.31*** | (0.22, 0.43) | |
| No peer disapproval, any confidant | 0.90 | (0.67, 1.21) | |
| Peer disapproval, any confidant | 0.14*** | (0.11, 0.19) | |
p-value < 0.05,
p-value < 0.01,
p-value < 0.0001
aOR: adjusted odds ratio; CI: confidence interval
Comparison group: no confidant
Comparison group: non-parental confidant
Comparison group: non-peer confidant
Note: All models adjusted for age, gender, past-year major depressive episode, and survey year. Sample sizes for models investigating effect moderation by relationship with confidant vary due to exclusion of adolescents reporting no confidant. Cross-product terms noted where significant; non-significant cross-product terms excluded.
The relationships between both parental and peer disapproval and tobacco use were modified by the presence of a confidant. Those who perceived parental disapproval and reported no confidant had a 72% decrease in the odds of using tobacco (aOR: 0.28; 95% CI: 0.18, 0.44); those who perceived parental disapproval and identified a confidant had an 86% decrease in the odds of using tobacco (aOR: 0.14; 95% CI: 0.10, 0.21). Among those reporting peer disapproval, adolescents with no confidant had a 69% decrease in the odds of using tobacco (aOR: 0.31, 95% CI: 0.22, 0.43) and those with a confidant had an 86% decrease in the odds of use (aOR: 0.14, 95% CI: 0.11, 0.19). Among adolescents reporting parental disapproval, having a non-peer confidant was more protective against past-month tobacco use than having a peer confidant (Table 2).
3.2.2. Past-Month Alcohol Use
Alcohol was the most frequently used substance among adolescents (9.7%, N=5334, Table 1). Similar to tobacco use, the impacts of parental and peer disapproval were comparable. Adolescents who perceived parental disapproval had a 66% decrease in the odds of using alcohol (aOR: 0.34; 95% CI: 0.29, 0.40) and those who perceived peer disapproval had a 69% decrease in the odds of using alcohol (aOR: 0.31; 95% CI: 0.28, 0.35) (Table 3).
Table 3.
Multivariable logistic regression models quantifying the association between parental and peer perceptions regarding drinking and actual adolescent alcohol use with effect modification by presence of a confidant and relationship of adolescent with confidant.
| Exposure | Alcohol use (Yes vs. no) | ||
|---|---|---|---|
| N | aOR | 95% CI | |
| Main effects | |||
| Parental disapproval towards drinking daily | |||
| Yes | 51407 | 0.34*** | (0.29, 0.40) |
| Peer disapproval towards drinking daily | |||
| Yes | 51407 | 0.31*** | (0.28, 0.35) |
| Presence of a confidanta | 51407 | 0.73*** | (0.60, 0.88) |
| Parental confidantb | 48960 | 0.62*** | (0.56, 0.68) |
| Peer confidantc | 48960 | 1.83*** | (1.65, 2.03) |
| Interaction terms | |||
| Parental disapproval towards drinking daily × Presence of a confidant* | 51407 | ||
| Parental disapproval, no confidant | 0.55*** | (0.37, 0.83) | |
| No parental disapproval, any confidant | 1.17 | (0.78, 1.75) | |
| Parental disapproval, any confidant | 0.42*** | (0.28, 0.62) | |
| Parental disapproval towards drinking daily × Peer confidant* | 48960 | ||
| Parental disapproval, non-peer confidant | 0.33*** | (0.27, 0.40) | |
| No parental disapproval, peer confidant | 1.18 | (0.82, 1.71) | |
| Parental disapproval, peer confidant | 0.61*** | (0.48, 0.77) | |
| Peer disapproval towards drinking daily × Presence of a confidant*** | 51407 | ||
| Peer disapproval, no confidant | 0.43*** | (0.32, 0.58) | |
| No peer disapproval, any confidant | 1.06 | (0.79, 1.41) | |
| Peer disapproval, any confidant | 0.34*** | (0.25, 0.45) | |
| Peer disapproval towards drinking daily × Parental confidant*** | 48960 | ||
| Peer disapproval, non-parental confidant | 0.36*** | (0.31, 0.41) | |
| No peer disapproval, parental confidant | 0.80* | (0.67, 0.96) | |
| Peer disapproval, parental confidant | 0.21*** | (0.18, 0.24) | |
p-value < 0.05,
p-value < 0.01,
p-value < 0.0001
aOR: adjusted odds ratio; CI: confidence interval; NS: not significant.
