Post Hoc tests for ANOVA

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Post Hoc tests for ANOVA

Amartya
Hi,

I was trying to figure out how to do  post-hoc tests for Two Way ANOVAs and found the following 2 approaches:

a. Do pairwise t-tests (bonferroni corrected) if one finds significance with the ANOVA.
Link- http://rtutorialseries.blogspot.com/2011/01/r-tutorial-series-two-way-anova-with.html

b. Do TukeyHSD  on an aov model
Link- http://www.r-bloggers.com/post-hoc-pairwise-comparisons-of-two-way-anova/

Running the data set given in the first example in SPSS gives significant pairwise difference for Treatment and Age (Treatmen and Age were the independent variables) , while using the directions given in the first link didn't give me significant pairwise different for Treatment (only gave for Age).

I have a few questions:

a. Is the first method completely incorrect as hinted in the second link?
b. What is the right way to do Bonferroni corrected post hoc tests for Two Way ANOVA in R?
c. Does anyone know how post hoc tests for SPSS work in the case of Two Way ANOVAs (Univariate analysis)? Especially for Bonferroni corrected tests.

I am new to R, so please let me know if I made a mistake in framing the question; I will try to elucidate as much as I personally can. Thanks for your help.

Reply | Threaded
Open this post in threaded view
|

Re: Post Hoc tests for ANOVA

Richard M. Heiberger
Please look at the ?mmc example for two-way ANOVA in library(HH).

If you don't already have HH you can get it with
install.packages("HH")
library(HH)
?mmc

mmc uses the glht function in the multcomp package for its calculations
and then draws the MMC graph.

Rich

On 10/19/12, Amartya <[hidden email]> wrote:

> Hi,
>
> I was trying to figure out how to do  post-hoc tests for Two Way ANOVAs and
> found the following 2 approaches:
>
> a. Do pairwise t-tests (bonferroni corrected) if one finds significance
> with
> the ANOVA.
> Link-
> http://rtutorialseries.blogspot.com/2011/01/r-tutorial-series-two-way-anova-with.html
>
> b. Do TukeyHSD  on an aov model
> Link-
> http://www.r-bloggers.com/post-hoc-pairwise-comparisons-of-two-way-anova/
>
> Running the data set given in the first example in SPSS gives significant
> pairwise difference for Treatment and Age (Treatmen and Age were the
> independent variables) , while using the directions given in the first link
> didn't give me significant pairwise different for Treatment (only gave for
> Age).
>
> I have a few questions:
>
> a. Is the first method completely incorrect as hinted in the second link?
> b. What is the right way to do Bonferroni corrected post hoc tests for Two
> Way ANOVA in R?
> c. Does anyone know how post hoc tests for SPSS work in the case of Two Way
> ANOVAs (Univariate analysis)? Especially for Bonferroni corrected tests.
>
> I am new to R, so please let me know if I made a mistake in framing the
> question; I will try to elucidate as much as I personally can. Thanks for
> your help.
>
>
>
>
>
> --
> View this message in context:
> http://r.789695.n4.nabble.com/Post-Hoc-tests-for-ANOVA-tp4646710.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> [hidden email] 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.
>

______________________________________________
[hidden email] 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.
Reply | Threaded
Open this post in threaded view
|

Re: Post Hoc tests for ANOVA

rent0009
In reply to this post by Amartya
Amartya: This is probably more of a statistics question than an R question.
I would go with the Tukey HSD, were I you as that is its intention.

Is there anyone in your organization (school, company, etc) that does
statistical consulting? I have had wonderful experience working with the
consulting clinic at the U of MN. If nothing else, buying a stats grad
student some coffee might save you hours of headache in reading and
searching message boards.

Mike.

On Oct 19 2012, Amartya wrote:

>Hi,
>
>I was trying to figure out how to do  post-hoc tests for Two Way ANOVAs and
>found the following 2 approaches:
>
 a. Do pairwise t-tests (bonferroni corrected) if one finds significance
with
>the ANOVA.
>Link-
 
http://rtutorialseries.blogspot.com/2011/01/r-tutorial-series-two-way-anova-with.html

>
>b. Do TukeyHSD  on an aov model
>Link-
>http://www.r-bloggers.com/post-hoc-pairwise-comparisons-of-two-way-anova/
>
>Running the data set given in the first example in SPSS gives significant
>pairwise difference for Treatment and Age (Treatmen and Age were the
>independent variables) , while using the directions given in the first link
>didn't give me significant pairwise different for Treatment (only gave for
>Age).
>
>I have a few questions:
>
>a. Is the first method completely incorrect as hinted in the second link?
>b. What is the right way to do Bonferroni corrected post hoc tests for Two
>Way ANOVA in R?
>c. Does anyone know how post hoc tests for SPSS work in the case of Two Way
>ANOVAs (Univariate analysis)? Especially for Bonferroni corrected tests.
>
>I am new to R, so please let me know if I made a mistake in framing the
>question; I will try to elucidate as much as I personally can. Thanks for
>your help.
>
>
>
>
>
>--
 View this message in context:
http://r.789695.n4.nabble.com/Post-Hoc-tests-for-ANOVA-tp4646710.html
>Sent from the R help mailing list archive at Nabble.com.
>
>______________________________________________
>[hidden email] 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.
>

______________________________________________
[hidden email] 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.