Weighted average for multiple factors?

I am working on creating a weighted average based off of multiple studies in a literature review. However, I would like to base the weights off of the number of participants as a slightly more subjective scoring system on the quality of the data. Is this possible and if it is, would this approach be statistically sound?


Well-Known Member
Weights are commonly estimated as proportional to 1/variance of the of the contributing numbers. The variance is often proportional to 1/the number of points used. This would suggest that the weights used in a weighted mean should be simply the number of participants contributing towards the various means.
Your suggestion sounds reasonable to me.