Comparison group: no confidant
Comparison group: non-parental confidant
Comparison group: non-peer confidant
Note: All models adjusted for age, gender, past-year major depressive episode, and survey year. Sample sizes for models investigating effect moderation by relationship with confidant vary due to exclusion of adolescents reporting no confidant. Cross-product terms noted where significant; non-significant cross-product terms excluded.
Parental and peer disapproval regarding alcohol were both modified by the presence of a confidant. Adolescents reporting parental disapproval and no confidant had a 45% decrease in odds of using alcohol (aOR: 0.55, 95% CI: 0.37, 0.83), whereas those reporting parental disapproval and any confidant had a 58% decrease in the odds of using alcohol (aOR: 0.42, 95% CI: 0.28, 0.62). Perceptions of peer and parental disapproval of alcohol use were additionally modified by the nature of the confidant relationship. Among adolescents reporting parental disapproval, noting a parental confidant was more protective than reporting a peer confidant; similarly, peer disapproval and a parental confidant was more protective than peer disapproval and a non-parental confidant (Table 3).
3.2.3. Past-Month Marijuana Use
Marijuana was used in the past month among 7.2% (N=3924) of respondents (Table 1). Adolescents who perceived parental disapproval had an 87% decrease in the odds of using marijuana (aOR: 0.13; 95% CI: 0.12, 0.15) (Table 4); those who perceived peer disapproval had a 93% decrease in the odds of using marijuana (aOR: 0.07; 95% CI: 0.06, 0.08).
Table 4.
Multivariable logistic regression models quantifying the association between parental and peer perceptions regarding marijuana use and actual adolescent marijuana use with effect modification by presence of a confidant and relationship of adolescent with confidant.
| Exposure | Marijuana use (Yes vs. no) | ||
|---|---|---|---|
| N | aOR | 95% CI | |
| Main effects | |||
| Parental disapproval towards monthly marijuana use | |||
| Yes | 51355 | 0.13*** | (0.12, 0.15) |
| Peer disapproval towards monthly marijuana use | |||
| Yes | 51355 | 0.07*** | (0.06, 0.08) |
| Presence of a confidanta | 51355 | 0.54*** | (0.46, 0.64) |
| Parental confidantb | 48914 | 0.62*** | (0.55, 0.68) |
| Peer confidantc | 48914 | 2.13*** | (1.93, 2.34) |
| Interaction terms | |||
| Parental disapproval towards monthly marijuana use × Presence of a confidant*** | 51355 | ||
| Parental disapproval, no confidant | 0.27*** | (0.18, 0.40) | |
| No parental disapproval, any confidant | 1.23 | (0.88, 1.72) | |
| Parental disapproval, any confidant | 0.16*** | (0.11, 0.22) | |
| Parental disapproval towards monthly marijuana use × Peer confidant* | 48914 | ||
| Parental disapproval, non-peer confidant | 0.12*** | (0.10, 0.14) | |
| No parental disapproval, peer confidant | 1.45** | (1.10, 1.90) | |
| Parental disapproval, peer confidant | 0.25*** | (0.21, 0.31) | |
| Peer disapproval towards monthly marijuana use × Presence of a confidant*** | 51355 | ||
| Peer disapproval, no confidant | 0.19*** | (0.13, 0.28) | |
| No peer disapproval, any confidant | 0.94 | (0.76, 1.15) | |
| Peer disapproval, any confidant | 0.06*** | (0.05, 0.08) | |
| Peer disapproval towards monthly marijuana use × Parental confidant* | 48914 | ||
| Peer disapproval, non-parental confidant | 0.08*** | (0.07, 0.09) | |
| No peer disapproval, parental confidant | 0.83** | (0.73, 0.95) | |
| Peer disapproval, parental confidant | 0.05*** | (0.04, 0.06) | |
| Peer disapproval towards monthly marijuana use × Peer confidant*** | 48914 | ||
| Peer disapproval, non-peer confidant | 0.06*** | (0.05, 0.07) | |
| No peer disapproval, peer confidant | 1.39*** | (1.22, 1.57) | |
| Peer disapproval, peer confidant | 0.16*** | (0.14, 0.19) | |
p-value < 0.05,
p-value < 0.01,
p-value < 0.0001
aOR: adjusted odds ratio; CI: confidence interval; NS: not significant.
