[R-sig-ME] [R] lme nesting/interaction advice

[Rolf Turner]

Thanks very much for your long, detailed, patient, and lucid
response to my cri de coeur.  That helps a *great* deal.

I'm not sure that I have a solid understanding of the issues yet
--- I never am! --- but I think I'm getting there.  I'll need to
chew over the posting a bit more and try some examples before
I feel comfortable.

One point that I'd like to spell out very explicitly, to make sure
that I'm starting from the right square:

The model that you start with in the Machines/Workers example is

>> fm1 <- lmer(score ~ Machine + (1|Worker) + (1|Machine:Worker),  
>> Machines)


My understanding is that this is the ``only'' model that could be fitted
by the Package-Which-Must-Not-Be-Named.  I.e. this package *could  
not* fit
the second, more complex model:

>> fm2 <- lmer(score ~ Machine + (Machine|Worker), Machines)

(At least not directly.)  Can you (or someone) confirm that I've got  
that
right?

It seems to me to be the key to why I've had such trouble in the past
in grappling with mixed models in R.  I.e. I've been thinking like
the Package-Which-Must-Not-Be-Named --- that the simple, everything- 
independent
model was the only possible model.

Thanks again.

	cheers,

		Rolf

######################################################################
Attention:\ This e-mail message is privileged and confid...{{dropped:9}}