[R-sig-ME] Multilevel weighted regression
brant.inman at me.com
Mon Oct 27 16:38:48 CET 2008
The mima function is indeed very useful for meta-regression and I thank you for kindly sharing it. The problem is that the example I have to work with in real life has covariates both at the study level (Level 2) and the patient level (Level 1), and to my knowledge this is not a scenario that mima can handle.
My main interest with lmer and the BCG example is that I might be able to generalize it to the scenario where I have 2 different levels of covariates. However, prior to doing that, I need to figure out how to get lmer to work right with just one level of covariates.
Now we are two to wait for some guidance on lmer in the context of meta-analysis!
>This is not a direct answer to your question, but something that still may be useful for you. Using the normal approximation to the log relative risk:
>yi <- c(-0.89, -1.59, -1.35, -1.44, -0.22, -0.79,
> -1.62, 0.01, -0.47, -1.37, -0.34, 0.45, -0.02)
>vi <- c(0.326, 0.195, 0.415, 0.020, 0.051, 0.007,
> 0.223, 0.004, 0.056, 0.073, 0.012, 0.533, 0.071)
>ablat <- c(44, 55, 42, 52, 13, 44, 19, 13, 27, 42, 18, 33, 33)
>year <- c(1948, 1949, 1960, 1977, 1973, 1953,
> 1973, 1980, 1968, 1961, 1974, 1969, 1976)
>mima(yi, vi, mods=cbind(ablat, year), method="REML")
>This is for the random-effects model. To get the same results as your model f1 (without having to go through the adjustment step for the SEs):
>mima(yi, vi, mods=cbind(ablat, year), fe="yes")
>There is a short tutorial about the function at: http://www.wvbauer.com/downloads.html
>I am curious as well to hear how lmer can be used in this context.
> Department of Methodology and Statistics
> University of Maastricht, The Netherlands
More information about the R-sig-mixed-models