2
$\begingroup$

I am attempting to calculate Orwin's (1983) modification of Rosenthal's (1979) Fail-safe N for my meta-analysis of Odds Ratios.

However, all the equations I am finding are using Cohen's d, which I cannot calculate (I don't have two groups).

I have STATA, SPSS, and MedCalc at my disposal.

$\endgroup$
2
$\begingroup$

Normally you always find ways to convert effect size. For example you can calculate $r$ from $d$ and back. So you surely will be able to converts odds ratio to Cohen's $d$.
One book I usually found a good resource for stuff like that is the one by Rosenthal & Rosnow. I think it was this one:
Rosenthal & Rosnow - Essentials of Behavioral Research: Methods and Data Analysis

But in your case you should take a look at his paper, it will probably solve your problem (but I haven't looked at it):
Chinn, S. (2000). A simple method for converting an odds ratio to effect size for use in meta-analysis. Statistic in Medicine, 19, 3127-3131.

$\endgroup$
1
  • $\begingroup$ +1 Great link, I missed this reference -- so I guess now it can be handled within R without any problem :) I put the PDF at this address in case someone want to read it: j.mp/bnnwsu $\endgroup$ – chl Sep 17 '10 at 9:08
1
$\begingroup$

So, perhaps check these additional resources: http://j.mp/d8znoP for SPSS. Don't know about Stata. There is some R code about fail-safe N in the following handout: Tests for funnel plot asymmetry and failsafe N, but I didn't check on the www.metaanalysis.com website.

Otherwise, ClinTools Software may be an option (I hope the demo version let you do some computation on real data), or better the MIX software.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy