An examination of the acquired preparedness model among college student marijuana users

Abstract

Objective:

To examine the Acquired Preparedness Model using a behavioral impulsivity facet and positive marijuana expectancies to examine direct and indirect effects on marijuana use and related problems.

Participants:

250 college students (61.7% female, 54% white) recruited from a southeastern university.

Methods:

Participants completed an online survey of delay reward discounting, marijuana expectancies, consideration of future consequences, and marijuana-related outcomes.

Results:

Delay reward discounting and consideration of future consequences related to marijuana-related problems, but not marijuana use. However, positive marijuana expectancies did not mediate the relation between impulsivity and marijuana outcomes.

Conclusions:

These results emphasize delay reward discounting and consideration of future consequences as important factors associated with marijuana-related problems. Interventions aimed at decreasing delay reward discounting and augmenting future orientation may be effective in college students who report light to moderate marijuana use. Future studies would benefit from longitudinal study designs using multiple impulsivity measures among light and heavy users.

Introduction

Marijuana is the most widely used drug in the United States,¹ particularly among young adults.² Recent estimates from the National Survey on Drug Use and Health (NSDUH) suggests that approximately 7.5 million young adults have used marijuana in the past month and 1.8 million have met criteria for a marijuana use disorder.¹ Among young adult college students in particular, past-month marijuana use is common³ and the prevalence of past-year marijuana use has reached the highest point in the past 30 years at 41.3%.⁴ This may be due, in part, to changing attitudes in the perceived risk of marijuana use^5–7 as well as increased support for, and the legalization of, recreational and medical marijuana in many states within the U.S.^8–10 Further, evidence suggests that marijuana use is associated with increased risk for adverse mental health outcomes such as depression and other psychiatric conditions,¹¹ impaired driving and vehicular crashes,^12,13 and decreased academic performance.¹⁴ Thus, in order to help inform prevention and intervention strategies, it is vital to understand factors associated with marijuana use and related problems.

Acquired preparedness model (APM)

The Acquired Preparedness Model (APM)¹⁵ integrates aspects of impulsivity and psychosocial learning to explain high-risk patterns of substance use. In particular, the APM posits that impulsive individuals respond to and learn more from rewarding experiences compared to less rewarding and/or aversive experiences.¹⁵ By this model, impulsive individuals are more likely to form stronger positive, and lesser negative, alcohol expectancies (i.e., expectations of the reinforcing effects due to drinking)¹⁶ compared to less impulsive individuals. In other words, more impulsive individuals have a tendency to form more positive, rewarding expectations regarding substance use, which in turn leads to subsequent use. Indeed, prior cross-sectional and prospective studies found that positive alcohol expectancies mediated the relation between facets of impulsivity and alcohol use,^17–20 such that greater levels of impulsivity predicted heightened positive alcohol expectancies, which in turn predicted alcohol-related outcomes (e.g., alcohol use, alcohol-related problems).

The majority of studies testing the APM have focused on alcohol outcomes, however an emerging literature has attempted to apply the APM to other addictive behaviors such as cigarette smoking²¹ and marijuana use.^22–27 Of the studies examining marijuana use, three found that positive expectancies (e.g., social facilitation, perceptual enhancement) partially or fully mediated the relation between different generalized impulsivity measures and marijuana outcomes (e.g., marijuana use, marijuana use disorder symptoms).^22–24 These three studies included samples consisting of young adult twins with and without lifetime substance use disorder,²³ young adult females,²⁴ and adolescents and adults recruited from a marijuana policy listserv.²²

Studies examining the mediational role of negative expectancies have produced mixed results.^24,27 One study found that negative expectancies mediated the relation between generalized impulsivity and past-year marijuana use among college students, such that greater impulsivity was related to less negative expectancies, which was related to greater use.²⁷ Conversely, Hayaki and colleagues (2011)²⁴ found that more impulsive young adult females endorsed greater negative expectancies, which were robustly associated with marijuana problem severity and marijuana dependence, but less marijuana use. Equivocal findings between these two studies may be due the heterogeneous study populations and the assessment of different marijuana outcomes. Vangsness and colleagues (2005)²⁷ analyzed data from male and female undergraduate students, some of whom were marijuana naïve whereas Hayaki and colleagues (2011)²⁴ analyzed data from young adult females engaged in motivational interviewing to reduce marijuana use. As such, it may be that more impulsive individuals in the latter study²⁴ endorsed more negative marijuana expectancies due to frequent and/or recurring marijuana problems. On the other hand, perhaps psychosocial risk factors for marijuana use operate differently for lighter and heavier users, which is an empirical question requiring additional research. Further, Vangsness et al. (2005)²⁷ assessed marijuana use but not problems, making comparisons across studies harder to generalize. Thus, disparate findings linking the APM to different marijuana outcomes warrant additional research among young adult college students with moderate marijuana use patterns and will address a gap in the literature.

