Self-Reported Perceived Negative Consequences of Marijuana Use Among U.S. Young Adult Users, 2008-2019

Abstract

Purpose.

This study estimated self-reported perceived negative marijuana use consequences among a national sample of U.S. young adults, examining consequence prevalence differences by use frequency, college attendance, living situation, employment, sex, and race/ethnicity; and use frequency/sociodemographic characteristic interactions.

Methods.

A subsample of 1,212 respondents from the 2004-2018 class cohorts of 12^th grade students participating in the nationally-representative Monitoring the Future study was surveyed up to two times from modal ages 19 through 22 (in 2008-2019). Respondents self-reported negative consequences related to their own past 12-month marijuana use. Bivariate and multivariable models examined subgroup differences in consequence prevalence.

Results.

Approximately 60% of those using frequently (20+ use occasions in the past 30 days) and 35% of those using non-frequently reported negative consequences. Among all young adult marijuana users, 31.1% reported emotional/physical consequences, 12.9% performance/financial consequences, and 12.3% relational consequences. Use frequency was positively associated with consequence likelihood, excluding regret and unsafe driving. Among college students, frequent use was more strongly associated with any and performance/financial consequences. Controlling for use frequency, men reported more performance/financial consequences; relational consequences were higher among Hispanic (vs. White) respondents, and those living with parents, employed full-time, and not attending 4-year colleges.

Conclusion.

Young adults using marijuana reported a wide range of negative use consequences; likelihood of most consequences increased with higher use frequency. Perceived consequences varied by college attendance, living situation, employment, sex, and race/ethnicity. Efforts to reduce negative marijuana consequences may be strengthened by recognizing and addressing the different types of negative consequences users perceive.

1. Introduction

While medical marijuana (or cannabis) use may be effective for treating some health conditions,^1,2,3 recreational use is associated with negative mental, physical, and behavioral outcomes.^{1,4,5,6,7,8,9,10} Research indicates positive associations between non-medical use (hereafter referred to simply as use) frequency and the overall risk of negative consequences.^{11,12,13,14,15} Among those using marijuana, subgroups with higher use frequency are likely at higher risk for overall negative consequences and the public health burdens of such outcomes. Yet, several components of use/consequence associations are still unclear, including if use frequency is associated equally across consequence types, if frequency/consequence association strength varies across subgroups, and if particular subgroups have higher consequence likelihood controlling for use frequency.

Young adults are at high risk for marijuana use: past 12-month marijuana prevalence has been highest among young adults aged 18-25 for more than 15 years.¹⁶ The percentage of young adults reporting frequent use (use on 20+ occasions in the past 30 days) has been increasing; 2019 frequent use prevalence estimates were the highest observed in over 30 years.¹⁷ Use frequency increases have occurred across young adult subgroups including both those attending and not attending college,¹⁷ as well as both men and women.¹⁸ Although most subgroups have seen historical increases in marijuana use frequency, such frequency is higher among those not attending college full-time (vs. attending full-time),¹⁷ those not employed (vs. employed),¹⁹ and men (vs. women).¹⁸

Research highlights the importance of understanding different types of perceived negative consequences in order to inform intervention methods such as motivation enhancement²⁰ and community-level efforts shown to have promise in reducing marijuana use, such as mass media campaigns.²¹ One young adult college student study found neither perceived academic nor social consequences were associated with marijuana frequency, nor was the actual experience of academic consequences associated with perceived likelihood of future academic consequences.²⁰ However, experiencing social consequences was associated with a higher level of anticipating such consequences in the future.²⁰ College-based studies have found higher marijuana frequency was associated with increased risk of some social/interpersonal consequences, academic/occupational or financial consequences, and/or risky behaviors.^11,12,22 Two studies found frequency was not associated with feeling bad about use.^11,22 Knowledge about which consequences are most often reported by users themselves, to what degree consequences are associated with use frequency, and if consequences vary across subgroups after accounting for use frequency, may provide important information for intervention development.

