Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Apr 1.
Published in final edited form as: Exp Clin Psychopharmacol. 2020 Oct 1;30(2):159–171. doi: 10.1037/pha0000397

Behavioral Economic Interactions Between Cannabis and Alcohol Purchasing: Associations With Disordered Use

Sean B Dolan 1, Tory R Spindle 1, Ryan Vandrey 1, Matthew W Johnson 1
PMCID: PMC8209692  NIHMSID: NIHMS1707005  PMID: 33001691

Abstract

As cannabis policy changes, there is an urgent need to understand interactions between cannabis and alcohol couse. An online sample of 711 adult past-month cannabis and alcohol users completed both single-item hypothetical purchasing tasks for cannabis and alcohol and cross-commodity purchasing tasks assessing adjusting-price cannabis with concurrently available, fixed-price alcohol, and vice versa. Participants provided information about cannabis and alcohol use patterns, and completed the Alcohol and Cannabis Use Disorder Identification Tests (AUDIT and CUDIT, respectively). Group data showed that cannabis and alcohol served as complements (as the price of the adjusting-price commodity increased, consumption of both commodities decreased). However, individual data showed substantial variability with nontrivial proportions showing patterns of complementarity, substitution, and independence. More negative slopes (greater complementarity) for fixed-price cannabis and alcohol were both associated with greater self-reported drug consumption and CUDIT and AUDIT scores. The negative relation between cross-price slope and CUDIT/AUDIT score indicates that individuals who treat cannabis and alcohol more as complements are more likely to experience disordered use. Based on these cross-commodity purchasing data, when both cannabis and alcohol are concurrently available at low prices, both may be used at high levels, whereas limiting consumption of one commodity (e.g., through increased price) may reduce consumption of the other. These data show the importance of examining individual participant analyses of behavioral economic drug interactions and suggest that manipulation of cost (e.g., through taxes) or cosale restrictions are potential public health regulatory mechanisms for reducing alcohol and cannabis use and couse behaviors.

Keywords: cannabis, alcohol, behavioral economics, cross-price elasticity, coconsumption


The 1996 legalization of medicinal cannabis in California initiated a wave of cannabis legislation reform across the United States and abroad. As of this writing, cannabis is legal for medicinal use in 33 U.S. states and many developed countries, and legalization of cannabis for nonmedicinal (“recreational”) use has occurred in 11 U.S. states, Canada, and Uruguay. As a cause and/or consequence of these policy changes, a trend toward increased cannabis use and dependence has been observed (Hasin et al., 2015; Hasin et al., 2017; Hasin, 2018), and perceptions of harm related to cannabis use have decreased (Carliner, Brown, Sarvet, & Hasin, 2017; Hughes, Lipari, & Williams, 2013).

One of the greatest public health concerns with the advent of widespread cannabis legalization relates to couse of cannabis and alcohol. Alcohol use alone is a substantial public health burden (Bonomo et al., 2019; Nutt, King, & Phillips, 2010; van Amsterdam, Opperhuizen, Koeter, & van den Brink, 2010), but studies indicate that couse of cannabis and alcohol may result in greater risks of harm than use of either substance alone (Substance Abuse and Mental Health Services Administration, 2018). In controlled laboratory studies, cannabis and alcohol couse has been shown to increase subjective intoxication (e.g., rating of “high”; Hartman et al., 2016), and additively impair driving performance (Hartman, Brown, Milavetz, Spurgin, Pierce, et al., 2015), compared to when these drugs are used alone (Hartman, Brown, Milavetz, Spurgin, Gorelick, et al., 2015). Additionally, epidemiological studies have indicated that individuals who use cannabis and alcohol concurrently are more likely to drive while intoxicated, binge drink, and experience a greater number of cannabis- and alcohol-related personal consequences (e.g., job-, health-, and social-related consequences; Arterberry, Treloar, & McCarthy, 2017; Gunn et al., 2018; Metrik, Gunn, Jackson, Sokolovsky, & Borsari, 2018; Subbaraman & Kerr, 2015). Although polysubstance use has been an important and growing research focus in recent years, there is still much to learn about cannabis and alcohol couse preferences and patterns and how couse relates to cannabis and alcohol use consequences.

Behavioral economic demand provides a framework for evaluating drug reinforcement and interactions by assessing consumption of one or more drugs across varying prices (Aston & Cassidy, 2019; Bickel, Johnson, Koffarnus, MacKillop, & Murphy, 2014). In demand analysis, consumption of a drug, whether real or hypothetical, is determined across a series of prices to determine effects on purchasing (or consumption), resulting in a demand curve. Regression analyses of demand curves can produce two reinforcement-related variables, demand intensity (purchasing at prices approaching zero) and demand elasticity (price-sensitivity), to provide a multidimensional assessment of drug reinforcement. In addition to evaluating demand for a single drug, cross-commodity tasks can be used in assessments of polysubstance use to evaluate the interactive effects of two concurrently available drugs, in which the price of one drug increases while another is kept fixed at a constant price. By analyzing the change in consumption of the fixed-price drug as the price of the other drug increases, it can be determined if each of the two drugs functions as a substitute (fixed-price drug consumption increases), complement (fixed-price drug consumption decreases), or independent commodity (no change in fixed-price drug consumption) in relation to the other drug (Bickel et al., 2014).

Epidemiological analyses have attempted to evaluate population-level economic changes in cannabis and alcohol consumption as a function of external influences, such as regulation and taxation, on cannabis and alcohol availability. These epidemiological analyses have yielded somewhat conflicting results regarding how the availability of cannabis influences alcohol use and vice versa. Behavioral economic analyses of epidemiological data sets have found evidence for substitution, complementarity, and independence between cannabis and alcohol across studies (Guttmannova et al., 2016; Subbaraman, 2016). One recent study evaluating adolescent drug and alcohol use demonstrated that adolescent binge drinking does not differ between U.S. states with or without medicinal cannabis laws, suggesting economic independence, but that legalization of medicinal cannabis use at the state level was followed by a significant decrease in binge drinking among eighth graders (but not 10th or 12th), suggesting substitution in this subgroup (Cerda et al., 2018). A second study also showed a significant decrease in adolescent alcohol consumption following legalization of medicinal cannabis at the state level, which also suggests a commodity substitution pattern (J. K. Johnson et al., 2018). One interpretation of these data is that the two drugs may function as substitutes in younger samples, but behaviors related to couse may change with age. A limitation of these analyses is that the determination that cannabis can substitute for alcohol among younger adolescents is contingent on the assumption that legalization of medicinal cannabis increased cannabis availability to adolescents, which may not have been the case. In Colorado, the association between binge drinking and cannabis use did not change when nonmedicinal (aka recreational) use of cannabis was legalized (Jones, Nicole Jones, & Peil, 2018). In contrast, in Oregon, the only significant increase in cannabis use following legislation of nonmedicinal cannabis use was among college students who binge drank (Kerr, Bae, Phibbs, & Kern, 2017). Thus, a clear economic relation between cannabis and alcohol is not evident from epidemiological studies. It is uncertain whether this is due to limitations with respect to assumptions about commodity access that result from legislative changes in these markets, or if there are differences in cannabis and alcohol use behavior between subpopulations or individuals that explain the divergent outcomes of these prior studies.

Due to the aforementioned limitations of analysis of epidemiological data sets to understand the behavioral economics of cannabis and alcohol couse behavior, research that the price and access of these drugs is needed to measure drug taking behavior at the individual participant level. Although demand assessment of laboratory-based drug self-administration provides the optimal means for determining drug reinforcement, ethical and logistical boundaries can be prohibitive for running these studies, in which case, evaluation of hypothetical drug consumption can serve as a viable substitute (Jacobs & Bickel, 1999). Hypothetical purchasing tasks, in which participants simply indicate how much of a drug they would purchase across various prices, have been used to evaluate drug reinforcement across numerous drug classes, and the demand metrics derived from analysis of purchasing patterns frequently relate to measures of drug use and dependence (Strickland, Campbell, Lile, & Stoops, 2020; Zvorsky et al., 2019). Cannabis and alcohol consumption have been extensively studied using hypothetical purchasing tasks (Aston, Metrik, Amlung, Kahler, & MacKillop, 2016; Aston, Metrik, & MacKillop, 2015; MacKillop & Murphy, 2007; Morris et al., 2017; Morris et al., 2018; Murphy & MacKillop, 2006; Ramirez, Cadigan, & Lee, 2019; Strickland, Alcorn, & Stoops, 2019; Strickland, Lile, & Stoops, 2017; Strickland, Lile, & Stoops, 2019); however, few studies have evaluated cross-commodity purchasing related to either cannabis or alcohol. In one such study, illicit cannabis was found to serve as a substitute for legal cannabis and vice versa (Amlung et al., 2019), and, in another, cannabis functioned as an independent reinforcer for concurrently available cigarette puffs (Peters, Rosenberry, Schauer, O’Grady, & Johnson, 2017). Despite the widespread couse of cannabis and alcohol, no studies to date have used hypothetical purchasing tasks to evaluate the behavioral economics of concurrently available cannabis and alcohol. The goal of this study was to measure demand for concurrently available cannabis and alcohol to provide novel information related to how individuals make decisions related to use of both drugs, and how these purchasing patterns relate to self-reported actual cannabis and alcohol use patterns and associated behavioral health consequences.

