[R] Comparison of demographics between 2 study samples

[Andrew Robinson]

On Thu, Aug 14, 2008 at 07:46:41PM +1000, Jim Lemon wrote:
> On Wed, 2008-08-13 at 19:14 -0700, Mark Home wrote:
> > Dear All:
> > 
> > I have a clinical study where I would like to compare the demographic information for 2 samples in a study.  The demographics include both categorical and continuous variables.  I would like to be able to say whether the demographics are significantly different or not.
> > 
> > The majority of papers that I have read use multiple techniques to achieve this (e.g., t-test for the continuous variables and either Fischer exact or Chi-square for categorical).  I wonder whether this might lead to spurious differences due to multiple significance tests.  Is there a better way to do this?
> > 
> Hi Mark,
> Most of these comparisons are uncorrected, as the aim is to demonstrate
> that the samples come from the same population. Therefore, you aren't
> worried about making a Type I error, but ignoring a sampling difference
> that might bias your results.
> 
> Jim

Hi Mark,

just following up on Jim's point, if your goal is to demonstrate that
the samples come from the same population then you probably should
take a look at equivalence testing.

Andrew
-- 
Andrew Robinson  
Department of Mathematics and Statistics            Tel: +61-3-8344-6410
University of Melbourne, VIC 3010 Australia         Fax: +61-3-8344-4599
http://www.ms.unimelb.edu.au/~andrewpr
http://blogs.mbs.edu/fishing-in-the-bay/