For my graduate reserach, I am designing an experiment with plants that may or may not survive, given a certain treatment. I am trying to decide on a statistical analysis - so I know how to collect my data - that informs me how many plants survive for each treatment.

Survival analysis sounds like a good method, but the concept is new to me, so I am not sure. Wikipedia states that:

Survival analysis is a branch of statistics for analyzing the expected duration of time until one or more events happen

However, I am not interested in the time-until-death, but the chance of survival (or the proportion of plants that ultimately survives).

Can survival analysis be used

  • when not all individuals eventually die?
  • to estimate the proportion of individuals that ultimately survives?

If not, what other method should be used?

  • $\begingroup$ see logistic regression $\endgroup$
    – ocram
    Commented Aug 22, 2018 at 10:27

1 Answer 1


In all realistic problems, observation time $T_O$ and survival times $T_i$ are finite, so "never dies" is defined as the event $T_i > T_O$. If:

  1. you really care only about that event, and not about any differences in $T_i$s below $T_O$, AND
  2. all plants have been observed for the same duration,

then logistic regression provides the same results while being conceptually easier. Otherwise the answer to both of your questions is cure fraction models - they explicitly model a fraction of population as "cured", and allow to test whether your exposure has an effect on this fraction.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.