Authors' replyBMJ 1996; 313 doi: https://doi.org/10.1136/bmj.313.7059.753c (Published 21 September 1996) Cite this as: BMJ 1996;313:753
- S Harding, Senior research analyst,
- R Balarajan, Professor
- Longitudinal Study, Office for National Statistics, London WC2B 6JP
- Institute of Public Health, University of Surrey, Guildford GU2 5YL
EDITOR,—J K Cruikshank raises the important issue of the likely contribution of smoking to the increased mortality from all cancers, and especially lung cancer. Mortality from respiratory disease was also raised in both men (non-significantly) and women. Unfortunately, we do not have data on smoking in the Office for National Statistics' longitudinal study.
We believe that, though testing for significance is central to the analysis, the observation of patterns that are consistent across sex, age, and time is equally important in epidemiological studies. Small numbers are …