Open RajarshiBhadra opened 8 years ago
The standard error calculations are getting affecting due to the presence of NA s in the ucm function.
The current code for standard error calculation is se <- sapply(find.indep.var, function(x) mean(sqrt(out$V[x, x, ])))
se <- sapply(find.indep.var, function(x) mean(sqrt(out$V[x, x, ])))
It has to be modified to se <- sapply(find.indep.var, function(x) mean(sqrt(out$V[x, x, ]), na.rm=T))
se <- sapply(find.indep.var, function(x) mean(sqrt(out$V[x, x, ]), na.rm=T))
So that NAs are ignored. This takes care of the situations where estimates are available but standard errors are NA in the current code
The standard error calculations are getting affecting due to the presence of NA s in the ucm function.
The current code for standard error calculation is
se <- sapply(find.indep.var, function(x) mean(sqrt(out$V[x, x, ])))
It has to be modified to
se <- sapply(find.indep.var, function(x) mean(sqrt(out$V[x, x, ]), na.rm=T))
So that NAs are ignored. This takes care of the situations where estimates are available but standard errors are NA in the current code