. 2022 Sep 1:238:109570.

doi: 10.1016/j.drugalcdep.2022.109570. Epub 2022 Jul 15.

The reciprocal relationships of social norms and risk perceptions to cigarette, e-cigarette, and cannabis use: Cross-lagged panel analyses among US young adults in a longitudinal study

Katelyn F Romm¹, Yan Wang², Yan Ma³, Christina N Wysota⁴, Melissa D Blank⁵, David M Huebner⁶, Kathleen M Roche⁷, Carla J Berg⁸

Affiliations

¹ Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: kromm@gwu.edu.
² Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: yanwang20@gwu.edu.
³ Department of Biostatistics and Bioinformatics, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: yanma@gwu.edu.
⁴ Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: cwysota@gwu.edu.
⁵ Department of Psychology, West Virginia University, Morgantown, WV 26506, USA. Electronic address: melissa.blank@mail.wvu.edu.
⁶ Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: davidhuebner@gwu.edu.
⁷ Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: kroche@gwu.edu.
⁸ Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA; George Washington Cancer Center, George Washington University, Washington, DC 20052, USA. Electronic address: carlaberg@gwu.edu.

PMID: 35926300
PMCID: PMC11365042
DOI: 10.1016/j.drugalcdep.2022.109570

The reciprocal relationships of social norms and risk perceptions to cigarette, e-cigarette, and cannabis use: Cross-lagged panel analyses among US young adults in a longitudinal study

Katelyn F Romm et al. Drug Alcohol Depend. 2022.

. 2022 Sep 1:238:109570.

doi: 10.1016/j.drugalcdep.2022.109570. Epub 2022 Jul 15.

Authors

Katelyn F Romm¹, Yan Wang², Yan Ma³, Christina N Wysota⁴, Melissa D Blank⁵, David M Huebner⁶, Kathleen M Roche⁷, Carla J Berg⁸

Affiliations

¹ Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: kromm@gwu.edu.
² Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: yanwang20@gwu.edu.
³ Department of Biostatistics and Bioinformatics, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: yanma@gwu.edu.
⁴ Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: cwysota@gwu.edu.
⁵ Department of Psychology, West Virginia University, Morgantown, WV 26506, USA. Electronic address: melissa.blank@mail.wvu.edu.
⁶ Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: davidhuebner@gwu.edu.
⁷ Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA. Electronic address: kroche@gwu.edu.
⁸ Department of Prevention and Community Health, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA; George Washington Cancer Center, George Washington University, Washington, DC 20052, USA. Electronic address: carlaberg@gwu.edu.

PMID: 35926300
PMCID: PMC11365042
DOI: 10.1016/j.drugalcdep.2022.109570

Abstract

Introduction: Given the distinct and evolving social norms, research on health implications, and regulations regarding cigarettes, e-cigarettes, and cannabis, it is important to understand the interplay between social norms, risk perceptions, and use of these products.

Methods: We analyzed 3 waves of longitudinal data (Fall 2018, 2019, 2020) from 3006 young adults (M_age=24.56, 54.8% female, 31.6% sexual minority, 39.6% racial/ethnic minority) from 6 US metropolitan statistical areas. Cross-lagged panel models (CLPMs) examined reciprocal relationships of (a) perceived social norms (i.e., peer use, social acceptability) and risk perceptions (i.e., harm, addictiveness) to (b) number of days of cigarette, e-cigarette, and cannabis use in the past 30 days, respectively.

Results: At baseline, lifetime and past 30-day use prevalence was: 61.8% and 26.9% for cigarettes, 57.7% and 37.7% for e-cigarettes, and 70.7% and 39.2% for cannabis. Perceived social norms and use of cigarettes and e-cigarettes decreased over time, and risk perceptions increased (except cigarettes showed stable perceived harm). Regarding cannabis, perceived social norms and use increased, yet perceived harm and addictiveness also increased. CLPM indicated that greater perceived social norms predicted greater cigarette, e-cigarette, and cannabis use over time, and vice versa. While greater perceived risk predicted less e-cigarette and cannabis use and vice versa, this did not hold true for cigarettes: use predicted lower perceived risk, but risk perceptions did not predict later use.

Conclusions: Tobacco and cannabis intervention and regulatory efforts should address health risks of use, particularly of e-cigarettes and cannabis, as well as denormalizing use.

Keywords: Cannabis use; Cross-lagged panel modeling; Risk perceptions; Social norms; Tobacco use; Young adults.

