Are Text-Only E-Cigarette Warnings About Health Harms More Effective Than the FDA’s Nicotine Addiction Warning?

This meta-analysis asked whether text-only e-cigarette warnings about health harms are more effective than the FDA’s nicotine addiction warning. Researchers examined 24 experimental studies including 22,549 participants, measuring outcomes such as attention, risk beliefs, and intentions to vape or quit. Results showed text-only warnings increased attention (d = 0.52), negative emotions (d = 0.65), risk beliefs (d = 0.26), and intentions to quit vaping (d = 0.34), while reducing intentions to vape (d = −0.14). Health harm warnings performed better than addiction warnings on most measures, suggesting that adding clear, health-based text warnings could strengthen public understanding and discourage vaping.

The Policy Scientist’s Perspective

This study addresses an increasingly important public health policy question: whether simple, text-based warnings on e-cigarettes can effectively shape perceptions of harm and influence vaping behavior. The topic matters broadly because regulatory agencies worldwide are grappling with how to communicate evolving scientific evidence about e-cigarettes without overstating risks or undermining smoking cessation efforts. The meta-analysis synthesizes 24 randomized experiments totaling over 22,000 participants, producing effect sizes that are moderate but statistically reliable. The study’s reliance on experimental data provides stronger causal leverage than observational work, although the brief, online exposures may limit ecological validity. Despite being U.S.-centered, the findings likely generalize to other developed regulatory settings. Overall, this article makes a meaningful and timely contribution by integrating dispersed evidence to inform the next phase of e-cigarette warning policy.

Full Citation and Link to Article

Jang, Y., Shaw, J., Wackowski, O. A., & Noar, S. M. (2025). Effectiveness of text-only e-cigarette warnings: A meta-analysis. JAMA Internal Medicine, 185(8), 955–964. https://doi.org/10.1001/jamainternmed.2025.1380

Extended Summary

Central Research Question

This meta-analysis asked whether text-only e-cigarette warnings—particularly those emphasizing specific health harms—are more effective than the U.S. Food and Drug Administration’s (FDA) current nicotine addiction warning in influencing public perceptions and behavioral intentions. The authors sought to determine not only whether such warnings attract attention and increase perceived risk but also whether they lead to measurable changes in intentions to use or quit vaping. The question carries broad policy relevance given the rapid expansion of e-cigarette use, persistent misperceptions about relative harm, and global debates over how to regulate nicotine products without unintentionally discouraging harm reduction among smokers.

Previous Literature

Prior studies on tobacco communication have consistently shown that health warnings can influence risk perceptions and behaviors, but much of this evidence comes from traditional cigarette research. Pictorial warnings for combustible tobacco have been shown to increase attention, evoke emotional responses, and motivate cessation across diverse populations. Meta-analyses by Noar, Brewer, and others demonstrated that visual health warnings on cigarette packaging generate moderate to large effects on quitting intentions and risk awareness. However, research on e-cigarettes remains more fragmented and primarily focused on text-only messages, which are the regulatory standard in the United States and several other jurisdictions.

Individual experiments over the past decade examined the FDA-mandated text-only warning—“This product contains nicotine. Nicotine is an addictive chemical”—and alternative warnings describing health harms such as lung or brain damage. Results were inconsistent, with some studies showing modest effects on perceived risk and others finding little behavioral impact. Unlike cigarettes, where graphic warnings dominate international policy, e-cigarettes have remained subject to less visible and less varied textual messages. The literature also revealed uncertainty over whether emphasizing health harms might inadvertently lead consumers to overestimate e-cigarettes’ relative danger compared with smoking. These mixed findings created a need for synthesis. The current study addresses this gap by integrating data from two dozen experimental studies to evaluate the overall impact and relative performance of text-only warning types.

Data

The analysis pooled data from 24 randomized experimental studies encompassing 22,549 participants. Mean participant age was 27.9 years, and 53.9 percent were female. Roughly one-fifth reported vaping within the past 30 days, and over one-third were current cigarette smokers. The studies were conducted mainly in the United States (88 percent), with smaller contributions from other high-income countries. Most experiments targeted adult or young adult populations, typically using online samples or controlled laboratory settings.

The studies included a variety of warning contexts: e-cigarette packaging, advertisements, social media posts, and stand-alone warning cards. They compared health harm warnings (e.g., “Vaping damages your lungs” or “Vaping exposes you to toxic chemicals”) with addiction warnings (the FDA-required text or close variants) and, in most cases, control conditions without any warning. Each study measured outcomes related to message processing (attention, emotional response, perceived message effectiveness), risk perception (beliefs about addiction, health harms, or relative risk versus smoking), and behavioral intentions (intentions to vape or quit).

The authors coded each study for sample characteristics, experimental design, and outcomes, achieving high intercoder reliability (Cohen’s κ = 0.87). They also conducted formal risk-of-bias assessments using the JBI randomized trial appraisal tool. Most studies demonstrated low risk of bias on randomization, blinding, and treatment consistency, though some were unclear on baseline comparability or measurement reliability due to reliance on single-item measures.

