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

Re: st: Effect size / sample size / power calculation

From   Ronan Conroy <[email protected]>
To   "[email protected]" <[email protected]>
Subject   Re: st: Effect size / sample size / power calculation
Date   Thu, 7 Jan 2010 19:22:56 +0000

On 7 Ean 2010, at 09:41, Miranda Kim wrote:

For example, previous research on a given association yields a
correlation coefficient 0.41 with p-value 0.131 and n=15.

Initially I was looking at what sample size n would be required to have
90% power to detect a correlation coefficient 0.41 using a test at the
5% level of significance.

The question is not what the actual correlation is likely to be, but the size of the smallest correlation which is of real-life importance. This should be the basis of your power calculation. In medicine we refer to the smallest effect which is of real life importance as the smallest clinically significant difference. I imagine that it has other names in other disciplines (any volunteers?).

I see this confusion a lot as an ethics committee reviewer.

Ronan Conroy

[email protected]
Royal College of Surgeons in Ireland
Epidemiology Department,
Beaux Lane House, Dublin 2, Ireland
+353 (0)1 402 2431
+353 (0)87 799 97 95
+353 (0)1 402 2764 (Fax - remember them?)

P    Before printing, think about the environment

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index