Re: [R] Summary values from Glm function (rms package)

2012-03-22 Thread Frank Harrell
The smearingEst function in the Hmisc package, given a set of residuals, can compute the smearing estimator of the mean. To get the confidence interval for the mean or difference in the mean will take some work (unless you have no covariates to adjust for; in that case you can just bootstrap the o

[R] Summary values from Glm function (rms package)

2012-03-21 Thread Yonghao
Dear fellow R-users, I’m using the Glm function (gamma family of distributions) from the rms package to compare 2 groups on costs data. Although the summary function does provide the mean cost difference and standard errors, I believe these values were in the (natural) log ratio format. Is there

[R] Summary values from Glm function (rms package)

2012-03-21 Thread Yonghao
Dear fellow R-users, I’m using the Glm function (gamma family of distributions) from the rms package to compare 2 groups on costs data. Although the summary function does provide the mean cost difference and standard errors, I believe these values were in the (natural) log ratio format. Is there