I can't seem to find a simple resource that explains what I can use the I2 (derived from Cochrane's Q), often used in meta-analyses for, exactly. Most, if not all resources I find on the subject go on some from of tangent, mostly ending in "it depends".
If I find a published meta-analysis, citing an I squared value to assess heterogeneity, with values low as 25% or less, can I, as a non-statistician (physician) say that this means the studies are homogeneous and whatever intervention they were using is probably working across different demographies (assuming that's what their results are)?
