Abstract
Purpose:
US homeschooling increased by 50% over 2007–2016. Homeschooled adolescents may have lower substance use rates, but past research treated other adolescents as homogeneous despite within-group differences. We used the 2015–18 National Survey on Drug Use and Health to compare adolescent substance use and psychopathology by homeschooled/educational status.
Methods:
Data were from 52,089 adolescents, classified by educational status (i.e., homeschooled; public/private school, low dropout risk; public/private school, at-risk for dropout; not in school) and compared on substance use and psychopathology variables.
Results:
Substance use rates were lowest in adolescents at low dropout risk, with significantly lower past-year prescription opioid misuse, tobacco use, non-marijuana illicit drug use, and nicotine dependence rates than homeschooled adolescents. Psychopathology treatment prevalence was lowest in homeschooled adolescents. Those at-risk for dropout had the highest rates of substance use.
Conclusions:
While homeschooled adolescents have relatively low substance use rates, they exceed those of low dropout risk adolescents.
Keywords: prescription misuse, substance use, adolescent, homeschool, education
INTRODUCTION
In 2016, 3.3% of United States (US) school-aged youth were homeschooled (1), with rates increasing, as only 2.2% were homeschooled in 2007 (2). Homeschooling occurs for many reasons, including concerns about negative school and peer influences, concerns about instruction, and for religious reasons (1).
Homeschooled adolescents may have lower substance use rates. Vaughn et al. (3) found that homeschooled adolescents (12–17 years) had lower rates of past-year tobacco, alcohol, and marijuana use than non-homeschooled adolescents, using the 2002–13 National Survey on Drug Use and Health (NSDUH). Similarly, Thomson and Jang (4) found elevated alcohol use and intoxication among public and private school adolescents versus homeschooled adolescents (≤13 years), via the 2002–05 National Study of Youth and Religion. Homeschooled adolescents are heterogeneous, though, and less religious homeschooled adolescents may have higher substance use rates (2).
Major questions remain about homeschooled adolescents. First, their prescription drug misuse (PDM) rates are unstudied, despite notable recent changes in adolescent PDM, including increased PDM-involved overdose (5). Second, their mental health functioning warrants further examination, with the only study, from Israel, suggesting differences in externalizing but not internalizing psychopathology (6). Third, past studies treated non-homeschooled adolescents as a homogeneous group, despite substance use differences based on dropout and risk for dropout (7). Our aim was to address these questions using the 2015–18 NSDUH, with multiple years aggregated to increase the homeschooled sample.
METHODS
The NSDUH is an annual US survey, with an independent, multistage area probability design and weighting for nationally representative estimates. Further information is available elsewhere (8). Participants were 52,089 adolescents (Table 1). 2,777 adolescents had missing educational data (5.0%); they were more likely to be male, younger, black/Latinx, and of lower income. The first author’s IRB exempted this research from further oversight.
Table 1:
Adolescent Sociodemographics by Educational Status
| Homeschooled (a) | Public/Private School, Low Dropout Risk (b) | Public/Private School, At-Risk for Dropout (c) | Not in School (d) | |
|---|---|---|---|---|
| Sample Size (weighted %) | 307 (0.6%) | 41,898 (80.8%) | 5,350 (9.8%) | 4,534 (8.8%) |
| Mean (95% CI) | Mean (95% CI) | Mean (95% CI) | Mean (95% CI) | |
| Age (in years) | 14.1 (13.9–14.4) | 14.6 (14.5–14.6) | 14.8 (14.7–14.9) | 14.7 (14.7–14.8) |
| Religiosity | 11.3 (10.7–11.8) | 10.8 (10.7–10.9) | 9.5 (9.4–9.6) | 9.9 (9.7–10.2) |
| % (95% CI) | % (95% CI) | % (95% CI) | % (95% CI) | |
| Male Sex | 54.9 (46.1–63.4) | 49.2 (48.5–49.9) | 53.3 (51.6–55.0) | 60.4 (58.5–62.3) |
| Race/Ethnicity | ||||