While the aforementioned studies provide some support for the APM and various marijuana outcomes, the majority of this work has been conducted among young adults from the community with heavy marijuana use (i.e., meeting substance use disorder criteria)^22,24 or included a large proportion of participants who did not use marijuana at all.^26,27 As rates of marijuana use initiation²⁸ and overall frequency⁴ occurring on college campuses are increasing, it is necessary to test the APM among college students who engage in moderate marijuana use patterns. If the proposed relations of the APM are observable among lighter marijuana users, it would suggest that impulsive individuals form strong positive expectancies relatively early after marijuana initiation that may strengthen as use patterns increase, which is consistent with prior studies of individuals with heavier marijuana use patterns. College students often report marijuana-related problems,²⁹ yet few seek out or receive formal treatment,³⁰ further compounding the need for additional research among this population.

Relevance of behavioral impulsivity and related constructs

Additionally, prior studies testing the APM for marijuana^22–27 have focused exclusively on generalized (i.e., trait) measures of impulsivity, while other facets of impulsivity and related constructs are under addressed. Impulsivity is a heterogeneous construct that contains generalized impulsivity measured via self-report, poor inhibitory control (i.e., behavioral response impulsivity), and delay reward discounting (i.e., behavioral choice impulsivity),^31,32 which broadly refers to a strong preference for immediate reward, coupled with rash action and a diminished regard for future consequences.^33–35 While some studies have found that behavioral response impulsivity relates to substance use outcomes (e.g., Corbin et al., 2020),³⁶ generalized impulsivity assessed via self-report is more reliably related to actual substance use behaviors.³¹

Delay reward discounting (i.e., inordinate preference for smaller, sooner relative to larger, later rewards)³⁷ is typically viewed as a form of behavioral choice impulsivity.^31,38 However, given delay reward discounting’s emphasis on immediate reward, perhaps individuals with greater delay reward discounting also learn more of the rewarding aspects of marijuana use, thereby increasing the strength of one’s positive marijuana expectancies even among lighter users. One study of young adults found that the test-retest reliability on a delay reward discounting task (r = .88) was similar to that of generalized impulsivity on the Barratt Impulsiveness Scale³⁹ (r = .92) and the UPPS-P (rs ranging from .81 to .93),⁴⁰ thus lending further support that, like generalized impulsivity, rates of delay reward discounting are stable over time.⁴¹ Moreover, like generalized facets of impulsivity, greater delay reward discounting is related to marijuana dependence symptomology, alcohol use and alcohol use disorder risk, and other drug use.^33,42,43

Relatedly, one study found that among heavy drinkers, greater delay reward discounting on the Monetary Choice Questionnaire⁴³ was associated with greater positive sexual-related alcohol expectancies (e.g., beliefs about alcohol’s ability to increase sexual performance and pleasure).⁴⁴ Moreover, Celio and colleagues (2016)⁴⁴ found that delay reward discounting moderated the relation between sexual-related alcohol expectancies and percentage of past 90-day alcohol-related condomless sex. Thus, there is reason to believe that delay reward discounting is related to more positive expectancies across other domains (e.g., marijuana use), and may be an important dimension of impulsivity that that has often been unaddressed when testing the APM.

Similar to delay reward discounting, consideration of future consequences (CFC)⁴⁵ refers to the extent that individuals contemplate immediate versus distal consequences of current behaviors. Impulsivity is often reflective of a proclivity for reward coupled with rash action and a diminished regard for future consequences.^33–35 Thus, impulsivity is variously characterized in different ways and shares conceptual similarities with CFC, yet only a few empirical studies have attempted to link CFC and impulsivity.^46,47 Likewise, CFC assesses one’s belief about how current behaviors influence more distal outcomes.⁴⁵ Higher levels of CFC are associated with less alcohol use⁴⁸ and lower levels of impulsivity.⁴⁶ Thus, in line with the APM, it is plausible that greater CFC is inversely related to positive marijuana expectancies given the emphasis on distal, rather than immediate, consequences of behaviors.

