That study is great for illustrating how deceptive percentage of 
variance explained statistics can be.  They terminated this study prematurely 
because the early returns showed an effect so large (odds ratio of about 1.8) 
that it was deemed unethical to continue the study.  In terms of variance 
explained, the treatment accounted for about one tenth of one percent of the 
variance in MI.

Cheers,
[Karl L. Wuensch]<http://core.ecu.edu/psyc/wuenschk/klw.htm>
From: John Kulig [mailto:[email protected]]
Sent: Tuesday, January 15, 2013 3:39 PM
To: Teaching in the Psychological Sciences (TIPS)
Subject: Re: [tips] Flu vaccine and p.6 level of significance


 Nice replies (Jim C, Karl W and Mike P and others ..) so I won't repeat what 
has been said except to note - as a tangent to the original posts - that in 
some of my classes I spend time with the "relative risk" Karl W discusses. I 
use the example of aspirin and MI (heart attack) in the 1988 (New England 
Journal of Med?? if I remember) article of 22,000+ physicians who took aspirin 
vs. placebo. My chi square calculated on their frequencies reveal p < .01, yet 
the risk of MI only drops from 1.7% to .9% in the sample over the years 
studied. As an absolute value, the % decrease is very small, but expressed as 
relative risk we can say we cut the risk in half. Of course, any "significant" 
decrease will be championed as the stakes are very high with MI .. and 
sometimes high with flu as well ..

At any rate, I got MY flu shot! So I am OK. p < .05 :-)
==========================
John W. Kulig, Ph.D.
Professor of Psychology
Coordinator, University Honors
Plymouth State University
Plymouth NH 03264
==========================


---
You are currently subscribed to tips as: [email protected].
To unsubscribe click here: 
http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=23047
or send a blank email to 
leave-23047-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu

<<inline: image001.jpg>>

Reply via email to