Textual and pictorial enhancement of cannabis warning labels: An Online experiment among at-risk U.S. young adults

https://doi.org/10.1016/j.drugalcdep.2022.109520Get rights and content

Highlights

  • Textual CWL enhancements improved information recall among US young adults.

  • Pictorial CWL enhancements improved information recall and perceived effectiveness.

  • Negative emotions mediated pictorial enhancements' effects on perceived effectiveness.

  • Currently required CWLs in the U.S. are overdue for a redesign.

Abstract

Background

This study experimentally examines whether enhanced cannabis warning labels (CWLs) outperform those currently required in the U.S. in improving recall of health risks, emotional responses, and perceived message effectiveness among at-risk young adults.

Method

We conducted an online national survey-based experiment in October 2020. Young adults aged 18–26 years old and at-risk for cannabis use (N = 523) were randomly assigned in an online experiment, to view either currently required CWLs in California with small font and a composite health risk statement, or enhanced single-theme CWLs with varying textual and pictorial components. We performed linear regression analyses to compare the enhanced with existing CWLs on information recall, negative emotions, and perceived message effectiveness. Furthermore, information recall and negative emotions were examined as parallel mediators to better understand the mechanisms underlying effective textual and pictorial enhancement of CWLs.

Results

Compared with currently required CWLs in California, both textually (b = 0.30, p = .011) and pictorially (b = 0.59, p < .001) enhanced CWLs increased recall accuracy. Pictorially enhanced CWLs outperformed their text-only counterparts (b = 0.28, p = .019) in improving information recall. Only pictorially enhanced CWLs improved perceived message effectiveness (b = 0.31, p = .008), which was mediated by negative emotions but not by information recall.

Conclusions

Given rapid expansion of the cannabis industry and declining perception of harm, currently required CWLs in the U.S. such as California’s, would benefit from redesign to improve public understanding of health risks and to prevent youth use.

Introduction

After decades of prohibition, the adult use of cannabis products containing tetrahydrocannabinol (THC) is now legal in 19 U.S. states, two territories, and in the District of Columbia (National Conference of State Legislatures, 2022). Since successful referendums in Colorado and Washington in 2012 legalized non-medicinal cannabis use (Hall and Lynskey, 2020), a multi-billion-dollar legal cannabis industry has emerged (Spindle et al., 2019). Parallel to the rapid growth of the cannabis industry, young adults’ perceptions of harm from cannabis have fallen, and perceptions of no-risk have risen (Azofeifa et al., 2016, Reboussin et al., 2019). For example, among those aged 18–25, perceived great risk of harm from smoking cannabis weekly declined from 19.1% in 2015 to 15% in 2019 (Substance Abuse and Mental Health Services Administration (SAMHSA), 2020). Furthermore, in 2019, 35.4% of young adults were past-year users of cannabis and 5.8% had a cannabis use disorder (CUD) (SAMHSA, 2020). Past-year use rates have risen significantly since 2002, and rates of CUD have risen since their lowest level in 2014 (SAMHSA, 2020). For both measures, rates are notably higher among young adults than other age groups. Against the backdrop of increased legalization efforts, declining risk perceptions, and rapid market transformation, effective communication of health risks is essential to inform and protect at-risk young adults. Globally, legalization of medical use is expanding rapidly, including many countries in the Americas and Europe, while recreational cannabis use, though expanding, is prohibited in most countries apart from a few exceptions (e.g., Canada, Mexico, South Africa, Uruguay, 18 states and several districts in the U.S.). Experience from tobacco control and early cannabis research suggests that robust approaches for consumer information should form part of these new regulatory frameworks. Evidence on CWLs’ educational and persuasive effects from at-risk young adults in the U.S. can inform policy both domestically and in other countries as legalization expands.

