[R] Why don't formulas that work for lm() work for plotmeans?

[Peter Ehlers]

Did lm(y ~ x | T) work for you? Let's see a reproducible example.

Peter

jrflanders wrote:
> I know that simple line plots showing a point representing the mean + an
> error bar are difficult in R, but I  am clearly missing something.  I simply
> want to show how a the mean and 95% CI of a chemical concentration (y) vary
> over space (x) under four temperature regimes (T; a categorical factor).
> plotmeans, in the gplots library, should allow for such a figure, right?
> After all, it invokes lm() in the help file for the formula, so any formula
> that works in lm() (e.g., y~x | T) should work in plotmeans.  Obviously this
> is not so - can anyone suggest an alternative  command or help me understand
> how  the syntac differs? Thanks...