PubMed Disclaimer

Conflict of interest statement

Declaration of Competing Interest

The authors declare no conflicts of interests.

Figures

**Fig. 1.**
a. Coefficients represent β. Bolded values denote statistical significance at p < .01 Autoregressive effects: Cig1 > Cig2: β = 0.64, p < .001; Cig2 > Cig3: β = 0.56, p < .001; Peer1 > Peer2: β = 0.59, p < .001; Peer2 > Peer3: β = 0.42, p < .001. Covariance associations: Cig1 > Peer1: β = 0.52, p < .001; Cig2 > Peer2: β = 0.22, p < .001; Cig3 > Peer3: β = 0.19, p < .001., Fig. 1b. Coefficients represent β. Bolded values denote statistical significance at p < .01. Autoregressive effects: Cig1 > Cig2: β = 0.69, p < .001; Cig2 > Cig3: β = 0.58, p < .001; Soc1 > Soc2: β = 0.46, p < .001; Soc2 > Soc3: β = 0.44, p < .001. Covariance associations: Cig1 > Soc1: β = 0.24, p < .001; Cig2 > Soc2: β = 0.09, p < .001; Cig3 > Soc3: β = 0.08, p = .002., Fig. 1c. Coefficients represent β. Bolded values denote statistical significance at p < .01. Autoregressive effects: Cig1 > Cig2: β = 0.69, p < .001; Cig2 > Cig3: β = 0.58, p < .001; Harm1 > Harm2: β = 0.39, p < .001; Harm2 > Harm3: β = 0.40, p < .001. Covariance associations: Cig1 > Harm1: β = −0.13, p < .001; Cig2 > Harm2: β = −0.08, p = .005; Cig3 > Harm3: β = −0.01, p = .714., Fig. 1d. Coefficients represent β. Bolded values denote statistical significance at p < .01. Autoregressive effects: Cig1 > Cig2: β = 0.70 p < .001; Cig2 > Cig3: β = 0.59, p < .001; Add1 > Add2: β = 0.48, p < .001; Add2 > Add3: β = 0.31, p < .001. Covariance associations: Cig1 > Add1: β = 0.05, p = .001; Cig2 > Add2: β = 0.01, p = .974; Cig3 > Add3: β = −0.01, p = .713., Fig. 1e. Coefficients represent β. Bolded values denote statistical significance at p < .01. Autoregressive effects: Cig1 > Cig2: β = 0.61, p < .001; Cig2 > Cig3: β = 0.51, p < .001; Intent1 > Intent2: β = 0.52, p < .001; Intent2 > Intent3: β = 0.41, p < .001. Covariance associations: Cig1 > Intent1: β = 0.73, p < .001; Cig2 > Intent2: β = 0.55, p < .001; Cig3 > Intent3: β = 0.54, p < .001. Note. Cig=Cigarette use; Peer=Peer use; Soc=Social acceptability; Harm=Perceived harm; Add=Perceived addictiveness; Intent=Use intentions. Covariates were included in each model, but are not illustrated in figures. Bolded lines are significant estimates (p < .01) in expected directions; dashed lines are non-significant.