Purpose of present study

The current study extends the literature on the APM by examining marijuana use and related problems in a sample of young adult college students that endorsed light to moderate marijuana use. College students are an important population to study given they are at increased risk to initiate marijuana use,²⁸ are often more impulsive,⁴⁹ and are less likely to seek treatment³⁰ than their non-college attending peers. Moreover, young adulthood is a salient period where substance use either decreases or when higher-risk patterns of use continues.⁵⁰ The main aim of this study was to examine the APM by investigating the relations among a facet of impulsivity (i.e., delay reward discounting), consideration of future consequences, marijuana expectancies, and marijuana outcomes. Similar to prior studies,^22,24 we hypothesized that greater delay reward discounting and reduced consideration of future consequences would relate to greater positive expectancies, which would in turn relate to greater marijuana use and related problems.

Method

Participants and procedures

Participants were undergraduate students from a large Southeastern University recruited through flyers, online advertisements, and university-wide email solicitation. The current sample included 250 students (61.7% Female; 54% Caucasian) who were between the ages of 18 and 26 years old, reported using marijuana at least three times in the past month, and provided complete data on all outcome variables (i.e., marijuana use, expectancies, problems). Following informed consent procedures, eligible participants completed an online survey that took approximately 60 minutes to complete. Participants were compensated $20 for their participation. All procedures were approved by the University’s Institutional Review Board (IRB).

Measures

Demographic characteristics

Demographic variables included age, sex, and race.

Marijuana initiation, duration of regular use, and current use

Age at first marijuana use was assessed using a single item (i.e., “About how old were you when you first tried marijuana?”). Duration of regular use was assessed using a single item (i.e., “About how long have you been using marijuana regularly (at least once a month)?”) with response options ranging from 1 to 3 months to 20+ years. Past-month marijuana use frequency was assessed using an open-response single item (i.e., “In the past month, on how many days did you use marijuana?”).

Delay reward discounting

The Delay Reward Discounting Task (DRD)⁵¹ is an 8-item measure of behavioral choice impulsivity. Participants made eight binary choices between a smaller monetary reward ($10 to $99) available “today” or a larger later reward ($100) available after four delays (2 weeks, 1 month, 6 months and 1 year). Similar to Acuff, MacKillop, & Murphy (2018),⁵² the ratio between immediate to delayed responses was utilized, such that higher values denote greater delay reward discounting (i.e., greater behavioral impulsivity). Internal consistency was acceptable (Cronbach’s α = .75).

Consideration of future consequences

The Consideration of Future Consequences Scale (CFC)⁴⁵ is a 12-item measure that examines the degree to which one considers the proximal versus distal consequences of their current behaviors (e.g., “I consider how things might be in the future, and try to influence those things with my day to day behavior”). Participants indicated whether each of the 12 statements were characteristic of them on a 5-point Likert type scale from “Extremely uncharacteristic” to “Extremely characteristic”. Seven items were reverse scored (e.g., “I only act to satisfy immediate concerns, figuring the future will take care of itself”) and summed with the remaining five items to create a total score. Higher scores indicate a greater consideration of future consequences. Internal consistency was acceptable (Cronbach’s α = .80).

Marijuana problems

Past-month marijuana-related problems were assessed using a 25-item version of the Rutgers Marijuana Problem Index (RMPI).⁵³ Participants indicated how often they experienced different consequences associated with their marijuana use (e.g., failure to meet responsibilities in work or school, jeopardizing relationships with friends or family) on a 5-point Likert type scale ranging from “Never” to “More than 10 times”. Sum scores were used for analyses. Internal consistency was excellent (Cronbach’s α = .93).

Marijuana expectancies

Marijuana expectancies were measured using the Marijuana Effect Expectancy Questionnaire – Brief (MEEQ-B).⁵⁴ The MEEQ-B was used to assess the extent participants agreed or disagreed with six statements regarding possible positive or negative marijuana effects (e.g., “Marijuana has effects on a person’s body and gives a person cravings [get the munchies/hungry; have a dry mouth; hard to stop laughing]”) on a 5-point Likert type scale from “Strongly Disagree” to “Strongly Agree”. We used four items from the MEEQ-B based upon exploratory factor analysis (described below), which provided sufficient internal consistency (Cronbach’s α = .75).

Depression, anxiety, and stress

Depression, anxiety and stress were assessed using the 21-item Depression, Anxiety, and Stress Scale (DASS-21).⁵⁵ Participants indicated whether each statement applied to them over the past week on a 4-point scale ranging from “Did not apply to me at all” to “Applied to me very much, or most of the time”. The total score was calculated as the sum of all items multiplied by two. Internal consistency was excellent (Cronbach’s α = .94).

