I sent this to Medstats with no response, and I would be grateful if somebody could offer me any guidance.
I
have a set of results on the failure stress of 3 groups of orthopaedic
implants. The data conforms to a Weibull distribution as assessed by
the Anderson-Darling test. However the Anderson Darling test shows that
the data also conforms to a normal distribution.
My quandary is should I use the Weibull distribution as this
fundamental to the analysis of failure, or a normal which will allow me
to use analysis of variance and contrasts to compare the groups which I
cannot using a Weibull analysis.
Is it ethical to analyse the same set of data using two different distributions?
Thank you,
Janet
___________________________________________________________
What kind of emailer are you? Find out today - get a free analysis of your email personality. Take the quiz at the Yahoo! Mail Championship.
http://uk.rd.yahoo.com/evt=44106/*http://mail.yahoo.net/uk
|