Method

Participants

Participants were recruited through the online crowdsourcing platform Amazon Mechanical Turk (MTurk). Individuals were eligible for the study if they met the following criteria: ≥21 years of age, used cannabis in the past 30 days, used alcohol in the past 30 days, indicated that inhalation was their primary cannabis consumption method, purchased cannabis in the past 30 days, had ever purchased cannabis plant material (to ensure translatability to the plant-based cannabis purchasing task), resided in the United States, were fluent in reading and writing English, had >95% MTurk approval rating, and had completed at least 100 approved MTurk Human Intelligence Tasks (consistent with criteria outlined in Strickland & Stoops, 2019). We included individuals with a preference for cannabis inhalation because this is the most popular method of administration (Borodovsky, Crosier, Lee, Sargent, & Budney, 2016; Doggett, Battista, & Leatherdale, 2020; Knapp et al., 2019), meaning these findings would translate to the vast majority of cannabis users. Eligibility was determined through a prescreening survey, and participants who met the inclusion criteria were provided a link to the full survey where they were provided with an overview of the study and provided consent. Participants who completed the survey and passed the two embedded attention checks (described below) were paid $4.00 USD. All study procedures were approved by the Johns Hopkins University School of Medicine Institutional Review Board.

Procedure

Following the prescreening survey, participants completed the full survey, which took on average 35.2 (SD = 23.4) minutes to complete. Participants provided basic demographic information and information related to their cannabis and alcohol use and couse patterns. Alcohol/cannabis use questions measured overall frequency of use (days of use in past 30-days, proportion of drinking/cannabis-use episodes in which they used both) and typical quantities used (i.e., grams of cannabis/session, alcoholic drinks/episode). Participants also indicated in separate questions whether they had driven under the influence of cannabis or alcohol in the past 30 days. Participants completed the Alcohol Use Disorder Identification Test (AUDIT; Saunders, Aasland, Babor, de la Fuente, & Grant, 1993) and Cannabis Use Disorder Identification Test (CUDIT; Adamson & Sellman, 2003).

Hypothetical Purchasing Tasks

Participants completed four separate hypothetical purchasing tasks for cannabis and alcohol. Participants first completed two separate, single-item purchasing tasks: one for cannabis and one for alcohol when each was the only drug available. For the cannabis purchasing task, participants purchased individual hits of cannabis, which were defined as “a single puff from a standard joint, blunt, or pipe (one hit = ~0.09 grams).” This definition is consistent with previous purchasing tasks (Aston et al., 2015; Strickland et al., 2017). For the alcohol purchasing task, participants purchased standard drinks of alcohol, which were defined as “a U.S. standard drink (one 5 oz. glass of wine; 1.5 oz. of liquor as a shot or mixed drink; one 12 oz. beer at 5% alcohol by volume).” In both tasks, participants indicated how much cannabis (“puffs”) or alcohol (“standard drinks”) they would purchase for use in 24 hr at the following prices: $0.25, $0.50, $0.75, $1.00, $1.50, $1.75, $2.00, $2.50, $3.00, $4.00, $5.00, $6.00, $7.00, $8.00, $9.00, and $10.00 (per hit or drink, depending on task). The full instructional sets for each task are presented in the online supplemental materials. Briefly, participants were asked to imagine that they could not get cannabis or alcohol elsewhere, for cheaper, or saved from previous episodes. They were to make choices as if they were using their own money and could not spend more than they had. In this scenario, they had not used any other drugs before making these decisions, the cannabis and alcohol being purchased were of their usual quality, and they would have to consume all the cannabis or alcohol that they purchased within the next 24 hr. There was no limit to the cannabis or alcohol they could purchase provided they consume it in the next 24 hr, and to consider each price individually, such that any cannabis or alcohol purchased at one price was not available at another. All prices in each task were presented on the same page. During the cannabis purchasing task, participants were instructed to type “1812” into a designated entry space to signify they were paying attention.

Participants also completed two separate cross-commodity purchasing tasks in which both cannabis and alcohol were concurrently available. The instructions were generally the same for the cross-commodity tasks as the single-item ones; however, participants were asked to imagine now that for each question (price), both cannabis and alcohol were available for purchasing and they were asked to select how much cannabis and alcohol they would purchase and consume in the next 24 hr. In one task, the price of cannabis hits varied from $0.25 to $10 as described above for the single-item task, and alcohol drinks were available for $1 at each cannabis price point. In the second task, alcohol prices varied from $0.25 to $10 and puffs of cannabis were concurrently available for $1 each. All prices were presented on the same page, with cannabis and alcohol in separate, side-by-side columns. The prices were listed with either “per hit” or “per drink” for clarity, and at the top of the fixed-price column, it was made explicit that hits or drinks cost $1. In the fixed-price alcohol cross-commodity task, an attention check required participants to type “36” in two designated places to ensure data integrity.

Data Analysis

Orderliness of purchasing task responses for the adjusting-price commodities across each task was determined using established methods for evaluating systematic demand responses (Bruner & Johnson, 2014; Stein, Koffarnus, Snider, Quisenberry, & Bickel, 2015). Responses were considered systematic if (a) consumption at one price did not exceed consumption at the previous, lower price by more than 20%, (b) consumption at the highest price was at least 10% less than consumption at the lowest price, and (c) consumption greater than 0 was not reported at higher prices after 0 consumption was endorsed at a lower price. Individual responses were modeled in GraphPad Prism Version 8 for macOS (GraphPad software, La Jolla, CA) using the exponential demand equation (Hursh & Silberberg, 2008): logQ=logQ0+k(e(αQ0C)1). This equation models how the dependent variable Q (consumption) changes as a function of the independent variable C (price) according to the free parameters α and Q0, representing demand elasticity (price-sensitivity) and demand intensity (consumption at prices approaching 0), respectively, and the constant k, which describes the range of log-transformed consumption across all prices among the four purchase tasks and was set to 5 for all applications of this equation. To allow for analysis in log space, the first instance of 0 consumption was converted to 0.1 and subsequent prices were not analyzed (Aston et al., 2015; M. W. Johnson, Johnson, Rass, & Pacek, 2017; MacKillop et al., 2012). Reported consumption at $0.25 per drink/puff was used as a measure of demand intensity instead of the curve-derived metric. Data Sets with 0 or unchanging consumption across prices still provide valuable information regarding consumption (intensity), but were excluded from demand elasticity analyses as nonlinear curves cannot be fit to these response patterns. Because of the extreme skew of the demand metrics, demand intensity (Q0) was log-transformed and demand elasticity (α) was natural log-transformed, which increased normality, for subsequent parametric analyses.

As with the single-item tasks, 0 consumption of the fixed-price commodity was transformed to 0.1 to allow for log transformation; however, all instances of 0 consumption were used for calculation of cross-price elasticity of the fixed-item commodity. Cross-price elasticity was determined at the individual level by calculating the slope of the log-transformed consumption of the fixed-price commodity as a function the log-transformed price of the adjusting-price commodity. Individual linear regression analyses with extra sum-of-squares F tests were performed on each fixed-price cannabis and alcohol demand curve to determine if the cross-price slope differed from 0. If the slope did not statistically differ from 0, the response pattern was considered independent. If the slope statistically differed and was positive, the response pattern was indicative of substitution, whereas if it was negative, it was indicative of complementarity. Linear regression was used to assess substitutability of cannabis for alcohol, and vice versa, at the group level with extra sum-of-squares F tests to determine if either slope significantly differed from 0.