**Fig. 2.**
a. Coefficients represent β. Bolded values denote statistical significance at p < .01 *Autoregressive effects:* ECIG1 > ECIG2: β = 0.73, p < .001; ECIG2 > ECIG3: β = 0.59, p < .001; Peer1 > Peer2: β = 0.52, p < .001; Peer2 > Peer3: β = 0.41, p < .001. *Covariance associations:* ECIG1 > Peer1: β = 0.25, p < .001; ECIG2 > - Peer2: β = 0.27, p < .001; ECIG3 > Peer3: β = 0.53, p < .001., Fig. 2b. Coefficients represent β. Bolded values denote statistical significance at p < .01. *Autoregressive effects:* ECIG1 > ECIG2: β = 0.76, p < .001; ECIG2 > ECIG3: β = 0.59, p < .001; Soc1 > Soc2: β = 0.45, p < .001; Soc2 > Soc3: β = 0.43, p < .001. *Covariance associations:* ECIG1 > Soc1: β = 0.09, p < .001; ECIG2 > Soc2: β = 0.10, p < .001; ECIG3 > Soc3: β = 0.28, p < .001., Fig. 2c. Coefficients represent β. Bolded values denote statistical significance at p < .01. *Autoregressive effects:* ECIG1 > ECIG2: β = 0.72, p < .001; ECIG2 > ECIG3: β = 0.58, p < .001; Harm1 > Harm2: β = 0.48, p < .001; Harm2 > Harm3: β = 0.41, p < .001. *Covariance associations:* ECIG1 > Harm1: β = −0.41, p < .001; ECIG2 > Harm2: β = −0.17, p < .001; ECIG3 > Harm3: β = −0.09, p < .001. Fig. 2d. Coefficients represent β. Bolded values denote statistical significance at p < .01. *Autoregressive effects:* ECIG1 > ECIG2: β = 0.75, p < .001; ECIG2 > ECIG3: β = 0.59, p < .001; Add1 > Add2: β = 0.45, p < .001; Add2 > Add3: β = 0.36, p < .001. *Covariance associations:* ECIG1 > Add1: β = 0.02, p = .376; ECIG2 > Add2: β = 0.01, p = .651; ECIG3 > Add3: β = 0.02, p = .268., Fig. 2e. Coefficients represent β. Bolded values denote statistical significance at p < .01. *Autoregressive effects:* ECIG1 > ECIG2: β = 0.65, p < .001; ECIG2 > ECIG3: β = 0.53, p < .001; Intent1 > Intent2: β = 0.50, p < .001; Intent2 > Intent3: β = 0.38, p < .001. *Covariance associations:* ECIG1 > Intent1: β = 0.80, p < .001; ECIG2 > Intent2: β = 0.57, p < .001; ECIG3 > Intent3: β = 0.60, p < .001. *Note.* ECIG=E-cigarette use; Peer=Peer use; Soc=Social acceptability; Harm=Perceived harm; Add=Perceived addictiveness; Intent=Use intentions. Covariates were included in each model, but are not illustrated in figures. Bolded lines are significant estimates (p < .01); dashed lines are non-significant.

**Fig. 3.**
a. Coefficients represent β. Bolded values denote statistical significance at p <.01 *Autoregressive effects:* Can1 > Can2: β = 0.74, p < .001; Can2 > Can3: β = 0.60, p < .001; Peer1 > Peer2: β = 0.50, p < .001; Peer2 > Peer3: β = 0.45, p < .001. *Covariance associations:* Can1 > Peer1: β = 0.48, p < .001; Can2 > Peer2: β = 0.26, p < .001; Can3 > Peer3: β = 0.22, p < .001., Fig. 3b. Coefficients represent β. Bolded values denote statistical significance at p < .01. *Autoregressive effects:* Can1 > Can2: β = 0.78, p < .001; Can2 > Can3: β = 0.62, p < .001; Soc1 > Soc2: β = 0.45, p < .001; Soc2 > Soc3: β = 0.37, p < .001. *Covariance associations:* Can1 > Soc1: β = 0.20, p < .001; Can2 > Soc2: β = 0.09, p < .001; Can3 > Soc3: β = 0.10, p < .001., Fig. 3c. Coefficients represent β. Bolded values denote statistical significance at p < .01. *Autoregressive effects:* Can1 > Can2: β = 0.76, p < .001; Can2 > Can3: β = 0.62, p < .001; Harm1 > Harm2: β = 0.59, p < .001; Harm2 > Harm3: β = 0.40, p < .001. *Covariance associations:* Can1 > Harm1: β = −0.27, p < .001; Can2 > Harm2: β = −0.11, p < .001; Can3 > Harm3: β = −0.06, p = .008., Fig. 3d. Coefficients represent β. Bolded values denote statistical significance at p < .01. *Autoregressive effects:* Can1 > Can2: β = 0.79, p < .001; Can2 > Can3: β = 0.62, p < .001; Add1 > Add2: β = 0.54, p < .001; Add2 > Add3: β = 0.54, p < .001. *Covariance associations:* Can1 > Add1: β = −0.07, p < .001; Can2 > Add2: β = −0.04, p = .122; Can3 > Add3: β = −0.05, p = .040., Fig. 3e. Coefficients represent β. Bolded values denote statistical significance at p < .01. *Autoregressive effects:* Can1 > Can2: β = 0.71, p < .001; Can2 > Can3: β = 0.55, p < .001; Intent1 > Intent2: β = 0.63, p < .001; Intent2 > Intent3: β = 0.49, p < .001. *Covariance associations:* Can1 > Intent1: β = 0.67, p < .001; Can2 > Intent2: β = 0.43, p < .001; Can3 > Intent3: β = 0.47, p < .001. *Note.* Can=Cannabis use; Peer=Peer use; Soc=Social acceptability; Harm=Perceived harm; Add=Perceived addictiveness; Intent=Use intentions. Covariates were included in each model, but are not illustrated in figures. Bolded lines are significant estimates (p < .01); dashed lines are non-significant.

