• Definition of ‘hardcore’ smoker;
  • hardening hypothesis;
  • longitudinal telephone survey;
  • Ontario Tobacco Survey (OTS);
  • predictive ability;
  • quitting;
  • quitting behaviour;
  • resistant smoker


Aims  It has been hypothesized that the smoking population is represented by an increasingly ‘hardcore’ group of smokers who are resistant to quitting. Many definitions of ‘hardcore smokers’ have been used, but their predictive validity is unknown. To evaluate whether ‘hardcore smoker’ definition components predict quitting behaviours and which combinations of ‘hardcore’ components are most predictive.

Design, setting and participants  Longitudinal, random telephone survey of a representative sample of adult smokers in Ontario, Canada (n = 4130, recruited 2005–08 and followed for 1 year).

Measurements  Multiple logistic regression models were compared to evaluate the predictive ability of ‘hardcore’ components (high daily cigarette consumption, high nicotine dependence, being a daily smoker, history of long-term smoking, no quit intention and no life-time quit attempt) on three outcomes [continued smoking, not attempting to quit and having unsuccessful quit attempt(s)].

Findings  All ‘hardcore’ components predicted having no quit attempt and continued smoking during follow-up (P < 0.05), except for history of long-term smoking and no life-time quit attempt (for continued smoking). Among respondents who made 1 + quit attempts during follow-up, only high nicotine dependence, high daily cigarette consumption and being a daily smoker were predictive of quitting failure (P < 0.01). The best combination of components depended on the outcome.

Conclusions  Measures of ‘hardcore’ include a mixture of motivational, dependence and behavioural variables. As found previously, motivational and behavioural measures, such as intention to quit, predict failure to make quit attempts. However, dependence components best predicted continued smoking and thus would be best for further exploring the hardening hypothesis.