To assess the influence of concurrent availability of an alternative reinforcer on cannabis and alcohol demand, separate repeated-measures analyses of variance (ANOVAs) were conducted for log-transformed demand intensity and natural log-transformed demand elasticity for both cannabis and alcohol. To address how individual cross-price elasticity category (substitution vs. complementarity vs. independence) for both fixed-price cannabis and alcohol was associated with a variety of cannabis- and alcohol-use related variables, a series of one-way ANOVAs were conducted testing for differences in demand metrics (log-transformed Q0 and natural log-transformed α of cannabis and alcohol when available alone or concurrently), cannabis, alcohol, couse variables (frequency and quantity), and dependence measures (CUDIT and AUDIT) were conducted with fixed-price cannabis or alcohol cross-price elasticity category as a between-subjects factor. When a significant effect was determined in the ANOVA, Tukey’s significant difference post hoc tests were used to determine which groups differed. Additionally, to determine potential differences in the prevalence of driving while under the influence of cannabis or alcohol according to cross-price elasticity category (substitute, complement, or independent), four separate chi-squared tests of independence were conducted comparing incidence of past-30-day cannabis- or alcohol-impaired driving by fixed-price alcohol or cannabis category. Demand intensity and elasticity from the cannabis and alcohol alone conditions, CUDIT scores, and AUDIT scores were compared between individuals who reported driving under the influence of cannabis or alcohol in the past 30 days using independent-samples t tests (presented in Tables 3 and 4 in the online supplemental materials).

Intercorrelations between demand metrics (Q0 and α of cannabis and alcohol when available alone or concurrently), cannabis, alcohol, couse variables (frequency and quantity), and dependence measures (CUDIT and AUDIT) were evaluated with bivariate Spearman rank correlations. Similarly, the relation between cross-price slope of fixed-price cannabis and alcohol and cannabis, alcohol, couse variables (frequency and quantity), and dependence measures (CUDIT and AUDIT) were evaluated with bivariate Spearman rank correlations.

Results

Participants

A total of 1,152 individuals completed the survey, and 711 participants passed the embedded attention checks. Of the 711 participants who passed the attention checks, 196 participants provided nonsystematic responses on at least 1 of the 4 purchasing tasks and were excluded from analysis, yielding a total sample size of N = 515. Participants were predominately white (71.1%) and male (56.1%) with a mean age of 34.1 (SD = 9.5) years. Per use session, participants reported consuming a mean 0.63 (SD = 1.2) grams of cannabis or 4.0 (SD = 3.2) drinks, and in the past 30 days, participants reported a mean cannabis use frequency of 15.91 (SD = 10.5) days and a mean alcohol use frequency of 9.27 (SD = 8.3) days. Participants reported using cannabis and alcohol together at a mean frequency of 39.9% (SD = 33.4) of all drinking/ smoking episodes. Eighty-four percent of participants reported some degree of couse (i.e., >0% of all drinking/smoking episodes involved use of both), and 45% of participants reported couse more than 50% of the time. Participant demographics according to fixed-price cannabis and alcohol cross-price elasticity category are presented in Tables 1 and 2 in the online supplemental materials, respectively.

Group Level Analyses of Demand

Demand curves for the 4 tasks are illustrated in Figure 1. There were no significant differences in demand elasticity between cannabis and alcohol when available alone, F(1, 503) = 0.488, p = .485, ηp2=0.001; however, the elasticity of adjusting-price alcohol when cannabis was concurrently available was greater than the elasticity of cannabis when alcohol was concurrently available, F(1, 457) = 11.272, p = .001, ηp2=0.024. When both cannabis and alcohol were available, a repeated-measures ANOVA indicated that there was a significant reduction in demand intensity (log-transformed) for both cannabis, F(1, 514) = 94.146, p < .001, ηp2=0.087 and alcohol, F(1, 514) = 48.958, p < .001, ηp2=0.155 relative to when they were available alone. A repeated-measures ANOVA evaluating natural log-transformed demand elasticity determined that concurrent-availability of cannabis significantly increased demand elasticity for alcohol relative to when alcohol was available alone, F(1, 464) = 73.710, p < .001, ηp2=0.137; however, cannabis demand elasticity was unaltered by concurrently available alcohol, F(1, 495) = 0.307, p = .580, ηp2=0.001. Linear regression analyses on group-level, fixed-price alcohol consumption yielded a significant negative slope for fixed-price alcohol (slope = −.256), F(1, 9326) = 222.4, p < .001, indicating that, at the group level, alcohol served as a complementary commodity for cannabis. Similarly, linear regression analyses on group-level, fixed-price cannabis consumption yielded a significant negative slope for fixed-price cannabis (slope = −.241), F(1, 9117) = 188.5, p < .001, indicating that, at the group level, cannabis also served as a complementary commodity for alcohol.

Figure 1.

Figure 1.

MSEM) consumption in single-item (empty circles) purchasing tasks and cross-commodity purchasing tasks of the adjusting-price commodity (filled circles) and fixed-price commodity (squares) for the cannabis purchasing tasks (left) and alcohol purchasing tasks (right).

Individual Level Analyses of Demand

Spearman rank correlations between cannabis demand metrics and cannabis use variables are presented in Table 1, and correlations between alcohol demand metrics and alcohol use variables are presented in Table 2. For both drugs, demand intensity (from the alone and adjusting-price conditions) was positively correlated with quantity and frequency of self-reported actual cannabis and alcohol consumption of study participants, whereas demand elasticity (from the alone and adjusting-price conditions) was negatively correlated with these variables. The negative correlations between demand elasticity (α) and self-reported use means that individuals who exhibit less elasticity (reduced price-sensitivity) generally use cannabis or alcohol more frequently (and in greater quantities) than individuals with greater elasticity (greater sensitivity to price).

Table 1.

Cannabis Purchasing Tasks

Variable Alone Q0 Adjusting Q0 Alone α Adjusting α

Quantity per session (g) .283** .261** −.211** −.214**
Quantity per day (g) .477** .449** −.276** −.290**
Quantity per month (g) .441** .417** −.289** −.273**
Past-30-day cannabis use (days) .334** .335** −.177** −.200**
CUDIT score .367** .388** −.301** −.311**

Note. CUDIT Cannabis Use Disorder Identification Test. Spearman rank correlations (rho) between cannabis demand metrics (Q0 = demand intensity [consumption at lowest price]; α = demand elasticity [price-sensitivity]) from single-item (alone) and cross-commodity (adjusting) cannabis purchasing tasks.

**

p < .01.

Table 2.

Alcohol Purchasing Tasks

Variable Alone Q0 Adjusting Q0 Alone α Adjusting α

Quantity per episode (drinks) .656** .572** −.378** −.382**
Past-30-day alcohol use (days) .326** .347** −.261** −.256**
Past-30-day binge drinking (days) .629** .548** −.380** −.401**
AUDIT score .629** .604** −.398** −.449**

Note. AUDIT Alcohol Use Disorder Identification Test. Spearman rank correlations (rho) between cannabis demand metrics (Q0 = demand intensity [consumption at lowest price]; α = demand elasticity [price-sensitivity]) from single-item (alone) and cross-commodity (adjusting) cannabis purchasing tasks.

**

p < .01.

Relation Between Cross-Price Elasticity and Drug Use Variables

Spearman rank correlations between fixed-price alcohol slope, fixed-price cannabis slope, and drug use variables are presented in Table 3. Fixed-price alcohol slope was negatively correlated with CUDIT score, AUDIT score, per-session cannabis consumption, past-30-day alcohol use, and couse frequency. Fixed-price cannabis slope was negatively correlated with CUDIT score, per-session and per-day cannabis consumption, and couse frequency. This means that more complementary hypothetical use choices were associated with greater severity of actual problematic use behavior.

Table 3.

Correlations Between Cross-Price Elasticity and Use Behaviors

Use variable Alcohol slope Cannabis slope

Past-30-day cannabis use    .017    .015
CUDIT score  −.128**  −.132**
AUDIT score  −.120**  −.080
Cannabis quantity per session  −.170**  −.180**
Cannabis quantity per day  −.053  −.102*
Cannabis quantity per month    .043  −.006
Drinks per episode  −.048  −.028
Past-30-day alcohol use  −.094*  −.056
Past-30-day binge drinking  −.025  −.036
Past-30-day couse  −.142**  −.120**
Couse percentage  −.179**  −.116**

Note. CUDIT Cannabis Use Disorder Identification Test; AUDIT = Alcohol Use Disorder Identification Test. Spearman rank correlations (rho) between cannabis and alcohol use variables and fixed-price alcohol (left) and cannabis (right).