| White | 61.5 (53.3–69.0) | 53.8 (53.0–54.6) | 52.7 (50.9–54.5) | 52.2 (50.1–54.3) |
| Black | 12.2 (7.9–18.3) | 12.8 (12.2–13.3) | 15.5 (14.3–16.8) | 16.6 (15.2–18.2) |
| Hispanic/Latinx | 19.5 (14.3–26.1) | 23.4 (22.7–24.1) | 24.3 (22.7–26.0) | 22.9 (21.2–24.6) |
| Asian-American | 2.9 (1.1–7.6) | 5.9 (5.5–6.3) | 2.5 (2.0–3.0) | 4.2 (3.3–5.3) |
| Multiracial | 2.4 (1.1–5.2) | 3.2 (3.0–3.4) | 3.4 (2.9–4.1) | 2.8 (2.3–3.4) |
| American Indian/Native American | 0.8 (0.3–2.4) | 0.5 (0.5–0.6) | 1.1 (0.8–1.6) | 0.7 (0.5–1.0) |
| Hawaiian/Pacific Islander | 0.7 (0.1–4.1) | 0.4 (0.3–0.5) | 0.4 (0.2–0.8) | 0.6 (0.4–1.0) |
| Household Income | ||||
| Less than $20,000 | 20.6 (14.6–28.3) | 13.4 (12.9–14.0) | 23.7 (22.1–25.3) | 21.2 (19.3–23.2) |
| $20,000–49,999 | 28.2 (22.5–34.6) | 26.6 (26.0–27.2) | 33.0 (31.3–34.8) | 30.0 (28.0–32.0) |
| $50,000–74,999 | 11.2 (7.9–15.8) | 14.6 (14.1–15.0) | 13.3 (12.3–14.5) | 13.5 (12.0–15.1) |
| $75,000 and greater | 40.0 (31.6–49.0) | 45.4 (44.5–46.3) | 30.0 (28.1–32.0) | 35.4 (33.2–37.7) |
| Population Density | ||||
| CBSA of 1 million or more persons | 46.2 (39.6–53.0) | 55.7 (54.8–56.6) | 48.9 (47.0–50.7) | 53.9 (51.9–55.9) |
| CBSA of under 1 million persons | 42.2 (35.0–49.8) | 38.8 (38.0–39.7) | 44.0 (42.1–45.9) | 39.3 (37.4–41.2) |
| Not in a CBSA | 11.6 (7.7–17.1) | 5.5 (5.0–6.0) | 7.1 (6.3–8.1) | 5.8 (5.3–6.3) |
Data: 2015–2018 NSDUH
Acronyms: CBSA = Core-based Statistical Area
The independent variable was past-year educational status: [1] homeschooled; [2] public/private school, low dropout risk; [3] public/private school, at-risk for dropout; [4] not in school. Homeschooled adolescents were “home-schooled at any time” in the past year. Dropout risk was having one or more of: grades ≤ D+ at the last grading period; ≥ one year older than typical for grade, and the adolescent “hated going to school”; these have been linked to elevated dropout risk (9) and PDM (7).
Substance use variables included past-year prescription use and misuse of opioid, benzodiazepine, or stimulant medication, past-month binge alcohol use, past-year tobacco use, past-year marijuana use, past-year non-marijuana illicit drug use (i.e., cocaine, heroin, methamphetamine, inhalant and/or hallucinogen), past-year substance use disorder (SUD), past-month nicotine dependence, and past-year SUD treatment. Past-year SUD is DSM-IV alcohol, marijuana, cocaine, heroin, hallucinogen, inhalant, methamphetamine, opioid, stimulant, tranquilizer, and/or sedative abuse/dependence.
Mental health variables were past-year major depression and past-year mental health treatment. Sociodemographics included age, sex, race/ethnicity, household income, population density, and religiosity. Religiosity was from Grucza et al. (10) and captured frequency of religious service attendance, importance and influence of beliefs, and importance that friends share beliefs.
Analyses employed STATA 15.1, incorporating the NSDUH complex survey design. Prevalence rates of substance use or mental health variables by educational status were established via weighted cross-tabulations, and logistic models evaluated pairwise differences. Logistic models controlled for sociodemographics, with pairwise differences Bonferroni-corrected (a priori p≤0.0083, or 0.05/6 comparisons).
RESULTS
The lowest substance use rates were generally in adolescents at low dropout risk (see Table 2). While homeschooled adolescents had lower past-year marijuana use rates (9.0%) than low dropout risk adolescents (11.1%), this difference was non-significant. In contrast, homeschooled adolescents had significantly higher rates of past-year prescription opioid misuse, benzodiazepine use, tobacco use and non-marijuana illicit drug use, and past-month nicotine dependence than low dropout risk adolescents. Notably, homeschooled adolescents had the lowest rates of mental health variables, with significantly lower mental health treatment rates than all other groups. Adolescents at-risk for dropout had the highest prevalence of all variables, followed by those not in school; both had significantly higher rates than low dropout risk adolescents.