Comparison group: no confidant
Comparison group: non-parental confidant
Comparison group: non-peer confidant
Note: All models adjusted for age, gender, past-year major depressive episode, and survey year. Sample sizes for models investigating effect moderation by relationship with confidant vary due to exclusion of adolescents reporting no confidant. Cross-product terms noted where significant; non-significant findings included for completion.
Parental and peer disapproval were modified by the presence of a confidant. Adolescents reporting parental disapproval and any confidant had an 84% decrease in the odds of using marijuana (aOR: 0.16; 95% CI: 0.11, 0.22), whereas those describing parental disapproval and no confidant reported a 73% decrease in the odds of using (aOR: 0.27; 95% CI: 0.18, 0.40). Among those perceiving peer disapproval, adolescents with no confidant had an 81% decrease in the odds of using (aOR: 0.19; 95% CI: 0.13, 0.28); those perceiving peer disapproval who identified a confidant had a 94% decrease in the odds of using (aOR: 0.06; 95% CI: 0.05, 0.08). Peer confidant relationships moderated parental disapproval and marijuana use, wherein identifying no parental disapproval and a peer confidant increased the odds of using by 45% (aOR: 1.45, 95%CI: 1.10, 1.90). Peer confidant relationships also moderated peer disapproval and marijuana use; those reporting no peer disapproval and a peer confidant had 39% increased odds of using (aOR: 1.39, 95%CI: 1.22, 1.57) (Table 4).
4. DISCUSSION
This study explored the impact of peer and parental attitudes on adolescent substance use. We further investigated the effect of having a confidant on the association between perceived parental and peer attitudes and adolescent substance use. We found that both parental and peer disapproval were protective against substance use, across substances. Effect modification by presence of a confidant was observed for associations between peer and parental disapproval and tobacco use, alcohol use and marijuana use. The association between parental disapproval and substance use was moderated by a peer confidant, which was less protective in comparison to a parental confidant, for tobacco, alcohol, and marijuana use. The association between peer disapproval and substance use was moderated by peer confidant relationships in reference, wherein no peer disapproval and a peer confidant increased odds of marijuana use.
Our results highlight several important findings. Overall, perceived disapproval was more protective against substance use outcomes than the presence of a confidant or confidant relationships, with peer confidant relationships increasing odds of substance use. This was consistent across all substances, suggesting that disapproval is more salient than the presence of a confidant alone. Identifying a parental confidant was protective, whereas a peer confidant increased odds of substance use, highlighting the importance of having a parental confidant. Additionally, the degree to which peer and parental attitudes influence substance use can potentially be dependent on the substance in question. While perceived peer and parental disapproval were consistently protective against all substances included in these analyses, the protective effect of perceived peer disapproval was greater than that of parental disapproval only in relation to marijuana use. This suggests that peer attitudes towards marijuana are more salient than parental attitudes, which could potentially result from differences in risk perception.37 Marijuana is classified as a controlled substance at the federal level, but legalized in the form of medical use in 36 states and the District of Columbia (DC) and in the form of recreational use for adults 21 years of age and older in 15 states and Washington, DC.38 It is possible that in cases where the legality of substance use for adults is difficult to interpret, adolescents might turn to their peers instead of parents for cues as to whether a substance is appropriate to use.