*

p < .05.

**

p < .01.

In further analyses, participants were categorized as individuals whose data indicated cannabis and alcohol were complements, substitutes, or independents based on cannabis and alcohol cross-price elasticity (fixed-price slope); each individual occupied two groups (one for alcohol, one for cannabis). Demand curves for cannabis and alcohol according to cross-price elasticity category are illustrated in Figure 2. Based on fixed-price alcohol hypothetical purchasing, 85 participants were categorized as using cannabis and alcohol as substitutes (16.5%), 205 as complements (39.9%), and 224 as independents (43.6%). Based on fixed-price cannabis hypothetical purchasing, 78 participants were categorized as using cannabis and alcohol as substitutes (15.1%), 177 as complements (34.4%), and 260 as independents (50.5%). Three hundred forty-eight participants (67.6%) demonstrated the same type of cross-commodity purchasing pattern with both cannabis and alcohol (46 both substitutes [8.9%], 137 both complements [26.6%], 165 both independents [32.0%]). The remaining 32.4% of participants responded differently for each fixed-price commodity (five alcohol substitute–cannabis complement [1.0%], 34 alcohol substitute–cannabis independent [6.6%], eight alcohol complement–cannabis substitute [1.6%], 60 alcohol complement–cannabis independent [11.7%], 35 alcohol independent–cannabis substitute [6.8%], 25 alcohol independent–cannabis complement [4.9%]). Table 4 illustrates one-way ANOVA results addressing fixed-alcohol cross-price group differences across cannabis and alcohol use variables. Group differences were determined for cannabis demand intensity and elasticity when available alone, alcohol demand intensity when available alone, alcohol demand intensity with concurrently available cannabis, alcohol demand elasticity with concurrently available cannabis, CUDIT and AUDIT scores, per-session and per-day cannabis consumption, and couse frequency, which generally indicated greater severity of problematic use behavior in individuals treating alcohol as a complementary drug. Table 5 illustrates one-way ANOVA results addressing fixed-cannabis cross-price group differences across cannabis and alcohol use variables. Group differences were determined for cannabis and alcohol demand intensity and elasticity, CUDIT and AUDIT scores, per-session and per-day cannabis consumption, per-episode alcohol consumption, cannabis and alcohol use frequency, and couse frequency, which generally indicated greater severity of problematic use behavior in individuals treating cannabis as a complementary drug.

Figure 2.

Figure 2.

MSEM) consumption of hypothetical cannabis (top) and alcohol (bottom) in single-item (empty circles) purchasing tasks and cross-commodity purchasing tasks of the adjusting-price commodity (filled circles) and fixed-price commodity (squares) for participants categorized as substitutes (left), complements (middle), or independents (right).

Table 4.

Comparisons of Cannabis and Alcohol Use Variables According to Fixed-Price Alcohol Purchasing Category

Fixed-price alcohol category
One-way ANOVA results
Use variable Substitute (n = 85) Complement (n = 205) Independent (n = 224) F p ηp2

Cannabis alone demand intensity (log Q0) 1.16 (0.04) 1.35 (0.05)# 1.16 (0.04) 5.581 .004 0.021
Cannabis alone demand elasticity (ln α) −5.43 (0.13) −6.10 (0.16)## −5.19 (0.09) 14.855 <.001 0.056
Cannabis concurrent demand intensity (log Q0) 1.07 (0.04) 1.09 (0.04) 0.97 (0.03)* 3.491 .031 0.013
Cannabis concurrent demand elasticity (ln α) −5.58 (0.11) −5.80 (0.11) −5.41 (0.08)* 4.749 .009 0.019
Alcohol alone demand intensity (log Q0) 0.76 (0.03) 0.94 (0.04)## 0.65 (0.02) 20.730 <.001 0.075
Alcohol alone demand elasticity (ln α) −5.67 (0.10) −6.02 (0.20) −6.04 (0.32) 0.358 .699 0.001
Alcohol concurrent demand intensity (log Q0) 0.67 (0.03) 0.87 (0.04)## 0.57 (0.02) 26.35 <.001 0.093
Alcohol concurrent demand elasticity (ln α) −5.28 (0.11) −5.63 (0.11) −5.28 (0.07)* 4.792 .009 0.020
CUDIT score 10.11 (0.51) 12.96 (0.51)## 10.25 (0.39) 11.720 <.001 0.044
AUDIT score 10.67 (0.56) 14.19 (0.64)## 10.91 (0.43) 12.312 <.001 0.046
Per-session cannabis quantity (g) 0.42 (0.03) 0.95 (0.14)## 0.43 (0.03) 10.964 <.001 0.045
Daily cannabis quantity (g) 0.96 (0.09)* 1.24 (0.11) 0.96 (0.07) 3.395 .034 0.014
Monthly cannabis quantity (g) 13.33 (1.60) 13.71 (1.37) 16.96 (1.73) 1.486 .227 0.006
Per-session alcohol quantity (drinks) 4.02 (0.29) 4.38 (0.27) 3.64 (0.18) 2.853 .059 0.011
Past 30-day cannabis use frequency (days) 15.19 (1.21) 15.54 (0.70) 16.48 (0.72) 0.656 .519 0.003
Past 30-day alcohol use frequency (days) 7.62 (0.74) 10.16 (0.59) 9.09 (0.58) 2.891 .056 0.011
Past 30-day couse frequency (days) 3.06 (0.50)** 5.88 (0.52) 4.41 (0.46) 5.690 .004 0.022
Couse frequency (percentage of episodes) 33.09 (3.35) 46.83 (2.35) 36.11 (2.21) 7.812 <.001 0.030
Past 30-day binge-drinking frequency (days) 1.76 (0.30) 2.52 (0.34) 2.30 (0.20) 0.824 .439 0.003

Note. ANOVA analysis of variance; CUDIT Cannabis Use Disorder Identification Test; AUDIT Alcohol Use Disorder Identification Test. Results from one-way ANOVAs comparing cannabis and alcohol use patterns by fixed-price alcohol category. MSEM) values for each variable are presented for each category.

#

p < .05

##

p < .01 relative to all other groups

*

p < .05

**

p < .01 relative to complement group in Tukey’s honest significant difference test.

Table 5:

Comparisons of Cannabis and Alcohol Use Variables According to Fixed-Price Cannabis Purchasing Category

Fixed-price alcohol category
One-way ANOVA results
Use variable Substitute (n = 78) Complement (n = 177) Independent (n = 260) F p ηp2

Cannabis alone demand intensity (log Q0) 1.12 (0.05) 1.38 (0.06)## 1.16 (0.03) 8.077 <.001 0.031
Cannabis alone demand elasticity (ln α) −5.42 (0.13) −6.16 (0.18)## −5.26 (0.08) 14.500 <.001 0.054
Cannabis concurrent demand intensity (log Q0) 0.98 (0.05) 1.13 (0.04) 0.99 (0.03)** 5.244 .006 0.020
Cannabis concurrent demand elasticity (ln α) −5.50 (0.13) −5.94 (0.12)# −5.38 (0.07) 9.825 <.001 0.038
Alcohol alone demand intensity (log Q0) 0.79 (0.03) 0.97 (0.05)# 0.66 (0.02)& 25.343 <.001 0.091
Alcohol alone demand elasticity (ln α) −5.42 (0.13) −6.16 (0.18)## −5.26 (0.08) 14.500 <.001 0.054
Alcohol concurrent demand intensity (log Q0) 0.73 (0.03) 0.90 (0.04)## 0.57 (0.02)& 33.13 <.001 0.115
Alcohol concurrent demand elasticity (ln α) −5.15 (0.12) −5.82 (0.12)## −5.23 (0.06) 14.410 <.001 0.059
CUDIT score 9.78 (0.56) 13.29 (0.57)## 10.40 (0.35) 13.809 <.001 0.051
AUDIT score 11.83 (0.66) 15.11 (0.73)## 10.26 (0.35) 23.170 <.001 0.083
Per-session cannabis quantity (g) 0.39 (0.32) 0.97 (0.15)## 0.49 (0.04) 9.714 <.001 0.040
Daily cannabis quantity (g) 0.88 (.10) 1.33 (0.12)# 0.95 (0.06) 6.041 .003 0.025
Monthly cannabis quantity (g) 11.83 (1.57) 15.03 (1.62) 16.07 (1.49) 1.091 .337 0.005
Per-session alcohol quantity (drinks) 4.37 (0.34) 4.67 (0.32) 3.42 (0.13)** 8.902 <.001 0.033
Past 30-day cannabis use frequency (days) 13.05 (1.18) 15.04 (0.74) 17.37 (0.67)&& 6.074 .002 0.023
Past 30-day alcohol use frequency (days) 9.58 (0.99) 10.40 (0.64) 8.42 (0.50)* 3.081 .047 0.012
Past 30-day couse frequency (days) 3.57 (0.66) 6.09 (0.58)## 4.23 (0.40) 5.332 .005 0.020
Couse frequency (percentage of episodes) 35.37 (3.60) 45.98 (2.52) 37.01 (2.07)* 4.681 .010 0.018
Past 30-day binge-drinking frequency (days) 2.56 (0.44) 2.82 (0.37) 1.85 (0.28) 2.589 .076 0.010