Table 2:
Adolescent Substance Use and Mental Health Characteristics by Educational Status
| Homeschooled (a) | Public/Private School, Low Dropout Risk (b) | Public/Private School, At-Risk for Dropout (c) | Not in School (d) | |
|---|---|---|---|---|
| Sample Size (weighted %) | 307 (0.6%) | 41,898 (80.8%) | 5,350 (9.8%) | 4,534 (8.8%) |
| % (95% CI) | % (95% CI) | % (95% CI) | % (95% CI) | |
| Prescription Use and Misuse | ||||
| Past-Year Opioid Use | 20.0 (14.3–27.2) | 17.9 (17.5–18.4)c, d | 23.9 (22.2–25.7)b | 23.6 (22.0–25.3)b |
| Past-Year Opioid Misuse | 5.3 (3.2–8.8)b | 2.8 (2.6–3.0)a, c, d | 6.2 (5.5–7.0)b | 5.2 (4.5–5.9)b |
| Past-Year Stimulant Use | 6.8 (3.6–12.5) | 6.4 (6.1–6.7)c, d | 11.1 (10.0–12.1)b | 10.1 (9.1–11.2)b |
| Past-Year Stimulant Misuse | 1.6 (0.6–4.4) | 1.5 (1.4–1.7)c, d | 3.1 (2.6–3.7)b | 2.6 (2.1–3.2)b |
| Past-Year Benzodiazepine Use | 5.3 (3.1–9.2)b | 2.5 (2.4–2.7)a, c, d | 6.7 (5.9–7.6)b | 5.9 (5.1–6.9)b |
| Past-Year Benzodiazepine Misuse | 2.4 (1.0–5.6) | 1.3 (1.2–1.4)c, d | 3.7 (3.1–4.4)b | 3.1 (2.6–3.7)b |
| Other Substance Use | ||||
| Past-Month Binge Alcohol Use | 6.9 (3.8–12.2) | 4.6 (4.3–4.8)c, d | 8.8 (8.0–9.7)b | 8.0 (7.2–8.9)b |
| Past-Year Tobacco Use | 14.1 (9.7–20.0)b | 8.5 (8.2–8.9)a, c, d | 18.5 (16.9–20.1)b, d | 15.9 (14.6–17.3)b, c |
| Past-Year Marijuana Use | 9.0 (5.7–14.0) | 11.1 (10.8–11.5)c, d | 21.4 (20.0–22.9)b, d | 17.8 (16.5–19.1)b, c |
| Past-Year Non-Marijuana Illicit Drug Use | 7.7 (4.8–12.2)b | 3.7 (3.4–3.9)a, c, d | 7.9 (7.0–8.9)b | 6.6 (5.9–7.5)b |
| Past-Year Any SUD | 3.9 (2.1–7.0) | 3.4 (3.2–3.6)c, d | 9.6 (8.5–10.8)b, d | 7.2 (6.4–8.1)b, c |
| Past-Year Nicotine Dependence | 1.8 (0.8–3.6)b | 0.3 (0.2–0.3)a, c, d | 2.1 (1.7–2.6)b | 1.4 (1.0–1.9)b |
| Past-Year SUD Treatment | 0.9 (0.4–2.2) | 0.5 (0.4–0.6)c, d | 1.9 (1.5–2.5)b | 1.7 (1.3–2.2)b |
| Mental Health | ||||
| Past-Year Major Depression | 10.8 (7.2–15.8)c | 12.6 (12.1–13.2)c | 23.4 (21.9–24.9)a, b, d | 11.7 (10.3–13.3)c |
| Past-Year Mental Health Treatment | 11.3 (7.1–17.7)b, c, d | 22.7 (22.1–23.3)a, c | 35.2 (33.5–36.9)a, b, d | 23.6 (21.9–25.4)a, c |
Data: 2015–2018 NSDUH
Acronyms: SUD = DSM-IV Substance Use Disorder
Notes: Past-Year Non-Marijuana Illicit Drug Use includes cocaine, heroin, methamphetamine, inhalant and/or hallucinogen use; Superscript letters denote differences from the group with the letter (i.e., a denotes a significant difference from homeschooled adolescents), with comparisons adjusted for age, race/ethnicity, sex, income, population density, and religiosity, and Bonferroni corrected for multiple comparisons (i.e., p-value for significance is 0.0083, or 0.05/6 comparisons).
Homeschooled adolescents had the highest mean religiosity (11.3), followed by low dropout risk adolescents (10.8), those not in school (9.9), and those at-risk for dropout (9.5; Table 1).
DISCUSSION
Public/private school adolescents at low dropout risk had the lowest prevalence rates of substance use, contrasting with prior research (3, 4). This difference was likely because Vaughn et al. (3) aggregated all adolescents, and Thomson and Jang (4) did not differentiate public/private school-attending adolescents by dropout risk. Adolescents not in school and those at-risk for dropout have significantly higher substance use rates (7, 9), highlighting the importance of comparison groups in studying homeschooled adolescents.