Data analytic plan

Descriptive statistics for all variables of interest were analyzed using SPSS version 23. When necessary, extreme values on any independent variable (i.e., z score greater than 3.29) were recoded to one unit higher than the next highest value.⁵⁶

Given that the MEEQ-B subscales have shown poor internal consistency in prior research assessing marijuana expectancies among college students⁵⁷ and because some items are double-barreled (e.g., “Marijuana has effects on a person’s body and gives a person cravings [get the munchies/hungry; have a dry mouth; hard to stop laughing]”), we examined the factor structure of the six MEEQ-B items using exploratory factor analysis. Due to concerns of basing factor structure solely on eigenvalues greater than one,⁵⁸ we also examined the scree plot⁵⁹ to determine the number of factors to retain in the analyses.

Next, the hypothesized path model was estimated using Mplus version 7.4.⁶⁰ We used maximum-likelihood (ML) estimation with bootstrapping and missing data adjustment for independent variables^60–62 in Mplus to avoid potential biases resulting from non-normality for parameter estimates in the model and the estimates of potential mediated effects obtained by the product of path coefficients (See p.5).⁶³ Acceptable model fit was supported by a root mean square error of approximation (RMSEA) value less than or equal to 0.08, Tucker-Lewis Index (TLI) and comparative fit index (CFI) value greater than or equal to 0.95, and standardized root mean square residual (SRMR) value less than or equal to 0.05.⁶⁴ Bivariate correlations, means, and standard deviations of the main study variables and demographics are reported in Table 1.

Table 1.

Sample Descriptives.

N = 250
Age, in years, M (SD), range	19.93 (1.41), 18–25
Caucasian, %	54 %
Female, %	61.7 %
Age of marijuana initiation, years	16.46 (1.80), 12–21
Duration of regular marijuana usea	3.77 (1.84), 1–8
Past-month marijuana use, in days	10.59 (8.13), 3–31
MEEQ-B Positive Expectanciesb	4.56 (1.01), 1–6
CFCS: Total Scorec	42.83 (7.39), 21–59
Delay Reward Discounting (DRD)d	.39 (.23), 0–1
Marijuana Problemse	10.81 (11.31), 0–52
DASS: Total Scoref	31.17 (26.25), 0–113

^aDuration of regular marijuana use M score equivalent to approx. 9–12 months, full possible range is 1–12.

^bMarijuana Expectancy Effect Questionnaire – Brief.

^cConsideration of Future Consequences Scale full range is 12–60.

^dReflects the number of immediate choices divided by total number of items.

^eRutgers Marijuana Problems Index full range is 0–100.

^fDepression, Anxiety, Stress Scale full range is 0–126.

Based on mediation analysis involving sequential multiple mediators,⁶⁵ consideration of future consequences (X₁) and delay reward discounting (X₂) were modeled to predict marijuana problems (Y₁) through positive marijuana expectancies (M₁) and past month marijuana use (M₂). In addition, direct relations between the exogenous variables (X₁ and X₂) and marijuana problems (Y₁) were estimated. Similar to prior studies, sex, age of marijuana initiation, and duration of regular marijuana use were included as covariates in estimating relationships among the main study variables.²⁴ Further, given prior research linking depression, marijuana use and related problems,^66,67 DASS-21 total scores were also included as a covariate. All exogenous variables were allowed to correlate with each other. Potential mediated effects were tested using the bootstrap method with 1,000 resamples, along with MODEL IND implemented in Mplus. Potential mediated effects of interest are shown in Figure 1 and included all possible pathways. Standardized 95% confidence intervals (CIs) around the estimates were examined, with CIs that do not include zero indicating significant indirect effects.^65,68,69

Figure 1.

Model for impulsivity and consideration of future consequences on past-month marijuana use and marijuana problems. Notes: Standardized point estimates were reported; Significant direct effects are reported with solid arrows; *p < 0.05, **p < 0.01, ***p < 0.001. Covariates include: Marijuana Initiation; Duration of regular marijuana use; sex; Depression, Anxiety, Stress Scale (DASS). All covariates were allowed to correlate with each other. DASS did not predict Use outcomes. Variables included in model: Future Consequences = Consideration of Future Consequences Scale (CFCS); Behavioral Choice Impulsivity = Delay Reward Discounting (DRD); Positive Expectancies = Marijuana Expectancy Effect Questionnaire – Brief (MEEQ-B); Marijuana Problems = Rutgers Marijuana Problems Index (RMPI); Marijuana Use (past month) = past month marijuana use. RMPI on sex=.109 (.053) p=.038; Past month use on duration of regular use=.298 (.060) p ≤ 0.001; RMPI on DASS=.333 (.054) p ≤ 0.001; all other covariate paths not significant (n.s.).

