Skip to content
Prev 138284 / 398506 Next

Problem plotting curve on survival curve (something silly?)

Calum wrote:
Customarily not marked differently from other kinds of censoring. (The
censoring date has to be "last day known alive", of course.)
Now there's the first problem... For survival models, the R^2 measures
which I have heard of, measure the predictability of the outcome, rather
than the discrepancy between observed and expected survival curves. I'm
not aware of nice goodness of fit procedures for survival curves (others
on the list might be) .