Most research on young adult negative marijuana consequences has focused exclusively on college samples.^{11,12,23,13,20,22,24,25,26,27,28,29,30,31,32,33,34,35,36,37} Compared with those attending college, young adults not attending college are more likely to have drug use and conduct disorders as well as some personality disorders,³⁸ which are associated positively with marijuana use and use disorder.³⁹ Further, some occupational groups (e.g., food preparation and serving; building and grounds maintenance) that may be more likely to be filled by young adults not attending college are associated with higher overall or workplace illicit drug use.⁴⁰ Thus, those not in college may experience a higher likelihood of perceived negative emotional and physical consequences, and possibly differences in performance/financial consequences. However, studies have found college-attending young adults with substance use disorders were significantly less likely to receive past year treatment for alcohol and drug use than those not attending who also had substance use disorders;³⁸ if high-risk use is occurring with less treatment, college students may report higher perceived negative consequences. Of the limited number of young adult marijuana consequences studies not restricted to college students, one involved 336 respondents in the Seattle metropolitan area¹³ and the other 288 respondents in Rhode Island.⁴¹ Lee et al.¹³ did not examine college-based differences; Moitra et al.⁴¹ found no differences in either cannabis disorder or the Marijuana Problem Scale based on “enrolled in school”, but did not examine differences in consequence type. Further, research examining college-related differences in alcohol consequences has raised the importance of exploring a range of factors related to college attendance, including type of college and living situation. Specifically, those attending 4-year colleges have been found to have higher risk than individuals attending 2-year colleges for alcohol-related consequences; those living with parents (vs. living on campus, or off-campus without parents) have been found to have lower risk for such consequences.^42,43 Data examining negative consequences from marijuana use in both college and non-college samples are needed.

Potential consequence differences also may be associated with employment, sex and race/ethnicity. Marijuana use has been associated with unemployment^44,45 and job loss.^46,47 Yet, use consequences may be high among the employed; marijuana use before/during work has been associated with negative job performance.⁴⁸ Sex differences in marijuana use and mental health associations⁴⁹ may indicate a potential for more emotional consequences among women. Regarding racial/ethnic differences, marijuana frequency does not differ strongly by race/ethnicity during the early 20s,⁵⁰ but misuse and dependence are more likely among adult Black and Hispanic versus White individuals who use.⁵¹ Available studies have observed few additional racial/ethnic use consequences differences.⁵²

Based only on overall use frequency differences, young adults not in college, the unemployed, and men would have the greatest likelihood of marijuana consequences. There are at least three potential challenges to this hypothesis and its implications for prevention and intervention services. First, use frequency may not be associated consistently across consequence types. If some consequences are not associated with use frequency, then alerting users to shared likelihood—irrespective of how frequently they use—may be an important public health message. Alternatively, describing which consequences are reported more by frequent users themselves may help prevent use progression among those who believe marijuana is relatively consequence-free. Second, higher use frequency may be associated with stronger increases in negative consequence likelihood for individuals in particular subgroups versus others; if so, targeted efforts to reduce high frequency use should be particularly focused on relevant high-frequency groups. Third, some subgroups may have a higher likelihood of particular marijuana consequences regardless of how frequently they use, and if so, prevention and intervention efforts should focus on all users among those subgroups. Knowledge of how consequences are reported by users themselves based on use frequency and sociodemographic characteristics can help improve targeting of prevention messages and intervention programming.

The current study used data from young adults across the U.S. aged 19-22 to address four research aims: (1) estimate overall prevalence of specific self-reported perceived negative marijuana consequences; (2) examine which consequences were more or less likely to change with use frequency; (3) examine if use frequency made more of a difference in consequence likelihood for some subgroups versus others; and (4) examine college attendance, living situation, employment, sex, and race/ethnicity differences in consequence types, controlling for use frequency.

2. Methods

2.1. Study sample and survey administration

All data were obtained from adults participating in the longitudinal Monitoring the Future (MTF) study at modal ages 19-22. Detailed methods are presented elsewhere.^17,53,54 In brief, MTF surveys annual, nationally representative cross-sectional samples of approximately 15,000 12th grade students at modal age (hereafter “age”) 18 from 130-140 public and private schools in the coterminous United States (school samples revised annually). A representative subsample of approximately 2,450 students is selected from each annual sample for longitudinal follow-up with oversampling of drug users (weighted analyses account for oversampling). A random half of each year’s selected sample begins biennial follow-up the year after 12th grade (age 19), and the other half begins the following year (age 20). Data were collected using paper and electronic surveys. Marijuana consequence measures were added in 2008. Thus, data were drawn from the 2004–2018 12th grade cohorts who were eligible to participate in first follow-up (age 19/20) and/or second follow-up (age 21/22) surveys administered from 2008-2019 (see Supplement Table 1). A University of Michigan Institutional Review Board approved the study.

Marijuana consequence measures were included on one of six randomly-distributed questionnaire forms (used consistently across biennial follow-ups); a total of 5,747 respondents from eligible cohorts received that form. Of these, 2,784 (48.4% of 5,747) responded to at least one of the age 19/20 and 21/22 surveys: 771 responded at age 19/20 only, 807 at age 21/22 only, and 1,206 at both, resulting in a total of 3,990 cases. Of these, 2,318 (58.8%) reported no past 12-month marijuana use, making them ineligible for the current analyses; 126 (3.2%) additional cases were removed due to missing data on 30-day marijuana use or use consequences, leaving 1,546 cases from 1,212 respondents for analysis.

