[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Wing on TM, on dose-response?

> I was surprised the other day to see one of the commentators suggest 
> that there was no showing of dose-response by Wing et al.

That was probably me.
> Don't take my word for it; check out Tables 2 and 3 in the paper. 

Yes, the Tables show O/E ratios by relative dose.  But exposure-response 
analysis requires some type of formal trend analysis (e.g., Mantel extension 
test, Cox proportion hazards).  I don't see that this has been done.  As a 
matter of fact, I cannot even determine whether any of these O/E ratios are 
significantly different from one.

When dealing with very small relative risks (and these are very small RRs by 
epidemiological standards), compulsive attention to experimental design and to 
statistical significance are required.

John Moulder (jmoulder@its.mcw.edu)