2

If you have the risk ratio plus its confidence interval then (a) take logs of the RR and both limits of the CI(b) work out the standard error of the log RR from the width of the CI on the log scale, dividing by 4 should be close enough. That is it, you now have what metafor calls yi and sei If you have further more programming type questions it would be ...


1

I would at least start out with multinomial regression. In R there are multiple packages implementing this, but the simplest in use is maybe nnet with function multinom. With your example data I get mod0.nnet <- nnet::multinom(cbind(NumAsian, NumBlack, NumWhite) ~ 1, data=ExData) mod1.nnet <- nnet::multinom(cbind(NumAsian, NumBlack, NumWhite) ~ ...


Only top voted, non community-wiki answers of a minimum length are eligible