The saliency of peer attitudes in relation to tobacco, alcohol, and marijuana use could reflect substance accessibility. Peer attitudes can reflect the substance use context that an individual is immersed in, such as using substances for socializing39 and partying,40,41 the accessible supply through social distribution,42 and friendship networks and selection.43–46 Perceiving no disapproval from peers suggests potentially heightened availability and usage among peer groups, assuming that peers engaging in either tobacco, alcohol or marijuana use are surrounded by individuals who similarly sanction use, as is consistent with the peer selection theory.9 This is evidenced through our findings demonstrating that no peer disapproval and the presence of a peer confidant increased the odds of marijuana use, while the presence of peer disapproval decreased odds of marijuana use. While the interaction between peer disapproval and identifying a peer confidant was statistically significant only when considering marijuana use, it should be noted that identifying a peer confidant increased odds of use across substances. This further emphasizes the importance of peers in adolescent social networks and their role in different substance use patterns.
Our findings further emphasize the intricacies of social networks among adolescents; the presence of a confidant moderates the relationship between both parental and peer perceptions and adolescent substance use in relation to tobacco, alcohol and marijuana use. The combined effect of identifying any confidant in conjunction with perceived disapproval consistently exerted a greater protective effect against substance use than the combined effect of no confidant and perceived disapproval, which highlights the importance of social connections during adolescence47. The relationship between adolescents and confidants was also influential; identifying a parental confidant was consistently more protective than identifying a non-parental confidant. In regard to marijuana use, identifying no disapproval in conjunction with a peer confidant increased the odds of substance use. This supports theories in relation to access to substances through social distribution networks39–41,43. These findings highlight the importance of multi-dimensional support networks and the concept that adolescent social networks impact the health and behaviors of adolescents.
Limitations are noted. As this is a cross-sectional analysis, we are unable to establish temporality; reverse causation, wherein substance use influences perceived disapproval or the presence and type of confidant, may also explain our results. As this analysis cannot disentangle perceived disapproval from actual disapproval, it is possible that misperceptions (either over- or underestimation) of disapproval are captured.48,49 However, perceived disapproval and actual disapproval are distinct constructs, and this study aimed to explore the impact of perceived disapproval rather than actual disapproval. Measurement error in regard to adolescents reporting substance use is also a concern. Were adolescents to misreport substance use, as has been previously suggested50, the effect estimates observed could be biased. NSDUH attempts to address underreporting through the use of Audio Computer Assisted Self Interviewing (ACASI).51 Our reported prevalence of smoking is likely underestimated due to the inability to assess e-cigarette use among adolescents. Nevertheless, NSDUH perception questions refer to smoking one pack or more of cigarettes, which aligns more closely with combustible cigarette smoking. Further, while this may have resulted in undercounting adolescents who smoke given that electronic cigarette use has increased in recent years, prior work suggests that smoking has decreased more rapidly.52 Additionally, selection bias as a result of excluding adolescents with missing values could introduce bias into the effect estimates. Unmeasured confounders could additionally introduce bias, though considering the size of the estimates obtained, this is unlikely to fully explain observed associations. Further, this analysis did not explore reasoning for lack of support (i.e. via social exclusion or social withdrawal). Lastly, the measures of disapproval for the substances considered assessed differing levels of frequency and intensity (such as smoking one or more packs of cigarettes per day and monthly marijuana use); therefore, we may not be able to compare the relative importance of disapproval and having a confidant across substances.
Notwithstanding these limitations, our study has several strengths, including the large nationally representative sample and the distinction of variables regarding adolescent and confidant relationships. Additionally, data collection procedures have been standardized within NSDUH21, providing reliability when pooling sample years.
