Ioulia Bespalova
2012-May-11 06:32 UTC
[R] Additional info: help with SMATR: help with pairwise comparisons using MA regression?
Also, this works (taking out multcomp=TRUE, multcompmethod="adjusted"): com.test=ma(Head.W1~Leg.3.1+Site, type="elevation", data=queens) print(com.test) ....so for some reason it will do an MA regression on all my data point together, but shows an error when I try to do pairwise comparisons between groups. Thank you, Ioulia -------- On Fri, May 11, 2012 at 2:09 AM, Ioulia Bespalova <bespa20i at mtholyoke.edu> wrote:> Hi there, > > I've been using the SMATR package to do standardized major axis (SMA) > regression on allometric data, and I've been able to to pairwise > comparisons of slope & elevation between multiple study sites. > > Now I'm trying to do the same thing using major axis (MA) regression, > but I'm getting errors (SMATR should be able to do MA as well as SMA > regression for everything, right?) > > I've attached the data file I'm using, the code looks like this: > > library(smatr) > > queens <- read.table("queens.txt", header=TRUE, sep="\t") > > # This works: > com.test=sma(Head.W1~Leg.3.1+Site, type="elevation", data=queens, > multcomp=TRUE, multcompmethod="adjusted") > print(com.test) > > # This doesn't (all I did was change sma to ma, the package manual > makes it seem that easy...): > com.test=ma(Head.W1~Leg.3.1+Site, type="elevation", data=queens, > multcomp=TRUE, multcompmethod="adjusted") > print(com.test) > > This is the error I get when I try to do MA regression: > Error in model.frame(formula = ..1, data = list(Site = c(2L, 2L, 2L, 2L, ?: > ?'nthcdr' needs a list to CDR down > > Any idea as to why it won't work? > > Thank you for this resource! > > Ioulia