Note. ANOVA analysis of variance; CUDIT Cannabis Use Disorder Identification Test; AUDIT Alcohol Use Disorder Identification Test. Results from one-way ANOVAs comparing cannabis and alcohol use patterns by fixed-price alcohol category. MSEM) values for each variable are presented for each category.

#

p < .05

##

p < .01 relative to all other groups

*

p < .05

**

p < .01 relative to complement group

&

p < .05

&&

p < .01 relative to substitute group in Tukey’s HSD test.

Relations With Intoxicated Driving

One hundred ninety-four participants (37.7%) endorsed driving under the influence of cannabis in the past 30 days. Chi-square tests of independence evaluating likelihood of driving under the influence of cannabis by cross-price cannabis or alcohol category determined that the likelihood of driving under the influence of cannabis did not differ by fixed-price alcohol category, χ2(2, N = 515) = 1.261, p = .532 or fixed-price cannabis category, χ2(2, N = 515) = 5.146, p = .076. Individuals who reported driving under the influence of cannabis demonstrated significantly greater alcohol and cannabis demand intensity, significantly greater CUDIT and AUDIT scores, and significantly lower cannabis and alcohol demand elasticity relative to those who did not (all ps < .001; Table 3 in the online supplemental materials). Seventy-four participants (14.1%) endorsed driving under the influence of alcohol in the past 30 days. Chi-square tests of independence evaluating likelihood of driving under the influence of alcohol by fixed-price cannabis or alcohol category determined that likelihood of driving under the influence of alcohol was significantly greater in the fixed-price alcohol complement group, χ2(2, N = 515) = 7.597, p = .022 and fixed-price cannabis complement group, χ2(2, N = 515) = 24.119, p < .001 relative to the substitute or independent groups. Individuals reporting driving under the influence of alcohol demonstrated significantly greater alcohol and cannabis demand intensity, significantly greater CUDIT and AUDIT scores, and significantly lower cannabis and alcohol demand elasticity relative to those who did not (all ps < .05; Table 4 in the online supplemental materials). Fifty-seven participants (11.1%) endorsed driving under the influence of both cannabis and alcohol (although not necessarily concurrently) in the past 30 days, and a chi-square test of independence demonstrated that participants who endorsed driving under the influence of alcohol were significantly more likely to also drive under the influence of cannabis, χ2(1, N = 515) = 57.010, p < .001.

Discussion

Summary of Findings and Novel Analysis Validity

To our knowledge, this is the first study to formally evaluate individual behavioral economic interactions between cannabis and alcohol use. Demand curves were generally orderly for both adjusting-price drugs when available alone and when the other drug was also available at a fixed price, such that consumption was generally defended at lower prices followed by monotonic decreases in consumption at higher prices. Similarly, consumption of the concurrent reinforcer (i.e., the fixed-priced drug in the cross-commodity tasks) typically followed a linear function across prices, and participants’ responses could readily be categorized according to classic patterns of commodity interaction. As in several previous reports evaluating hypothetical drug demand (reviewed in Aston & Cassidy, 2019; González-Roz, Jackson, Murphy, Rohsenow, & MacKillop, 2019; Zvorsky et al., 2019), we found significant correlations between demand intensity and elasticity and clinically relevant outcomes, including CUDIT and AUDIT dependence scores and frequency and quantity of cannabis and alcohol use, for both single-item and cross-commodity purchase tasks. Although consumption of cannabis appears to exceed that of alcohol, puffs of cannabis and drinks of alcohol are not likely equivalent or comparable. Thus, direct comparisons between cannabis and alcohol were limited to the unitless elasticity-related parameter α (Gilroy, Kaplan, & Reed, 2020), and comparisons between the different units in demand intensity, like apples and oranges, were not performed. These findings highlight that our cross-commodity approach was valid and similarly applicable as single-item tasks to a variety of behaviors.

Cross-Price Elasticity Categorization and Relations to Substance Use

Although several previous reports have suggested overlap between hypothetical purchasing patterns and real-world drug use behavior, the novelty in this study lies in the cross-commodity analyses and analyses of drug use behaviors according to how participants treat the two drugs under hypothetical situations of concurrent availability (e.g., cross-price elasticity category). At the individual level, participants generally indicated independent purchasing patterns across drugs; however, a substantial number of participants endorsed steep complementary purchasing leading to an overall negative slope for the fixed-price items at the group level. The nontrivial representation of all three cross-price elasticity patterns at the individual-level is comparable to the lack of a clear economic pattern of couse at the population-level (Guttmannova et al., 2016; Subbaraman, 2016), such that all three patterns are represented and vary depending on the population being studied, and highlights the importance of assessing couse patterns at the individual level. Additionally, the substantial representation of all three categories of cross-price elasticity runs contrary to traditional considerations of fixed pharmacological interactions between drugs (e.g., always a substitute) such that cannabis and alcohol may not interact in the same way across all users, despite their fixed pharmacological mechanisms. Although the substitution cross-price elasticity group, comprising approximately 15% of participants between tasks, represented the smallest proportion of participants, when considered at the population level of past-month cannabis users in the U.S., this 15% translates into more than 4 million people (Substance Abuse and Mental Health Services Administration, 2018), further demonstrating the importance of individual-level assessments in these analyses.

The most-important findings from these analyses involve the interactive behavioral-economic effects of cannabis and alcohol and the relation between these interactive effects and behaviors related to cannabis and alcohol use. Cross-price elasticity was generally negatively correlated with drug-use outcomes for both fixed-price cannabis and alcohol, suggesting that greater complementarity was generally associated with greater drug use frequency and dependence. Importantly, couse frequency was significantly negatively correlated with cross-price elasticity, providing greater confidence in the measure, as individuals endorsing complementary-purchasing patterns also endorsed a greater frequency of couse. However, despite these significant correlations demonstrating more severe outcomes in complementary consumers, the correlations were generally weak (ρ <.20). The weak effects may have come from the large number of independent consumption patterns (43.6% alcohol; 50.5% cannabis) with zero or near-zero slopes.

To address this skew in the continuous variable of slope, participants were categorized according to cross-price elasticity status. As in the correlational analyses, the group comparisons generally demonstrated greater demand (higher demand intensity and/or lower elasticity), greater cannabis consumption quantity, more frequent couse, and greater dependence in participants endorsing complementary purchasing of fixed-price cannabis or alcohol relative to other cross-price patterns. Additionally, we determined that driving under the influence of alcohol, but not cannabis, was significantly greater in individuals endorsing complementary purchasing of fixed-price cannabis or alcohol. Notably, self-reported couse frequency was typically lowest in the substitution groups for both drugs, suggesting participants indicating price-related replacement of one drug with the other are also less likely to demonstrate couse behavior. These findings are comparable to reports indicating that more frequent couse of cannabis and alcohol in self-report measures is associated with greater dependence, drug-related consequences, and intoxicated driving (Arterberry et al., 2017; Subbaraman & Kerr, 2015). Additionally, these data corroborate the population-level findings in which cannabis use increased in binge drinkers, but not the rest of the population, following state-level cannabis legalization (Jones et al., 2018; Kerr et al., 2017). These binge-drinking, subgroup specific changes not seen at the overall population level mirror the findings in the current dataset in which complementarity of cannabis and alcohol was associated with greater alcohol dependence. In previous analyses of single-item alcohol demand, demand for alcohol was greater in individuals reporting couse of cannabis and alcohol (Morris et al., 2018; Ramirez et al., 2019). Curiously, there were no differences among groups in frequency of driving under the influence of cannabis, which is contrary to previous reports indicating an association between couse and driving under the influence of cannabis (Arterberry et al., 2017; Subbaraman & Kerr, 2015). The lack of difference in cannabis-influenced driving rates may have resulted from the relatively large proportion of participants endorsing past-30-day intoxicated driving in the overall sample. The consistently worse outcomes in the complementary group and the typically better health outcomes in the substitution group should inform public health messaging campaigns surrounding cannabis and alcohol to encourage that, if either drug is to be used at all, only a single drug should be used at a time, instead of simultaneous use. Even a shift of a few percentage points from complementary purchasing patterns to substitution might reduce the severity of cannabis and alcohol dependence for millions of people globally. Additionally, these data indicate the importance of price in influencing use patterns when both drugs are concurrently available, which identifies price (e.g., taxation) as a viable regulatory strategy for reducing use/couse of these substances.

