R help - Jul 2017 - How to make a figure plotting p-values by range of different adjustment values?

Hi all,

Thank you for taking the time to read my message. I'm trying to make a
figure that plots p-values by a range of different adjustment values.

(Using the **logit** function in package **car**)

My Statistical analyses were conducted on probability estimates ranging
from 0% to 100%. As it's not ideal to run linear models on percentages that
are bounded between 0 and 1, these estimates were logit transformed.

However, this introduces a researcher degree of freedom. In Package
**Car**, the logit transformation code is

        logit(p = doc$value, adjust = 0.025)

    logit definition/Description

    Compute the logit transformation of proportions or percentages.

    Usage

         logit(p, percents=range.p[2] > 1, adjust)

    Arguments

    p  a numeric vector or array of proportions or percentages.
    percents TRUE for percentages.
    adjust   adjustment factor to avoid proportions of 0 or 1; defaults to
0 if there are no such proportions in the data, and to .025 if there are.)

I chose the default adjustment factor of .025, but I need to determine at
what point my values are greater than .05 to show I did not choose an
ajustment value that makes my results significant.

Ultimately, I want to find the range of adjustment factors do we get P <
0.05?And at what point do we get P > 0.05?

    ## The final product I'm looking for is a figure with the following
features:
      ## 1) Adjustment factor on the x-axis
      ## 2) P value on the y-axis

    Does anyone know how to do this? Thank you so much in advance.

	[[alternative HTML version deleted]]

Hi Kirsten,
Perhaps this will help:

set.seed(3)
kmdf<-data.frame(group=rep(1:4,each=20),
 prop=c(runif(20,0.25,1),runif(20,0.2,0.92),
 runif(20,0.15,0.84),runif(20,0.1,0.77)))
km.glm<-glm(prop~group,kmdf,family=quasibinomial(link="logit"))
summary(km.glm)
pval<-0.00845
padjs<-NA
npadj<-1
# assume you have five comparisons in this family
for(method in p.adjust.methods) {
 padjs[npadj]<-p.adjust(pval,method=method,n=5)
 npadj<-npadj+1
}
plot(padjs,xaxt="n",main="P
plot",xlab="Method",ylab="adjusted p values")
abline(h=0.05,col="lightgray")
library(plotrix)
staxlab(1,at=1:8,labels=p.adjust.methods)

Jim


On Thu, Jul 13, 2017 at 12:53 AM, Kirsten Morehouse
<kmoreho1 at swarthmore.edu> wrote:
> Hi all,
>
> Thank you for taking the time to read my message. I'm trying to make a
> figure that plots p-values by a range of different adjustment values.
>
> (Using the **logit** function in package **car**)
>
> My Statistical analyses were conducted on probability estimates ranging
> from 0% to 100%. As it's not ideal to run linear models on percentages
that
> are bounded between 0 and 1, these estimates were logit transformed.
>
> However, this introduces a researcher degree of freedom. In Package
> **Car**, the logit transformation code is
>
>         logit(p = doc$value, adjust = 0.025)
>
>     logit definition/Description
>
>     Compute the logit transformation of proportions or percentages.
>
>     Usage
>
>          logit(p, percents=range.p[2] > 1, adjust)
>
>     Arguments
>
>     p  a numeric vector or array of proportions or percentages.
>     percents TRUE for percentages.
>     adjust   adjustment factor to avoid proportions of 0 or 1; defaults to
> 0 if there are no such proportions in the data, and to .025 if there are.)
>
> I chose the default adjustment factor of .025, but I need to determine at
> what point my values are greater than .05 to show I did not choose an
> ajustment value that makes my results significant.
>
> Ultimately, I want to find the range of adjustment factors do we get P <
> 0.05?And at what point do we get P > 0.05?
>
>     ## The final product I'm looking for is a figure with the following
> features:
>       ## 1) Adjustment factor on the x-axis
>       ## 2) P value on the y-axis
>
>     Does anyone know how to do this? Thank you so much in advance.
>
>         [[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.

Hi Jim,

Thanks for your help, I really appreciate it.

Perhaps I'm misunderstanding, but does this formula run different ajustment
values for this function?

  logit(p = doc$value, adjust = 0.025)

I'm looking to plot the p-values of different adjustment values.

Thanks so much,

Kirsten

On Wed, Jul 12, 2017 at 8:49 PM, Jim Lemon <drjimlemon at gmail.com>
wrote:
> Hi Kirsten,
> Perhaps this will help:
>
> set.seed(3)
> kmdf<-data.frame(group=rep(1:4,each=20),
>  prop=c(runif(20,0.25,1),runif(20,0.2,0.92),
>  runif(20,0.15,0.84),runif(20,0.1,0.77)))
>
km.glm<-glm(prop~group,kmdf,family=quasibinomial(link="logit"))
> summary(km.glm)
> pval<-0.00845
> padjs<-NA
> npadj<-1
> # assume you have five comparisons in this family
> for(method in p.adjust.methods) {
>  padjs[npadj]<-p.adjust(pval,method=method,n=5)
>  npadj<-npadj+1
> }
> plot(padjs,xaxt="n",main="P
plot",xlab="Method",ylab="adjusted p values")
> abline(h=0.05,col="lightgray")
> library(plotrix)
> staxlab(1,at=1:8,labels=p.adjust.methods)
>
> Jim
>
>
> On Thu, Jul 13, 2017 at 12:53 AM, Kirsten Morehouse
> <kmoreho1 at swarthmore.edu> wrote:
> > Hi all,
> >
> > Thank you for taking the time to read my message. I'm trying to
make a
> > figure that plots p-values by a range of different adjustment values.
> >
> > (Using the **logit** function in package **car**)
> >
> > My Statistical analyses were conducted on probability estimates
ranging
> > from 0% to 100%. As it's not ideal to run linear models on
percentages
> that
> > are bounded between 0 and 1, these estimates were logit transformed.
> >
> > However, this introduces a researcher degree of freedom. In Package
> > **Car**, the logit transformation code is
> >
> >         logit(p = doc$value, adjust = 0.025)
> >
> >     logit definition/Description
> >
> >     Compute the logit transformation of proportions or percentages.
> >
> >     Usage
> >
> >          logit(p, percents=range.p[2] > 1, adjust)
> >
> >     Arguments
> >
> >     p  a numeric vector or array of proportions or percentages.
> >     percents TRUE for percentages.
> >     adjust   adjustment factor to avoid proportions of 0 or 1;
defaults
> to
> > 0 if there are no such proportions in the data, and to .025 if there
> are.)
> >
> > I chose the default adjustment factor of .025, but I need to determine
at
> > what point my values are greater than .05 to show I did not choose an
> > ajustment value that makes my results significant.
> >
> > Ultimately, I want to find the range of adjustment factors do we get P
<
> > 0.05?And at what point do we get P > 0.05?
> >
> >     ## The final product I'm looking for is a figure with the
following
> > features:
> >       ## 1) Adjustment factor on the x-axis
> >       ## 2) P value on the y-axis
> >
> >     Does anyone know how to do this? Thank you so much in advance.
> >
> >         [[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.
>
	[[alternative HTML version deleted]]