2.2. Measures

2.2.1. Perceived consequences

Respondents were asked, “Think back over the last 12 months. Did your use of marijuana cause you any of the following problems, even just a little?” Fourteen items (Table 1) were listed; respondents marked all that applied. Analyses examined perceived consequences as individual measures, and grouped within categories to emphasize conceptual differences. Five category types were coded: (1) emotional/physical consequences (five consequences; e.g., less energy); (2) performance/financial consequences (two consequences; e.g., hurt performance in school/on the job); (3) relational consequences (five consequences; e.g., hurt relationship with parents); (4) behave in ways later regretted (single consequence); and (5) driving unsafely (single consequence). Respondents were coded “yes” if they said yes to any of the category consequences. An overall binary measure indicating any perceived negative consequences (yes/no) was coded.

Table 1.

Prevalence of Perceived Marijuana Use Consequences among U.S. Past 12-month Users Aged 19-22

	% (95% CI)
Any consequence endorsement	39.8 (36.8, 42.9)
Any emotional/physical consequences	31.1 (28.2, 33.9)
Caused you to have less energy	24.0 (21.3, 26.6)
Caused you to be less interested in other activities	13.4 (11.3, 15.4)
Made you feel bad (e.g., depressed, anxious, ashamed)	7.5 (6.0, 9.0)
Caused you to be less stable emotionally	5.7 (4.5, 7.0)
Caused your physical health to be bad	2.6 (1.8, 3.4)
Any performance/financial consequences	12.9 (10.9, 14.8)
Hurt performance in school/on the job	8.3 (6.8, 9.9)
Caused financial difficulties	7.5 (6.0, 9.0)
Any relational consequences	12.3 (10.4, 14.2)
Hurt relationship with parents	6.4 (5.0, 7.7)
Hurt relationship with spouse/fiancee/girlfriend/boyfriend	5.7 (4.4, 7.0)
Hurt relationships with friends	2.8 (1.9, 3.6)
Caused you to get into angry argument	2.8 (1.9, 3.8)
Hurt relationship with children	0.3 (0.0, 0.7)
Caused you to behave in ways that you later regretted	8.5 (6.8, 10.2)
Caused you to drive unsafely	4.2 (2.9, 5.5)

Notes: Unweighted n=1,546.

2.2.2. Predictors and covariates

For marijuana use, respondents were asked, “On how many occasions (if any) have you used marijuana (weed, pot) or hashish (hash, hash oil) during the last 30 days?” [0 occasions, 1-2, 3-5, 6-9, 10-19, 20-39, and 40 or more]; responses were recoded as 0, 1.5, 4, 7.5, 15, 29.5, and 40. For descriptive purposes, an additional measure indicating frequent marijuana use (sometimes referred to as daily use) was coded defined as use on 20+ occasions in the past 30 days versus other.^17,55,56 College attendance was coded as 1=not attending, 2=attending a 2-year college or a vocational or technical school (2-year/votech), or 3=attending a 4-year college. Living situation was coded as living with parents versus other. Employment was coded as 1=not employed (no outside job/laid-off or waiting to start a job/no paid employment), 2=employed part-time (2 or more jobs/one part-time job), or 3=one full-time job. Sex was coded as male or female. Race/ethnicity was coded as White, Black/African American, Hispanic, or Other (other groups, including multiracial, were combined due to sample size limitations). Sex and race/ethnicity were reported at age 18; remaining measures were reported at each follow-up. Age (based on modal age of the cohort at time of follow-up survey) was coded in integers from 19 to 22.

2.3. Statistical analysis

Descriptive analyses used SAS 9.4 (SAS Institute Inc., Cary, NC) survey commands (SURVEYFREQ) and clustering to account for repeated individual participation. Bivariate and multivariable logistic regression modeling used Mplus 7.4 (Muthén & Muthén, Los Angeles, CA), clustering by individual, and specifying type=complex, estimator=MLR, and integration=montecarlo. Missing predictor and covariate data were addressed by using full-information maximum-likelihood estimation.⁵⁷ All analyses were weighted to adjust for sampling probability and non-response (based on extensive 12^th grade measures). Aim 2 (use frequency/consequence associations) multivariable models simultaneously included use frequency, college status, living situation, employment, sex, race/ethnicity, and age. Aim 3 (use frequency by subgroup interactions with consequence likelihood) multivariable models simultaneously included marijuana frequency, the specific subgroup of interest (e.g., sex), the interaction of frequency by subgroup, and age. Aim 4 (subgroup differences in consequences, controlling for use frequency) multivariable models simultaneously included marijuana frequency, the specific subgroup of interest (e.g., sex), and age. Due to large sample sizes and number of statistical tests examined, discussion will be limited to results significant at p<.01.

