Mixed model parameterization

Apologies for my insistence:

"...I agree that using the mean you miss a lot of information. The simplest
case would be no interaction between random effects. Say,
(1|month)+(1|day)+(1|hour)+(1|plot).?..."

...and that also assumes that the only effect of random factors is on the intercept, but not on the slope of the relationship between the fixed factor(s) and the response. Overall, if you want to consider (isolate) all the possible random effects that might be involved in your design, you would need a very complex model. But you previously suggested that some of those factors might not be that important and perhaps may be dismissed; in my view, I wonder if some of your dat are actually pseudoreplicates (e.g., measurements at different times of the same day) and, if so, a parsimonious way to handle them is by averaging across them and analyzing the average values.

Best regards,

Salvador S?NCHEZ-COL?N

En Mar, 21 Mayo, 2019 en 14:37, Joaqu?n Aldabe <joaquin.aldabe at gmail.com> escribi?:
?

Para: Manuel Sp?nola; r-sig-mixed-models at r-project.org
I agree that using the mean you miss a lot of information. The simplest
case would be no interaction between random effects. Say,
(1|month)+(1|day)+(1|hour)+(1|plot). As far as I understand, if you think
that, for example, hour effect depends on the day, then you should consider
a nested structure of randoms effects.

Let us know your progress!

joaquin

<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
Libre
de virus. www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

El mar., 21 may. 2019 a las 14:01, Manuel Sp?nola (<mspinola10 at gmail.com>)
escribi?:

Mixed model parameterization

Thread (4 messages)