Coding Group Equivalency on Pre-Treatment Measures for Meta-Analysis?

Hello- I am constructing a code book for a meta-analysis. I only want to include evaluations where program and comparison groups were reasonably well matched on pre-existing differences in key variables, such as age, prior educational achievement, etc. Any recommendations on what statistic to use so that I can fairly consistently calculate the similarity between groups on specific variables for a meta-analysis? I have seen that some metas will only include evaluations where the treatment and comparison groups are
are matched within about 1/20th of a standard deviation at baseline on certain variables. Any ideas as to how they are calculating that standard deviation? I am sorry if this is a dumb question- I am stats newbie! Thank you in advance for any help you can provide!!