Methods

Following PRISMA reporting guidelines, the authors used a systematic search strategy covering five databases: PubMed, Embase, Scopus, PsycINFO, and Business Source Premier. They included studies that (1) randomized participants into at least two conditions involving text-only e-cigarette warnings; (2) included either an FDA-style addiction warning, a health harm warning, or a control message; and (3) reported at least one relevant outcome among the three major categories—message processing, risk beliefs, or behavioral intentions.

Effect sizes were calculated using standardized mean differences (Cohen’s d), adjusted for small-sample bias. For studies with multiple relevant warning conditions, separate effect sizes were computed to avoid data loss while adjusting control group weights to maintain independence. Random-effects models were used throughout to account for heterogeneity across study contexts and samples. The authors also reported 95% confidence intervals, prediction intervals, and heterogeneity statistics (Q and I²) for each outcome.

Moderator analyses explored whether warning type (addiction vs. health harms), warning medium (packaging vs. advertisements), or exposure frequency (single vs. repeated) influenced results. Publication bias was assessed using funnel plots and the trim-and-fill method. The analytic procedures were implemented with Comprehensive Meta-Analysis version 4 software (Biostat).

The methodological approach is notably strong for communication research. By restricting inclusion to randomized experiments, the authors achieved a level of causal inference that exceeds most observational or cross-sectional designs commonly seen in this field. While not randomized controlled trials in the clinical sense, these experiments isolate causal effects of message exposure on proximate psychological outcomes, such as risk perception and intention formation. The large cumulative sample size provides substantial statistical power, though external validity remains limited by the predominance of short-term, U.S.-based online studies.

Findings/Size Effects

The meta-analysis found that text-only e-cigarette warnings significantly improved most outcomes compared with no-warning controls. Message processing outcomes showed the strongest effects: attention (d = 0.52; 95% CI, 0.33–0.70), negative affect (d = 0.65; 95% CI, 0.49–0.81), and perceived message effectiveness (d = 0.95; 95% CI, 0.69–1.21). Risk-related beliefs also improved: addiction beliefs (d = 0.24; 95% CI, 0.05–0.42) and general risk beliefs (d = 0.26; 95% CI, 0.16–0.36). However, there was no measurable change in relative risk beliefs (d = 0.00; 95% CI, −0.11 to 0.12), indicating that warnings did not cause people to believe e-cigarettes are more dangerous than cigarettes—a key policy concern.

Behavioral intention measures revealed smaller but meaningful effects. Intentions to vape declined (d = −0.14; 95% CI, −0.27 to −0.01), while intentions to quit vaping increased (d = 0.34; 95% CI, 0.09–0.58). When comparing health harm warnings to addiction warnings directly, health harm messages were superior on five of eight outcomes: attention (d = 0.37), negative affect (d = 0.44), perceived message effectiveness (d = 0.36), risk beliefs (d = 0.07), and intentions to quit (d = 0.17). These effects are modest in size but consistent across studies and statistically robust.

Moderator analyses showed larger effects when warnings were placed on packaging rather than advertisements (d = 0.48 vs. 0.21) and when participants saw a single exposure instead of repeated ones (d = 0.31 vs. 0.11), suggesting that novelty and context enhance message salience. Publication bias analyses showed minimal influence on the main results, with adjusted effect sizes remaining statistically significant.

Overall, the results demonstrate that even simple text-only warnings can increase cognitive and emotional engagement, alter risk perceptions, and slightly shift behavioral intentions. Importantly, these effects are achieved without producing the unintended belief that vaping is more dangerous than smoking—an outcome that could deter harm reduction efforts.

Conclusion

This meta-analysis provides the strongest synthesis to date on the effectiveness of text-only e-cigarette warnings. The findings show that such warnings, though less vivid than pictorial ones, can still meaningfully affect attention, emotional response, perceived risk, and behavioral intentions. Health harm warnings outperform addiction warnings across multiple outcomes, indicating that broadening the content of mandated warning statements could enhance public understanding of e-cigarette risks.

From a methodological standpoint, the study demonstrates strong internal validity and credible causal inference for message effects. Its limitation lies mainly in external validity—nearly all studies were brief experiments using online panels rather than field or longitudinal designs. Thus, the durability of these effects in real-world settings remains uncertain. Additionally, the findings primarily reflect U.S. samples, which may limit generalizability to regions with different regulatory environments, media exposure, or cultural perceptions of tobacco risk. Nevertheless, given that many high-income countries use similar text-only approaches, the results likely extend to comparable jurisdictions.

The policy implications are clear. Regulators should consider expanding the range of e-cigarette warnings to include rotating health harm statements supported by current toxicological and epidemiological evidence. The results suggest that even absent imagery, well-crafted text warnings can influence public understanding and behavior. By integrating 24 randomized studies, this analysis also establishes a new empirical baseline for evaluating the next generation of vaping warnings, including those that may eventually combine text and imagery.

In sum, this article represents a methodologically rigorous, timely, and policy-relevant contribution. It consolidates fragmented experimental evidence into a coherent picture showing that text-only warnings—especially those addressing specific health harms—can advance public health communication goals without distorting relative risk perceptions. As e-cigarette use continues to grow and governments reassess labeling requirements, the findings offer strong empirical guidance for designing warnings that balance accuracy, salience, and behavioral impact.