Cannabis,research ethics,and a duty of care

Abstract

Despite growing evidence to the contrary, researchers continue to posit causal links between cannabis, crime, psychosis, and violence. These spurious connections are rooted in history and fueled decades of structural limitations that shaped how researchers studied cannabis. Until recently, research in this area was explicitly funded to link cannabis use and harm and ignore any potential benefits. Post-prohibition cannabis research has failed to replicate the dire findings of the past. This article outlines how the history of controlling cannabis research has led to various harms, injustices, and ethical complications. We compare commonly cited research from both the prohibition and post-prohibition eras and argue that many popular claims about the dangers of cannabis are the result of ethical lapses by researchers, journals, and funders. We propose researchers in this area adopt a duty of care in cannabis research going forward. This would oblige individual researchers to establish robust research designs, employ careful analytic strategies, and acknowledge limitations in more detail. This duty involves the institutional recognition by funders, journals, and others that cannabis research has been deliberately misconstrued to criminalize, stigmatize, and pathologize.

Introduction

On successive evenings in prime time in June 2022, Fox News television host Laura Ingraham linked cannabis to psychosis and psychosis to mass shootings, an all too common feature of American society. ¹ Legalization, she argued, was fueling cannabis use. Consuming cannabis resulted in adverse mental health outcomes and, ultimately, gun violence. The notion that increased cannabis use will lead to more cases of cannabis-induced psychotic disorder (CIPD) emerged from a meta-analysis conducted by the National Academy of Science, Engineering, and Medicine (NASEM) in 2017. The work of a cluster of psychiatric researchers in the United Kingdom (UK) led by Robin Murray (Murray et al., 2017) and Marta Di Forti (Di Forti et al., 2019) was particularly prominent. We have argued (Heidt and Wheeldon, 2022a) that this research should be seen alongside other dubious correlational research linking cannabis, use disorders, and public safety risks (Green et al., 2010; Hasin et al., 2015; Pedersen and Skardhamar, 2010; Wagner and Anthony, 2002). It might also be seen in ethical terms.

Research ethics has proceeded in the 20th century from an effort to constrain moral depravity (Weindling, 2001). A series of trigger events, including but not limited to Nazi experimentation on Jewish people and the Tuskegee experiments, led to the Nuremberg Code (1947) and later recommendations by the U.S. National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, established in 1974. Charged with identifying the basic ethical principles that should underlie the conduct of biomedical and behavioral research involving human subjects, the Commission published the Belmont Report in 1979. Finally, in 1991, a broad set of U.S. Federal regulatory provisions governing human subjects and required ethical protections, known colloquially as the Common Rule, was adopted in the U.S. after a decade of inter-agency negotiations.

The Common Rule specifies how research that involves human subjects is to be conducted and reviewed, including specific rules for obtaining informed consent and the requirements for review by institutional review boards (IRB). ² In many countries, the combination of informed consent from participants and independent review before commencing research is common. For example, in Canada, any institution that receives funding from one of the three federal granting agencies—the Canadian Institutes of Health Research, the Social Sciences and Humanities Research Council, and the Natural Sciences and Engineering Research Council of Canada—must ensure that all research involving humans conducted at that institution complies with the Tri-Council Policy Statement. In addition, researchers should design studies in ways that demonstrate respect for persons, concern for welfare, and justice.

By contrast, in the UK prior to 2004, researchers largely governed their research and were offered “advice” instead of approval or rejection. Following 2004 and the enactment of the European Union’s (EU) Clinical Trials Directive 2002, institutions that hosted research had to produce policies based on rules rather than rely on older profession-based ethical norms (Kerrison and Pollock, 2005). In 2020, the EU’s transition from the Clinical Trials Directive to the Clinical Trials Regulation further detailed the research ethics assessment required for funding. This involves ethical considerations at multiple stages in the research project, informed consent, protections for vulnerable populations, and demonstrating “respect for people and for human dignity and fair distribution of the benefits and burden of research, and that you must protect the values, rights, and interests of the research participants.” ³ Critics contend research ethics is increasingly at odds with practice (Friesen et al., 2017). While some steps have improved transparency by ensuring research funding is acknowledged, questions persist (Horbach et al., 2022). Perhaps the most troubling observation is that because ethical guidelines are individualistic, they fail to account for potential risks to communities (Emanuel and Weijer, 2005).

This analysis is relevant to cannabis (and other drug policy) given that these laws are often differentially enforced, resulting in unequal arrest and incarceration rates in communities of color (Alexander, 2010). Exaggerating the risks of cannabis has been a strategic feature of the War on Drugs (Fisher, 2021). The policing cultures which emerged promoted aggressive police tactics and have furthered divisions by targeting neighborhoods based on their racial and ethnic makeup (Geller and Fagan, 2010). Cannabis reform is linked to ethical questions about criminalization, police practices, prosecutorial discretion, and the collateral consequences of incarceration (Wheeldon and Heidt, 2022a). Liberalizing cannabis policy can potentially reduce the harms associated with the justice system, including racial disparities in enforcement (Sheehan et al., 2021), and perhaps provoke meaningful police reform (Brown, 2022).

While there are few risks associated with the responsible use of legal cannabis, the harms associated with its prohibition are profound. The enforcement of cannabis prohibitions has been costly (Kaplan, 1970), racially biased (Mize, 2020), and based on ideology rather than evidence (Ritter, 2021). While the media has long been complicit in presenting cannabis as an existential risk (Conrad et al., 1992; Levine, 2003), research journals also bear some responsibility for publishing dubious research that led to harm borne mainly by Black, Indigenous, and other people of color (BIPOC). Researchers in the emerging post-cannabis prohibition period must consider why such methods persist and take steps to ensure these errors do not endure.

This article outlines the history of controlling and restricting cannabis research. We define two broad eras of cannabis research and compare research conducted during these periods. During the prohibition era (1961–2017), many researchers suggested decriminalizing cannabis would increase youth use, crime rates, and significant adverse mental health outcomes. While not universal, research in this era often appeared to support longstanding tropes and unconscious biases. These findings have not been replicated in what we call the emerging post-prohibition research era (2018–present). We argue the variances between findings in prohibition and post-prohibition eras of research can be traced to structural realities, methodological choices, and individual interests. In the past, ethical quandaries were resolved in ways that benefited academic researchers and harmed those being studied. Reframing ethical obligations in research requires a fulsome engagement with researchers’ goals, assumptions, and practices.

We present one route, which we define as a duty of care. Adopting a duty of care in cannabis research would oblige researchers to establish robust research designs, employ careful analytic strategies, and acknowledge limitations in more detail. While consistent with the Belmont Report and the principle of beneficence, this duty demands recognizing how research may be manipulated and misconstrued to continue criminalizing, stigmatizing, and pathologizing people who use cannabis. Institutionally, it may require funders, journal editors, and peer reviewers to engage more fully with ethical concerns about cannabis research and demand that extraordinary claims about the harms of cannabis use be based on clear and compelling evidence.

Controlling cannabis and restraining research

Efforts to control cannabis are more than 200 years old (Beweley-Taylor et al., 2014). In the 20th century, many national developments were informed by international developments. For example, during the Second Opium Conference in 1924, Mohamed El Guindy, a delegate from Egypt, proposed the inclusion of cannabis within the Convention, supported by Brazil, Greece, South Africa, ⁴ and Turkey. ⁵ El Guindy painted a worrying picture of the adverse effects of hashish and led to international concerns that cannabis led to addiction and insanity. Western cannabis policy was initially influenced by non-Western discourses (Beweley-Taylor et al., 2014; Campos, 2012; Shamir and Hacker, 2001). These views were adopted by politicians and practitioners and later refashioned to focus on racial and ethnic “threats” (Wheeldon and Heidt, 2022b).

