Hi everyone! I am a biologist from Argentina and have to solve this problem. I have an insect population obtained from 10 different nests and need to know its sex ratio. But as I cannot ensure insects independence I need to run a model where I can include the variable “nest” as with a random effect. The response variable has a binomial distribution (males or females). I’ve been reading for a while and found the MASS and lmer packages that will allow me to do such a thing with my data. I found the script I should write “lmer(y~fixed+(time | random), family=binomial)” but, despite of using the R commander cannot make it work. Can you help me with this? I’ll appreciate that! Thanks, Martín PS: mi data has to columns (nest & sex) and about 200 rows. [[alternative HTML version deleted]]
Mart?n Quiroga <mquiroga <at> ssdfe.com.ar> writes:> The response variable has a binomial distribution (males or females). > I?ve been reading for a while and found the MASS and lmer packages that will > allow me to do such a thing with my data. I found the script I should write > ?lmer(y~fixed+(time | random), family=binomial)? but, despite of using the R > commander cannot make it work. > Can you help me with this?No, because "cannot make it work" could be everything from a bad hardware, a bug in XP/Linux to wrong data format or forgetting library(). Post a sample of the data including the code you tried. Dieter
On Fri, Mar 13, 2009 at 2:39 PM, Mart?n Quiroga <mquiroga at ssdfe.com.ar> wrote:> Hi everyone! I am a biologist from Argentina and have to solve this problem. > I have an insect population obtained from 10 different nests and need to > know its sex ratio. But as I cannot ensure insects independence I need to > run a model where I can include the variable ?nest? as with a random effect. > The response variable has a binomial distribution (males or females). > I?ve been reading for a while and found the MASS and lmer packages that will > allow me to do such a thing with my data. I found the script I should write > ?lmer(y~fixed+(time | random), family=binomial)? but, despite of using the R > commander cannot make it work. Can you help me with this? I?ll appreciate > that!The package is called lme4 and the function is called lmer. To be able to use the lmer function you must first enter library(lme4) in a script or console session (I'm not sure how this would be done in R Commander). There is a mailing list, R-SIG-Mixed-Models at R-project.org specifically for discussion of mixed models. I am cc:ing that list on this reply.