Hi everyone

I am doing my first meta-analysis using a three-level model.

Looking at the information of study heterogeneity, a very (and seemingly unusually) large percentage of variance is accounted for at level 1 using the I^2 statistics:

When examining heterogeneity in the three-level model, sampling error heterogeneity (level 1) accounted for 54.67% of the variance in effect size; with 26.32% being accounted for by variance at the study level (level 2) and between-study heterogeneity accounting for 19.01% of the variance (level 3).

I have tried to make some sense of this, in relation to discussion of results - what to make of this, what does it signify, and are there any precautions I should take when interpreting the results?

All the best,

Katrine