In the U.S., anti-drug campaigners of this period warned against the encroaching “Marijuana Menace” and linked the use of cannabis with violence, crime, and other socially deviant behaviors. It was quickly linked to so-called racially inferior underclass communities (Tosh, 2019). As a result, many Americans became persuaded that cannabis was foreign (Halpern, 2018), evil (Vitiello, 2021: 449), or both. By the time the global prohibition regime emerged in earnest in the 1960s, cannabis was widely seen as dangerous and in need of international control (Collins, 2021). This took on different dimensions. In the U.S., cannabis was linked to immigrants, insanity, Mexicans, Black people, crime, anti-war protesters, and hippies internally. Externally, cannabis is considered an example of the colonization of drug control policy, which refers to:

. . .the use of drug control by states in Europe and America to advance and sustain the systematic exploitation of people, land and resources and the racialized hierarchies, which were established under colonial control and continue to dominate today. (Daniels et al., 2021: 1)

This was achieved, in part, by linking the dangers of distinct drugs and pursuing widespread prohibition through international agreements.

By the 1970s, the overall architecture of the current drug policy was established. As Newhart and Dolphin (2019: 51) detail, global policy developed “under the guidance and recommendations of experts at the World Health Organization (WHO).” Although regulation was formalized at the international level through the Single Convention on Narcotic Drugs in 1961, the amendments in 1972 introduced the concept of drug scheduling involving three domains. These include (1) accepted medical use, (2) abuse potential, and (3) potential for public health harms through risk to safety or risk of dependence. Failing to distinguish between opium, cocaine, and cannabis ensured a common approach to their regulation (Collins, 2021; Mills, 2003; Sinha, 2001). In 2017, the U.S. Committee on the Health Effects of Marijuana of the National Academies of Sciences (the HEMNAS Committee) noted that despite significant changes in state policy and the increasing prevalence of cannabis use, cannabis’ legal status limited research as a result of restrictive policies and regulations. ⁶

The “global drug prohibition regime” (Collins, 2021) that emerged ensured was designed to uncover any and all potential harms of cannabis use while constraining research suggesting benefits (Newhart and Dolphin, 2019). As we detail below, efforts to study any potential benefits were limited in some cases and prevented in others. Scholars publicly noted that U.S. state and Federal laws prevented scientific investigation of cannabis and its uses (Mikuriya, 1969). By 1975 they could point to reports from four different countries on cannabis, and criminal justice policy suggested the continued criminalization of cannabis created more problems than it solved. As we have explored, these include the Indian Hemp Drugs Commission (IHDC, 1894), the Panama Zone Report (1925), the La Guardia Report (New York Academy of Medicine (NYAM), 1944), the Wootton Report on Cannabis (1969), and the Le Dain Commission final report (1973). Although little evidence emerged from key governmental reports to justify the continued criminalization of cannabis, criminal justice systems worldwide embraced increasingly punitive policies (Wheeldon and Heidt, 2022b).

Cannabis research that can inform public health and keep pace with changes in cannabis policy and patterns of use requires funding. In the U.S., the National Institute for Health (NIH) is responsible for funding research across many health domains. However, because cannabis was historically perceived to have only negative effects, most cannabis research has been conducted under the auspices of the National Institute of Drug Abuse (NIDA). This research focuses on studying factors related to substance abuse and dependence and the adverse health outcomes and behavioral consequences associated with cannabis use. As Newhart and Dolphin (2019: 27) note:

Scientific discourse has trickled down to set the terms of popular culture discourse in many respects. Topics about physical harm, mental harm, gateway theories, and amotivational effects dominated media and common understandings of cannabis among the public. . .

Many misunderstandings persist.

Despite the changing status of cannabis, present research in this area continues to be funded by organizations that view the substance in purely negative terms. These structural constraints adversely impacted cannabis research. As cannabis prohibition has receded, research has increasingly focused on the benefits of cannabis legalization and the adverse consequences of criminalization. To better understand these dynamics, below we summarize some examples of cannabis research conducted before and after 2018. Considering cannabis research through these eras may require some justification.

On eras of cannabis research

Efforts to restrict the use of cannabis cannot be easily divorced from the historical antecedents related to colonialism, racism, and xenophobia (Wheeldon and Heidt, 2022b). Despite periods of liberalization on cannabis policy, such as in the 1960s, early to mid-1970s, and the 1990s (Dufton, 2017), today, in many parts of the world, cannabis remains illegal. Thus, efforts to define a post-prohibition cannabis policy era remain complicated by overlapping and discontinuous timelines, geographies, and policy choices. It is essential to acknowledge the potential for ethnocentrism, in which changes in the global north are presumed to be an augur of things to come everywhere else. While cannabis decriminalization appears to be the most likely scenario, suggesting a post-prohibition cannabis policy era is upon us is presumptuous and problematic. Persistent obstacles include resilient illicit markets, supply shortages in some cases, and oversupply in others. A significant challenge is banking, access to capital, and commercial point-of-sale protections (Wheeldon and Heidt, 2022c).

Just as defining cannabis policy eras is difficult, organizing cannabis research into distinct eras may suffer from similar challenges. Indeed, seeking conceptual clarity may sacrifice some historical coherence. However, we assume drawing a distinction between prohibition and post prohibition cannabis research is valuable. Indeed, it may be essential given that defining policy eras is intimately connected to different periods of cannabis research. For example, both the prohibition research and policy era are tied to the United Nations (UN) 1961 Single Convention on Narcotic Drugs, which introduced the concept of scheduling and continues to inform the complex relationships between funding, research, and acceptance (Newhart and Dolphin, 2019; Nutt, 2022). As a result, research was constrained in a variety of ways. Limited by dubious theoretical antecedents, insufficient data, and inappropriate methodologies, it often served political interests based on prohibition mindsets.

However, by the end of 2018, policy changes ushered in a new research environment. Four developments emerged that altered how cannabis was studied. First, after 5 years of legal cannabis, reports from Colorado, Washington, and Uruguay had begun to document the impact of legal cannabis. Early reports suggested the dire warnings of the prohibitionists were exaggerated. Second, Canada legalized the recreational use of cannabis in 2018. Although the second country, after Uruguay, to do so, the Canadian approach expanded access to legal cannabis, allowed provincial variation, and established new country-level approaches to studying cannabis (Health Canada, 2016). A third development is increasing international pressure to de-penalize cannabis possession. Indeed, following international meetings in 2018, the UN formally called on member states to promote “alternatives to conviction and punishment in appropriate cases, including decriminalizing drug possession for personal use.” ⁷ Fourth and finally was a resurgent interest in qualitative research to explore the perceptions and experiences of people who use cannabis.

Qualitative researchers, skeptical of large data sets devoid of depth and meaning, used cannabis’ shifting status to explore the role of cannabis in the lives of people. By focusing on the consequences of policies designed to criminalize, stigmatize, and prohibit its use, studies from around the world (Agoff et al., 2022; Feltmann et al., 2021; Nelson, 2021; Wanke et al., 2022), offer profound insights. These include how changes in cannabis’ legal status have led to blended, shifting, and often paradoxical policies that co-exist alongside each other. This shapes interactions among people in jurisdictions in which cannabis is legal and regulated, places where it has been decriminalization and de-emphasized, and countries where it remains prohibited and punished. Prohibition as a policy option seems to be waning based on changing public attitudes about drug use.

While the Netherlands began reforming cannabis policies in the 1970s, that approach serves as an uneasy example of tolerance for cannabis (Korf, 2020). In recent years, other examples of cannabis liberalization have occurred around the world. Court decisions have provoked legislative and policy changes in Argentina, Germany, and Italy, and specific laws led to cannabis decriminalization in 20 countries, from Belize to Switzerland (Seddon and Floodgate, 2020: 40–43; 53–56). Legalization and regulation are now a fixture in half of U.S. states and emerging in countries like Germany, Malta, Mexico, South Africa, and Thailand. While a new era of cannabis policymaking may be upon us, old ideas about the dangers of cannabis persist. These can be linked with research from the prohibition-era and findings which are difficult to dislodge in the minds of policymakers and the public.