Due to the cross-sectional nature of the data, the current study was unable to account for temporal relations between exogenous and endogenous variables. As such, an alternative model was tested that included all variables from the hypothesized model, however in this alternative model positive marijuana expectancies (X₁) were modeled to predict marijuana problems (Y₁) through delay reward discounting (M₁), consideration of future consequences (M₂) and marijuana use (M₃).

Results

Descriptive statistics and correlation analyses

Sample descriptive statistics (n = 250) were analyzed (Table 1) and bivariate correlations were tested for possible issues with multicollinearity (Table 2). Four extreme values on the RMPI, one extreme value on the CFCS, and one extreme value on the DASS were recoded to unit above the greatest, non-extreme value.⁵⁶ Most participants self-reported marijuana initiation at approximately 16 years old, self-reported using marijuana regularly for less than 1 year, and generally endorsed positive marijuana expectancies. Consideration for future consequences and delay reward discounting were moderately, inversely correlated (r = −0.24, p< .01). Positive marijuana expectancies were moderately correlated with past month marijuana use (r = .222, p< .01).

Table 2.

Bivariate Correlations of study variables.

	1.	2.	3.	4.	5.	6.	7.	8.
1. Marijuana Initiation	-
2. Duration of regular marijuana use	−.415**	-
3. Past-month marijuana use	−.199**	.333**	-
4. MEEQ-B Positive Expectancies	−.075	.110	.222**	-
5. CFCS	.111	−.029	.049	.071	-
6. Delay Reward Discounting	−.026	−.011	.037	.031	−.239**	-
7. Marijuana Problems	−.143*	.155*	.336**	.111	−.318**	.216**	-
8. DASS-21	−.074	−.006	.048	.136*	−.212**	.141*	.399**	-

Notes: MEEQ-B = Marijuana Effect Expectancy Questionnaire – Brief; CFCS = Consideration of Future Consequences Scale; Marijuana Problems = Rutgers Marijuana Problems Index; DASS-21 = Depression, Anxiety, Stress Scale.

^*p < .05;

^**p < .01.

Factor analysis

Upon examination of the factor structure of the MEEQ-B, the scree plot suggested that the MEEQ-B contained one factor. Four of the six items loaded onto one factor (i.e., positive marijuana expectancies) with all factor loadings above .50 and an eigenvalue of 2.51. One item (“Marijuana generally has bad effects on a person [you become angry or careless; after feeling high you feel down]”) did not have a factor loading above .40 on either of the two factors, and the final item (“Marijuana makes it harder to think and do things [harder to concentrate or understand; slows you down when you move]”) had cross loadings of .36 and .89 on the first and second factor, respectively. Thus, we retained only the four items from the positive marijuana expectancies factor given that factors with less than three items are typically unstable.⁵⁸ The mean positive expectancy score was used in all analyses.

Path analysis

Hypothesized model: Delay reward discounting and consideration of future consequences → expectancies → use → problems

The hypothesized model (Figure 1) fit the data well (X²[5] = 7.344, p = 0.196, CFI= 0.983, TLI = 0.930, RMSEA = 0.043, 90% CI (0.000, 0.105); probability of RMSEA ≤ .05 = .493; SRMR=.024). Variance of the endogenous variables explained by their predictors can be found in Figure 1.

Direct effects

Direct effects in the full model are depicted in Figure 1. Consideration for future consequences (β = −0.231, SE= 0.49, p < 0.001) had a negative direct effect on marijuana problems. Conversely, delay reward discounting (β = 0.116, SE= 0.042, p = 0.005) had a positive direct effect on marijuana problems. Neither consideration for future consequences nor delay reward discounting were significantly related to past month marijuana use. Further, consideration for future consequences, delay reward discounting, and marijuana problems were not significantly related to positive expectancies. However, positive marijuana expectancies had a positive direct effect on past month marijuana use (β = 0.176, SE= 0.055, p = 0.003). Similarly, past month marijuana use had a positive direct effect on marijuana problems (β = 0.330, SE = 0.056, p < 0.001).

Given the moderate sample size and potential multicollinearity between exogenous and endogenous variables, a post hoc model was tested to verify the robustness of these effects. Thus, age of marijuana initiation, duration of regular marijuana use, and DASS-21 scores were removed as covariates in the model. However, removing these covariates did not change the results concerning impulsivity domains, positive marijuana expectancies, and marijuana outcomes (Supplemental Table 1).

Indirect effects on marijuana use and related problems

Tests of potential mediated effects are listed in Table 3 with total, direct, and indirect effects. Delay reward discounting had a positive total effect on marijuana problems, whereas consideration for future consequences had a negative direct effect on marijuana problems. Positive expectancies did not have any significant effects on our tested model.

