

BASEHAZARD vs. ESTBASELINE under SAVE... 

Message/Author 

Hanjoe Kim posted on Thursday, January 28, 2016  11:18 pm



What is the difference between these two commands? I ran both and found: BASELHAZARD file: the first column seemed to match the survival time values in the raw data and I assume that the second column is the estimated baseline hazards. ESTBASELINE file: I am not sure what this is...I thought the first column would be the survival time values and the second would be the estimated baseline hazards. However, the first column is showing numbers that do not match the survival time values in the raw data (although the metric seems right) and the second column values are monotonically decreasing (could it be the survival function?). Also, please tell me how the baseline hazard rates are estimated in Mplus. I am guessing it is related to Dr. Breslow's comments in the Cox (1972) paper. Ultimately, what I am trying to do is get an estimate of E(T) = Integral of the survival function, showed in the following link: http://thirdorderscientist.org/homoclinicorbit/2013/6/25/thedarthvaderrulemdashorcomputingexpectationsusingsurvivalfunctions Is there a way I can do this in Mplus? Thank you. Hanjoe. 


BASELHAZARD is the baseline hazard function, ESTBASELINE is the baseline survival probability, which is evaluated at 100 equally spaced points between 0 and the max survival time. Take a look at formulas (610) in http://statmodel.com/download/Survival.pdf The above paper also has references to the estimation procedure  see the two Larsen's papers. To get E(T) approximately/numerically (at baseline) you can sum up the survival probabilities values of ESTBASELINE and multiply by the time interval. 

Hanjoe Kim posted on Monday, February 01, 2016  1:52 pm



Thank you very much! I think it would be most helpful if the Mplus user guide is updated with this information. At the moment, the description for BASEHAZARD and ESTBASELINE is not clear. Hanjoe. 

Back to top 