Prohibition-era research findings (1961–2017)

Cannabis is globally the most used psychoactive substance under international control. The UN Office of Drug Control (UNDOC) reports cannabis enforcement is undertaken in almost all countries worldwide. In the period 2010–2019, sanctions were reported by 151 countries, covering 97% of the global population. ⁸ One important finding relates to Cannabis Use Disorder (CUD). The Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-V) combines two older stand-alone concepts related to cannabis abuse and cannabis dependence (Jutras-Aswad et al., 2019). ⁹ In a recent book, noted drug policy expert, Alison Ritter (2021: 21) suggests that “. . .between 10% and 30% of people who consume cannabis will develop a cannabis use disorder.” In support of this claim, she cites two studies (Hasin et al., 2015; Wagner and Anthony, 2002), suggesting they are authoritative. These studies haven’t aged well.

Wagner and Anthony (2002) focused on the transition from drug use to drug dependence. They presented evidence on risk based on the initiation of cannabis, cocaine, and alcohol use and risks for progression from first drug use to the onset of drug dependence. The data for this study were gathered for the National Comorbidity Survey (NCS), and drug dependence was assessed according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R) diagnosis criteria, published in 1987. In studies like this, the analysis works backward from the most severe drugs to the least severe drugs and infers that one came first (Andresen, 2012). This approach has been used in an analysis of Australian high school students, where the increased frequency of cannabis use supposedly led to increased rates of illicit and licit (tobacco) drug uptake (Swift et al., 2012). In France, Mayet et al. (2012) concluded that cannabis use significantly increased the risk of illicit and licit (tobacco and alcohol) uptake. Some suggest cannabis is a trigger for cocaine uptake (O’Brien et al., 2012).

Another oft-cited study (Hasin et al., 2015) presented nationally representative data on the past-year prevalence rates of cannabis use, cannabis use disorder, and cannabis use disorder among adults who used cannabis in the U.S. between 2001–2002 and 2012–2013. They conclude cannabis use “doubled over a decade,” and . . . “there was a large increase in marijuana use disorders during that time” (Hasin et al., 2015: 1236). There are two interesting findings here, given that several states and countries have now had legal recreational cannabis for over 5 years. The first is the question of the doubling of cannabis use in a decade, which the study authors note contrasts with numerous other studies (Hasin et al., 2015: 1241). This is a problem, presumably, because increased cannabis liberalization is assumed to lead to increased availability for youth. While a paper by Shi et al. (2015) suggested that cannabis liberalization is significantly associated with higher odds of adolescent cannabis use, Stevens (2019) shows that the paper relies upon omitting inconvenient outliers and insufficient predictive models. Consistent findings where cannabis is legal suggest no significant increase in adolescent cannabis use (Dilley et al., 2019; Haines-Saah and Fischer, 2021; Nguyen and Mital, 2022).

The second claim is that while not all people who use cannabis experience problems, “. . .nearly 3 of 10 marijuana users manifested a marijuana use disorder in 2012–2013” (Hasin et al., 2015: 1326). However, subsequent studies have yet to establish a massive influx of CUD diagnoses in areas with liberal cannabis policies (Mennis and Stahler, 2020; Mennis et al., 2021; Philbin et al., 2019). Cannabis researchers during the prohibition era made other alarming claims. Some suggested youth who use cannabis 20 times or more have 1.5 times the risk of being arrested for a property crime and 1.8 times the risk of self-reported property crime by midlife compared to light/non-users (Green et al., 2010). Pedersen and Skardhamar (2010) suggested that cannabis use at ages 15 and 20 years predicted subsequent offending. A common worry is “heavy” cannabis use by young people.

Heavy adolescent cannabis use has been associated with an increased risk of being poor, unmarried, and experiencing heightened anxiety in midlife. Other researchers have gone further, suggesting that “cannabis use predicts subsequent violent offending, suggesting a possible causal effect (our emphasis) . . .” (Schoeler et al., 2016: 1663). In a section that appears especially problematic in 2023, researchers suggested, without explanation, that heavy cannabis use seemed to be a specific risk for Black urban youth, setting them “. . .on a long-term trajectory of disadvantage that persists into midlife” (Green et al., 2010: 567). The confluence of violence, race, and cannabis has seeped into public consciousness in ways that cannot be justified by even a cursory examination of studies used to justify cannabis prohibition on public health and public safety grounds.

Of course, not all research in this era conformed to simplistic categorization. Qualitative research has long suggested a number of problems with studies such as these. Incorporating the insights of those who use drugs and are familiar with the subculture(s) is not new (Becker, 1963; Young, 1971). As jurisdictions liberalized their cannabis policies, new approaches emerged. For example, Sifaneck and Kaplan (1995) explored coffeeshop culture in the Netherlands, where cannabis was informally decriminalized. By creating a culture of responsible use and self-regulation, cannabis policies have not led to more people misusing cannabis. Moreover, these policies have assisted people who use other drugs. Sifaneck and Kaplan (1995: 500) describe this process as one in which: “. . .cannabis served as a means of breaking the cycle of hard-drug use and addiction.”

Other international work offered important nuance and highlighted some of the complications with how cannabis was framed (Sandberg, 2012a; Sandberg, 2012b) and considered cannabis culture (Sandberg, 2013). Prioritizing people who use cannabis through research began to challenge assumptions about cannabis and risk. For example, Dahl (2015) conducted interviews with 25 young Norwegian adults and suggested that people who use cannabis held little in common with those who used other illicit substances. Based on in-depth interviews with people who recreationally use cannabis in Alberta, Canada, Osborne and Fogel (2017) found the vast majority supported legalization. They viewed cannabis prohibition as unjust and wasteful, resulted in missed economic opportunities, and perpetuated crime and violence from the drug trade. Similar to Sifaneck and Kaplan’s (1995) findings, respondents in this study did not see cannabis legalization as leading to more use of other illegal substances, such as cocaine or heroin. Research like this provided important nuance and has shaped the nascent post-prohibition cannabis era.

Studying legal cannabis (2018–Present)

Starting in 2015, three years since the first U.S. states legalized cannabis, reports from Washington (ADAI, 2017), Colorado (Monte et al., 2015), and Uruguay (Walsh and Ramsey, 2016) began to document the benefits of cannabis legalization. This research took several years to develop, validate, and publish. The result of this research consistently paints a very different picture from the risk-laden framing of the cannabis prohibition era. This has led to what some suggest is an emerging post-prohibition cannabis research era. On this view, post-prohibition is a conceptual category for research that is self-reflexive about the entanglements among cannabis, prohibition, and legalization. Even where legal, cannabis policy is often guided by prohibitionist mindsets (Corva and Meisel, 2021; Heidt and Wheeldon, 2022a). Since 2018, cannabis liberalization has allowed researchers to test previous findings. This includes examining how changes in legal status impacted cannabis use, property, and violent crime, cannabis use disorders, and treatment admissions for schizophrenia. The results do not support past findings.