Although the between-groups patterns were largely similar for both fixed-price drugs, there were some notable differences. For alcohol demand elasticity (both own- and cross-price), individuals with complementary cannabis purchasing patterns endorsed significantly lower demand elasticity relative to other groups; however, there were no fixed-price alcohol group differences in own-price alcohol elasticity relative to other groups and only complementary and independent alcohol purchasing groups differed. Complementary fixed-price cannabis purchasers demonstrated greater daily cannabis consumption quantity, but only the alcohol substitution group differed from the alcohol complementary group in consumption quantity. Cannabis and alcohol use frequency did not differ according to fixed-price alcohol purchasing category, whereas cannabis use frequency was highest and alcohol use frequency was lowest in individuals endorsing independent cannabis purchasing patterns. This outcome suggests that an unchanging consumption of $1.00/puff cannabis may be indicative of generally greater cannabis use relative to alcohol and, perhaps, a greater cannabis preference. Altogether, the group-based analyses indicated that cross-commodity purchasing patterns when hypothetical cannabis and alcohol were concurrently available were indicative of differing drug-use patterns, including potentially disordered use, with individuals endorsing complementary purchasing patterns of either drug demonstrating generally greater use and dependence. Additionally, patterns of fixed-price cannabis consumption may be more sensitive to differences in actual drug-use behavior relative to fixed-price alcohol consumption patterns; however, replication of these findings with alternative, market-based prices for cannabis and alcohol are needed to determine if this is indicative of drug-based or price-based differences in task utility.

Although there was substantial representation of the three types of cross-commodity categories (e.g., substitute, complement, independent), there were consistent decreases in consumption of either adjusting-price commodity (cannabis or alcohol) when the other drug was also available at a fixed price, indicative of one definition of substitution that is distinct from cross-price elasticity, that is, the ability of the presence of one commodity to reduce the consumption of another commodity. These findings are consistent with previous reports indicating significant reductions in real (M. W. Johnson & Bickel, 2003; M. W. Johnson, Bickel, & Kirshenbaum, 2004) and hypothetical (Johnson et al., 2017) cigarette consumption when alternative sources of reinforcement (nicotine and money) were concurrently available, and reductions in consumption of licit or illicit cannabis when the opposite legal-type of cannabis was available (Amlung et al., 2019). Although there were significant relative decreases in consumption of both drugs in the presence of an alternative than when alone, we found asymmetrical changes in consumption of alcohol and cannabis when the availability of the other drug was added, such that, despite overall reductions in consumption for both drugs, demand elasticity increased for alcohol, but remained constant for cannabis, when the alternative product was present at a fixed price. The effect sizes for the reduction in cannabis consumption (demand intensity) was nearly double that for the changes in alcohol consumption, indicating users are likely to consume substantially more cannabis when it is the only available drug compared to alcohol. Additionally, there were significant reductions in demand intensity and increases in elasticity of alcohol in the presence of cannabis, whereas the reduction in cannabis was a parallel shift limited to decreased intensity. This multidimensional reduction in alcohol reinforcement in the presence of cannabis and significantly greater cannabis consumption when no other drug is available may be interpreted to mean that cannabis is generally the more reinforcing of the two drugs in this population. Finally, the reductions in consumption of either drug in the presence of the other drug, and demand intensity for both drugs across tasks (single- or cross-price) were greatest in the complement groups, suggesting that not only is overall consumption highest in this group, but participants are much more likely to demonstrate compensatory cannabis or alcohol use when only one of the two drugs is available, providing additional behavioral economic evidence for the problematic drug use demonstrated in this group.

Limitations and Future Directions

The large sample size, parametric assessment of cannabis and alcohol consumption and interaction, and novel analytical approach involving individual participants assessments of drug interactions were major strengths of the current study; however, the findings presented herein should be considered in the context of certain limitations. Two interrelated issues comprise the most noteworthy limitation: the online sample and the high degree of participant exclusion. Workers on MTurk are incentivized to maximize income by completing as many tasks in as short a time as possible and may consequently rush through the tasks and/or skim the instructions, thereby leading to relatively high rates of missed attention checks and nonsystematic responding. In the laboratory, research staff can interface with participants to ensure understanding of instructions or monitor participant responding to detect misunderstandings or unusual responses in real time; whereas, in online surveys, nonsystematic responding or suspected rushed responses can only be determined after survey completion. Despite utilizing best practices for ensuring data quality on MTurk (Chandler & Shapiro, 2016; Strickland & Stoops, 2019) and incorporating multiple attention checks into the survey, including two within primary outcome tasks, numerous participants provided nonsystematic data and were subsequently removed from analysis. Given that the majority of these participants provided nonsystematic responses across all four tasks, it is possible that they were rushing through the survey while maintaining sufficient vigilance to detect the attention checks. By only analyzing data from participants who passed all attention checks and provided systematic responses across all four tasks, we aimed to counteract some of the potential limitations of using MTurk convenience samples through exclusion of potentially faulted data; however, the high exclusion rate represents a limitation of the current study and online versus laboratory-based behavioral studies.

Although puffs have frequently been assessed as the unit of consumption in hypothetical cannabis purchasing tasks (e.g., Aston et al., 2015; Aston et al., 2016; Strickland et al., 2017), recent focus group data suggest cannabis users find grams to be a more reasonable and familiar unit (Aston, Metrik, Rosen, Swift, & MacKillop, 2020). We chose puffs and drinks as the purchase units as they seemed the most appropriate units that would be understood by participants as meaningful to use during bouts of consumption (as opposed to grams and drinks; puffs and sips/small liquid volumes) for potential market interaction; however, whether these units are comparable remains unknown and is fertile ground for future research to determine equivalent units across substances for behavioral economic comparison.

These findings open up novel avenues for potential future research. The current study was limited to evaluation of cannabis plant material consumption; however, other methods of cannabis consumption, such as edibles or concentrates, are also popular and may interact differently with alcohol consumption relative to plant material. There is also additional work needed in terms of determining pharmacologically comparable units of cannabis and alcohol for cross-commodity analyses (e.g., using amounts of each substance that are judged to by individual participants to be equally valued). Finally, given that differential patterns of disordered cannabis and alcohol use emerged across cross-price elasticity categories, future research determining whether framing effects (e.g., advertising, public health messaging campaigns, etc.) can shift an individual’s purchasing patterns from complementary to substitutive or independent will be useful for informing how to successfully bring about these behavioral changes.

Conclusions

The novel categorization by individual cross-price elasticity represents a unique application of behavioral economic analyses to influence public health. These data could be useful for public health campaigns intended to highlight the negative effects of couse of cannabis and alcohol and to encourage responsible use by promoting substitution-like patterns and single use instead of couse. More importantly, these results are essential for consideration in development of state and national policies related to cannabis sales and advertising, such that promotion of couse through cannabis-alcohol pairing suggestions, cannabis-infused alcoholic beverages, or geographic proximity of cannabis and alcohol retailers should be discouraged. Additionally, increasing the price of one drug (e.g., through taxation) may serve as a viable regulatory strategy to reduce coconsumption in areas where both drugs are available. Given the higher ratings of cannabis and alcohol dependence, incidence of drunk driving, and quantity of cannabis consumption in the complementary groups, individuals preferring coconsumption may be at greater risk for cannabis- and alcohol-related consequences relative to those preferring to use either drug alone, and public health messaging should be geared toward discouraging couse and promoting responsible use of cannabis or alcohol alone.

