R help - Aug 2010 - basic question about t-test with adjusted p value

I have read the R manual and help archives, sorry but I'm still stuck. 

How would I do a t-test with an adjusted p-value?

Suppose that I use t.test ( ) , with the function argument alternative = 
"two.sided",  and data such that degrees of freedom = 20.  The
function
calculates a t-statistic of 2.086, and p-value =0.05

How do I then adjust the p-value?  My thought is to do
p.adjust (pt(2.086, df=20),"BH") 
but that doesn't change anything (returns 0.975)

what is the procedure?  I'm sorry if there is a basic concept that I am 
missing here...








	[[alternative HTML version deleted]]

On Sat, Aug 7, 2010 at 1:08 PM,  <Josef.Kardos at phila.gov>
wrote:
> I have read the R manual and help archives, sorry but I'm still stuck.
>
> How would I do a t-test with an adjusted p-value?
>
> Suppose that I use t.test ( ) , with the function argument alternative >
"two.sided", ?and data such that degrees of freedom = 20. ?The
function
> calculates a t-statistic of 2.086, and p-value =0.05
>
> How do I then adjust the p-value? ?My thought is to do
> p.adjust (pt(2.086, df=20),"BH")
> but that doesn't change anything (returns 0.975)
>
> what is the procedure? ?I'm sorry if there is a basic concept that I am
> missing here...
These adjustments are generally designed to control a family wise or
experiment wise error rate.  If you only have one p-value, then the
adjustment is equivalent to the unadjusted.  The argument n defaults
to length(p) (which is the length of the first argument).  Here are
some examples:
> p.adjust(p = 0.975, method = "BH")
[1] 0.975
> p.adjust(p = 0.975, method = "BH", n = 1)
[1] 0.975
> p.adjust(p = 0.975, method = "BH", n = 5)
[1] 1

HTH,

Josh
>
>
>
>
>
>
>
>
> ? ? ? ?[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org 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.
>


-- 
Joshua Wiley
Ph.D. Student, Health Psychology
University of California, Los Angeles
http://www.joshuawiley.com/

On Sat, Aug 07, 2010 at 04:08:40PM -0400, Josef.Kardos at phila.gov
wrote:
> I have read the R manual and help archives, sorry but I'm still stuck. 
> 
> How would I do a t-test with an adjusted p-value?
> 
> Suppose that I use t.test ( ) , with the function argument alternative = 
> "two.sided",  and data such that degrees of freedom = 20.  The
function
> calculates a t-statistic of 2.086, and p-value =0.05
> 
> How do I then adjust the p-value?  My thought is to do
> p.adjust (pt(2.086, df=20),"BH") 
> but that doesn't change anything (returns 0.975)
> 
> what is the procedure?  I'm sorry if there is a basic concept that I am
> missing here...
I'm confused - what result where you expecting? p.adjust will need to
know the number of test you are trying to adjust for - either by
giving explicitly giving the number or by handing a vector of p-values
to the function.

cu
	Philipp

-- 
Dr. Philipp Pagel
Lehrstuhl f?r Genomorientierte Bioinformatik
Technische Universit?t M?nchen
Wissenschaftszentrum Weihenstephan
Maximus-von-Imhof-Forum 3
85354 Freising, Germany
http://webclu.bio.wzw.tum.de/~pagel/

On 08/07/2010 03:08 PM, Josef.Kardos at phila.gov wrote:
> I have read the R manual and help archives, sorry but I'm still stuck.
>
> How would I do a t-test with an adjusted p-value?
Please be more specific about what you mean by 'adjusted p-value'. See
below...
>
> Suppose that I use t.test ( ) , with the function argument alternative >
"two.sided",  and data such that degrees of freedom = 20.  The
function
> calculates a t-statistic of 2.086, and p-value =0.05
>
> How do I then adjust the p-value?  My thought is to do
> p.adjust (pt(2.086, df=20),"BH")
> but that doesn't change anything (returns 0.975)
A couple things here.

1) You can get the p-value returned from t.test by inspecting the object with 
?str, and noting that it's stored in an element called p-value.

So,

p.val <- t.test(extra ~ group, data = sleep)$p.value

assigns the p-value to an object called p.val.

2) You're calling p.adjust with a single p-value, so what kind of adjustment
did
you have in mind? You would normally be adjusting p-values because of a multiple
comparison issue (i.e., multiple tests performed).  You'd pass p.adjust the 
*vector* of p-values, and a method to adjust them by.  Your vector is of length 
1, so in p.adjust's opinion, there is nothing to adjust for.

So, what do *you* mean by 'adjusted p-value'?  `Why are you looking to
adjust?`
is another way of stating that.

ANOVA will give you an adjusted t-test. For example:
fit<-lm(height~sex+age)
summary(fit)
Will give an age-adjusted t-test of the heights of men and women.
John
John Sorkin
Chief Biostatistics and Informatics
Univ. of Maryland School of Medicine
Division of Gerontology and Geriatric Medicine
JSorkin at grecc.umaryland.edu 
-----Original Message-----
From: Erik Iverson <eriki at ccbr.umn.edu>
To:  <Josef.Kardos at phila.gov>
Cc:  <r-help at r-project.org>

Sent: 8/7/2010 4:24:35 PM
Subject: Re: [R] basic question about t-test with adjusted p value

On 08/07/2010 03:08 PM, Josef.Kardos at phila.gov wrote:
> I have read the R manual and help archives, sorry but I'm still stuck.
>
> How would I do a t-test with an adjusted p-value?
Please be more specific about what you mean by 'adjusted p-value'. See
below...
>
> Suppose that I use t.test ( ) , with the function argument alternative >
"two.sided",  and data such that degrees of freedom = 20.  The
function
> calculates a t-statistic of 2.086, and p-value =0.05
>
> How do I then adjust the p-value?  My thought is to do
> p.adjust (pt(2.086, df=20),"BH")
> but that doesn't change anything (returns 0.975)
A couple things here.

1) You can get the p-value returned from t.test by inspecting the object with 
?str, and noting that it's stored in an element called p-value.

So,

p.val <- t.test(extra ~ group, data = sleep)$p.value

assigns the p-value to an object called p.val.

2) You're calling p.adjust with a single p-value, so what kind of adjustment
did
you have in mind? You would normally be adjusting p-values because of a multiple
comparison issue (i.e., multiple tests performed).  You'd pass p.adjust the 
*vector* of p-values, and a method to adjust them by.  Your vector is of length 
1, so in p.adjust's opinion, there is nothing to adjust for.

So, what do *you* mean by 'adjusted p-value'?  `Why are you looking to
adjust?`
is another way of stating that.

______________________________________________
R-help at r-project.org 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.

Confidentiality Statement:
This email message, including any attachments, is for th...{{dropped:6}}