Impact of Survey Item Wording and Response Options on Prevalence of Beliefs About Nicotine Causing Cancer: A Randomized Survey Experiment

Nicotine Tob Res. 2026 Mar 24;28(4):634-641. doi: 10.1093/ntr/ntaf215.

Abstract

Introduction: Misperceptions of nicotine's role in health harms among US adults have increased. This study examined how survey item wording affects estimates of the misperception that nicotine itself causes cancer.

Methods: Two thousand five hundred twenty-six US adults aged 18-45 completed a randomized split-sample experiment through Cloud Research in August 2022. After receiving one of 10 existing and novel survey items about the role of nicotine in causing cancer, participants completed an open-ended question explaining their response. Analyses examined prevalence of beliefs by study condition and open-ended responses.

Results: Overall, 36% of participants believed that nicotine does not cause cancer (coded as "correct"), but this varied by condition (10%-81%). Prevalence of this belief was similar when using magnitude-of-harm items from existing national surveys (HINTS, PATH; 44%) but lower with more direct wording (ie, "nicotine is a cause of cancer"; 23%). Changing response options (including/excluding don't know; true/false versus Likert scale) did not alter prevalence of a correct belief. Open-ended responses were distributed across four themes: (1) EXPOSURE to other chemicals or smoke cause cancer, not nicotine (n = 598, 23.7%), (2) Nicotine on its own DIRECTLY causes cancer (n = 705, 28.0%), (3) Nicotine only causes cancer by keeping people ADDICTED to tobacco (n = 796, 31.6%), and (4) Other (n = 419, 16.6%); study condition was associated with theme (p < .001).

Conclusions: Question wording influences the prevalence of agreement that nicotine itself causes cancer. Including "don't know," "probably yes," and "probably not" options may identify a broader group of people who could benefit from nicotine education.

Implications: Ensuring accuracy in estimating misperceptions about nicotine has implications for tobacco control in two important ways: first, these measures may identify the people or groups whose misperceptions may put them at greater risk for tobacco-related disease and death, and second, the prevalence of nicotine misperceptions and people affected may determine the prioritization and tailoring of interventions aimed at correcting misperceptions.

Publication types

  • Randomized Controlled Trial

MeSH terms

  • Adolescent
  • Adult
  • Female
  • Health Knowledge, Attitudes, Practice*
  • Humans
  • Male
  • Middle Aged
  • Neoplasms* / chemically induced
  • Neoplasms* / psychology
  • Nicotine* / adverse effects
  • Prevalence
  • Surveys and Questionnaires
  • United States
  • Young Adult

Substances

  • Nicotine