5. CONCLUSIONS
Our findings highlight the intricacies of peer and parental relationships, and the complex interactions between social and familial networks and adolescent substance use. We found that the impact of peer and parental disapproval was consistently protective, with effect moderation observed by the presence of a confidant and the relationship between adolescents and their confidant. When considering policy geared towards preventing and mitigating adolescent substance use, both peer and parental relationships should be considered as influential. The impact of social support and confidant relationships should not be underestimated as a contextual factor that influences adolescents’ substance use behaviors. In light of rapidly changing policies regarding substance use, particularly marijuana, further investigations might consider how the influence of peer and parental attitudes, moderated by source of support, on substance use differ by state to investigate how state laws may be affecting adolescent substance use and the role of peer and parental attitudes in adolescent substance use decisions. Future research should additionally explore the interaction between either age or gender, perceived disapproval, and confidant relationships on adolescent substance use to guide intervention programs.
Supplementary Material
Acknowledgements:
This work would not be possible without all NSDUH participants, to whom we are grateful for providing their time and effort
Role of Funding Source:
This work was supported by National Institutes of Health, National Institute on Drug Abuse grant R01DA037866 (PI: Martins) and T32DA031099 (PI: Hasin).
Footnotes
Disclosure Statement: None of the authors report a conflict of interest.
REFERENCES
- 1.Leung RK, Toumbourou JW, Hemphill SA. The effect of peer influence and selection processes on adolescent alcohol use: A systematic review of longitudinal studies. Health Psychol Rev. 2014;8(4):426–457. [DOI] [PubMed] [Google Scholar]
- 2.D’Amico EJ, McCarthy DM. Escalation and initiation of younger adolescents’ substance use: The impact of perceived peer use. J Adolesc Health. 2006;39(4):481–487. [DOI] [PubMed] [Google Scholar]
- 3.Mundt MP. The impact of peer social networks on adolescent alcohol use initiation. Acad Pediatr. 2011;11(5):414–421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Branstetter SA, Low S, Furman W. The influence of parents and friends on adolescent substance use: A multidimensional approach. J Subst Use. 2011;16(2):150–160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Beardslee J, Datta S, Byrd A, et al. An examination of parental and peer influence on substance use and criminal offendingduring the transition from adolescence to adulthood. Crim Justice Behav. 2018;45(6):783–798. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bandura A. A sociocognitive analysis of substance abuse: An agentic perspective. Psychol Sci. 1999;10(3):214–217. [Google Scholar]
- 7.Bandura A. Self-efficacy: Toward a unifying theory of behavioral change. Psychol Rev. 1977;84(2):191–215. [DOI] [PubMed] [Google Scholar]
- 8.Giovazolias T, Themeli O. Social learning conceptualization for substance abuse: Implications for therapeutic interventions. The European Journal of Counselling Psychology. 2014;3(1):69–88. [Google Scholar]
- 9.Becker S, Marceau K, Hernandez L, Spirito A. Is it Selection or Socialization? Disentangling Peer Influences on Heavy Drinking and Marijuana Use Among Adolescents Whose Parents Received Brief Interventions. Subst Abus. 2019;13:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Simons-Morton B. Social influences on adolescent substance use. Am J Health Behav. 2007;31(6):672–684. [DOI] [PubMed] [Google Scholar]
- 11.Cheng S-T, Chan ACM. The multidimensional scale of perceived social support: dimensionality and age and gender differences in adolescents. Pers Individ Dif. 2004;37(7):1359–1369. [Google Scholar]
- 12.Giordano PC. Relationships in Adolescence. Annu Rev Sociol. 2003;29(1):257–281. [Google Scholar]
