Hello,
I have come across an error message when using cem with a list of imputed datasets from Amelia:
> match <- cem(treatment = "iccpr.ratified", data = original, datalist = imputed, drop = ignore, cutpoints = list(lag.lg.pop.gled = pop.cut, lag.lg.rgdppc.gled = gdp.cut, lag.polity2.p4 = polity.cut))
Error in rbind(bigdata, datalist[[i]]) :
number of columns of matrices must match (see arg 2)
I checked the original dataset and the 5 imputed ones (imputed$imputations$imp1, etc), and they all have the same dimensions. Also, when I specify one of the imputed as the the data option, it works:
> match <- cem(treatment = "iccpr.ratified", data = imputed$imputations$imp1, drop = ignore, cutpoints = list(lag.lg.pop.gled = pop.cut, lag.lg.rgdppc.gled = gdp.cut, lag.polity2.p4 = polity.cut))
> match
G0 G1
All 1443 3350
Matched 1161 2331
Unmatched 282 1019
I will go ahead and do matching and analysis on the individual imputed datasets, and combine them with the mi.meld command, but I thought I should raise this issue here in case this is a bug.
Thank you for making Amelia and cem available.
Best wishes,
Timo