3. Results

Table 2 provides sample characteristics.

Table 2.

Sample Descriptives: U.S. Past 12-month Marijuana Users Aged 19-22

	%/Mean	(SE)
Past 30-day marijuana use (n=1,546)
Mean frequency	3.44	(0.19)
Non-frequent use (%)	81.4	(1.29)
Frequent use (%)	18.6	(1.29)
College attendance (n=1,535) %
Not attending	36.8	(1.61)
2-year/votech	17.7	(1.26)
4-year college	45.6	(1.64)
Living situation (n=1,528) %
Without parents	57.3	(1.66)
With parents	42.7	(1.66)
Employment (n=1,515) %
Not employed	35.3	(1.58)
Employed part-time	44.1	(1.57)
Employed full-time	20.6	(1.39)
Sex (n=1,546) %
Female	49.1	(1.75)
Male	50.9	(1.75)
Race/ethnicity (n=1,525) %
Black/African-American	12.0	(1.44)
Hispanic	14.4	(1.34)
White	60.6	(1.82)
Other	13.0	(1.27)

3.1. Perceived consequence prevalence

Approximately 40% (39.8%) of all users perceived any negative use consequences (Table 1): 31.1% emotional/physical consequences, 12.9% performance/financial consequences, 12.3% relational consequences, 8.5% behavior later regretted, and 4.2% unsafe driving. Among emotional/physical consequences, having less energy was reported most frequently (24.0%), followed by less interest in other activities (13.4%). Among performance/financial consequences, 8.3% reported hurt school or job performance; 7.5% reported financial difficulties. Among relational consequences, hurting relationships with parents (6.4%) or spouse/fiancee/girlfriend/boyfriend (5.7%) were most frequently reported.

3.2. Direct frequency/consequence associations

Table 3 shows the odds of any consequences, and emotional/physical, performance/financial, and relational consequences, increased with marijuana frequency. Additional use occasions were associated with an adjusted increase of 2% in the odds of relational consequences, 3% for emotional/physical consequences, 4% for any consequences, and 6% for performance/financial consequences. Among those reporting frequent use, 60.5% reported any consequences (vs. 35.1% among those reporting non-frequent use), 45.0% (vs. 27.9%) emotional/physical consequences, 30.2% (vs. 8.9%) performance/financial consequences, and 21.2% (vs. 10.3%) relational consequences. Use frequency was not associated with regret or unsafe driving at the p<.01 level. Some individual consequences were not associated with use frequency at the p<.01 level: three emotional/physical consequences (made you feel bad; less stable emotionally; and—in the multivariable model—caused physical health to be bad); and three relational consequences (hurt relationship with spouse/fiancée/girlfriend/boyfriend; caused angry arguments; hurt relationships with children).

Table 3.

Associations Between Use Frequency and Prevalence of Perceived Marijuana Use Consequences Among U.S. Past 12-month Users Aged 19-22