For example, in jurisdictions with legal cannabis, the number of people using cannabis has not significantly increased (Grucza et al., 2018; Hawke and Henderson, 2021; Johnson et al., 2019; Rotermann, 2020; Ta et al., 2019). In addition to use, another measure of interest is treatment admissions for cannabis use disorder (CUD). Using data from the 2002 to 2019 National Survey on Drug Use and Health, 43,307 individuals who met past-year DSM-5-proxy CUD criteria were identified. Askari et al. (2021) report between 2002 and 2019, 6.1% of people reported using a CUD treatment. However, treatment use decreased by more than 50% during this time. Based on data from 2004 to 2012, researchers found no increase in treatment admissions for problem cannabis use in states with liberal cannabis policies (Mennis and Stahler, 2020; Philbin et al., 2019). In stark contrast to the repeated claims that up to 30% of people who use cannabis will develop a CUD (Ritter, 2021), in 2020, the Substance Abuse and Mental Health Services Administration (SAMHSA) reported that just over 5% of people, 12 and older, who use cannabis had a cannabis use disorder. ¹⁰

Other studies where cannabis was legalized have now repeatedly determined that cannabis liberalization policies have no impact on crime or public disorder (Dragone et al., 2019; Huber et al., 2016; Hunt et al., 2018; Morris, 2018). For example, Lu et al. (2021) used a quasi-experimental, multi-group interrupted time-series design to determine if and how Uniform Crime Reports (UCR) crime rates in Colorado and Washington were influenced by legalization. They concluded cannabis legalization and sales have had minimal to no effect on violent or property crimes in Colorado or Washington. In Canada, the implementation of the Cannabis Act in Canada (2018) is associated with a decrease of between 55% and 65% in cannabis-related crimes among youth (Callaghan et al., 2021). A specific example may be useful.

As introduced above, concerns about cannabis, psychosis, and schizophrenia appeared in the National Academy of Science, Engineering, and Medicine’s (NASEM) study in 2017. They concluded, based on their review, that:

. . .cannabis use is likely to increase the risk of developing schizophrenia, other psychoses, and social anxiety disorders, and to a lesser extent depression. . . . Heavy cannabis users are more likely to report thoughts of suicide than non-users, and in individuals with bipolar disorder, near-daily cannabis users show increased symptoms of the disorder than non-users. ¹¹

As discussed, these studies emerged from an era of cannabis prohibition. As a result, these studies suffer from numerous methodological problems (Heidt and Wheeldon, 2022a). Since liberalization, the longstanding notion that use is a causal factor for schizophrenia is not only unsupported (Hamilton and Monaghan, 2019) but has been turned on its head. People with schizophrenia are more likely to develop a CUD (Ahmed et al., 2021), not the other way around. Researchers long associated with studies concluding that cannabis use represented significant dangers to those who consumed it now explicitly state that cannabis use is neither necessary nor sufficient to cause psychosis (D’Souza et al., 2022).

This is not to say that there is no reason to continue to study cannabis and its impact on the people who use it. New products and techniques of consumption require study, as do the socio-medico-legal realities which shape cannabis use in different countries, cities, and towns. Beyond the purview of this article is interest in and research on medicinal cannabis, including therapeutic properties, benefits associated with harm reduction, and opiate replacement. While medical cannabis remains a paradox (Wheeldon and Heidt, 2022c), establishing a credible means to measure positive health outcomes, assess therapeutic effects, and catalog potential side effects is needed. Long constrained by limitations on cannabis research, restrictions have undermined evidence-based efforts to educate people about the responsible use of cannabis, including where it can enhance some experiences and complicate others. Future work in this area might seek to replicate this analysis to better understand the variance between prohibition-era and post-prohibition research on medicinal and therapeutic cannabis.

Ethics and variances in cannabis research pre/post legalization

To understand the ethics of cannabis research, it is useful to look at what gets published and where. Between 1829 and 2021, 29,802 journal articles were published on cannabis (Ng and Chang, 2022). Most of these papers focus on the plant’s chemistry, pharmacology, molecular biology, and genetics, including forensic analysis, molecular markers, and traceability (Liu et al., 2021; Matielo et al., 2018). An increasing number of these articles were published during the last 30 years and focus on cannabis, addiction, and crime. The journals which have published the largest number of journal articles on cannabis are Drug and Alcohol Dependence and Addictive Behaviors, and Ng and Chang (2022: 29) note most are published by US-based researchers, funded by, and affiliated with, the NIH and the National Institute on Drug Abuse (NIDA).

In a recent analysis of the 100 most cited publications focused on drug use and addiction, Zurián et al. (2021) note that only a few publications provide most of the citations. They note that an ongoing challenge is the concentration of publications from the U.S. and the intersection of institutions, funding agencies, and individual authors. In their analysis, the most cited articles were associated with funding institutions including the National Institute on Drug Abuse, the National Institute on Alcohol Abuse and Alcoholism, and the National Institute of Mental Health (Zurián et al., 2021). Other funding agencies, including the Medical Research Council and the Wellcome Trust, both from the UK, were also referenced. In criminology, the ways in which funding agencies can influence research is not in dispute (Savelsberg et al., 2004). Its application to cannabis research is worthy of more attention, given the potential that this research can shape policy in ways that undermine liberty, justify intrusions, and perpetuate systems of social control.

During the cannabis prohibition era (1961–2017), most cannabis research in criminology was characterized by three features. First, the research tended to be quantitative. An early example includes an empirical study of people who use drugs and compared crime rates between Muslims and Jewish people living in Israel (Drapkin and Landau, 1966). In the 1970s, cannabis use was used as an example of delinquent characteristics of youth (Mott, 1973), and by the 1980s, Keane et al. (1989) selected cannabis use as a means to operationalize deviance in journals like the British Journal of Criminology. An example of the focus on quantitative analysis in this area is the journal Addictive Behavior. A popular outlet for addiction research, it expressly requires that submissions include experimental designs and quantitative measures of assessment and evaluation as a “requisite for acceptance.” ¹² Since the 1990s, publications based on secondary analysis of large data sets have become more common. In addition, case-control studies that attempt to draw generalizable inferences based on very small samples gained prominence (NASEM, 2017).

A second feature of cannabis research in this era was that most suffered from definitional issues. This was first noted in two book reviews (Kaiser, 1969; Leech, 1969) which itemized numerous problems with definitions and descriptions of illicit drugs. These substances, including cannabis, were inaccurately defined and insufficiently explained. This resulted in confusion when operationalizing variables, especially related to dose and variables such as frequency of use and adverse outcomes. These definitional issues sometimes led to questionable ethical decisions. For example, in the prestigious journal Criminology, Jensen and Brownfield (1983: 547) published a study based on survey data from a school in which:

Several influential parents believed that the school studied had a particularly serious drug problem, and the student body was actually under secret surveillance by the police (our emphasis) for a period of time. A student organization at the school received permission from the superintendent of the district to conduct a survey under the direction of University of Arizona faculty in order to assess the opinions, perceptions and reported behavior of the student body regarding drug use.

Children were asked to:

. . .“circle all of those drugs you think your parents have used,” including tobacco (pipe, cigarettes, etc.), alcohol (beer, wine, liquor), marijuana (grass), hard narcotics (like heroin), pep pills, psychedelic drugs such as cocaine (“coke”), tranquilizers, and sleeping pills.

Of specific interest was cannabis.

To explore whether perceived parental drug use undermined social bonds in children and led to future drug use, researchers relied on basic correlational analysis. Based on a review of the analysis tables in the published article, “perceived cannabis use by parents” was not associated with any measure of adverse attachment, nor was it a significant correlate of children’s cannabis use. Nevertheless, the authors concluded that attachment to parents who don’t use drugs (called “straight parents” in the article) was a significant barrier to student drug use. This conclusion was made possible by using a “total drug score,” which combined perceived parental use of cannabis, alcohol, cocaine, and other narcotics (Jensen and Brownfield, 1983: 549). In the same journal, five years later, researchers suggested cannabis was one of several narcotics which contributed to a so-called addiction-fueled crime wave (Nurco et al., 1988). Funded by NIDA, the authors conclude policymakers ought to consider how to “intervene at the earliest signs of deviance. . . to control factors that increase the likelihood of later criminal involvement . . .” (Nurco et al., 1988: 420).

