Meta analysis involves a) finding a summary statistics for each study
and b) combining the summary statistics to obtain a weighted average.
With different types of data, the summary statistic would differ. Some
averaging process (e.g. inverse variance method) can be applied to both
continuous or binary data.
For a quick introduction, read Chapter 15 of "Systematic Reviews in
Health Care : Meta-analysis in context" (ISBN : 072791488X).
Another good book to read is "Methods for Meta-analysis in Medical
Research" (ISBN : 0471490660).
Google would probably be able to find you something useful if you do not
have access to these books.
Regards, Adai
On Mon, 2005-04-18 at 07:32 +0100, Neville Calleja wrote:
> Dear colleagues
>
> I have been asked to help out on a meta-analysis to pool a number of means
> and standard errors. Unfortunately, during my statistical training, I only
> tackled meta-analysis using odds ratios.
>
> Can anyone of you point out any differences? Maybe some can point me to
> some good notes on the net/journals etc. as to which variables do I need to
> collect and how to go about it?
>
> Thanking you in advance
>
> Yours truly
>
> Neville Calleja
>
> ==================
> Dr Neville Calleja
> 12 Mon Nid
> Ganni Faure Str
> Tarxien PLA12
> MALTA
>
|