[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: fair reporting of the Hosmer Lemeshow statistic
Due to a skewed data set, the standard lfit, group(10) command
divides my data into only 7 separate groups.
I can report the statistic in that way, or if I try lfit, group(18) (or so)
then I end up with 10 groups (and a statistic which is still fine, but
somewhat less impressive (I know, I know...))
Any thoughts as to which it should be?
* For searches and help try: