Multilevel survival models and Small-sample inference
Stata 14 introduces new command
mestreg which fits parametric mixed-effects survival models with single or multiple-record st data. The new command supports a variety of distributions: exponential, loglogistic, Weibull, lognormal and gamma and proportional hazards and accelerated failure-time parametrizations.
Using the new option
dfmethod() in the known command
mixed users can now implement denominator degrees of freedom (DDF) adjustment methods for small-sample inference: these methods include Kenward-Roger, Satterwhaite, ANOVA, repeated-measures ANOVA, and residual DFF adjustments. Users can also report the DDF for each coefficient using estat
ddf command.
Available Estimators |
Available Features |
Five distributions are supported: exponential, loglogistic, Weibull, lognormal and Gamma |
Graph survivor, cumulative and hazard functions after multilevel mixed-effects survival models |
Two parameterizations: proportional hazards and accelerated failure-time |
Linear mixed-effects models with small-sample denominator degrees of freedom: Kenwar-Roger, Satterwaite, ANOVA, repeated-measures ANOVA, residual |
Single or multiple-record st data allowed |
test, testparm and lincom permit small-sample inference after mixed |
|
Survey features now supported by all multilevel mixed-effects GLM |
Return to Stata 14 New Features