Meta-Analysis Resources

Tools for Those Who Summarize the Evidence Base

Resources and networking for those who conduct or interpret meta-analyses related to any phenomenon that is gauged in multiple studies.

A beginners guide you could recommend?

Hi, I hope you can help. I'm new to meta-analysis but slightly older to systematic literature reviewing. I'm trying to learn about the types of comparisons I can use in meta-analysis with newbie questions like

a) can I do meta-analysis where the outcome measures are different measures but of the same construct?

b) can you mix up study types in a meta-analysis / is it even possible to meta-analyse pre-post designs?

c) when you have RCT with pre and follow-up data - what information do you use to calculate an effect size - in same cases I have the ANOVA F statistic in others just means and standard deviations of the groups - are they ok to use?

d) In controlled studies, I've been told to extract mean outcome score at follow-up for the treatment and control group, with s.d. and n - don't I need some information about baseline too?

 

As you can see -I'm very new and confused and would be really grateful if someone could help me out and point me in the direction of a good beginners guide for this. I have at this stage no-one available in person to offer me support so any virtual-support is greatly appreciated!

 

Many thanks!

Views: 887

Reply to This

Replies to This Discussion

Dear F M,

Thanks for your questions.

a) can I do meta-analysis where the outcome measures are different measures but of the same construct?

Yes, but usually you will want to conduct a moderator analysis to see if results replicate on the different measures. If there are differences, they may be difficult to explain but at least you have this bit of knowledge to offer others who study the same issues.

b) can you mix up study types in a meta-analysis / is it even possible to meta-analyse pre-post designs?

Same answer: Do sensitivity analyses to see if the patterns differ between the types of studies. It is possible to meta-analyze studies with pre-post designs (see Becker, 1988). I discuss these issues in a recent chapter that I will email you.

c) when you have RCT with pre and follow-up data - what information do you use to calculate an effect size - in same cases I have the ANOVA F statistic in others just means and standard deviations of the groups - are they ok to use?

The chapter I mention offers equations for these comparisons. And there is a spreadsheet here on this website that makes a lot of these comparisons.

d) In controlled studies, I've been told to extract mean outcome score at follow-up for the treatment and control group, with s.d. and n - don't I need some information about baseline too?

Usually our group will correct for differences that appear at baseline, so yes, your instructions sound too simple. Tania Huedo-Medina and I recently finished a Monte Carlo simulation that offers best-practice guidelines for these situations, and in the one you mention a strategy Becker (1988) offered will work. You'll find a technical report about these, written for AHRQ, here.

Check here to see a group on this website meant for beginner meta-analysts (or those who want the big picture).

Hope that helps! -Blair

Dear Blair and F M,

I am new to meta-analysis, too. Like F M, I am working with pre-post design RCTs that report findings in various statitstics (mostly means and SD, but some F-statistic). I am using the Comprehensive Meta-Analysis software program with some difficulty. In particular, I am unsure of how to estimate the pre-post correlation if I don't have the raw data (it is not reported in any of the studies with which I'm working). I'll take a look at the links you've posted above and I'd love to see the chapter you mention, Blair.

Thank you, Blair and F M for posting this thread.

Laurie

HI Laurie, thanks for your interest! I'll email you separately with the chapter. -Blair

Thanks!



Blair T. Johnson said:

HI Laurie, thanks for your interest! I'll email you separately with the chapter. -Blair

Dear Blair

May I also please have a copy of this chapter?

Many thanks

Dan

This chapter is now available on ResearchGate (click here).

You can find the chapter I mention in earlier posts here:

http://digitalcommons.uconn.edu/chip_docs/35/

Reply to Discussion

RSS

© 2017   Created by Blair T. Johnson.   Powered by

Badges  |  Report an Issue  |  Terms of Service