Study members have been asked about their smoking habits at ages 20, 25, 31, 36, 43, 53, 60-64 and 68-70 years. The smoking questions distinguish between cigarettes, cigars and pipe smoking. Records are made of the ages of starting and stopping smoking and the quantities smoked.

Derived variables named 'cigsta…' are available which describe smoking status based on cigarette smoking at ages 25, 31, 36, 43, 53, 60-64 and 68-70 years. They show whether a study member was a current smoker, an ex smoker or had never smoked at each age.

There are also measures of the number of cigarettes smoked, including longitudinal 'pack years' variables, which give a measure of lifetime (up to age 60-64) smoking intake.

Lifetime Smoking Trajectory

Current cigarette-smoking status (‘yes’, ‘no’) and the number of cigarettes smoked per day that were obtained in the first six such waves (at ages 20, 25, 31, 36, 43 and 53 years) were used to create a lifetime smoking trajectory variable.

Study members who provided an affirmative response to being current cigarette smokers, regardless of the quantity of cigarettes smoked per day, were classified as ‘smokers’. Those who provided negative responses were classified as ‘non-smoker’.

From age 36 onwards, research nurses obtained information on smoking during a home interview, while data at earlier years were collected by postal questionnaire. 

Study members who provided data for at least three waves and for whom missing data are not sequential were included in the smoking trajectory variable (SMOCatBU). Study members were classified into one of four smoking trajectories:

  1. Never smoker: a non smoker at all available data collections

  2. Predominantly non-smoker: a non smoker for at least three data collections

  3. Predominantly smoker: a smoker at four or more of the data collections

  4. Lifelong smoker: a smoker at all available data collections

In cases where data were missing from one or more waves, classification was based on the smoking status at the majority of data collections. Thus, if at three of four data collections the cohort member reported being a smoker, then they were classified as ‘predominantly smoker’.

Key publications

Clennell, S., Kuh, D., Guralnik, J. M., Patel, K. V., Mishra, G. D., Characterisation of smoking behaviour across the life course and its impact on decline in lung function and all-cause mortality: evidence from a British birth cohort. J Epidemiol Community Health 2008;62:1051-1056

Obtaining the standard smoking variables

  • You can obtain a list of the standard topic variables to use in an NSHD data sharing request by selecting the link at the end of this page.

Please Note that if you are interested in smoking status at various years, the generally-used variables are the 'c' ones, e.g. cigsta82c, cigsta09c - however these may not be the best variables for your project, be sure to investigate the smoking variables further on Skylark before using them.

More information is available on this topic - including the main variable naming conventions and some value labels.

HTML version of the standard variables to view

  • mrepo/topics/smoking.txt
  • Last modified: 3 months ago
  • by director