• Epidemiological measurements;
  • internet;
  • nicotine;
  • smoking;
  • substance withdrawal syndrome;
  • tobacco use disorder;
  • validation studies


Aims  To compare the psychometric properties of three cigarette withdrawal scales.

Design  An internet cohort study.

Participants  Each of 4644 current (44%), former (49%) and never smokers (7%) completed the three scales via the internet. A subsample completed the scales again after 14 days (n = 1309), and indicated their smoking status after 42 days (n = 1431).

Measurements  The Cigarette Withdrawal Scale (CWS), the Wisconsin Withdrawal Scale (WWS) and the Minnesota Withdrawal Form (MWF).

Findings  All three scales covered the main elements in the Diagnostic and Statistical Manual version IV (DSM-IV) and the International Classification of Diseases version 10 (ICD-10) definitions of tobacco withdrawal, but WWS did not cover weight gain. Factor analyses indicated that only six factors were present in WWS, instead of the expected seven factors. Cronbach's alpha coefficients (0.76–0.93) were high for all scales. Test–retest coefficients were in the range of 0.66–0.86 for CWS and WWS, but were somewhat lower for some MWF items (range 0.52–0.80). In 324 ex-smokers who had quit smoking 31 days or less before baseline, craving predicted relapse at 14-day follow-up (CWS: odds ratio = 1.53 per point, P = 0.003; WWS: odds ratio = 1.40, P = 0.04; MWF: odds ratio = 1.49, P = 0.002). In 34 baseline smokers who had quit smoking by 14-day retest, an increase in craving (WWS and MWF), depressed mood (MWF) and difficulty concentrating (WWS) between baseline and retest predicted relapse at 42-day follow-up. In terms of construct validity, scales performed similarly, but performance on some key tests (e.g. withdrawal will increase post-cessation) were inadequate, due perhaps to methodological limitations.

Conclusions  No scale showed a decisive advantage over the others. MWF has the advantage of brevity.