While cannabis has proven medicinal benefits, emerging evidence also highlights a host of associated harms and risks among both adolescents and adults. In addition to increased risk of motor vehicle crashes and of low birth weight in offspring of mothers who use cannabis while pregnant, the National Academies of Science, Engineering and Medicine (NASEM) has concluded that there is substantial evidence of an association of cannabis use with the development of schizophrenia and psychoses among frequent users (National Academies of Sciences, Engineering, and Medicine (NASEM), 2017). The relationship with psychosis and schizophrenia is supported by other recent studies (Hasan et al., 2020, Sami et al., 2020). The Surgeon General has noted that ongoing brain development through the mid-20 s renders young adults more vulnerable to the effects of addictive substances (Office of the Surgeon General, 2019), and frequent cannabis use during adolescence has been associated with neurological changes in areas responsible for attention, memory, decision-making, motivation, impaired learning, declines in IQ, school performance, high school, and college graduation rates and life satisfaction (Meier et al., 2012, Silins et al., 2014). Research is needed to examine effective communication strategies to improve young adults’ awareness and understanding of such health risks to prevent early initiation, habitual use, and CUD.

Mandating warning labels on cannabis products is a promising communication-based regulatory strategy, as it can increase exposure to educational health messages among potential consumers en masse and with minimal expenditures to government. Extensive research has shown that tobacco control warning labels, particularly prominent labels enhanced with graphic visuals, can better attract attention, facilitate information recall, inform smokers about the health hazards of smoking, trigger strong affective responses, and encourage cessation. They may also prevent initiation by nonsmokers (Brewer et al., 2016, Noar et al., 2020). Whereas Canada has begun to translate the principles of designing effective tobacco warning labels to cannabis warning labels (CWLs) (Goodman et al., 2019, Mutti-Packer et al., 2018, Pepper et al., 2020, Winstock et al., 2021), the U.S. lags behind. Currently required CWLs in U.S. states are predominantly composite textual messages in hard-to-read fonts and often lack information on key evidence-based health risks (Goodman et al., 2019, Silver et al., 2020).

Joining a burgeoning body of literature that tests improved CWLs, this study sought to: 1) design improved CWLs with single themes, a larger font, potentially more effective language adapted from theme-matched Canadian CWLs, and visual enhancements; 2) test enhanced CWLs against a currently required U.S. CWL, specifically, California’s, in relation to information recall, negative affect, and perceived message effectiveness (PME) among at-risk young adults; and 3) examine the respective mediating contributions of recall versus negative affect on PME for the best-in-show CWLs. Our focus on assessing these three outcomes to evaluate enhanced CWLs and establish an evidence base is motivated by past legal challenges surrounding the Food and Drug Administration (FDA)’s mandate to implement graphic tobacco control warning labels and more broadly, issues regarding the regulation of “compelled” speech that includes mandatory disclosures such as health warning messages. We focused on California’s existing CWL, first because it is the nation’s largest legal cannabis market, and second, because it exemplifies design weaknesses observed in currently required CWLs across states in which only a long and hard-to-read textual message in a small font on cannabis packaging is required (Silver et al., 2020). Additionally, because a bill for graphic front-of-pack cannabis warning labels was being developed in that state (introduced as S.B. 1097, 2022), our research had the potential to immediately inform translation efforts by providing scientific evidence for policy design, in California and beyond.

Because CWLs represent a form of “compelled speech”, empirical evidence is needed to justify that the CWLs can directly and materially advance the government’s substantial interest and are no more extensive than necessary (Kraemer and Baig, 2013, McKeon, 1980). To assess the ability of CWLs to advance these interests, we assessed the effects of alternate designs on three outcomes: message recall, perceived message effectiveness (PME), and the generation of negative emotions.

