A more appropriate statistical method for analyzing mortality data in shock research.

Much of biological research deals with analysis of survival times such as time until death of an organism, length of survival of a skin graft, time of remission of an illness, or time until relief of a particular symptom. When analyzing data that consists of time until occurrence of an event, the commonly used parametric t-tests for comparing mean survival times of treatment groups is inappropriate when 1) the survival times do not follow the bell-shaped Gaussian (normal) distribution and/or 2) when some of the organisms are still "alive" (censored) at termination of the study period. As an alternative to the t-test, a relatively simple-to-use statistical procedure for comparing survival times of two groups is described. This procedure was developed to deal specifically with data consisting of survival times and to take into account the presence of censored observations.