Third, this research suffers from a modesty deficit. The role of cannabis’ legal status, the related impacts on research design, and the challenges within existing data sets were rarely considered when reporting findings. Instead, researchers described alarming results with little context. Near the end of the prohibition era of research, Pacula and Sevigny (2014: 209) observed researchers would need “a bit more time and a lot better data” to answer the most important cannabis questions. They suggested:

Existing policy experiments have not been used to answer what we see as the most important questions, namely are public health harms from marijuana use a function of the person consuming it (age, polysubstance user, or other identifying factor)? Amount consumed? Activities engaged in while under the influence? Method of consumption? Potency? Or duration of use?

These concerns remain important.

Contrast this with the aims of the journal responsible for the most articles on the adverse impact of cannabis, namely Drug and Alcohol Dependence. The journal is focused on substance abuse treatment and prevention and:

. . .serves as an interface among governmental, industrial and academic communities maintaining liaisons with regulatory and research agencies as well as educational, treatment, and prevention facilities in the drug abuse field. It also functions as a collaborating center of the World Health Organization. ¹³

While not all researchers who published articles in this journal can be said to embrace prohibition, the assumption that cannabis represents a significant risk appears to be baked into the journal’s aims and focus. The failure to consider the serious and significant complexities related to cannabis research has complicated research, policy, and practice. Although these are a function of the long shadow of prohibitionist ideas, research findings that now comprise the cannabis “evidence base” cannot be divorced from the structural arrangements related to funding (Newhart and Dolphin, 2019), methodological design (Wheeldon and Heidt, 2022b), and individual careerist choices in the neo-liberal academy (Wheeldon et al., 2014). Each deserves more attention and specific analysis.

Structural realities: Legal status, funding, and research decisions

The Controlled Substance Act (CSA) defines Schedule I substances as those considered the most dangerous and addictive, with a “high potential for abuse” with “no currently accepted medical use in treatment in the United States.” ¹⁴ Cannabis research remains constrained by its Schedule I status. Since the 1961 Single Convention on Narcotic Drugs, which attempted to control cannabis through coordinated international action (United Nations, 1962), funding for cannabis research was organized through global drug prohibition. As Newhart and Dolphin (2019: 26–27) noted:

. . .scientific research on cannabis has been skewed, because research whose aim is to prove harm has been well funded, but research meant to discount harm or show benefits has been successfully delayed or blocked and has not been eligible for financial support from most funding.

Cannabis research during the prohibition era is a function of the number of studies conducted that specifically explore the adverse effects of cannabis. For example, in the U.S., the HEMNAS Committee summarized several NIDA reports. They state:

. . . less than one-fifth of cannabinoid research funded by NIDA in fiscal year 2015 concerns the therapeutic properties of cannabinoids. . . Moreover, because NIDA funded the majority of all the National Institutes of Health (NIH)-sponsored cannabinoid research in fiscal year 2015. . . its focus on the consequences of drug use and addiction constitutes an impediment to research on the potential beneficial health effects of cannabis and cannabinoids. ¹⁵

Between 2000 and 2018, cannabis research funding by the U.S. National Institute on Drug Abuse (NIDA) focused almost exclusively on cannabis misuse and its adverse effects (O’Grady, 2020). ¹⁶ Researchers funded on these terms might consider to what extent they surrender their ethical objectivity and allow the perceived risks of cannabis use to frame their research.

Cannabis’ legal status is relevant to ethics in another way. In an era of cannabis prohibition, psychopharmacological risks were conflated with the risks associated with engaging with illegal markets (Goldstein, 1985). Indeed, most harms related to cannabis arise from its illegality (Fischer et al., 2021) and the associated risks of seeking out an illicit substance. For example, criminalizing cannabis has led to territorial disputes, the constant fear of informants, and implicated people who use cannabis with other criminal behaviors (Kaplan, 1970). Research that focuses on those engaged in criminal activity requires that ethical principles of confidentiality and anonymity be observed. This tends to favor observational studies based on large, anonymized data sets. Qualitative cannabis research is certainly not unknown (Wheeldon and Heidt, 2022c). However, it is more difficult to secure permission for this kind of research from institutional review boards (IRBs) in many countries. In the U.S., for example, any research that might disclose criminal behavior requires additional approvals from the NIH.

In addition to administrative and bureaucratic hurdles, cannabis’s legal status and scheduling created profound research challenges. Best (2001) describes one difficulty known as “defining.” Problems arise when researchers attempt to define social problems, especially when formulating concepts for studies about people who use drugs (Heidt and Wheeldon, 2022b). Mining large data sets for convenient correlates requires accepting some ethical elasticity. One problem is defining cannabis itself. Most observational studies still fail to define cannabis strains, quality, dose, or consumption type. ¹⁷ However, efforts to operationalize relevant variables have suffered in other, more specific ways.

Methodological challenges: Defining and operationalizing variables

To establish that cannabis causes problem cannabis use, one must define “cannabis use” and “problem cannabis use.” This has complicated efforts to demarcate cannabis use within the Diagnostic and Statistical Manual of Mental Disorders (DSM). The DSM–V-R has revised past formulations, and cannabis use disorder is now defined by nine pathological patterns. Although problem cannabis use has undergone revision, few researchers consider how changing definitions impact past conclusions (Leung et al., 2020; Ritter, 2021). For example, Wagner and Anthony (2002) assess drug dependence according to the DSM-III-R diagnosis criteria, published in 1997. However, cannabis dependence is not a category retained in later revisions. ¹⁸ Likewise, Hasin et al. (2015) used the concept of “marijuana use disorder” from the DSM-IV, which has since been abandoned. ¹⁹ Those who continue to cite these studies are ethically obliged to note these findings are based on conceptual categories that no longer exist. Unfortunately, few choose to do so.

As we have noted (Heidt and Wheeldon, 2022a), another issue concerns how to group people who use cannabis. This varies widely and appears disconnected from real-world cannabis use. For example, so-called heavy cannabis users include people smoking more than five joints a day (Yücel et al., 2008), consuming cannabis ten times a month (Cousijn et al., 2011), 500 times a year (Pope and Yurgelun-Todd, 1996), daily for 2 years (Degenhardt et al., 2003), or a joint a day for more than 20 years (Gracie and Hancox, 2021). Another problem is how to define “occasional users” for the purposes of comparison. Ponto (2006) suggests that because it would be unethical to introduce a drug to non-users, studies generally make comparisons between occasional and heavy users resulting in variations between control and experimental groups. In other words, the definition of “cannabis,” “heavy users,” and “occasional users” are widely inconsistent across most cannabis studies. Meta-analyses focused on cannabis use, and adverse outcomes in this area should be approached with caution.

The wide variety of definitions associated with heavy cannabis use renders it essentially meaningless as a category across different studies. Illustrated in the examples above, people who use cannabis “heavily” could include people who smoke less than daily. As a result, research focusing on heavy use may consist of people who had one puff of homegrown ditch weed today, alongside people who consume massive amounts of THC through dabbing cannabis concentrates such as amber, budder, butane hash oil (BHO), shatter, or wax daily. Studies that fail to define their terms properly perpetuate problematic classifications and normalize the use of deficient designs. Despite the obvious problems with much of the cannabis research, researchers often fail to report the non-generalizability of findings in clear and unambiguous ways.

Individual considerations: Causation, careerism, and modesty

Most cannabis research in the prohibition era was correlational. Using ecological correlations to draw causal, individual-level conclusions may amount to research malpractice. Despite the confident suggestions to the contrary, causation is very difficult to establish. Gage et al. (2013) identified various residual confounding variables and noted that drug use is often associated with other risk factors. Applied to psychosis, these include traits or contextual factors that occur earlier in life, such as certain personality types, early-life trauma, and family adversity. Other confounding variables include how to capture periods of non-use amongst people who use cannabis (Kroon et al., 2021), early alcohol use, early sexual activity, poor school performance, abuse, and trauma (Ksir and Hart, 2016). The relationships between tobacco use and disorders on the schizophrenia spectrum characterized by psychosis and violent outbursts further complicate simplistic conclusions (Scott et al., 2018).