Message recall is thought to indicate audience’s attention to, encoding and storage of persuasive messages (Johnson et al., 2005, McGuire, 1968, Meernik et al., 2016, Strasser et al., 2012), so it is often used as a proxy for comprehension and cognitive processing. This use is supported by associations of recall of tobacco warning labels with physiological measures of attention such as eye-tracking (Meernik et al., 2016). That said, successful encoding and storage precede, but do not necessarily guarantee, persuasion (Cappella, 2006; Fisher et al., 2018). Although high message recall is often found to predict persuasive outcomes (e.g., reduced tobacco cravings, Klein et al., 2017) including PME (Noar et al., 2020), recalled messages needed to be accepted before persuasion could occur. Acceptance of recalled messages depends upon moderating factors such as their argument strength, message recipients’ preexisting attitudes, and the degree of counterarguing (Cappella, 2006; Carpenter and Boster, 2013). Whether or not message recall subsequently prevents at-risk young adults from initiating or habitually consuming recreational cannabis products, improving recall of cannabis health risks is necessary to fulfill the government’s goal of educating vulnerable populations about health risks.

Government also has an interest in preventing cannabis initiation and habitual use among vulnerable populations such as minors and young adults. Towards this end, PME can be employed in individual-level analyses as a useful, though not perfect, proxy for warning labels’ persuasive impacts on prevention and cessation as demonstrated by extensive research on tobacco control messages including warning labels (Cappella, 2018, Noar et al., 2020, Noar et al., 2018). Given that the inconsistency in the design and implementation of earlier PME scales has contributed to doubts about PME’s measurement validity (Noar et al., 2018; O′Keefe, 2018), recent research has developed new PME measures focused on messages’ behavioral references and impacts (dubbed “effects perceptions PME”) (Baig et al., 2021; Rohde et al., 2021). These revised PME measures have shown superior psychometric properties when deployed to assess tobacco control messages and predict individual-level behavioral outcomes well (Baig et al., 2020, 2019; Rohde et al., 2020). Given our need to assess multiple CWLs per participant, and that one’s behavioral intentions tend to stay comparatively stable and are not likely to vary meaningfully across individual CWLs, repeatedly measuring behavioral intentions (i.e., one measurement per CWL) is not appropriate. In contrast, the PME scale is tied to each individual CWL (e.g., “The health information in the warning label I just saw discourages me from wanting to use marijuana.”) and hence is more sensitive to capture the impacts of individual CWLs when each participant needs to process multiple CWLs.

Second, the fact that pictorial tobacco control warning labels tend to produce negative emotions has made them a target of litigation by the tobacco industry on the grounds that they are unnecessarily emotional and not factual (Kraemer and Baig, 2013). This simplistic view that emotional and factual information are incompatible contradicts long-standing psychological and behavioral science, which has shown that emotions and facts are oftentimes intertwined (Popova et al., 2017) and emotional processing can trickle down to improve behavioral changes (Hammond, 2011, Yotam et al., 2019). Graphic tobacco control messages, for example, were perceived as equally informative as their text-only counterparts (Popova et al., 2017). Graphic warnings do tend to induce more negative emotions; however, stronger emotionality, as indicated by neural activity, is associated with reduced urge to smoke (Rubinstein, 2015). Other studies have found similar enhanced effectiveness of pictorial messages (Yotam et al., 2019), whose emotional appeal can reduce intention to smoke either directly or by updating health beliefs (Skurka et al., 2018).

Anticipating that similar legal pushback may arise with regards to enhanced CWLs and their emotional impact, we sought to examine whether induced negative emotions mediate the persuasive effects of CWLs on PME, even after factoring in mediation by information recall. Though not conclusive, use of the parallel mediation test can provide preliminary evidence to clarify whether negative emotions represent an important pathway leading to enhanced CWLs’ effectiveness in prevention efforts among at-risk young adults.

Section snippets

Participants

In fall 2020, we recruited a sample of 523 (Nfemale=246) young adults aged between 18 and 26 from the Qualtrics online panel. The sample was matched to national distributions on sex, race, and ethnicity (see Table 1 for demographic information). We focused on participants at elevated risk of cannabis use and screened out those reporting “definitely no” to all of the following: a) use of marijuana in the next 6 months, b) willingness to try marijuana if one of their best friends offered it, and

Negative emotions

As illustrated in Fig. 3a, participants experienced higher negative emotions when viewing pictorial enhanced CWLs than the CA-mock control (b = 0.30, p = .016, 95% CI [0.06, 0.55]). When comparing enhanced CWLs with either the CA-mock or CA-control CWLs, neither was statistically significant.