Similarly, homeschooled adolescents are diverse (1, 2), and Hodge et al. (2) found that highly religious homeschooled adolescents had significantly lower substance use rates. While the available sample size was not sufficient to replicate their latent classes, we accounted for self-reported religiosity in analyses, and homeschooled adolescents still had higher rates of past-year prescription opioid misuse, benzodiazepine use, tobacco use and non-marijuana illicit drug use, and past-month nicotine dependence. In contrast, homeschooled adolescents had lower rates of major depression and psychopathology treatment, highlighting a potential strength.
Limitations include self-selection and response bias, the inability to infer causality (i.e., does elevated substance use result in homeschooling or vice versa?), and limited available measures. Adolescents with missing data differed from those with complete data, and only 0.6% of the sample was homeschooled, consistent with past research (3). Still, 3.3% of adolescents nationwide are homeschooled, suggesting greater response bias among homeschooled adolescents.
Future research needs to account for within-group differences both in homeschooled and comparison adolescents, and investigations of whether the lower homeschooled rates of psychopathology treatment reflect better mental health or unmet treatment needs are warranted. Also, future longitudinal research is needed to capture potential causality in the relationships among educational status, substance use, and mental health. Policies are needed to limit adolescent dropout by addressing underlying factors, including substance use and academic factors (9); screening for psychopathology is recommended in at-risk adolescents. As some homeschooled adolescents have elevated substance use rates, careful substance use screening for all adolescents is warranted.
IMPLICATIONS AND CONTRIBUTION STATEMENT.
Past research suggests that homeschooled adolescents have lower substance use rates than other adolescents. This study found that risk for dropout or having left school moderates these past findings, with the lowest substance use rates among adolescents in school at low dropout risk, with homeschooled adolescents somewhat higher.
Acknowledgments
This work was funded by R01 DA043691 and R01 DA031160 from the National Institute on Drug Abuse (NIDA). The NSDUH is funded by the Substance Abuse and Mental Health Services Administration (SAMHSA). The content is the authors’ responsibility and does not necessarily represent the views of NIDA or SAMHSA. NIDA and SAMHSA had no role in the design of the study, the analyses, interpretation of results or the decision to submit the manuscript for publication.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Ty Schepis completed the first draft of the manuscript; no one was paid for work on this manuscript.
REFERENCES
- [1].Wang K, Rathbun A, Musu L. School Choice in the United States: 2019 (NCES 2019–106). Washington, DC: U.S. Department of Education, National Center for Education Statistics, 2019. [Google Scholar]
- [2].Hodge DR, Salas-Wright CP, Vaughn MG. Behavioral Risk Profiles of Homeschooled Adolescents in the United States: A Nationally Representative Examination of Substance Use Related Outcomes. Substance Use & Misuse 2017;52:273–285. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Vaughn MG, Salas-Wright CP, Kremer KP, et al. Are homeschooled adolescents less likely to use alcohol, tobacco, and other drugs? Drug and Alcohol Dependence 2015;155:97–104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Thomson RA Jr., Jang SJ. Homeschool and underage drinking: Is it more protective than public and private schools? Deviant Behavior 2016;37:281–301. [Google Scholar]
- [5].Ford JA. Prescription Opioid Misuse Among Adolescents. Pediatric Clinics of North America 2019;66:1099–1108. [DOI] [PubMed] [Google Scholar]
- [6].Guterman O, Neuman A. Schools and emotional and behavioral problems: A comparison of school-going and homeschooled children. The Journal of Educational Research 2017;110:425–432. [Google Scholar]
- [7].Schepis TS, Teter CJ, McCabe SE. Prescription drug use, misuse and related substance use disorder symptoms vary by school enrollment status in U.S. adolescents and young adults. Drug Alcohol Depend 2018;189:172–177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Center for Behavioral Health Statistics and Quality. 2016 National Survey on Drug Use and Health: Methodological summary and definitions. Rockville, MD: Substance Abuse and Mental Health Services Administration, 2017. [Google Scholar]
- [9].DuPont RL, Caldeira KM, DuPont HS, et al. America’s dropout crisis: The unrecognized connection to adolescent substance use. Rockville, MD: Institute for Behavior and Health, Inc. Available at www.ibhinc.org, www.PreventTeenDrugUse.org, and www.cls.umd.edu/docs/AmerDropoutCrisis.pdf; 2013. [Google Scholar]
- [10].Grucza RA, Agrawal A, Krauss MJ, et al. Declining Prevalence of Marijuana Use Disorders Among Adolescents in the United States, 2002 to 2013. J Am Acad Child Adolesc Psychiatry 2016;55:487–494 e486. [DOI] [PMC free article] [PubMed] [Google Scholar]