See this image and copyright information in PMC

Cited by

Perceived risk of harm for different methods of cannabis consumption: A brief report.
Florimbio AR, Walton MA, Coughlin LN, Lin LA, Bonar EE. Florimbio AR, et al. Drug Alcohol Depend. 2023 Oct 1;251:110915. doi: 10.1016/j.drugalcdep.2023.110915. Epub 2023 Aug 9. Drug Alcohol Depend. 2023. PMID: 37597308 Free PMC article.
Direct and Indirect Effects of Cannabis Risk Perceptions on Cannabis Use Frequency.
Florimbio AR, Walton MA, Duval ER, Bauermeister JA, Young SD, McAfee J, Bonar EE. Florimbio AR, et al. Addict Res Theory. 2024;32(1):68-73. doi: 10.1080/16066359.2023.2221029. Epub 2023 Jun 16. Addict Res Theory. 2024. PMID: 38268741 Free PMC article.
Associations between Cannabis Consumption Methods and Cannabis Risk Perception.
Choi NG, Marti CN, Choi BY. Choi NG, et al. Int J Environ Res Public Health. 2024 Jul 27;21(8):986. doi: 10.3390/ijerph21080986. Int J Environ Res Public Health. 2024. PMID: 39200597 Free PMC article.
Differences in Online Descriptions and Marketing of Derived Intoxicating Cannabis Products.
LoParco CR, Bone C, Garg A, Berg CJ, Rossheim ME. LoParco CR, et al. J Psychoactive Drugs. 2025 Mar 25:1-11. doi: 10.1080/02791072.2025.2484376. Online ahead of print. J Psychoactive Drugs. 2025. PMID: 40131178
Exploring Perceptions About Paracetamol, Tramadol, and Codeine on Twitter Using Machine Learning: Quantitative and Qualitative Observational Study.
Carabot F, Donat-Vargas C, Santoma-Vilaclara J, Ortega MA, García-Montero C, Fraile-Martínez O, Zaragoza C, Monserrat J, Alvarez-Mon M, Alvarez-Mon MA. Carabot F, et al. J Med Internet Res. 2023 Nov 14;25:e45660. doi: 10.2196/45660. J Med Internet Res. 2023. PMID: 37962927 Free PMC article.

See all "Cited by" articles

References

1. Agarwal D, Loukas A, Perry CL, 2018. Examining college students’ social environment, normative beliefs, and attitudes in subsequent initiation of electronic nicotine delivery systems. Health Educ. Behav 45 (4), 532–539. 10.1177/1090198117739672. - DOI - PMC - PubMed
1. Ajzen I, Fishbein M, 1980. Understanding Attitudes and Predicting Social Behavior Prentice-Hall, Inc.
1. Ajzen I, 1991. The theory of planned behavior. Organ. Behav. Hum. Decis. Process 50, 179–211. 10.1016/0749-5978(91)90020-T. - DOI
1. Allison PD, 2003. Missing data techniches for structural equation modeling. J. Abnorm. Psychol 112 (4), 545–557. 10.1037/0021-843X.112.4.545. - DOI - PubMed
1. American Public Health Association, 2020, A public health approach to regulating commercially legalized cannabis Retrieved from 〈https://www.apha.org/policies-and-advocacy/public-health-policy-statemen...〉 (accessed 1 February 2022).

Publication types

Actions

MeSH terms

Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions

Grants and funding

LinkOut - more resources

Full Text Sources
Medical
- MedlinePlus Health Information
Miscellaneous
- NCI CPTAC Assay Portal

Save citation to file

Email citation

Add to Collections

Add to My Bibliography

Your saved search

Create a file for external citation management software

Your RSS Feed

The reciprocal relationships of social norms and risk perceptions to cigarette, e-cigarette, and cannabis use: Cross-lagged panel analyses among US young adults in a longitudinal study

Affiliations

The reciprocal relationships of social norms and risk perceptions to cigarette, e-cigarette, and cannabis use: Cross-lagged panel analyses among US young adults in a longitudinal study

Authors

Affiliations

Abstract

Conflict of interest statement

Figures

Similar articles

Cited by

References

Publication types

MeSH terms

Grants and funding

LinkOut - more resources

Full Text Sources

Medical

Miscellaneous