	Associations with Continuous Use Frequency		Prevalence by Frequent Usea Dichotomy
	OR (95% CI) p	AOR (95% CI) p	Non-frequent % (95% CI)	Frequent % (95% CI)
Any consequence endorsement	1.04 (1.03, 1.05) <.001	1.04 (1.03, 1.05) <.001	35.1 (31.8, 38.4)	60.5 (53.4, 67.6)
Any emotional/physical consequences	1.03 (1.02, 1.04) <.001	1.03 (1.02, 1.04) <.001	27.9 (24.7, 31.0)	45.0 (38.2, 51.8)
Caused you to have less energy	1.03 (1.02, 1.04) <.001	1.03 (1.02, 1.04) <.001	21.6 (18.7, 24.4)	34.3 (28.0, 40.6)
Caused you to be less interested in other activities	1.04 (1.03, 1.05) <.001	1.04 (1.03, 1.05) <.001	10.0 (8.0, 12.0)	28.2 (22.1, 34.4)
Made you feel bad (e.g., depressed, anxious, ashamed)	1.00 (0.98, 1.01) 0.621	1.00 (0.98, 1.01) 0.684	7.5 (5.9, 9.2)	7.5 (4.3, 10.7)
Caused you to be less stable emotionally	1.01 (1.00, 1.03) 0.116	1.01 (1.00, 1.03) 0.158	5.5 (4.1, 6.8)	7.0 (4.0, 9.9)
Caused your physical health to be bad	1.02 (1.01, 1.04) 0.006	1.02 (1.01, 1.04) 0.011	2.1 (1.3, 2.9)	4.7 (2.4, 7.1)
Any performance/financial consequences	1.05 (1.04, 1.06) <.001	1.06 (1.04, 1.07) <.001	8.9 (7.1, 10.7)	30.2 (24.1, 36.4)
Hurt performance in school/on the job	1.04 (1.02, 1.05) <.001	1.04 (1.02, 1.05) <.001	6.5 (5.0, 8.1)	16.1 (11.4, 20.9)
Caused financial difficulties	1.07 (1.05, 1.08) <.001	1.07 (1.05, 1.08) <.001	4.2 (3.0, 5.4)	21.8 (16.5, 27.2)
Any relational consequences	1.03 (1.01, 1.04) <.001	1.02 (1.01, 1.04) <.001	10.3 (8.3, 12.2)	21.2 (15.9, 26.5)
Hurt relationship with parents	1.03 (1.02, 1.05) <.001	1.03 (1.02, 1.05) <.001	4.6 (3.4, 5.9)	13.8 (9.2, 18.4)
Hurt relationship with spouse/fiancee/girlfriend/boyfriend	1.02 (1.00, 1.03) 0.050	1.02 (1.00, 1.03) 0.030	5.1 (3.7, 6.4)	8.4 (5.0, 11.8)
Hurt relationships with friends	1.04 (1.02, 1.06) <.001	1.04 (1.02, 1.06) <.001	2.0 (1.2, 2.8)	6.2 (3.4, 9.0)
Caused you to get into angry argument	1.03 (1.00, 1.05) 0.026	1.02 (1.00, 1.05) 0.076	2.3 (1.2, 3.3)	5.4 (2.6, 8.3)
Hurt relationship with children	1.04 (0.99, 1.10) 0.144	1.03 (0.98, 1.08) 0.303	0.2 (0.0, 0.5)	0.9 (0.0, 2.3)
Caused you to behave in ways that you later regretted	1.01 (0.99, 1.02) 0.353	1.01 (0.99, 1.02) 0.337	8.1 (6.2, 9.9)	10.4 (6.4, 14.5)
Caused you to drive unsafely	1.02 (1.00, 1.04) 0.040	1.02 (1.00, 1.04) 0.036	3.8 (2.4, 5.2	6.0 (2.9, 9.0

Notes: Unweighted n=1,546. OR = bivariate odds ratio. AOR = adjusted odds ratio from multivariable models simultaneously controlling for continuous past 30-day use frequency, sex, race/ethnicity, college status, living situation, employment, and age. Bold font indicates associations with p<.01 or stronger.

^aFrequent use defined as use on 20+ occasions in the past 30 days.

3.3. Subgroup differences in frequency/consequence associations

Significant interactions at the p<.01 level were observed only for use frequency by college status. For any consequences, the frequency by not attending (vs. attending at a 4-year college) interaction was Est. −0.033 (SE 0.012); p=0.006, indicating use frequency had a stronger association with any consequences for those attending a 4-year college versus those not attending. Similarly, for performance/financial consequences, the frequency by not attending (vs. attending at a 4-year college) interaction was Est. −0.032 (SE 0.012); p=0.009, again indicating that use frequency had a stronger association with performance/financial consequences among those attending a 4-year college than those not attending.

3.4. Subgroup consequence associations controlling for use frequency

Controlling for use frequency and age, subgroup differences at the p<.01 level were observed only for performance/financial and relational consequences (Table 4). Performance/financial consequences were more likely among men than women. Relational consequences differed by college status, living situation, employment, and race/ethnicity. Relational consequences were more likely among both those not attending school and 2-year/vo-tech students (vs. 4-year college students); those living with parents (vs. not); those employed full-time (vs. not employed), and Hispanic (vs. White) respondents.

Table 4.

Associations Between Sociodemographic Characteristics and Perceived Negative Consequences Controlling for Use Frequency and Age Among U.S. Past 12-month Marijuana Users Age 19–22.

