Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
median survival
Message posted by P Keane on December 14, 1999 at 12:00 AM (ET)
I have 9 data points in a survival analysis 6 dead and 3 alive. If I calculate a "manual" median I return a value of 10 months if I compute the median using a computer generated Kaplan-Meier analysis I get 17.87 months. The 10 month value is obviously correct Do I have a"large sample approximation" error. Any commnets would be gratefully received
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: median survival
Message posted by JG on December 14, 1999 at 12:00 AM (ET)
You probably have a 'small sample error'. I presume that you wish to estimate the median of the population from which your sample is drawn. To do this you plot the % surving verses time and thus estimate the comulative distribution for your data. Where the cumulative curve crosse .5 is you estimate of the population median. That is if you plot your data how long the 6 survived you might easily decide that you have too many small values and that a larger sample might suggest a bigger median. Business statistics
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.
