Skip to main content

Table 2 Minimal effect size that can be detected for the revised sample size given comparison arm risk behaviour prevalence, for a range of ICCs*

From: The prevention access and risk taking in young people (PARTY) project protocol: A cluster randomised controlled trial of health risk screening and motivational interviewing for young people presenting to general practice

   Effect size
Risk behaviours Control[7] ICC = 0.01 ICC = 0.04 ICC = 0.07
Alcohol use 41% 11% 13% 15%
Substance abuse 38% 11% 13% 15%
Tobacco use 24% 9% 10% 12%
  1. ICC = Intracluster Correlation Coefficient; Control = comparison arm.
  2. *Calculations assume an 80% power with alpha of 5% for a two-sided test.