Poor research may not itself be unethical, of course. It is worrying when research decisions appear designed to benefit researchers, competing for funding, prestige, and job security. Cohen (2012) has referred to the “dark side of careerism” when discussing emergent ethical complications of academic publishing. For example, the recent research phenomenon of “salami slicing” is defined as the trend by which career-driven scholars create several publications out of material that could have been one article. This might be connected to the tendency to exaggerate the importance of research findings through creative wordplay. For example, the conclusion that “cannabis use predicts subsequent violent offending, suggesting a possible causal effect . . .” (Schoeler et al., 2016: 1663) requires “possible” to do quite a bit of conceptual work. It is an example of the careerist impulse rather than the inferential modesty that should guide researchers working in this area (Wheeldon et al., 2014).

It is especially dispiriting when such claims further the “reefer madness” approach of the past by reinforcing damaging tropes linking ethnicity, cannabis, mental illness, and violence (Earleywine, 2010). Responsible cannabis research would clarify the concerns about the quality of the data on cannabis, crime, and social harm (Farley and Orchowsky, 2019). As Andresen (2012) notes in his satire of prohibition-era cannabis research, adopting the approach taken by some researchers to infer causal relationships would result in the conclusion that mother’s milk and infant formula are the “ultimate gateway drugs.” Does a deliberate choice to underplay research limitations while emphasizing findings that are politically convenient amount to an ethical lapse worthy of institutional attention? If so, it must doubly apply when research furthers public safety approaches to cannabis that exacerbate historic harms to vulnerable populations.

Criminal justice ethics, harm reduction, and a duty of care

The criminal justice system wields awesome powers to deprive individuals of their life, liberty, and a host of associated freedoms. While focusing on the ethical orientation of practitioners within the system is important (Bjerregaard and Lord, 2004; Stinson et al., 2016; Westmarland and Conway, 2020), the integrity of the system itself is complicated by the number of cases of criminal activity among criminal justice personnel. ²⁰ Criminalizing cannabis has long been criticized for resulting in a long list of moral outrages. Central among these are the ethical implications of the institutional complicity in decades of harm to young Americans, alienating them from society while justifying police intrusions and the deterioration of constitutional values (Kaplan, 1970).

Seddon and Floodgate (2020: 8–9) summarize the ethical problems of relying on public safety and cannabis prohibition in economic, social, and political terms. These include:

• Facilitating a large and untaxed income stream for groups and individuals involved in crime.

• Criminalizing millions of young people for behavior that, for most, is short-lived and causes minimal social harm.

• Enforcement activities fall most heavily on marginalized and excluded communities, undermining social justice.

• Enforcement is disproportionately targeted at minority ethnic groups, driving racial injustice.

• Incentivizing the creation of more potent and more dangerous synthetic cannabinoids.

In forthcoming work, we argue the ethical lapses which have guided cannabis policy require efforts to reduce the harm perpetuated on people who use cannabis by the system and those who work within it (Heidt and Wheeldon, 2023). This includes cannabis researchers (Solomon, 2020).

Harm reduction can be described as a pragmatic and compassionate set of strategies designed to reduce the harms associated with certain behaviors (Marlatt, 1998). This may include policies, programs, or practices that aim to minimize adverse health, social and legal impacts associated with certain behaviors, often associated with illicit drug use. ²¹ While reducing the potential harm for those participating in research is a central goal of ethical research, harm reduction, as a guiding philosophy, may offer a means to rethink why the many myths about cannabis persist. More than a century of evidence suggesting cannabis is relatively innocuous compared to legal drugs like alcohol and tobacco (Wheeldon and Heidt, 2022b). Cannabis does not appear to be a driver for mental health conditions, including psychosis (D’Souza et al., 2022; Johnstad, 2022).

Combating decades of myths may take years. Efforts to control cannabis are often framed based on the mental health risks represented by its use. Despite a lack of credible evidence establishing the scope of the risk, persistent worries about cannabis use can be connected to the role of medicalization, which pathologizes behaviors deemed unpalatable. Once a behavior is pathologized, it becomes even more difficult to confront (Newhart and Dolphin, 2019). Indeed, any defiance amounts to proof that the underlying behavior is problematic. The persistence of ideas, despite mountains of contrary evidence, has been described by John Roman as zombie theories. ²² It generally applies to simplistic and often binary views of criminal behavior based on the medico-psychological model of crime. For example:

Criminal behavior is primarily the result of long-standing criminal predispositions and psychopathologies that cause individuals to offend. Research should be invested in programs to prevent the development of criminality, which can be changed through appropriate treatments, and, these changes once made would then persist. ²³

The idea that cannabis causes criminal behavior is a “zombie theory” in that it refuses to die, shuffling through the pages of peer review research, feasting on the brains of naïve researchers and mercurial academics even after being repudiated. ²⁴ For example, even when dubious correlational research connecting cannabis liberalization to increase use (Shi et al., 2015) was corrected (Stevens, 2019), the flawed research continues to be cited at twice the rate of refutation. ²⁵ While there is increasing interest in correcting past research (Munsch, 2018; Zoorob, 2020), retraction is rare. Researchers may wish to consider the consequences when their work is cited to support absurd, invasive, and otherwise blatantly unconstitutional laws. ²⁶ The best way to reset the cannabis research agenda may be to take steps to improve publication standards and for researchers to adopt a duty of care. This could more responsibly guide post-prohibition research.

The Belmont report and a research duty of care

Consistent with calls to expand harm reduction in drug policy (Ritter, 2021), we present “a duty of care in cannabis research.” This duty of care is based on past formulations ²⁷ but specifically considers cannabis research and those responsible for its publication. This, of course, is not a legal obligation. It can be understood as researchers taking personal responsibility for the injury their flawed research might do and publishers considering the damage wrought by decades of spurious correlations reported in august journals. It can be defined as a duty to look after vulnerable individuals and communities and protect them from harm that would likely result from poor research. Media ecosystems allow untruths to re-emerge again and again. For example, just days after Laura Ingraham connected mass shootings to legal cannabis, the Wall Street Journal ran a piece repeating many of the same myths. ²⁸

Expecting media to fairly report research when some researchers take methodological shortcuts is folly. One approach is to rethink the role of ethics in research on cannabis and other drugs. Despite the value and import of the Belmont Report on research guidelines, critics observe several persistent issues in integrating ethics within research. These include the disconnect between research and practice (Friesen et al., 2017), the failure to account for the unique harms to communities that can occur due to research (Emanuel and Weijer, 2005), and the unwillingness to require transparency between research funders and research findings (Horbach et al., 2022). These critiques are pertinent to cannabis research. Addressing the disconnect between research and practice requires expanding the kinds of research undertaken, engaging in robust research designs based on reliable data, and doing more to acknowledge research limitations.

However, change involves rethinking who is part of the conversation. Ritter (2021: 10) notes the growing recognition that people who use drugs should “. . . have a central position in drug policy formation.” This includes addressing equity and inclusion in cannabis research. Martin-Willett and Bidwell (2021) argue that given the failure to include communities of color in past studies, researchers should actively work toward improving equity in cannabis research. This requires inclusive recruitment, giving a voice to those likely to be impacted, and expanding qualitative research (Grigoropoulou and Small, 2022).