	% (95% CI)	AOR (95% CI p	% (95% CI)	AOR (95% CI) p	% (95% CI)	AOR (95% CI) p
	Any Consequence Endorsement		Emotional/ Physical Consequences		Performance/ Financial Consequences
College attendance
Not enrolled	40.6 (35.3, 46.0)	0.98 (0.72, 1.33)	30.6 (25.6, 35.6)	0.87 (0.63, 1.20)	13.2 (9.9, 16.6)	0.76 (0.48, 1.19)
		0.878		0.396		0.226
2-year/votech	44.7 (37.7, 51.7)	1.27 (0.91, 1.78)	35.9 (29.0, 42.9)	1.26 (0.88, 1.80)	12.5 (8.1, 16.8)	0.80 (0.49, 1.31)
		0.163		0.200		0.377
4-year	37.4 (33.2, 41.6)	(ref)	30.0 (26.1, 33.9)	(ref)	12.9 (10.0, 15.7)	(ref)
Living situation
Not with parents	37.8 (34.1, 41.5)	(ref)	31.2 (27.6, 34.7)	(ref)	11.7 (9.4, 14.0)	(ref)
With parents	42.2 (37.4, 47.0)	1.10 (0.84, 1.43)	30.6 (26.1, 35.0)	0.90 (0.68, 1.18)	13.7 (10.5, 16.8)	0.98 (0.67, 1.43)
		0.502		0.451		0.912
Employment
Not employed	35.6 (30.7, 40.6)	0.70 (0.48, 1.03)	28.4 (23.8, 33.0)	0.82 (0.55, 1.22)	13.1 (9.9, 16.3)	1.35 (0.80, 2.25)
		0.068		0.319		0.260
Part-time	41.6 (37.3, 46.0)	1.00 (0.70, 1.41)	32.2 (28.1, 36.3)	1.04 (0.71, 1.51)	13.3 (10.4, 16.3)	1.56 (0.95, 2.57)
		0.979		0.854		0.078
Employed full-time	45.1 (37.9, 52.2)	(ref)	34.1 (27.1, 41.1)	(ref)	11.4 (7.3, 15.5)	(ref)
Sex
Female	40.4 (36.5, 44.4)	(ref)	33.4 (29.6, 37.1)	(ref)	9.5 (7.2, 11.8)	(ref)
Male	39.3 (34.7, 43.8)	0.84 (0.65, 1.10)	28.8 (24.5, 33.1)	0.72 (0.54, 0.95)	16.2 (13.1, 19.3)	1.63 (1.13, 2.35)
		0.197		0.022		0.009
Race/ethnicity
Black	41.6 (29.6, 53.5)	1.04 (0.63, 1.73)	33.4 (22.2, 44.6)	0.99 (0.58, 1.69)	9.9 (2.9, 16.9)	0.64 (0.29, 1.44)
		0.873		0.983		0.283
Hispanic	43.8 (35.1, 52.6)	1.32 (0.88, 1.98)	27.1 (19.3, 34.9)	0.82 (0.53, 1.27)	14.6 (8.6, 20.7)	1.29 (0.75, 2.20)
		0.183		0.375		0.357
Other	39.8 (30.5, 49.1)	1.03 (0.66, 1.61)	29.5 (20.3, 38.6)	0.86 (0.53, 1.42)	11.5 (6.2, 16.8)	0.82 (0.43, 1.53)
		0.907		0.566		0.526
White	38.9 (35.6, 42.3)	(ref)	32.2 (29.0, 35.4)	(ref)	13.2 (11.0, 15.4)	(ref)
	Relational Consequences		Regret		Drive Unsafely
College attendance
Not enrolled	15.2 (11.5, 18.9)	1.84 (1.21, 2.79)	7.5 (4.8, 10.1)	0.90 (0.56, 1.44)	3.8 (1.9, 5.6)	0.96 (0.50, 1.87)
		0.004		0.655		0.913
2-year/votech	16.0 (11.2, 20.8)	2.00 (1.27, 3.15)	12.0 (6.5, 17.6)	1.53 (0.85, 2.74)	6.5 (1.5, 11.5)	1.54 (0.64, 3.75)
		0.003		0.159		0.335
4-year	8.3 (6.2, 10.3)	(ref)	8.1 (6.0, 10.2)	(ref)	3.8 (2.4, 5.3)	(ref)
Living situation
Not with parents	9.7 (7.6, 11.8)	(ref)	8.1 (6.1, 10.1)	(ref)	3.5 (2.3, 4.7)	(ref)
With parents	15.9 (12.6, 19.2)	1.64 (1.16, 2.32)	9.0 (6.1, 11.9)	1.09 (0.70, 1.68)	5.3 (2.7, 7.9)	1.34 (0.74, 2.45)
		0.005		0.705		0.335
Employment
Not employed	9.5 (6.9, 12.2)	0.49 (0.30, 0.81)	8.2 (5.8, 10.5)	0.69 (0.36, 1.29)	3.5 (2.1, 5.0)	0.48 (0.20, 1.16)
		0.005		0.243		0.104
Part-time	12.0 (9.3, 14.7)	0.68 (0.42, 1.08)	7.6 (5.5, 9.7)	0.65 (0.35, 1.19)	3.8 (2.2, 5.4)	0.57 (0.24, 1.37)
		0.105		0.161		0.209
Employed full-time	17.9 (12.4, 23.3)	(ref)	11.4 (6.0, 16.8)	(ref)	6.7 (2.0, 11.4)
Sex
Female	11.2 (8.6, 13.8)	(ref)	8.7 (6.5, 10.9)	(ref)	4.3 (2.8, 5.9)	(ref)
Male	13.4 (10.6, 16.2)	1.14 (0.79, 1.65)	8.3 (5.8, 10.9)	0.94 (0.61, 1.44)	4.1 (2.1, 6.2)	0.90 (0.48, 1.71)
		0.469		0.768		0.758
Race/ethnicity
Black	5.3 (0.7, 9.8)	0.39 (0.15, 1.01)	10.5 (2.2, 18.8)	1.22 (0.49, 3.05)	6.1 (0.0, 13.3)	1.51 (0.42, 5.45)
		0.053		0.664		0.531
Hispanic	20.0 (13.1, 27.0)	1.94 (1.18, 3.18)	9.1 (4.4, 13.9)	1.06 (0.57, 1.96)	3.2 (0.3, 6.0)	0.74 (0.28, 2.00)
		0.009		0.848		0.557
Other	11.2 (5.9, 16.5)	0.91 (0.52, 1.60)	5.7 (2.3, 9.2)	0.64 (0.33, 1.25)	3.9 (0.9, 6.8)	0.92 (0.39, 2.16)
		0.745		0.194		0.849
White	12.1 (9.9, 14.2)	(ref)	8.7 (6.8, 10.5)	(ref)	4.2 (2.9, 5.5)	(ref)

