It appears that risk ratio (comparing p/(p+q) for each group) is preferred to odds ratio (using p/q for each group) in some medical journals. The <meta esize> command asks for numbers corresponding to p & q. If I enter p & (p+q) instead, it will calculate the ln RR but are there consequences I've not noticed?
Related Posts with Risk ratio meta analysis - naive question
marginsplot after running a multinomial logit: how to plot the cumulative probabilities as an area graph- Theoretical question about the importance of stationarity in panel data for OLS estimator
- Clustered standard errors with cross-sectional data
- Parsing a string variable of a dta file using a dictionary
- Codes gets highlighted in yellow when executed in Stata console - Execution impact
Subscribe to:
Post Comments (Atom)
0 Response to Risk ratio meta analysis - naive question
Post a Comment