I am using Stata 15 (64-bit). I am meta-analysing hedges g effect sizes (CI 95%).
When I transform the variables to log, with 'ln', and run the meta-analysis, stata tells me there is an error - "
Effect size and confidence intervals invalid: order should be {effect size, lower ci limit, upper ci limit}".
the command I am using is 'metan lnhg lnlci lnuci, random lcolss (Study TraumaSample NonTraumaSample)
As such I tried to run the command without transforming the data to log, and this results into a meta-analysis, however I don't understand why it does not do so when I transform the data to log
I browsed the data to see if there was an issue with the inputed the data, and my effect sizes for some studies are smaller than the lower CI. I am assuming that is why stata wont run the meta-analysis? however, I cant help my effect sizes being smaller than the lowerCI, is there anything I can do about this? could I run the meta analysis without transforming the variables to log?
As such I tried to run the command without transforming the data to log, and this results into a meta-analysis, however I don't understand why it does not do so when I transform the data to log
I browsed the data to see if there was an issue with the inputed the data, and my effect sizes for some studies are smaller than the lower CI. I am assuming that is why stata wont run the meta-analysis? however, I cant help my effect sizes being smaller than the lowerCI, is there anything I can do about this? could I run the meta analysis without transforming the variables to log?
0 Response to Stata command error - "Effect size and confidence intervals invalid: order should be {effect size, lower ci limit, upper ci limit}"
Post a Comment