Comments on: Random inequalities I: introduction http://www.johndcook.com/blog/2008/07/26/random-inequalities-i/ The blog of John D. Cook Fri, 19 Mar 2010 19:25:59 -0400 http://wordpress.org/?v=2.8.4 hourly 1 By: EastwoodDC http://www.johndcook.com/blog/2008/07/26/random-inequalities-i/comment-page-1/#comment-21629 EastwoodDC Sat, 25 Jul 2009 04:02:04 +0000 http://www.johndcook.com/blog/2008/07/26/random-inequalities-i/#comment-21629 John: I followed your link back to this post, and I'm glad I did - now I'm going to have to read the whole series. Good stuff! When working with someone who is planning a study, we (statisticians) need some information about the mean and variance of the outcome to estimate a sample size. This is often difficult because most people don't think about variance or variability. What I can do though is ask them to express the difference between two groups as the "odds" that given patient with treatment "A" will have a better outcome than a similar patient that receives treatment "B". This is something clinicians understand and can usually answer readily. Given the odds (and a few assumptions) I can work out P[X>Y] (where X and Y are the outcomes for group A and B), and calculate power for a Chi-square test. It's not as good as having pilot data to calculate means and standard deviations, but expressing the clinical "inequality" of treatments as "odds" is a good place to start when no better information is available. John: I followed your link back to this post, and I’m glad I did – now I’m going to have to read the whole series. Good stuff!

When working with someone who is planning a study, we (statisticians) need some information about the mean and variance of the outcome to estimate a sample size. This is often difficult because most people don’t think about variance or variability. What I can do though is ask them to express the difference between two groups as the “odds” that given patient with treatment “A” will have a better outcome than a similar patient that receives treatment “B”. This is something clinicians understand and can usually answer readily. Given the odds (and a few assumptions) I can work out P[X>Y] (where X and Y are the outcomes for group A and B), and calculate power for a Chi-square test.
It’s not as good as having pilot data to calculate means and standard deviations, but expressing the clinical “inequality” of treatments as “odds” is a good place to start when no better information is available.

]]> By: Logic Nest · Carnival of Mathematics #37 http://www.johndcook.com/blog/2008/07/26/random-inequalities-i/comment-page-1/#comment-3284 Logic Nest · Carnival of Mathematics #37 Mon, 28 Jul 2008 14:56:34 +0000 http://www.johndcook.com/blog/2008/07/26/random-inequalities-i/#comment-3284 [...] Cook from The Endeavour gives us an explanation of Random Inequalities in this three part series. Random inequalities are often used in Bayesian clinical trial methods, and should [...] [...] Cook from The Endeavour gives us an explanation of Random Inequalities in this three part series. Random inequalities are often used in Bayesian clinical trial methods, and should [...]

]]> By: Yoav http://www.johndcook.com/blog/2008/07/26/random-inequalities-i/comment-page-1/#comment-3174 Yoav Sat, 26 Jul 2008 15:45:06 +0000 http://www.johndcook.com/blog/2008/07/26/random-inequalities-i/#comment-3174 Basically, what you're saying here is that people have a hard time understanding what the p in the p(..) means Basically, what you’re saying here is that people have a hard time understanding what the p in the p(..) means

]]>