Notes: Unweighted n = 1,546. AOR = adjusted odds ratio from multivariable models simultaneously controlling for continuous past 30-day use frequency, noted sociodemographic characteristic, and age. Separate models run for each sociodemographic characteristic. Bold font indicates associations with p < .01 or stronger.

4. Discussion

In this national sample of young adults reporting past 12-month marijuana use, 39.8% of all users—and 60.5% of frequent users—reported perceiving negative consequences from their use. Perceived negative use consequences were not only a function of use frequency, but also reflected frequency by subgroup interactions, and subgroup-specific differences controlling for use frequency. Efforts to reduce harms associated with young adult marijuana use may be strengthened by developing prevention and intervention messaging addressing types of negative consequences users perceive from their own use, ensuring that such efforts reach across subgroups and address subgroup differences in perceived consequence likelihood.

Research has called for public education aimed at dispelling perceptions that marijuana use is without risk.⁵⁸ During the years included in this study (2008-2019), perceived risk of harm from regular marijuana use among U.S. young adults aged 19-22 decreased from 51.6% to 23.5%.¹⁷ Meanwhile, the percentage of 19-28 year-olds reporting past 12-month use increased from 28.6% to 40.1%.¹⁷ Ongoing state legalization of adult recreational marijuana use has been associated with reduced price and increased availability and potency.⁵ The percentage of the U.S. young adult population using marijuana—and experiencing negative consequences from use—will likely increase. Joining prior studies,^11,12,13,22 this study indicates young adults perceive negative outcomes associated with their marijuana use across daily life domains, especially emotional/physical consequences, performance/financial consequences, and relational consequences. In particular, the majority of frequent users acknowledge negative consequences of their use. Accurate, relevant broad-based marijuana prevention messaging providing data on negative consequences reported by young adult users may be helpful across adolescent population groups to prevent initiation or continuation, and among young adults to prevent progression to higher frequency use, combined with access to screening and brief intervention services found to help reduce marijuana use and consequences.⁵⁹

This study showed the likelihood of many—but not all—consequence types increased as use frequency increased. The strongest frequency/consequence associations were observed for performance/financial consequences, followed by emotional/physical and relational consequences. Results may help inform mass-media prevention efforts,⁶⁰ as well as development of effective brief intervention methods for marijuana use—an area of urgently needed research⁶¹—to help individual users address discrepancies between experienced consequences and perceived risk²⁰ and to employ effective behavioral strategies to reduce use and consequences.³⁷ Use frequency was not associated with some specific consequences (e.g., feel bad [depressed, anxious, ashamed]; less stable emotionally), indicating that both frequent and non-frequent users reported an equal likelihood of these consequences.