- 13.Chen E, Brody GH, Miller GE. Childhood close family relationships and health. Am Psychol. 2017;72(6):555–566. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Shakya HB, Christakis NA, Fowler JH. Parental influence on substance use in adolescent social networks. Arch Pediatr Adolesc Med. 2012;166(12):1132–1139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Cleverley K, Grenville M, Henderson J. Youths perceived parental influence on substance use changes and motivation to seek treatment. J Behav Health Serv Res. 2018;45(4):640–650. [DOI] [PubMed] [Google Scholar]
- 16.Umberson D, Montez JK. Social relationships and health: a flashpoint for health policy. J Health Soc Behav. 2010;51 Suppl(Suppl):S54–S66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Copeland M, Fisher JC, Moody J, Feinberg ME. Different kinds of lonely: Dimensions of isolation and substance use in adolescence. J Youth Adolesc. 2018;47(8):1755–1770. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kramer RA, Vaquera E. Who is really doing it? Peer embeddedness and substance use during adolescence. Sociol Perspect. 2011;54(1):37–58. [Google Scholar]
- 19.Ennett ST, Bauman KE, Hussong A, et al. The Peer Context of Adolescent Substance Use: Findings from Social Network Analysis. J Res Adolesc. 2006;16(2):159–186. [Google Scholar]
- 20.Osgood DW, Feinberg ME, Wallace LN, Moody J. Friendship group position and substance use. Addict Behav. 2014;39(5):923–933. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Center for Behavioral Health Statistics and Quality. 2018 National Survey on Drug Use and Health public use file codebook. Rockville, MD: Substance Abuse and Mental Health Services Administration; 2019. [Google Scholar]
- 22.Center for Behavioral Health Statistics and Quality. 2016 National Survey on Drug Use and Health public use file codebook. Rockville, MD: Substance Abuse and Mental Health Services Administration; 2018. [Google Scholar]
- 23.Center for Behavioral Health Statistics and Quality. 2017 National Survey on Drug Use and Health final analytic file codebook. Rockville, MD: Substance Abuse and Mental Health Services Administration; 2018. [Google Scholar]
- 24.Center for Behavioral Health Statistics and Quality. 2015 National Survey on Drug Use and Health Public Use File Codebook. Rockville, MD: Substance Abuse and Mental Health Services Administration; 2018. [Google Scholar]
- 25.Substance Abuse and Mental Health Services Administration. Comparing and Evaluating Youth Substance Use Estimates from the National Survey on Drug Use and Health and Other Surveys [Internet]. Substance Abuse and Mental Health Services Administration (US). https://www.ncbi.nlm.nih.gov/books/NBK533892/. Published 2012. Accessed Oct 25, 2021. [PubMed] [Google Scholar]
- 26.Mojtabai R, Olfson M, Han B. National Trends in the Prevalence and Treatment of Depression in Adolescents and Young Adults. Pediatrics. 2016;138(6). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Martins SS, Mauro CM, Santaella-Tenorio J, et al. State-level medical marijuana laws, marijuana use and perceived availability of marijuana among the general U.S. population. Drug and Alcohol Dependence. 2016;169:26–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Mariani AC, Williams AR. Perceived risk of harm from monthly cannabis use among US adolescents: National Survey on drug Use and Health, 2017. Prev Med Rep. 2021;23:101436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Fuller-Thomson E, Lacombe-Duncan A, Goodman D, Fallon B, Brennenstuhl S. From surviving to thriving: factors associated with complete mental health among childhood sexual abuse survivors. Soc Psychiatry Psychiatr Epidemiol. 2020;55(6):735–744. [DOI] [PubMed] [Google Scholar]
- 30.Hall-Lande J, Eisenberg M, Christenson S, Neumark-Sztainer D. Social isolation, psychological health, and protective factors in adolescence. Adolescence. 2007;42(166):265–286. [PubMed] [Google Scholar]
- 31.Sutherland I, Shepherd JP. Social dimensions of adolescent substance use. Addiction. 2001;96(3):445–458. [DOI] [PubMed] [Google Scholar]
- 32.Mennis J, Mason MJ. Social and geographic contexts of adolescent substance use: The moderating effects of age and gender. Social Networks. 2012;34(1):150–157. [Google Scholar]
