Meta-analysis

View previous topic View next topic Go down

Meta-analysis

Post by CatieWalsh on Tue Mar 07, 2017 3:16 pm

Hi All!
If you have been directed to this post, you stated that you were interested in, or are currently using, quantitative meta-analytic techniques in your analyses. Please use this forum to connect with each other either in person or online... It would be great if we could share knowledge and advice with each other!
avatar
CatieWalsh

Posts : 22
Reputation : 0
Location : OEH sometimes, but mostly wandering around Sennott

View user profile

Back to top Go down

Unsure if my comps topic lends itself to meta-analysis

Post by mbowdring on Fri Mar 17, 2017 9:16 am

Hey guys,


I'm in the process of figuring out the details of my comps proposal. I have a topic and was hoping to propose a meta-analysis, as I think it'd be a useful technique to learn. However, I'm not sure if it makes sense due to there being a limited number of studies (about 12) and the studies being split between experimental and observational designs. Furthermore, about half of the studies have a positive effect of alcohol, about a quarter have a negative effect, and about a quarter have no (significant) effect. 


So although there will likely not be a large (or potentially even significant) effect if I use meta-analysis, could it still make sense to do it? I plan to stick with my topic either way, as I hope to highlight the many methodological limitations of past studies and call for improvements. I'm just not sure how to best organize what's in the literature to date.


Thanks for any help you can give me!


-- Molly
avatar
mbowdring

Posts : 2
Reputation : 0

View user profile

Back to top Go down

Re: Meta-analysis

Post by Jeffrey Girard on Tue Mar 21, 2017 1:49 am

mbowdring wrote:I'm not sure if it makes sense due to there being a limited number of studies (about 12) and the studies being split between experimental and observational designs.
An ideal number for comps (in my opinion) would be around 20-40 studies, especially if you hope to publish it afterwards. I passed my comps with a meta-analysis of 11 studies, so it is possible, but I don't think I'd have much luck trying to publish it in that form. My plan is to loosen the inclusion criteria in order to expand it to ~25 studies and then try again after dissertation. Knowing what I know now, I would still pick this project and would still do a meta-analysis, but I would probably loosen the inclusion criteria from the start. If this is the topic you are set on and there are no ways to loosen the criteria to get more studies, I think it's still worth it to do a meta-analysis. It can be difficult and stressful to learn, but you'd have help from CAMEL and it really makes the write-up easier and your conclusions a lot more precise.

mbowdring wrote:Furthermore, about half of the studies have a positive effect of alcohol, about a quarter have a negative effect, and about a quarter have no (significant) effect. So although there will likely not be a large (or potentially even significant) effect if I use meta-analysis, could it still make sense to do it?
This is a great reason to do a meta-analysis! The overall effect size might not be significant, but if there is a lot of heterogeneity among the studies contributing to it, you would have the potential to do some really interesting and (hopefully) fruitful moderation analyses. The dream would be to identify the variables that explain why some studies have found negative results and others have found positive ones. You probably wouldn't have enough power to answer this question definitively, but you could reveal some important information that might motivate future work in the area (e.g., your dissertation).
avatar
Jeffrey Girard
Admin

Posts : 16
Reputation : 0
Location : 4325 Sennott Sq

View user profile http://www.jmgirard.com

Back to top Go down

Re: Meta-analysis

Post by mbowdring on Tue Mar 21, 2017 6:15 am

Thanks Jeff, this is really helpful!! I'm sure I'll have more questions as I move forward, but now I feel a bit more confident working toward proposing a meta-analysis.
avatar
mbowdring

Posts : 2
Reputation : 0

View user profile

Back to top Go down

Re: Meta-analysis

Post by Sponsored content


Sponsored content


Back to top Go down

View previous topic View next topic Back to top


 
Permissions in this forum:
You cannot reply to topics in this forum