Hi all !
I am helping a doctor with the statistics in her thesis.
They have made a 2x2 factorial experiment with 6 replicates
per factor combination.
They have 5 response variables which are concentrations of
different substances. In the analysis I used 2-way ANOVA
with the ordinary F-tests.
For 4 of the responses there seems not to be any serious
violations of the underlying assumptions.
The values of one of the variables, however, are rather close to zero.
All 6 values for one of the factor combinations are in fact zero.
What should I do with this variable ? We are interested in tests of as well
main effects as interactions.
Could someone give me some hint or some reference ?
Kjell Pettersson
Department of Statistics
Goteborg University
[log in to unmask]
|