R help - Jul 2016 - WRS2 package problems with post hoc test lincon

Hello,

	I am analyzing some data that has violated assumptions of ANOVA and am
using the WRS2 package in R. I am comparing three groups, Dx,
(schizophrenia, schizoaffective and control) on various variables, one of
which is premorbid IQ (preIQ).

	preIQdataSDx  #trimmed means > tapply(preIQdataSPreIQ, preIQdataSDx, mean,
tr=.2, na.rm=TRUE)  Schizophrenia Schizoaffective Control 89.29412 95.83333
95.54545
	>
	The post hoc test shows that the schizophrenia and controls group differ
significantly which makes sense when you look at their trimmed means;
however, what does not make sense is that the control and schizoaffective
disorder groups differ significantly from each other (their means are
almost identical). Additionally, if the schizophrenia and control group
differ significantly, shouldn't the schizophrenia group also differ
significantly from the control group too, as the control group and
schizoaffective disorder groups have very similar means. I am wondering if
the lincon post hoc test is flawed in some way or if I am not setting
something up correctly in R. Any help is greatly appreciated. Thank you.

	Best,

	Leah Hartman

	[[alternative HTML version deleted]]

1) HTML formatted email does not come through reliably. Please read the Posting
Guide.

2) It is very nearly always necessary to provide a reproducible example when
asking for help on this list to avoid complete failure to communicate.

3) Given the above limitations (meaning I may not be understanding you
correctly), this looks like a problem with your understanding of statistics, not
a problem with R. If so, you should be asking this question in a forum focused
less on the tool and more on the theory, such as stats.stackexchange.com.
-- 
Sent from my phone. Please excuse my brevity.

On July 7, 2016 5:43:58 PM PDT, lhartman at yorku.ca
wrote:
>
>
>	Hello,
>
>	I am analyzing some data that has violated assumptions of ANOVA and am
>using the WRS2 package in R. I am comparing three groups, Dx,
>(schizophrenia, schizoaffective and control) on various variables, one
>of
>which is premorbid IQ (preIQ).
>
>	preIQdataSDx  #trimmed means > tapply(preIQdataSPreIQ, preIQdataSDx,
>mean,
>tr=.2, na.rm=TRUE)  Schizophrenia Schizoaffective Control 89.29412
>95.83333
>95.54545
>	>
>	The post hoc test shows that the schizophrenia and controls group
>differ
>significantly which makes sense when you look at their trimmed means;
>however, what does not make sense is that the control and
>schizoaffective
>disorder groups differ significantly from each other (their means are
>almost identical). Additionally, if the schizophrenia and control group
>differ significantly, shouldn't the schizophrenia group also differ
>significantly from the control group too, as the control group and
>schizoaffective disorder groups have very similar means. I am wondering
>if
>the lincon post hoc test is flawed in some way or if I am not setting
>something up correctly in R. Any help is greatly appreciated. Thank
>you.
>
>	Best,
>
>	Leah Hartman
>
>	[[alternative HTML version deleted]]
>
>______________________________________________
>R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
>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.

> On Jul 7, 2016, at 5:43 PM, lhartman at yorku.ca wrote:
> 
> 
> 
> 	Hello,
> 
> 	I am analyzing some data that has violated assumptions of ANOVA
Violated? Violated how exactly?

> and am
> using the WRS2 package in R.

Where's the code?

> I am comparing three groups, Dx,
> (schizophrenia, schizoaffective and control) on various variables, one of
> which is premorbid IQ (preIQ).
> 
> 	preIQdataSDx  #trimmed means
> > tapply(preIQdataSPreIQ, preIQdataSDx, mean,
> tr=.2, na.rm=TRUE)  
CR's and minor editing added.
> Schizophrenia Schizoaffective Control
> 89.29412      95.83333        95.54545
> 	>
> 	The post hoc test
Test? Test of what (and how)?

> shows that the schizophrenia and controls group differ
> significantly which makes sense when you look at their trimmed means;
> however, what does not make sense is that the control and schizoaffective
> disorder groups differ significantly from each other (their means are
> almost identical). Additionally, if the schizophrenia and control group
> differ significantly, shouldn't the schizophrenia group also differ
> significantly from the control group too,
Did you really intend that sentence to say that?

> as the control group and
> schizoaffective disorder groups have very similar means. I am wondering if
> the lincon post hoc test is flawed in some way or if I am not setting
> something up correctly in R. Any help is greatly appreciated.
That's the test ( lincon? ) you didn't illustrate with code?

> Thank you.
> 
> 	Best,
> 
> 	Leah Hartman
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> 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.
David Winsemius
Alameda, CA, USA