Supplementary Material

supplementary materials

Public Health Significance.

We evaluated hypothetical purchasing of concurrently available cannabis and alcohol and its relation to cannabis and alcohol use and dependence. Individuals who prefer using cannabis and alcohol simultaneously (complementary purchasing patterns) tend to demonstrate greater disordered cannabis and alcohol use than those who use them separately (substitution or independent purchasing). Public health efforts to discourage couse may reduce problematic use patterns and dependence for either drug.

Acknowledgments

Funding was provided by National Institute on Drug Abuse Grants R01DA042527 (Matthew W. Johnson), R01DA043475 (Ryan Vandrey), and T32DA007209 (Sean B. Dolan and Tory R. Spindle). The funding sources had no role in the study other than financial support. Sean B. Dolan was supported by travel awards to present preliminary analyses of these data at the International Study Group Investigating Drugs as Reinforcers annual meeting (William L. Woolverton ISGIDAR travel award) and American Psychological Association annual convention (funded by R13AA022858) in 2019.

Footnotes

The authors have no conflicts of interest to report.

Supplemental materials: http://dx.doi.org/10.1037/pha0000397.supp

References

  1. Adamson SJ, & Sellman JD (2003). A prototype screening instrument for cannabis use disorder: The Cannabis Use Disorders Identification Test (CUDIT) in an alcohol-dependent clinical sample. Drug and Alcohol Review, 22, 309–315. 10.1080/0959523031000154454 [DOI] [PubMed] [Google Scholar]
  2. Amlung M, Reed DD, Morris V, Aston ER, Metrik J, & MacKillop J (2019). Price elasticity of illegal versus legal cannabis: A behavioral economic substitutability analysis. Addiction, 114, 112–118. 10.1111/add.14437 [DOI] [PubMed] [Google Scholar]
  3. Arterberry BJ, Treloar H, & McCarthy DM (2017). Empirical profiles of alcohol and marijuana use, drugged driving, and risk perceptions. Journal of Studies on Alcohol and Drugs, 78, 889–898. 10.15288/jsad.2017.78.889 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Aston ER, & Cassidy RN (2019). Behavioral economic demand assessments in the addictions. Current Opinion in Psychology, 30, 42–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Aston ER, Metrik J, Amlung M, Kahler CW, & MacKillop J (2016). Interrelationships between marijuana demand and discounting of delayed rewards: Convergence in behavioral economic methods. Drug and Alcohol Dependence, 169, 141–147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Aston ER, Metrik J, & MacKillop J (2015). Further validation of a marijuana purchase task. Drug and Alcohol Dependence, 152, 32–38. 10.1016/j.drugalcdep.2015.04.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Aston ER, Metrik J, Rosen RK, Swift R, & MacKillop J (2020). Refining the marijuana purchase task: Using qualitative methods to inform measure development. Experimental and Clinical Psychopharmacology. Advance online publication. 10.1037/pha0000355 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bickel WK, Johnson MW, Koffarnus MN, MacKillop J, & Murphy JG (2014). The behavioral economics of substance use disorders: Reinforcement pathologies and their repair. Annual Review of Clinical Psychology, 10, 641–677. 10.1146/annurev-clinpsy-032813-153724 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bonomo Y, Norman A, Biondo S, Bruno R, Daglish M, Dawe S, … Castle D (2019). The Australian drug harms ranking study. Journal of Psychopharmacology, 33, 759–768. 10.1177/0269881119841569 [DOI] [PubMed] [Google Scholar]
  10. Borodovsky JT, Crosier BS, Lee DC, Sargent JD, & Budney AJ (2016). Smoking, vaping, eating: Is legalization impacting the way people use cannabis? The International Journal on Drug Policy, 36, 141–147. 10.1016/j.drugpo.2016.02.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Bruner NR, & Johnson MW (2014). Demand curves for hypothetical cocaine in cocaine-dependent individuals. Psychopharmacology, 231, 889–897. 10.1007/s00213-013-3312-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Carliner H, Brown QL, Sarvet AL, & Hasin DS (2017). Cannabis use, attitudes, and legal status in the U.S.: A review. Preventive Medicine, 104, 13–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Cerda M, Sarvet AL, Wall M, Feng T, Keyes KM, Galea S, & Hasin DS (2018). Medical marijuana laws and adolescent use of marijuana and other substances: Alcohol, cigarettes, prescription drugs, and other illicit drugs. Drug and Alcohol Dependence, 183, 62–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Chandler J, & Shapiro D (2016). Conducting clinical research using crowdsourced convenience samples. Annual Review of Clinical Psychology, 12, 53–81. 10.1146/annurev-clinpsy-021815-093623 [DOI] [PubMed] [Google Scholar]
  15. Doggett A, Battista K, & Leatherdale ST (2020). Modes of cannabis use among Canadian youth in the COMPASS study; using LCA to examine patterns of smoking, vaping, and eating/drinking cannabis. Drugs Education Prevention & Policy. Advance online publication. 10.1080/09687637.2020.1769560 [DOI] [Google Scholar]
  16. Gilroy SP, Kaplan BA, & Reed DD (2020). Interpretation(s) of elasticity in operant demand. Journal of the Experimental Analysis of Behavior, 114, 106–115. 10.1002/jeab.610 [DOI] [PubMed] [Google Scholar]
  17. González-Roz A, Jackson J, Murphy C, Rohsenow DJ, & MacKillop J (2019). Behavioral economic tobacco demand in relation to cigarette consumption and nicotine dependence: A meta-analysis of cross-sectional relationships. Addiction, 114, 1926–1940. 10.1111/add.14736 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gunn RL, Norris AL, Sokolovsky A, Micalizzi L, Merrill JE, & Barnett NP (2018). Marijuana use is associated with alcohol use and consequences across the first 2 years of college. Psychology of Addictive Behaviors, 32, 885–894. 10.1037/adb0000416 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Guttmannova K, Lee CM, Kilmer JR, Fleming CB, Rhew IC, Kosterman R, & Larimer ME (2016). Impacts of changing marijuana policies on alcohol use in the United States. Alcoholism, Clinical and Experimental Research, 40, 33–46. 10.1111/acer.12942 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hartman RL, Brown TL, Milavetz G, Spurgin A, Gorelick DA, Gaffney G, & Huestis MA (2015). Controlled cannabis vaporizer administration: Blood and plasma cannabinoids with and without alcohol. Clinical Chemistry, 61, 850–869. 10.1373/clinchem.2015.238287 [DOI] [PubMed] [Google Scholar]
  21. Hartman RL, Brown TL, Milavetz G, Spurgin A, Gorelick DA, Gaffney G, & Huestis MA (2016). Controlled vaporized cannabis, with and without alcohol: Subjective effects and oral fluid-blood cannabinoid relationships. Drug Testing and Analysis, 8, 690–701. 10.1002/dta.1839 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hartman RL, Brown TL, Milavetz G, Spurgin A, Pierce RS, Gorelick DA, … Huestis MA (2015). Cannabis effects on driving lateral control with and without alcohol. Drug and Alcohol Dependence, 154, 25–37. 10.1016/j.drugalcdep.2015.06.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hasin DS (2018). US epidemiology of cannabis use and associated problems. Neuropsychopharmacology, 43, 195–212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hasin DS, Saha TD, Kerridge BT, Goldstein RB, Chou SP, Zhang H, … Grant BF (2015). Prevalence of marijuana use disorders in the United States between 2001–2002 and 2012–2013. JAMA Psychiatry, 72, 1235–1242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Hasin DS, Sarvet AL, Cerda M, Keyes KM, Stohl M, Galea S, & Wall MM (2017). US adult illicit cannabis use, cannabis use disorder, and medical marijuana laws: 1991–1992 to 2012–2013. JAMA Psychiatry, 74, 579–588. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Hughes A, Lipari RN, & Williams M (2013). State estimates of adolescent marijuana use and perceptions of risk of harm from marijuana use: 2013 and 2014. The CBHSQ Report, 2121, 1–12. Retrieved from https://www.samhsa.gov/data/sites/default/files/report_2121/ShortReport-2121.pdf [PubMed] [Google Scholar]
