Dear Alan:
I think the techniques presented in the book
Brunner, E., Domhof, S. and Langer, F. (2002). Nonparametric analysis of
longitudinal data in factorial experiments. John Wiley & Sons, New York.
may fit in your problem. The authors have SAS macros and R functions
available.
I have a paper with a simulation study, some other references and an
application
Singer, J.M., Poleto, F.Z. and Rosa, P. (2004). Parametric and
nonparametric analyses of repeated ordinal categorical data. Biometrical
Journal 46, 460-473.
Good luck,
Frederico
Frederico Zanqueta Poleto
[log in to unmask]
--
"An approximate answer to the right problem is worth a good deal more than an exact answer to an approximate problem." J. W. Tukey
Alan wrote:
>I have the data from a study which comprises a number of organisms (5)
>exposed to one of 4 concentrations of a test compound, with parameters of
>interest being measured at 4 separate time points.
>
>For each of 10+ parameters, I wish to find out whether there is a
>significant effect due to the compound, and does the effect vary
>significantly over time. Additionally, I am interested in the multivariate
>case in order to make a generalisation about the overall effect of the
>compound on the organisim
>
>The data do not meet the assumptions for ANOVA and so preclude the use of
>RM ANOVA or ANCOVA with time as a block or covariate respectively. For most
>of the parameters measured, there does appear to be an effect due to time,
>but it is not linear.
>
>Initially, I considered Friedmans test, but this would not tell me about
>the time effect and to do this, I would have to average the values at each
>treatment\time to get one measurement and hence lose information about the
>variability of the data.
>
>Any suggestions on how best to analyse this data gratefully received
>
>Thanks
>Alan
>
>
>
>
|