[prev in list] [next in list] [prev in thread] [next in thread]
List: r-sig-mixed-models
Subject: Re: [R-sig-ME] most conservative df for mixed effects anova (Carrie Perkins)
From: Nik Tuzov <ntuzov () ntuzov ! com>
Date: 2019-10-16 13:51:10
Message-ID: 738130388.1575956.1571233870040 () mail ! yahoo ! com
[Download RAW message or body]
Hello Carrie:
Strictly speaking, the term conservative or aggressive should apply to the p-values \
rather than to the degrees of freedom.I assume you are asking what approach generates \
the largest p-value for your fixed effect of interest.In theory, one can't answer \
that question in advance. If methods X and Y result in n and m df for the error \
term, where n > m, it doesn't imply that X will produce a smaller p-value than Y,even \
though it's often in case in practice. That being said, a method that is likely to be \
conservative should have df <= (count of experimental units less the count of \
fixedparameters in the model) <= (count of experimental units less one). Regards,Nik \
Tuzov
On Wednesday, October 16, 2019, 5:02:31 AM CDT, \
r-sig-mixed-models-request@r-project.org <r-sig-mixed-models-request@r-project.org> \
wrote:
Send R-sig-mixed-models mailing list submissions to
r-sig-mixed-models@r-project.org
To subscribe or unsubscribe via the World Wide Web, visit
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
or, via email, send a message with subject or body 'help' to
r-sig-mixed-models-request@r-project.org
You can reach the person managing the list at
r-sig-mixed-models-owner@r-project.org
When replying, please edit your Subject line so it is more specific
than "Re: Contents of R-sig-mixed-models digest..."
Today's Topics:
1. most conservative df for mixed effects anova (Carrie Perkins)
2. Re: most conservative df for mixed effects anova
(Thierry Onkelinx)
----------------------------------------------------------------------
Message: 1
Date: Tue, 15 Oct 2019 09:12:54 -0400
From: Carrie Perkins <cperk@terpmail.umd.edu>
To: r-sig-mixed-models@r-project.org
Subject: [R-sig-ME] most conservative df for mixed effects anova
Message-ID:
<CAPtr_T4Gn5zhD6xtvRR9ckTV+=nhDSi6xG_v0qiXUg2aw28s-w@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"
Hello!
I have data from an experiment and would like to run an anova with fixed
and random effects in R. Here is information on the data:
In the experiment, 3 replicates of 48 plant genotypes were planted into
each of 4 salinity treatments. This resulted in a total of 144 individuals
per treatment, amounting to a grand total of 576 individuals in the whole
experiment. Within each treatment, random sets of 24 plants were grouped
into a total of 6 pools to make it easier to monitor salinity levels. I
would like to model these pools as random Experimental Units.
I would like to make Experimental Unit the random effect and look at the
treatment X genotype interaction as fixed effects.
lmer_model_3 <- aov(Y~Genotype*Treatment + Error(1|Experimental Unit),
data=dataframe)
What would be the most conservative method for calculating degrees of
freedom for the random effects term of an anova? When I've tried
researching this question online, I find a lot of information on
calculating degrees of freedom for basic 1- and 2-way anovas (which I
understand) but I can't find clear guidance on how to calculate the degrees
of freedom for anovas with random effects.
Thank you!
Sincerely,
Carrie Perkins
[[alternative HTML version deleted]]
------------------------------
Message: 2
Date: Wed, 16 Oct 2019 09:31:14 +0200
From: Thierry Onkelinx <thierry.onkelinx@inbo.be>
To: Carrie Perkins <cperk@terpmail.umd.edu>
Cc: r-sig-mixed-models <r-sig-mixed-models@r-project.org>
Subject: Re: [R-sig-ME] most conservative df for mixed effects anova
Message-ID:
<CAJuCY5ydZvPE6w9cg1uuyPMTD=2_BGyKVARokVa6vxnhLoWSJQ@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"
Dear Carrie,
The most conservative number IMHO is the sum of the number fixed effects
parameters and the number of random effects parameters (in case of a random
intercept: 1 level = 1 parameter). Het most liberate number would replace
the number random effects parameters with the number of random effect
hyperparameters (a random intercept = 1 variance = 1 hyperparameter).
Best regards,
ir. Thierry Onkelinx
Statisticus / Statistician
Vlaamse Overheid / Government of Flanders
INSTITUUT VOOR NATUUR- EN BOSONDERZOEK / RESEARCH INSTITUTE FOR NATURE AND
FOREST
Team Biometrie & Kwaliteitszorg / Team Biometrics & Quality Assurance
thierry.onkelinx@inbo.be
Havenlaan 88 bus 73, 1000 Brussel
www.inbo.be
///////////////////////////////////////////////////////////////////////////////////////////
To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey
///////////////////////////////////////////////////////////////////////////////////////////
<https://www.inbo.be>
Op di 15 okt. 2019 om 15:20 schreef Carrie Perkins <cperk@terpmail.umd.edu>:
> Hello!
>
> I have data from an experiment and would like to run an anova with fixed
> and random effects in R. Here is information on the data:
>
> In the experiment, 3 replicates of 48 plant genotypes were planted into
> each of 4 salinity treatments. This resulted in a total of 144 individuals
> per treatment, amounting to a grand total of 576 individuals in the whole
> experiment. Within each treatment, random sets of 24 plants were grouped
> into a total of 6 pools to make it easier to monitor salinity levels. I
> would like to model these pools as random Experimental Units.
>
> I would like to make Experimental Unit the random effect and look at the
> treatment X genotype interaction as fixed effects.
>
> lmer_model_3 <- aov(Y~Genotype*Treatment + Error(1|Experimental Unit),
> data=dataframe)
>
> What would be the most conservative method for calculating degrees of
> freedom for the random effects term of an anova? When I've tried
> researching this question online, I find a lot of information on
> calculating degrees of freedom for basic 1- and 2-way anovas (which I
> understand) but I can't find clear guidance on how to calculate the degrees
> of freedom for anovas with random effects.
>
> Thank you!
>
> Sincerely,
> Carrie Perkins
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>
[[alternative HTML version deleted]]
------------------------------
Subject: Digest Footer
_______________________________________________
R-sig-mixed-models mailing list
R-sig-mixed-models@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
------------------------------
End of R-sig-mixed-models Digest, Vol 154, Issue 17
***************************************************
[[alternative HTML version deleted]]
_______________________________________________
R-sig-mixed-models@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic