Author Topic: Working with Meta-analyses  (Read 3674 times)

0 Members and 1 Guest are viewing this topic.

Offline Dev

  • Administrator
  • Sr. Member
  • *****
  • Posts: 60
  • Karma: 2
  • Forum Administrator
    • ClinTools
Working with Meta-analyses
« Reply #1 on: 21 July, 2010, 06:47:19 PM »
I've just had a question that I think others may benefit from seeing the answer. Here's what it said:

Quote
Hi there,
We are interested in purchasing the Effect Size Generator Pro in order to conduct a meta-analysis. We have downloaded the free version and we see that there are a limited number of categories in the database (e.g., study ID, study name, year of study, etc). We are wondering whether we can add more columns to the spreadsheet as we are interested in testing for moderators, or specific elements of the study.
Please let us know if this is possible.
Best,

Here's my answer:

Dear xxx,

There are a few ways of doing what you want but I'm not sure that all of them are easy (maybe even a little cumbersome). If you get a *.met file (let's call it metaanalysis1.met) and enter data you can put many elements from the same study into the file but under the 'include' heading change it to YES or NO (just write over) depending upon whether you want to include that element (i.e., that row) in the meta-analysis. I save the file under a new name before making any changes this way and it means I don't have to change back the yes and no's for each analysis (i.e., I keep a different .met file for each analysis).

If you want to include more columns, however, then you will need to export it as a microsoft access file (e.g., metaanalysis2.mdb) and work with it in Access. However, make sure you don't re-order the columns (you can add columns to the right though) and when it comes to wanting to work with the Effect Size Generator again, you'll need to cut-off those extra columns away and then just rename the file from, say, metaanalysis2.mdb to metaanalysis3.met (simply just rename the file and open it as normal).

Hope this helps,

Dev
I am logged on, therefore I am.