Table 3.

Total, direct, and indirect effects of impulsivity and consideration of future consequences on marijuana outcomes via positive marijuana expectancies (MEEQ-Bp).

Independent variable:	Behavioral Choice Impulsivity (Delay Reward Discounting)			Consideration of Future Consequences
	B	LL	UL	B	LL	UL
Dependent variable: marijuana use
Total indirect	0.007	−0.016	0.033	0.014	−0.012	0.046
Specific indirect effects
Positive Expectancies (MEEQ-Bp)	0.007	−0.016	0.033	0.014	−0.012	0.046
Direct effect	0.038	−0.088	0.160	0.057	−0.075	0.172
Total effect	0.045	−0.083	0.164	0.071	−0.050	0.187
Dependent variable: marijuana problems
Total indirect	0.016	−0.029	0.060	0.024	−0.018	0.070
Specific indirect effects
Positive Expectancies (MEEQ-Bp)	0.001	−0.008	0.012	0.001	−0.013	0.016
Marijuana use (Past Month)	0.013	−0.029	0.053	0.019	−0.024	0.061
Positive Expectancies (MEEQ-Bp) → marijuana use	0.002	−0.005	0.012	0.005	−0.004	0.017
Direct effect	0.116	0.033	0.193	−0.231	−0.331	−0.138
Total effect	0.132	0.035	0.223	−0.207	−0.315	−0.104

Note: Parameter estimates and significance tests based on 1000 bootstrapped samples. Significant effects (p < 0.05) denoted in bold typeface.

Alternate model: Expectancies → delay reward discounting and consideration of future consequences → use → problems

Given the possibility of a bidirectional relation between impulsivity domains and positive marijuana expectancies, an alternative model was tested where impulsivity domains and marijuana use mediated the relation between positive marijuana expectancies and marijuana problems. However, this did not fit the data well (X²[8] = 26.958, p = 0.001, CFI= 0.879, TLI = 0.608, RMSEA = 0.097, 90% CI (0.058, 0.139); probability of RMSEA ≤ .05 = .025; SRMR=.05).

Direct effects

Like in the hypothesized model, consideration of future consequences (β = −0.238, SE= 0.050, p < 0.001) had a negative effect, and delay reward discounting (β = .122, SE= .043, p = 0.004) had a positive effect on marijuana problems. Neither consideration of future consequences nor delay reward discounting were significantly related to past month marijuana use (Table 4). Positive marijuana expectancies had a positive direct effect on past month marijuana use (β = .175, SE= 0.055, p = 0.001). Similarly, past month marijuana use had a positive effect on marijuana problems (β = 0.339, SE = 0.057, p < 0.001).

Table 4.

Total, direct, and indirect effects of positive marijuana expectancies (MEEQ-Bp) on marijuana outcomes via delay reward discounting and consideration of future consequences.

Independent variable:	Positive Marijuana Expectancies (MEEQ-Bp)
	B	LL	UL
Dependent variable: marijuana use
Total indirect	0.005	−0.010	0.024
Specific indirect effects
Consideration of Future Consequences	0.004	−0.009	0.021
Delay Reward Discounting	0.001	−0.009	0.013
Direct effect	0.175	0.066	0.283
Total effect	0.179	0.065	0.286
Dependent variable: marijuana problems
Total indirect	0.047	−0.008	0.102
Specific indirect effects
Consideration of Future Consequences	−0.016	−0.055	0.018
Delay Reward Discounting	0.002	−0.014	0.020
Marijuana use (Past Month)	0.059	0.023	0.104
Delay Reward Discounting → Marijuana Use	0.000	−0.003	0.004
Consideration of Future Consequences → Marijuana Use	0.001	−0.003	0.007
Direct effect	0.014	−0.100	0.127
Total effect	0.061	−0.071	0.191

Note: Parameter estimates and significance tests based on 1000 bootstrapped samples. Significant effects (p < 0.05) denoted in bold typeface.

Positive marijuana expectancies were not significantly associated with delay reward discounting (β = 0.021, SE = 0.067, p = 0.760) or consideration of future consequences (β = 0.068, SE= 0.073, p = 0.347).

Indirect effects on marijuana use and related problems

Tests of potential mediated effects are listed in Table 4 with total, direct, and indirect effects. Positive marijuana expectancies had a positive total effect on marijuana use but not problems. However, there was a specific indirect effect of positive marijuana expectancies on marijuana problems via marijuana use (β = 0.059, SE= 0.021, p = 0.006). There were no specific indirect effects of positive marijuana expectancies on marijuana use or problems via consideration of future consequences or delay reward discounting.

