[R-sig-ME] lme random effects in additive models with interaction

[Ben Bolker]

 <elifnurdogruoz at ...> writes:

> 
> 
> Hello,I posted this
> message previously on R-help.

  Your data and code came through as almost hopelessly mangled,
at least when viewed in Gmane (a web portal for the mailing list).
Could you please try re-posting in plain text rather than HTML
(as you are requested by the posting guide) to improve the chances
that it gets through?

  Also,

 * some more context may be helpful.
 * As I said on r-help, it is going to be _extremely_ difficult to
generalize to a random population from two replicates.  Have you
read the suggested mailing list thread
https://stat.ethz.ch/pipermail/r-sig-mixed-models/2010q2/003709.html
?

> The experiment is repeated twice (Replication 1 and 2) ,
> hence I
> thinkthat Replication should be random effect. My
> replications are randomlychosen from a population and I
> should make inference about thepopulation.  I don't have a
> chance to take more replications.
>