Hello.
I have several questions regarding performing a survival analysis on a
small sample (sample size n between 10 an 20, for example). They are
Statistics questions but also SPSS specific ones.
(1) In case ALL observations are NON-CENSORED, could MEDIAN
estimation be done just by using SAMPLE MEDIAN, which is not always
equal to the estimation of MEDIAN survival time that SPSS uses in a
survival analysis?? (if I am not wrong, SPSS estimates MEDIAN survival
time as the smallest survival time for which the survival function is
less than or equal to 0.5, or something similar)
(2) Which specific equations does SPSS use in order to compute a
CONFIDENCE INTERVAL for MEDIAN survival time in a survival analysis??
Brookmeyer and Crowley's? Andersen's? Another?
(2bis) In the specific case of small sample size and non-censored
observations, is that CONFIDENCE INTERVAL for MEAN that SPSS computes
appropriate, or should I use another "exact" approach?
(3) Are Kaplan-Meier method for estimating survival and Log-Rank,
Breslow and Tarone-Ware tests appropriate when managing small sample
sizes? I think so, but I would like to be sure.
(3bis) Do you know an exact test for specifically testing difference
of medians between two independent populations, that goes well with
small sample sizes, and that could be then used as an alternative or
complement to Log-Rank, Breslow and Tarone-Ware tests in my case?
Any other hint, tip or piece of help will be also welcomed. By the
way, I am using SPSS 16 (base) on Windows XP.
Thank you in advance for your time and your answers!
--
vicent
estadis.webs.upv.es
@vginer_upv
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Locked