Discussion

The current study tested the Acquired Preparedness Model (APM)¹⁵ among a sample of young adult college students by examining the direct and indirect relations between delay reward discounting, consideration of future consequences, marijuana expectancies, marijuana use and related problems. Overall, our hypotheses were largely unsupported. Based on the hypothesized model, results suggested that delay reward discounting and consideration of future consequences were not related to greater positive expectancies. Interestingly, greater delay reward discounting and less consideration of future consequences were related to greater past-month marijuana problems but not use. Conversely, greater positive marijuana expectancies were related to greater past-month use but not problems. These results are similar to prior research in this area that found delay reward discounting was related to greater marijuana dependence symptomology but not marijuana use frequency,⁷⁰ and another study linking certain marijuana expectancies to use but not problems.⁷¹ Moreover, research has identified stronger relations between impulsivity domains and problems than with substance use.^72,73

Similarly, when an alternative model was tested to examine the indirect effects of positive marijuana expectancies on marijuana outcomes via impulsivity domains, impulsivity domains were not predicted significantly by marijuana expectancies. Further, there were no indirect effects of impulsivity domains on the relation between marijuana expectancies and marijuana outcomes. However, there was an indirect effect of marijuana expectancies on problems via past-month marijuana use, such that greater expectancies were related to past-month use, which in turn was related to more marijuana-related problems. Notably, the data did not fit this model well based on fit statistics. As such, this specific result should be interpreted with caution because, while statistically significant, the indirect effect may not be valid as it is unlikely to demonstrate accurate data generating processes. Alternatively, the hypothesized model provided better fit to the data, which provides some support that modeling the APM (i.e., impulsivity domains as predictors of marijuana expectancies, which in turn predict marijuana outcomes) is a more appropriate specification and should be explored further in future longitudinal studies.

Unlike prior studies examining the APM for marijuana use,^22–24 positive marijuana expectancies did not mediate the relationship between impulsivity and marijuana outcomes. One explanation for these disparate findings may be due to the study sample, specifically that previous studies examining the APM have utilized samples of adults with heavier marijuana use patterns. For example, one study found that greater expectations of positive marijuana effects (e.g., talkative, elated), fully mediated the relation between positive urgency (i.e., generalized impulsivity following intense positive emotional states⁷⁴) and marijuana use in a sample of older adults who were near daily users.²⁵ Further, participants in the study conducted by Bolles and colleagues (2014)²² reported using on average ¾ of an ounce of marijuana per month. Similarly, Hayaki and colleagues (2011)²⁴ found that participants reported marijuana use on average 57% of days in the past three months.

Conversely, participants in the current study were primarily light to moderate users that endorsed a shorter duration of regular use with a relatively later age of marijuana initiation compared to participants in prior studies.²⁴ For example, more than half of the current sample reported using marijuana two or less times per week. As such, it may be that while impulsive individuals have more positive reinforcing effects associated with substances (i.e., acquired preparedness), heavier and longer use is needed before their positive marijuana expectancies lead to subsequent use and problems. Indeed, one prior prospective study found that marijuana expectancies developed and become more pronounced as marijuana use escalated over time.⁷⁵

On the other hand, given the APM’s emphasis on personality characteristics and psychosocial learning, generalized measures of impulsivity like the Barratt Impulsiveness Scale (BIS-11)³⁹ or the UPPS-P⁷⁴ may be better predictors of marijuana expectancies than behavioral facets of impulsivity (i.e., delay reward discounting) or consideration of future consequences. Indeed, all prior studies testing the APM in relation to marijuana outcomes utilized a generalized measure of impulsivity.^22–27 However, consideration of future consequences is conceptually related to generalized impulsivity,⁴⁷ in particular non-planning and/or lack of premeditation (i.e., rash action with diminished regard for future consequences). One study of adolescents and young adults found higher consideration of future consequences to be negatively correlated with the non-planning subscale on the BIS-11 (r = −.41, p<.01).⁷⁶ Nevertheless, it is possible that consideration of future consequences is less relevant to marijuana expectancies given that the perceived risks of marijuana have been steadily decreasing.⁶ The current study also utilized a brief delay discounting measure, but other measures that include adjusting-amount procedures (e.g., Monetary Choice Questionnaire⁴³) may have provided better insight into potential relations between behavioral impulsivity and marijuana expectancies.