- 33.Jacobs W, Goodson P, Barry AE, McLeroy KR, McKyer EL, Valente TW. Adolescent Social Networks and Alcohol Use: Variability by Gender and Type. Subst Use Misuse. 2017;52(4):477–487. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Davis L, Uezato A, Newell JM, Frazier E. Major depression and comorbid substance use disorders. Curr Opin Psychiatry. 2008;21(1):14–18. [DOI] [PubMed] [Google Scholar]
- 35.Schelde JT. Major depression: behavioral markers of depression and recovery. J Nerv Ment Dis. 1998;186(3):133–140. [DOI] [PubMed] [Google Scholar]
- 36.Westreich D, Greenland S. The table 2 fallacy: presenting and interpreting confounder and modifier coefficients. Am J Epidemiol. 2013;177(4):292–298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Miech R, Johnston L, O’Malley PM. Prevalence and attitudes regarding marijuana use among adolescents over the past decade. Pediatrics. 2017;140(6). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.ProCon. Legal recreational marijuana states and DC. Encyclopedia Britannica, Inc. https://marijuana.procon.org/legal-recreational-marijuana-states-and-dc/. Published 2019. Updated June 25, 2019. Accessed Aug 20, 2020. [Google Scholar]
- 39.Valente TW. Social network influences on adolescsent substance use: An introduction. Connections. 2003;25(2):11–16. [Google Scholar]
- 40.Duff C. Party drugs and party people: Examining the ‘normalization’ of recreational drug use in Melbourne, Australia. Int J Drug Policy. 2005;16(3):161–170. [Google Scholar]
- 41.Wilson H, Bryant J, Holt M, Treloar C. Normalisation of recreational drug use among young people: Evidence about accessibility, use and contact with other drug users. Health Sociol Rev. 2010;19(2):164–175. [Google Scholar]
- 42.Hathaway AD, Mostaghim A, Erickson PG, Kolar K, Osborne G. “It’s really no big deal”: The role of social supply networks in normalizing use of cannabis by students at Canadian universities. Deviant Behavior. 2018;39(12):1672–1680. [Google Scholar]
- 43.Belackova V, Vaccaro CA. “A friend with weed is a friend indeed”. J Drug Issues. 2013;43(3):289–313. [Google Scholar]
- 44.Mundt MP. The Impact of Peer Social Networks on Adolescent Alcohol Use Initiation. Academic Pediatrics. 2011;11(5):414–421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Cheadle JE, Walsemann KM, Goosby BJ. Teen Alcohol Use and Social Networks: The Contributions of Friend Influence and Friendship Selection. J Alcohol Drug Depend. 2015;3(5):224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Hatzenbuehler ML, McLaughlin KA, Xuan Z. Social Networks and Sexual Orientation Disparities in Tobacco and Alcohol Use. Journal of Studies on Alcohol and Drugs. 2015;76(1):117–126. [PMC free article] [PubMed] [Google Scholar]
- 47.Stewart T, Suldo S. Relationships between social support sources and early adolescents’ mental health: The moderating effect of student achievement level. Psychol Sch. 2011;48(10):1016–1033. [Google Scholar]
- 48.Perkins HW. Misperception Is Reality: The “Reign of Error” About Peer Risk Behaviour Norms Among Youth and Young Adults. In: Xenitidou M, Edmonds B, eds. The Complexity of Social Norms. Springer, Cham; 2014. [Google Scholar]
- 49.Amialchuk A, Ajilore O, Egan K. The influence of misperceptions about social norms on substance use among school-aged adolescents. Health Economics. 2019;28(6):736–747. [DOI] [PubMed] [Google Scholar]
- 50.Murphy SM, Rosenman R. The “Real” Number of Washington State Adolescents Using Marijuana, and Why: A Misclassification Analysis. Subst Use Misuse. 2019;54(1):89–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.O’Reilly JM, Hubbard ML, Lessler JT, Biemer PP, Turner CF. AUDIO AND VIDEO COMPUTER-ASSISTED SELF INTERVIEWING: Preliminary Tests of New Technologies for Data Collection. J Off Stat. 1994;10(2):197–214. [PMC free article] [PubMed] [Google Scholar]
- 52.Meza R, Jimenez-Mendoza E, Levy DT. Trends in Tobacco Use Among Adolescents by Grade, Sex, and Race, 1991–2019. JAMA Network Open. 2020;3(12):e2027465–e2027465. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.