Recall sensitivity

As illustrated in Fig. 3b, compared with the CA-mock CWL, text enhanced CWLs with textual statements taken from theme-matched Canadian CWLs and a yellow background significantly improved recall accuracy

Discussion

Given the spread of legalization and the decline in young adults’ perceived harms of cannabis use (Azofeifa et al., 2016, Reboussin et al., 2019), it is in governments’ substantial interest to inform the public, and particularly vulnerable populations, about known health risks associated with cannabis use through means such as CWLs. Compared with currently required CA-control CWL and our CA-mock CWL (i.e., CA-control plus mental health risks), enhanced textual CWLs using single-themed health

Conclusions

Through an online experiment with a national sample of at-risk young adults, we gathered evidence that both textual and pictorial enhancements can improve information recall and perceived message effectiveness (PME). Furthermore, we found that negative emotions, but not information recall, emerged as an important pathway through which pictorial enhanced CWLs accrued gains in PME. Given the rapid expansion of the cannabis industry and declining perceptions of harm, currently required CWLs in the

Acknowledgements

This work was supported by the Office of the Vice Chancellor for Research and Graduate Education at the University of Wisconsin–Madison with funding from the Wisconsin Alumni Research Foundation, USA, under grants number MSN231886 and MSN249368 to Sijia Yang; and by The Getting it Right from the Start program at the Public Health Institute to Lynn Silver.

Contributors

Everyone who contributed significantly to the work has been listed as an author.

Contributors

All authors have read and approved the manuscript for submission to Drug and Alcohol Dependence; have made a substantial contribution to the conception, design, gathering, analysis and/or interpretation of data and a contribution to the writing and intellectual content of the article; and acknowledge that they have exercised due care in ensuring the integrity of the work.

Declaration of Competing Interest

No conflict declared. All authors declare no conflict of interest.

References (45)

  • J.M. Whitehill et al.

    Exposure to cannabis marketing in social and traditional media and past-year use among adolescents in states with legal retail cannabis

    J. Adolesc. Health

    (2020)
  • A.R. Winstock et al.

    Perceptions of cannabis health information labels among people who use cannabis in the U.S. and Canada

    Int. J. Drug Policy

    (2021)
  • An act to add Sections 26070.3, 26121, and 26151.5 to the Business and Professions Code, Relating to Cannabis, S.B....
  • A. Azofeifa et al.

    National estimates of marijuana use and related indicators - national survey on drug use and health, United States, 2002-2014

    Mmwr. Surveill. Summ.

    (2016)
  • S.A. Baig et al.

    Incremental criterion validity of message perceptions and effects perceptions in the context of anti-smoking messages

    J. Behav. Med.

    (2021)
  • E. Bigsby et al.

    Efficiently and effectively evaluating public service announcements: additional evidence for the utility of perceived effectiveness

    Commun. Monogr.

    (2013)
  • N.T. Brewer et al.

    Effect of pictorial cigarette pack warnings on changes in smoking behavior: a randomized clinical trial

    JAMA Intern. Med.

    (2016)
  • J.N. Cappella

    Perceived message effectiveness meets the requirements of a reliable, valid, and efficient measure of persuasiveness

    J. Commun.

    (2018)
  • C.J. Carpenter et al.

    The relationship between message recall and persuasion: More complex than it seems

    J. Commun.

    (2013)
  • W. Hall et al.

    Assessing the public health impacts of legalizing recreational cannabis use: the US experience

    World Psychiatry

    (2020)
  • D. Hammond

    Health warning messages on tobacco products: a review

    Tob. Control

    (2011)
  • A. Hasan et al.

    Cannabis use and psychosis: a review of reviews

    Eur. Arch. Psychiatry Clin. Neurosci.

    (2020)
  • Cited by (0)

    1

    Note: SJK and MM contributed equally to the manuscript hence sharing the first authorship.

    View full text