Regardless, results from this study have important clinical implications. Indicated prevention and intervention strategies may benefit by attempting to decrease delay reward discounting and increase consideration of future consequences as a way to reduce marijuana-related problems among young adults who have not escalated to more high-risk patterns of marijuana use. Using a behavioral economic approach, increasing the involvement in, and reinforcement from, substance-free alternatives is a viable way to reduce substance use and related problems,^77–80 and potentially increase consideration of future consequences.⁷⁹ Incorporating strategies from cognitive behavioral therapy or mindfulness-based intervention aimed at increasing self-regulation skills (e.g., urge surfing, acceptance vs. avoidance) may also be beneficial.^81–83 Further, there is some indication that certain behavioral (e.g., executive function training, episodic future thinking) and pharmacological (e.g., atomoxetine) interventions can reduce delay reward discounting.^77,84,85 While results from these studies typically produce short-term decreases in delay reward discounting, combining procedures that reduce delay reward discounting with strategies that increase substance-free reinforcement remain relatively unexplored. The current study highlights the importance of delay reward discounting and consideration of future consequences in relation to marijuana problems. Thus, interventions that target both simultaneously could be a viable strategy to reduce marijuana-related problems among young adult college students.

Limitations and future directions

Despite the various strengths of the current study, some limitations should be noted. As with all cross-sectional self-report data, reports are prone to recall bias and the temporal precedence between impulsivity domains, marijuana expectancies, and marijuana outcomes could not be determined. While this is the first study to our knowledge to test delay reward discounting under the APM framework, other facets of impulsivity were not tested such as generalized impulsivity measured using the Barratt Impulsiveness Scale,³⁹ which may help to explain non-significant results in the current study. Marijuana expectancies may also not have been optimally captured using the MEEQ-B⁵⁴ as some items likely represent an amalgam of negative and positive expectancies for some individuals (e.g. “marijuana has effects on a person’s body and gives a person cravings [get the munchies/hungry; have a dry mouth; hard to stop laughing”]). Specifically, the current study did not assess negative marijuana expectancies under the APM framework due to concerns regarding the factor structure of the MEEQ-B.⁵⁴ As such, this omission limited the ability to comprehensively test the APM by examining whether negative expectancies related to marijuana outcomes among more impulsive individuals. Additionally, the past-month marijuana use frequency item did not distinguish between individuals who use multiple times per day versus once a day, nor did it account for total quantity used. In the study conducted by Bolles and colleagues (2014),²² positive marijuana expectancies partially mediated the relation between generalized impulsivity and ounces of marijuana smoked in the past month. While a validated measure of marijuana use assessing both frequency and quantity of use (e.g., Daily Sessions, Frequency, Age of Onset, and Quantity of Cannabis Use Inventory⁸⁶) would have been optimal in the current study, prior research has also highlighted psychometric concerns regarding the measurement of marijuana quantity.^86,87 Lastly, this study was conducted in a state where recreational marijuana use is illegal, and it remains unclear whether these results are generalizable to college students in states where recreational marijuana use is legal. In light of these limitations, results from the current study add to the existing literature and suggest that greater delay reward discounting and lesser consideration of future consequences are related to more marijuana-related problems. Further, interventions should place an emphasis on increasing future oriented thinking among more impulsive individuals who experience marijuana-related problems whereas expectancy challenges may be more efficacious in reducing frequency of marijuana use.

Moving forward, studies would benefit from implementing longitudinal study designs to elucidate whether links between marijuana expectancies and facets of impulsivity develop earlier in adolescence, and if so, how these relations influence the initiation and maintenance of marijuana use over time. Moreover, it is vital to understand if, and to what extent, impulsivity relates to negative marijuana expectancies. Indeed, marijuana expectancies should be assessed using less ambiguous items to clarify the relationship between frequency of use and positive and/or negative expectancies. Further, future studies would also benefit by assessing marijuana use in more detail, such as asking about cannabinoid composition, route of administration, and number of times used in a given day. In addition, the APM should be tested among young adults enrolled in traditional four-year universities as well as community colleges and trade schools in states with different legal landscapes in order to help tailor interventions to curtail problematic marijuana use.

Conclusion

Findings from this study replicate prior work by demonstrating higher levels of impulsivity (as measured by delay reward discounting and consideration of future consequences) is related to marijuana related problems. Contrary to our hypothesis however, positive marijuana expectancies did not mediate this relationship. This lack of mediation may be due to this sample endorsing lighter marijuana use compared to prior work. Given the null findings on positive expectancies among our sample of light to moderate marijuana users, this study illuminates the importance of examining diverse samples of substance users. As we disentangle the complex relationships among risk-factors for marijuana use and problems, these results offer valuable insight for tailoring interventions to increase consideration for future consequences and reduce behavioral choice impulsivity.