View Single Post
Old 08-18-2009, 01:22 PM   #15
medifastwoman

Join Date
Oct 2005
Posts
490
Senior Member
Default
Not being a statistician, I have no idea if the sample size and number of cities sampled combined is enough to be demonstrably significant. Certainly the scientists conducting the test appeared to think it covered their needs.

What should the number of samples have been, to be considered valid?
No one, not even a Ph.D. in statistics can answer that question. Statistics is one of the few "fuzzy" sciences... which is surprising since it is highly mathematical.

As new statistical methods are developed, they do no replace the existing ones. There are multiple methods to produce similar statistics, it is really up to the statistician's interpretation on which methods to use, since different methods can produce some wildly different results.

To answer your question, a bit of real logic can be used here: A sample of 234 bills is what percentage of the total number of bills available? Even if they are selected truly at random, what does 90% of
medifastwoman is offline


 

All times are GMT +1. The time now is 11:03 AM.
Copyright ©2000 - 2012, Jelsoft Enterprises Ltd.
Design & Developed by Amodity.com
Copyright© Amodity