I am analyzing a dataset on the effects of six pesticides on population
growth rate of a predatory mite. The response variable is the population
growth rate of the mite expressed as ln(Nfinal/Nstarting) of the mite,
where N final the population of the mite at the end of the experiment
and N starting the population of the mite at the beginning of the
experiment. Each of the six treatments was ran in three time blocks with
2 replicates per block. In the third block I lost 1 replicate for four
of the six treatments. I am analyzing the data in the nlme package using
model<-lme(growth.rate~treatment,random=~1|block). When I ran
intervals(model), the confidence intervals of the variance of the random
factor range from 0 to inf. Any comments as to why I get unrealistic
confidence intervals for the random factor?
In another study, I am investigating the interactions between pesticides
in a two-way design: (pesticideA x no pesticide A) crossed with
(pesticideB x no pesticide B). The blocks and number of replicates is as
above, and the data are unbalanced again. The model is defined as
model<-lme(growth.rate~pestA*pestB,random=~1|block). When I run
intervals (model), I usually get the following error message: "Error in
intervals.lme(model) : Cannot get confidence intervals on var-cov
components: Non-positive definite approximate variance-covariance". I
have read on the mailing list that the error message appears when the
model is not well-specified, but I do not see an alternative way of
specifying the model.
Any ideas as to why I get wide confidence intervals or the error
message? Any recommendations on other possibilities for analyzing the
data are greatly appreciated. I cannot use aov because of the unbalanced
data. I have tried to use Type III sums of squares in the car package
[model<-aov(growth.rate~pestA*pestB+Error(block))], but when I run
Anova(model,type=c("3")) I get the following error message: Error in
as.data.frame.default(data, optional = TRUE) : cannot coerce class
'"function"' into a data.frame
Thank you very much,
Menelaos Stavrinides
Lecturer, Dept. of Agricultural Sciences
Cyprus University of Technology
P.O. Box 50329, 3603 Limassol, Cyprus
Tel.: + 357 25002186
Fax: + 357 25002767
Email: m.stavrinides@cut.ac.cy
[[alternative HTML version deleted]]
