[R-meta] A continuous moderator whose max can't be determined

Stefanou Revesz @te|@noureve@z @end|ng |rom gm@||@com
Fri Oct 1 20:59:06 CEST 2021


Dear Meta Community,

We are coding for a continuous, intervention-level moderator that can
take whole numbers 1, 2, 5 etc.

This moderator has to due with how targeted the intervention has been
in addressing the flaws that came up in a creative human product.

While many studies clearly state how many types of flaws (1 or 2 or 5)
their intervention was designed to target, there are some studies that
simply say *"our intervention has no parameter on the type of flaws"*
which means their intervention targeted ANY kind of flaw that came up.
So, for each such study, ANY kind of flaws might mean a different but
non-infereable number.

This is an important control moderator, but we're wondering given that
the max number of flaws in some studies (N = 10 out of 49) can't be
determined, can we still use this moderator?

Thank you for your consideration,
Stefanou



More information about the R-sig-meta-analysis mailing list