Endgames Statistical Question

Non-parametric statistical tests for two related groups: numerical data

BMJ 2012; 344 doi: http://dx.doi.org/10.1136/bmj.e2537 (Published 11 April 2012) Cite this as: BMJ 2012;344:e2537
  1. Philip Sedgwick, senior lecturer in medical statistics
  1. 1Centre for Medical and Healthcare Education, St George’s, University of London, Tooting, London, UK
  1. p.sedgwick{at}sgul.ac.uk

Researchers investigated the effects on the health of bar staff of legislation for smoke free workplaces implemented in the Republic of Ireland in 2004. They used a before and after study design. Participants were bar staff in public houses in the Republic of Ireland. Bar staff in public houses in Northern Ireland acted as a control group. The analyses included exposure of non-smokers to secondhand smoke. Outcome measures included the laboratory assessment of salivary cotinine concentration, which was compared before and after the law change.1

In total, 111 participants in the Republic of Ireland and 20 in Northern Ireland had cotinine concentrations measured at baseline and one year later at follow-up. Non-parametric statistical tests were used, as the distribution of salivary cotinine concentration was not normal. The median salivary cotinine concentration of bar staff in the Republic of Ireland was 29.0 nmol/L before the legislation and 5.1 nmol/L afterwards (median within-subject reduction of 22.7 nmol/L (P<0.001)). The median concentration in Northern Ireland was 25.3 nmol/L before the legislation and 20.4 nmol/L afterwards (median within-subject reduction of 5.7 nmol/L (P=0.05)). The researchers …

View Full Text

Sign in

Log in through your institution

Free trial

Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
Sign up for a free trial

Subscribe