As Solomon (2020) has pointed out, cannabis researchers have failed to consider not only the aggregate of harms to individuals within a community but also the harms suffered by communities themselves (Weijer, 1999; Weijer and Anderson, 2002). Prohibition-era cannabis research justified policies resulting in adverse and disparate outcomes. Racialized policing patterns regarding cannabis cannot be credibly doubted (Baker and Goh, 2004; Belackova et al., 2017; Hughes and Stevens, 2010; Lammy, 2017; Logan, 2014; Owusu-Bempah and Luscombe, 2021; Shiner, 2015). In 2020, the mere suspicion of cannabis at a school near Hackney led to a shocking strip search of a Black 15-year-old student by London Metropolitan Police officers. ²⁹ In 2022, more than 40 Black people in Georgia were awarded $900,000 ³⁰ following a racially motivated and unconstitutional search for cannabis. ³¹

Previously, we argued that the lack of rigorous cannabis research allowed “New Prohibitionists” to compound longstanding myths using methodologically problematic analysis (Heidt and Wheeldon, 2022a). However, the nexus between the politics of cannabis research funding and published findings is hard to ignore. Cannabis is a reminder that research is shaped by funding patterns, and because of this, researchers may unconsciously adopt assumptions that are defined by the prevailing political perspectives of the time (Savelsberg et al., 2004). The failure to take seriously the ethical problems that arise when funding, directly or indirectly, shapes research findings can no longer be ignored. One specific challenge is around transparency (Piller, 2015).

Adopting a “duty of care” can address the research/practice by providing a means to address the dichotomy between research protections on the one hand and meaningful participation on the other. Embracing such a duty can promote a

. . .move away from protectionism and toward participation, from subjects to partners, is supported by both the principle of beneficence, since it may lead to greater positive health outcomes, and the principle of respect for persons, since many individuals desire to be more involved in research. (Friesen et al., 2017: 19)

This approach is underway by qualitative researchers engaging people who use cannabis, the costs of criminalization and the complexity of liberalization are increasing. Studies in Canada (Heidt, 2021), Norway (Sandberg, 2008; Sandberg, 2013), Sweden (Feltmann et al., 2021), Nigeria (Nelson, 2021), Mexico (Agoff et al., 2022), and Poland (Wanke et al., 2022) offer profound insights. Prioritizing those who use cannabis and other insiders can breathe life into broad administrative categories, better define how people use cannabis, and privilege how they describe the consequences, both positive and negative.

Embracing a duty of care can also serve as an antidote to the Belmont Report’s failure to consider the potential for harm to communities. It provides a specific application of the principle of “beneficence,” understood as an obligation to (1) do not harm and (2) maximize possible benefits and minimize possible harms. Indeed, the Belmont Report (National Commission, 1979: 2) states that “persons are treated in an ethical manner not only by respecting their decisions and protecting them from harm, but also by making efforts to secure their well-being.” Within the EU, this approach is consistent with the obligation that research demonstrates “respect for people and for human dignity. . . values, rights and interests of the research participants.” ³² Adopting a duty of care is consistent with calls to ensure researchers respect “. . . the values and interests of the community in research and, wherever possible, to protect the community from harms” (Weijer and Anderson, 2002: 192).

Embracing a duty of care may also serve to justify a renewed focus on transparency. This is important both to research funding (Newhart and Dolphin, 2019), reporting on clinical studies (Piller, 2015), and limitations associated with cannabis research itself (Heidt and Wheeldon, 2022a). Despite significant challenges with cannabis research in the prohibition era, many studies that report limitations tend to downplay their impact. Cohort studies relying on imprecise measures of the substance under investigation are not generalizable. This applies to studies that define groups in inconsistent and atheoretical ways, case-control studies that report findings that cannot be reproduced using other research designs, and metanalyses that fail to distinguish between changing definitions of so-called problem cannabis use within the DSM over the last 20 years.

For example, instead of euphemistically reporting that substantial heterogeneity within metanalyses poses interpretive challenges (Imrey, 2020), researchers might consider more direct language. Given the risk that these findings could be used to justify racially disparate, punitive, and invasive cannabis policies (Solomon, 2020), explicitly noting that these research findings are not generalizable is essential. We are not suggesting that scholars never engage in cannabis-based research. Likewise we do not advocate that only pro-cannabis research be published. However, much more care must be taken going forward. The damage inflicted on individuals and communities can no longer be ignored. A duty of care in cannabis research would require researchers to take additional care to define research terms in meaningful and consistent ways, adopt research designs that are appropriate to the area of study, and, most importantly, present research limitations clearly and honestly.

Change requires engaging journal editors and others who make decisions about publishing research. This means prioritizing researchers who make a good faith effort to ensure the integrity of their research. It also means promoting accountability that is swift, certain, and proportionate. It may be time for journal editors to consider temporary bans on manuscript submissions for researchers who continue to make grandiose and untenable claims about their findings and fail to consider the consequences should the findings suggested by their research be implemented. It may require that peer reviewers, editors, and publishers consider fundamental questions about submitted cannabis research. In addition to noting who funded a study, other questions matter. A non-exhaustive but perhaps useful list includes:

• Does the funder of the research have a view on the risks of consuming cannabis?

• How are the risks of cannabis framed? Are risks of using other substances noted?

• Does the methodology avoid problems linked to cannabis’ legal status?

• Do researchers define terms consistently and in line with credible research?

• Is the method of cannabis preparation, dose, and means of ingestion described?

• How does the study operationalize adverse outcomes?

• Are clinical designations current, and are any inferred clinical diagnoses validated?

• Do researchers credible outline the limitations of their research?

• Do researchers consider the policy implications of their research?

• Are the included limitations consistent with the study’s methodological gaps?

• Are the conclusions consistent with the data?

These obligations are not one-sided, of course. Indeed, post-prohibition cannabis research that concludes cannabis presents no risks to anyone, anywhere, or which fails to consider the threats of unbridled commercial cannabis is also a problem. These include the implications for cannabis workers, the environmental consequences of expanding legal cannabis, and the complexities around mixing alcohol and cannabis. Indeed, a variety of important law-based criminological questions are emerging (Fischer et al., 2021). To proceed, however, credible research requires recognizing the gaps in data and the limitations of their methodological designs. It means a clear acknowledgment in all published research that no study, on its own, can be conclusive.

Ethics and post-prohibition cannabis research programs

In the last few years, the recognition that we are in a post-prohibition environment has led to renewed efforts to study cannabis and consider responsible regulation (Decorte et al., 2020; Fischer et al., 2021; Oldfield et al., 2022; Ritter, 2021; Wheeldon and Heidt, 2023). If post-prohibition cannabis research is to proceed responsibly, we agree with Corva and Meisel (2021) that researchers must be self-reflexive about how the entanglement of prohibition, decriminalization, racism, legalization, commercial interests, and regulatory models all shape research. These appear to align and disperse in enigmatic ways. For example, in 2022, U.S. President Joe Biden signed an executive order to pardon citizens and lawful permanent residents convicted of simple cannabis possession under federal law and Washington, DC statute. This follows the passage by the U.S. Senate of a bill to expand the cultivation of cannabis and expedite access for researchers. ³³ If passed and signed into law, S. 253 will reverse decades of policy by requiring the Department of Health and Human Services (HHS) to investigate cannabis health benefits. ³⁴ This reversal is based in part on the ethical incongruity of imprisoning people for possessing something which is legal in an increasing number of U.S. states.

While Biden has done more than other U.S. Presidents to reform cannabis laws, his administration has failed to push the SAFE Banking Act in Congress, which would allow legal cannabis businesses to access financial support, investment, and protections. ³⁵ Despite a request that the Health and Human Services and the Justice Department review whether cannabis should remain as a Schedule 1 substance under the Controlled Substances Act, ³⁶ Biden has not yet announced support for specific bills that would expand cannabis research. This piecemeal approach stokes fears that cannabis decriminalization may occur by replacing public safety approaches with paternalistic public health models focused on risk and harm (Wheeldon and Heidt, 2022a).

