Abstract Background Smoke-free policy aims to protect the health of the population by reducing exposure to environmental tobacco smoke (ETS), and World Health Organisation (WHO) guidance notes that these policies are only successful if there is full and proper enforcement. We aimed to investigate the problem of resistance to smoking restrictions and specifically compliance with smoke-free policy. We hypothesised that an explanation for non-compliance would lie in a measurable difference between the smoking behaviours of compliant and non-compliant smokers, specifically that non-compliance would be associated with nicotine dependence and different reasons for smoking. Methods We conducted a questionnaire-based, descriptive, cross-sectional study of hospital employees. Seven hundred and four members of staff at Addenbrooke's Hospital, Cambridge, UK, completed the questionnaire, of whom 101 were smokers. Comparison between compliant and non-compliant smokers was made based on calculated scores for the Fagerström test and the Horn-Waingrow scale, and level of agreement with questions about attitudes. For ordinal data we used a linear-by-linear association test. For non-parametric independent variables we used the Mann-Whitney test and for associations between categorical variables we used the chi-squared test. Results The demographic composition of respondents corresponded with the hospital's working population in gender, age, job profile and ethnicity. Sixty nine smokers reported they were compliant while 32 were non-compliant. Linear-by-linear association analysis of the compliant and non-compliant smokers' answers for the Fagerström test suggests association between compliance and nicotine dependence (p = 0.049). Mann-Whitney test analysis suggests there is a statistically significant difference between the reasons for smoking of the two groups: specifically that non-compliant smokers showed habitual smoking behaviour (p = 0.003). Overall, compliant and non-compliant smokers did not have significantly different attitudes towards the policy or their own health. Conclusion We demonstrate that those who smoke in this setting in contravention to a smoke-free policy do so neither for pleasure (promotion of positive affect) nor to avoid feeling low (reduction of negative affect); instead it is a resistant habit, which has little or no influence on the smoker's mood, and is determined in part by chemical dependence.