Are self-reports of smoking rate biased? Evidence from the Second National Health and Nutrition Examination Survey

J Clin Epidemiol. 1995 Oct;48(10):1225-33. doi: 10.1016/0895-4356(95)00020-5.


This study determined evidence for digit preference in self-reports of smoking in the Second National Health and Nutrition Examination Survey (NHANES II). Subjects were 4275 adult smokers. Self-reports of smoking showed a marked degree of digit preference, with the vast majority of smokers reporting in multiples of 10 cigarettes per day. When number per day was compared to an objective measure of smoking exposure (carboxyhemoglobin; n = 2070) the distribution was found to be significantly assymetrical. Analysis of the distribution of COHb and various levels of number per day indicates that the differences in distribution are not due to variability in COHb. Heavier smokers, Caucasians, and those with less education were more likely to report a digit preference than lighter smokers. African-Americans, and those with more education. Results suggest that self-reports of number of cigarettes per day may be biased towards round numbers (particularly 20 cigarettes per day). Implications for assessment of smoking behavior are discussed.

Publication types

  • Comparative Study
  • Research Support, Non-U.S. Gov't
  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Adolescent
  • Adult
  • Aged
  • Bias
  • Carboxyhemoglobin / analysis
  • Confidence Intervals
  • Female
  • Health Surveys*
  • Humans
  • Logistic Models
  • Male
  • Middle Aged
  • Odds Ratio
  • Reproducibility of Results
  • Smoking / blood
  • Smoking / epidemiology*
  • Smoking / psychology
  • Surveys and Questionnaires / standards
  • United States / epidemiology


  • Carboxyhemoglobin