time intervals

[fanny]

Hello,

in the robudt design, the time length affect the survival estimate. Is the only one?

[Bill Kendall]

Fanny,
Gary should confirm this, but I believe time scaling only applies to survival for the robust design using MARK. Whereas scaling makes complete sense for survival, for transitions to and from an unobservable state (which is what the gamma's estimate), it only makes sense in specific cases.

[gwhite]

Bill:

Correct -- only the survival rate is affected by the time interval. A user can verify this by running 2 datasets with different time intervals, and checking the real parameter estimates against each other for the 2 runs.