Use frequency was not associated with regretted behavior or unsafe driving, categories reported by the lowest percentages of young adult users. Items addressing behavioral regret are present on other marijuana consequences measures,^13,22 but exactly what is being measured is unclear. In the current study and prior studies including similar measures, relatively low percentages of users reported this item.^13,22 In contrast, research on alcohol consequences has found regretted behavior to be the most frequently reported negative consequence category.⁴³ Regarding unsafe driving, the measure read “caused you to drive unsafely.” Prior research using a similarly-worded item, “had your driving affected after using marijuana,” also had low prevalence in a Seattle-based young adult community sample (6.9%).¹³ However, the item, “I have driven a car when I was high” obtained high prevalence in a Midwest and Northeast college student sample (58.2%).²² The current study indicates frequent users do not believe their use causes them to drive unsafely more than non-frequent users. Whether or not this is true and whether frequent use is associated with actual driving impairment or the frequency of driving under the influence should be carefully examined in future research.

The current study indicated the impact of increasing frequency on the likelihood of reporting negative consequences was generally similar across subgroups. There was one key exception: higher frequency was more likely to result in any consequences, as well as performance/financial consequences among young adults attending 4-year colleges versus those not attending college. This finding may be associated with lower treatment usage among college-attending young adults,³⁸ resulting in high-frequency users being less likely to utilize available help to address the consequences of harmful use. There may be an increased need for 4-year college campuses to reach out to students who use marijuana frequently to help them reduce use and successfully complete academic and other goals.

After controlling for use frequency, consequence prevalence varied across subgroups. Relational consequences evidenced the highest number of subgroup differences; the finding is particularly interesting given the results of Kilmer et al.²⁰ who found experiencing social consequences increased the likelihood of anticipating future social consequences from use. The current study found perceived relational consequences were particularly likely for those not attending school and 2-year/vo-tech students (vs. 4-year college students); those living with parents (vs. not); those employed full-time; and Hispanic (vs. White) respondents. Administrators and other decision makers at 2-year/vo-tech schools should be aware of an increased likelihood of such consequences among their students in comparison with 4-year institutions; intervention efforts are needed on 2-year college campuses. The racial/ethnic differences may at least partially reflect racial/ethnic differences in the perceived acceptability of marijuana, which is lower among Hispanic than White respondents;^62,63 use may result in more interpersonal conflict for Hispanic than White young adults. Further, as Ash-Houchen and Lo⁶⁴ observed, family conflict is associated with higher risk of marijuana use for Hispanic than White young adults. Differences observed by college status and employment both point to higher perceived negative relational consequences among young adults in more “permanent” life roles (those not attending 4-year colleges or employed full-time), as well as those in day-to-day relationships that would be likely to experience higher conflict regarding use (long-term relationships, living closely with parents) if opinions on use differed.

4.1. Limitations

These findings are subject to limitations. All data were self-report; exactly what perceived negative consequences indicate requires further study, because the reported consequences may or may not reflect actual consequences of use. For example, there is a literature that suggests connections between marijuana use and some outcomes, such as mental health and cognitive ability, may be the result of shared confounds rather than direct causal linkages.^65,66 The sample was drawn from 12th graders; dropouts were not included. Dropping out is associated with higher marijuana use.⁶⁷ Differential attrition based on substance use likely means the marijuana consequence estimates presented here are somewhat underreported; use of attrition weights helped address this limitation. Sample size limitations prevented further differentiating college attendance (e.g., differences between community colleges and vocational/technical schools) or living situation (e.g., differences between living on campus in fraternities/sororities versus living on campus outside of Greek systems). Limitations notwithstanding, this study provides new information about the prevalence of perceived negative marijuana use consequences among young adults in and outside of college settings.

5. Conclusion

Young adults aged 19-22 reported a wide range of perceived negative marijuana consequences. Higher use frequency was associated with higher odds of some but not all perceived consequence types; the impact of higher frequency use was particularly strong among 4-year college students for any consequences and performance/financial consequences. Even after controlling for use frequency, performance/financial consequences were reported more often by men than women; relationship consequences were reported more often by non-college young adults, those living with parents, and those working fulltime. Efforts to reduce negative marijuana consequences may be strengthened by recognizing and addressing the different types of negative consequences perceived by users.

Highlights.

• 60.5% of frequent users perceived negative consequences from their marijuana use.

• Emotional/physical consequences were most frequently reported by all users (31.1%).

• Perceptions of causing unsafe driving were least frequently reported (4.2%).

• Use frequency/consequence associations were strongest for 4-year college students.

• Relational consequences were highest among 2-year college/vo-tech and non-students.