A robust and readily implementable method for the meta-analysis of response ratios with and without missing standard deviations.