  27. Hursh SR, & Silberberg A (2008). Economic demand and essential value. Psychological Review, 115, 186–198. 10.1037/0033-295X.115.1.186 [DOI] [PubMed] [Google Scholar]
  28. Jacobs EA, & Bickel WK (1999). Modeling drug consumption in the clinic using simulation procedures: Demand for heroin and cigarettes in opioid-dependent outpatients. Experimental and Clinical Psychopharmacology, 7, 412–426. 10.1037/1064-1297.7.4.412 [DOI] [PubMed] [Google Scholar]
  29. Johnson JK, Johnson RM, Hodgkin D, Jones AA, Matteucci AM, & Harris SK (2018). Heterogeneity of state medical marijuana laws and adolescent recent use of alcohol and marijuana: Analysis of 45 states, 1991–2011. Substance Abuse, 39, 247–254. 10.1080/08897077.2017.1389801 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Johnson MW, & Bickel WK (2003). The behavioral economics of cigarette smoking: The concurrent presence of a substitute and an independent reinforcer. Behavioural Pharmacology, 14, 137–144. 10.1097/00008877-200303000-00005 [DOI] [PubMed] [Google Scholar]
  31. Johnson MW, Bickel WK, & Kirshenbaum AP (2004). Substitutes for tobacco smoking: A behavioral economic analysis of nicotine gum, denicotinized cigarettes, and nicotine-containing cigarettes. Drug and Alcohol Dependence, 74, 253–264. 10.1016/j.drugalcdep.2003.12.012 [DOI] [PubMed] [Google Scholar]
  32. Johnson MW, Johnson PS, Rass O, & Pacek LR (2017). Behavioral economic substitutability of e-cigarettes, tobacco cigarettes, and nicotine gum. Journal of Psychopharmacology, 31, 851–860. 10.1177/0269881117711921 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Jones J, Nicole Jones K, & Peil J (2018). The impact of the legalization of recreational marijuana on college students. Addictive Behaviors, 77, 255–259. [DOI] [PubMed] [Google Scholar]
  34. Kerr DCR, Bae H, Phibbs S, & Kern AC (2017). Changes in undergraduates’ marijuana, heavy alcohol and cigarette use following legalization of recreational marijuana use in Oregon. Addiction, 112, 1992–2001. 10.1111/add.13906 [DOI] [PubMed] [Google Scholar]
  35. Knapp AA, Lee DC, Borodovsky JT, Auty SG, Gabrielli J, & Budney AJ (2019). Emerging trends in cannabis administration among adolescent cannabis users. The Journal of Adolescent Health, 64, 487–493. 10.1016/j.jadohealth.2018.07.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. MacKillop J, Few LR, Murphy JG, Wier LM, Acker J, Murphy C, … Chaloupka F (2012). High-resolution behavioral economic analysis of cigarette demand to inform tax policy. Addiction, 107, 2191–2200. 10.1111/j.1360-0443.2012.03991.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. MacKillop J, & Murphy JG (2007). A behavioral economic measure of demand for alcohol predicts brief intervention outcomes. Drug and Alcohol Dependence, 89, 227–233. [DOI] [PubMed] [Google Scholar]
  38. Metrik J, Gunn RL, Jackson KM, Sokolovsky AW, & Borsari B (2018). Daily patterns of marijuana and alcohol co-use among individuals with alcohol and cannabis use disorders. Alcoholism, Clinical and Experimental Research, 42, 1096–1104. 10.1111/acer.13639 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Morris V, Amlung M, Kaplan BA, Reed DD, Petker T, & MacKillop J (2017). Using crowdsourcing to examine behavioral economic measures of alcohol value and proportionate alcohol reinforcement. Experimental and Clinical Psychopharmacology, 25, 314–321. 10.1037/pha0000130 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Morris V, Patel H, Vedelago L, Reed DD, Metrik J, Aston E, … Amlung M (2018). Elevated behavioral economic demand for alcohol in co-users of alcohol and cannabis. Journal of Studies on Alcohol and Drugs, 79, 929–934. 10.15288/jsad.2018.79.929 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Murphy JG, & MacKillop J (2006). Relative reinforcing efficacy of alcohol among college student drinkers. Experimental and Clinical Psychopharmacology, 14, 219–227. [DOI] [PubMed] [Google Scholar]
  42. Nutt DJ, King LA, & Phillips LD, & the Independent Scientific Committee on Drugs. (2010). Drug harms in the U. K.: A multicriteria decision analysis. Lancet, 376, 1558–1565. 10.1016/S0140-6736(10)61462-6 [DOI] [PubMed] [Google Scholar]
  43. Peters EN, Rosenberry ZR, Schauer GL, O’Grady KE, & Johnson PS (2017). Marijuana and tobacco cigarettes: Estimating their behavioral economic relationship using purchasing tasks. Experimental and Clinical Psychopharmacology, 25, 208–215. 10.1037/pha0000122 [DOI] [PubMed] [Google Scholar]
  44. Ramirez JJ, Cadigan JM, & Lee CM (2019). Behavioral economic demand for alcohol among young adults who engage in simultaneous alcohol and marijuana use. Substance Abuse. Advance online publication. 10.1080/08897077.2019.1671939 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Saunders JB, Aasland OG, Babor TF, de la Fuente JR, & Grant M (1993). Development of the alcohol use disorders identification test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption—II. Addiction, 88, 791–804. 10.1111/j.1360-0443.1993.tb02093.x [DOI] [PubMed] [Google Scholar]
  46. Stein JS, Koffarnus MN, Snider SE, Quisenberry AJ, & Bickel WK (2015). Identification and management of nonsystematic purchase task data: Toward best practice. Experimental and Clinical Psychopharmacology, 23, 377–386. 10.1037/pha0000020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Strickland JC, Alcorn JL III, & Stoops WW (2019). Using behavioral economic variables to predict future alcohol use in a crowdsourced sample. Journal of Psychopharmacology, 33, 779–790. 10.1177/0269881119827800 [DOI] [PubMed] [Google Scholar]
  48. Strickland JC, Campbell EM, Lile JA, & Stoops WW (2020). Utilizing the commodity purchase task to evaluate behavioral economic demand for illicit substances: A review and meta-analysis. Addiction, 115, 393–406. 10.1111/add.14792 [DOI] [PubMed] [Google Scholar]
  49. Strickland JC, Lile JA, & Stoops WW (2017). Unique prediction of cannabis use severity and behaviors by delay discounting and behavioral economic demand. Behavioural Processes, 140, 33–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Strickland JC, Lile JA, & Stoops WW (2019). Evaluating nonmedical prescription opioid demand using commodity purchase tasks: Test-retest reliability and incremental validity. Psychopharmacology, 236, 2641–2652. 10.1007/s00213-019-05234-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Strickland JC, & Stoops WW (2019). The use of crowdsourcing in addiction science research: Amazon Mechanical Turk. Experimental and Clinical Psychopharmacology, 27, 1–18. 10.1037/pha0000235 [DOI] [PubMed] [Google Scholar]
  52. Subbaraman MS (2016). Substitution and complementarity of alcohol and cannabis: A review of the literature. Substance Use & Misuse, 51, 1399–1414. 10.3109/10826084.2016.1170145 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Subbaraman MS, & Kerr WC (2015). Simultaneous versus concurrent use of alcohol and cannabis in the National Alcohol Survey. Alcoholism, Clinical and Experimental Research, 39, 872–879. 10.1111/acer.12698 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Substance Abuse and Mental Health Services Administration. (2018). Results from the 2018 National Survey on Drug Use and Health: Detailed tables. Rockville, MD: Author. Retrieved from https://www.samhsa.gov/data/sites/default/files/cbhsq-reports/NSDUHDetailedTabs2018R2/NSDUHDetTabsSect1pe2018.htm [Google Scholar]
  55. van Amsterdam J, Opperhuizen A, Koeter M, & van den Brink W (2010). Ranking the harm of alcohol, tobacco and illicit drugs for the individual and the population. European Addiction Research, 16, 202–207. 10.1159/000317249 [DOI] [PubMed] [Google Scholar]
  56. Zvorsky I, Nighbor TD, Kurti AN, DeSarno M, Naude G, Reed DD, & Higgins ST (2019). Sensitivity of hypothetical purchase task indices when studying substance use: A systematic literature review. Preventive Medicine, 128, 105789. 10.1016/j.ypmed.2019.105789 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplementary materials

RESOURCES