[Minutes and Presentation] Fall 2016 Meeting #12 - Applied Meta-Analysis 11/29/16

Hi all,

Becca Emery gave a great overview of how to actually DO a meta-analysis, including how to frame your questions and your literature search, organize your data, and all the great nitty-gritty stuff that you usually only learn through trial and error. I've attached her slides from yesterday here as well as some notes from her presentation. Having not actually done a meta-analysis, these aren't my recommendations but rather just some of the big points that I picked up on that didn't seem to be explicitly stated in the slides (for those of you who may not have been able to make it this week):

- The most important thing with a meta-analysis is to keep EVERYTHING organized and record every single decision you make, such as what search terms you use and how many articles they yield, what exclusion criteria you add/remove and when you made those changes, why you excluded every article that was excluded, etc. If you need to make changes later or justify a decision to a journal, this will help immensely.

- A lot of journals have clear guidelines about what they want out of a meta-analysis, including what information they want you to report - check these early (especially if you are doing this for comps and have a journal in mind!). You should also check out the Meta-Analysis Report Standards for a sort of checklist of what information should be included (a section of this is shown in the slides).

- One other big take home to keep in mind is that a majority of the time, thought, and effort in these projects is on the forefront (or at least should be), as the analyses are often really straightforward. Smart planning from the beginning will pay off (e.g., deciding on eligibility criteria for what studies to include so that you read everything that's relevant but only that). 

- Some journals have recommendations for restricting studies based on region/language. If you have a journal in mind, check if this is the case, as you may plan to exclude articles that aren't published in English but the journal may not allow that. 

- One final thing, be prepared to contact authors, a lot. Becca mentioned several reasons why this might come up, such as not having the effect sizes (or stats to calculate effect sizes) that you need, not having enough information on the sample, or not being able to access a paper. This would also be a good way to get access to unpublished work! One recommendation she gave was to set deadlines for authors when you contact them so that they know you need the information by a certain date a few weeks away (and don't answer a year later).