R help - Jul 2010 - specifying an unbalanced mixed-effects model for anova

hi all - i'm having trouble using lme to specify a mixed effects
model.
i'm pretty sure this is quite easy for the experienced anova-er, which
i unfortunately am not.

i have a data frame with the following columns:
col 1 : "Score1" (this is a continuous numeric measure between 0 and
1)
col 2 : "Score2" (another continuous numeric measure, this time
bounded between 0 and 100)
col 3 : "Class" (a fixed-effects factor with 4 (ordered) levels)

i have ~2000 observations, but unbalanced w.r.t. the "Class" factor.
each observation has a distinct "Score1" and "Score2".

i'd like to try a mixed-effect anova model where "Score1" is the
dependent response, and "Score2" and "Class" are the
explanatory
covariates (with an expected interaction between them).

naively, i simply tried:
aov(Score1 ~ Score2 * Class)

but i'm pretty sure this treated "Score2" as a fixed-effects
covariate
with a single observation per level.

can anyone help guide me to the correct model specification using
lme(...)?

thanks much for any help!

You are confused. You have not specified a "grouping" random effect,
so  this is not a mixed effect model as it stands.

If this is a homework problem, ask a teacher or classmate for help.
Otherwise, try consulting your local statistician. You do not appear
to understand the concepts of mixed effects models, so just walking
you through a lme model specification is not likely to help.

Bert Gunter





On Tue, Jul 27, 2010 at 8:24 PM, Murat Tasan <mmuurr at gmail.com>
wrote:
> hi all - i'm having trouble using lme to specify a mixed effects
> model.
> i'm pretty sure this is quite easy for the experienced anova-er, which
> i unfortunately am not.
>
> i have a data frame with the following columns:
> col 1 : "Score1" (this is a continuous numeric measure between 0
and
> 1)
> col 2 : "Score2" (another continuous numeric measure, this time
> bounded between 0 and 100)
> col 3 : "Class" (a fixed-effects factor with 4 (ordered) levels)
>
> i have ~2000 observations, but unbalanced w.r.t. the "Class"
factor.
> each observation has a distinct "Score1" and "Score2".
>
> i'd like to try a mixed-effect anova model where "Score1" is
the
> dependent response, and "Score2" and "Class" are the
explanatory
> covariates (with an expected interaction between them).
>
> naively, i simply tried:
> aov(Score1 ~ Score2 * Class)
>
> but i'm pretty sure this treated "Score2" as a fixed-effects
covariate
> with a single observation per level.
>
> can anyone help guide me to the correct model specification using
> lme(...)?
>
> thanks much for any help!
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>