If this occurs, governments may expand juridical-medical control by reifying prohibitionist assumptions that people who use cannabis are “sick” and in need of “treatment” (Szalavitz, 2015; Taylor et al., 2016). As we have recently observed:

A significant criminological concern is how shifting patterns in cannabis policy suggest reform but may be merely illusory (Cohen, 1979, 1985). The overly punitive character of contemporary penal practice has infected therapeutic relationships and correctional programming through coerced treatment models (Spivakovsky et al., 2018). Outdated notions about cannabis are evident in our communities, among drug policy scholars (and journals), and entrenched within drug treatment curricula. (Wheeldon and Heidt, 2022a: 3)

As the number of jurisdictions that have liberalized cannabis laws grows, researchers should be aware that the public safety/public health nexus frame around cases of cannabis use disorder is likely to persist (Ritter, 2021). For example, in Scotland, diversion is often framed as beneficial without confronting the potentially damaging consequences of pathologization and stigmatization (Price et al., 2021: 123–124).

This may represent issues for otherwise law-abiding consumers and people who use medical cannabis. It may be a particular worry for Black, Indigenous, and Other People of Color (BIPOC), given its potential as a pretextual means for police to stop, frisk, search, or otherwise interfere with citizens (Geller and Fagan, 2010). For example, there is evidence that despite legalization in Canada, the historic over-representation of Black and Indigenous people in arrest statistics (Owusu-Bempah and Luscombe, 2021) persists despite cannabis’ legal status. ³⁷ Wiese et al. (2023: 25) describe the experience of racialized communities as “overpoliced and underrepresented.” How race intersects with concerns about impaired driving (Pearlson et al., 2021), public consumption (Wheeldon and Heidt, 2022c), and safe cannabis storage (Wang et al., 2016) remains an essential question. The ethical implications of deferring to policies that continue to impact people of color disproportionately can no longer be doubted (Sheehan et al., 2021).

Caveats and conclusion

In this article, we considered cannabis research eras during and following prohibition by defining essential differences between these periods. By distinguishing between prohibition-era research (1961 and 2017) and post-prohibition research (2018–present), we have shown that the increasingly apparent divide based on cannabis’ legal status can be linked to structural, methodological, and ethical concerns. However, we concede that not everyone will be convinced by the distinctions we have drawn between these eras. Indeed, more work is needed to delineate these eras further and in greater detail. For example, our conceptual model would benefit from empirical assessment, perhaps based on quantitative meta-analyses of cannabis research to compare findings in the eras we have defined. To further explore cannabis research eras, it may be useful to consider the period between 1973 and 2017 when drug scheduling emerged and cannabis medicalization was combined with fears of “amotivational syndrome” (Newhart and Dolphin, 2019). This analysis must consider to what extent the decision by Portugal to legalize the possession of small amounts of all drugs, including cannabis, in 2001 fits within our proposed eras.

Perhaps even more important from a historical lens is the period between 2012 and 2017. At this time, recreational cannabis was legalized in Washington and Colorado, and cannabis became more widely available commercially. Uruguay legalized cannabis soon after. By 2017, five years of post-legalization data on legal cannabis was available, ushering in what we call the post-prohibition era. However, we need not make the perfect the enemy of the good. We encourage further efforts to define and delineate eras of cannabis research. In this article, we conclude that while methodologically limited efforts of the past now comprise the evidence “base,” many of these findings cannot be replicated in jurisdictions where cannabis policy has been liberalized. Moving forward means abandoning the abstracted and exclusionary designs of the past. It also requires considering how cannabis research, even in jurisdictions where cannabis is legal, must contend with the ways in which criminalization, decriminalization, legalization, and regulation co-exist, complicating definitional certainty (Corva and Meisel, 2021).

Whatever historical periods emerge as categorically most useful many examples of research that do not conform to the broad characteristics of the period exist. Indeed, substantive and prescient critiques of criminalizing cannabis existed during the prohibition period (Brecher, 1972; Kaplan, 1970; Mikuriya, 1969). One useful example is qualitative research by Sifaneck and Kaplan (1995), which managed to explode the gateway theory of drug addiction, suggest the harm reduction properties of cannabis, and highlight the value of supporting responsible cannabis use. Just as skeptical accounts emerged during the prohibition era, research claiming cannabis continues to represent significant risks persists in the post-prohibition period. One profound question is related to whether individual researchers bear some moral responsibility for research that justifies policies that were expected to cause harm based on dubious research decisions. Such effort may support careerist imperatives (Cohen, 2012) while failing in the ethical obligation of all researchers to acknowledge relevant limitations.

This is important to consider as the carceral creep of public health models is framed as progressive reform (Ritter, 2021). The value of distinguishing research conducted where people who use cannabis are forced to engage in criminal interactions versus those who can simply visit a store is important. This observation is not new (Goldstein, 1985) but has been insufficiently framed as a distinct methodological limitation (Johnstad, 2022) with undertheorized ethical implications. One option is to reduce future harm by explicitly confronting the damage done by the criminal justice system and its supporting institutions in the past and present (Quinney, 1970). Another is to build on efforts to document the experience of people who use drugs using more creative, equitable, and participatory methods. Ideally, researchers will take further steps to limit how the use of their work to criminalize, stigmatize, and pathologize people who use cannabis. This can begin to address the failure of past cannabis research generally to promote inclusivity and greater diversity (Martin-Willett and Bidwell, 2021). Solomon (2020: 5) has suggested:

To move forward, we need to understand our own history and the false premise on which we have based this misguided policy. We need to treat the cannabis policy started in 1937 the same way we treat segregated schools, miscegenation, and other race-based policy. Our inquiry needs to start with an acknowledgment of the history of racial discrimination in our drug policy and move toward serious evidence-based research. If we fail to do so, we will remain the willing victims of our own racist history.

Few have clean hands.

In recent work, we argue that cannabis represents a case study in moral, legal, and cultural renegotiation (Wheeldon and Heidt, 2023). In this article, we demonstrated how this renegotiation could be framed in ethical terms and linked structural constraints to methodological limitations. Every cannabis research failing cannot be laid solely at the feet of individual researchers. However, good faith assumptions about researcher intentions must be tested based on to what extent they continue to inflate narrow research findings or fail to properly account for how limitations likely restrict the generalizability of their conclusions. To avoid any confusion, we feel obliged to state again that we are not suggesting that researchers avoid publishing findings that suggest the risks of using cannabis. We do not believe anyone should alter their findings based on external social and political considerations, including cannabis’ increasing acceptance.

Instead, the limitations of the study should be adequately discussed and made readily apparent. Researchers should more willingly acknowledge the difficulties associated with finding clear cut direct causal relationships rather than suggesting their correlations might be causal in nature without justification. Given the harms of criminalizing cannabis, the obligations outlined in this article also fall to peer reviewers, editors, and funding bodies. While presenting research in ways that can preempt media misrepresentations may not be possible, researchers can avoid simplistic research conclusions by asking what harms are done by cannabis policies and how they compare to the risks of cannabis use. The need for evidence to answer the questions that will allow for responsible cannabis policy has never been greater.

An emergent question is how to frame the ontology of cannabis legalization. One way is to detail a post-prohibition epistemology to guide future research. By focusing on these questions, coherent categorical eras of cannabis research may be less important than working to define and explore the nascent transdisciplinary field of post-prohibition cannabis research. However, although cannabis legalization may be inevitable, responsible research is not. As we have argued, the legacy of racism, the persistence of zombie theories, and the realities of institutional inertia mean the harms of cannabis prohibition remain. In this article, we confronted the ways in which previous eras of research influenced the existing evidence base. Embracing a duty of care in cannabis research going forward can begin to subvert the unconscious bias that has guided global drug policy for a century. While our approach in this article is focused on cannabis, a duty of care is something to consider when other once-prohibited drugs are de-penalized, decriminalized, and legalized in jurisdictions around the world.