Problem with Kruskal–Wallis test

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

Problem with Kruskal–Wallis test

mngon
Dear all,
I am a beginner with R (and also with the statistics) for which I hope to
be clear.
I should do this non-parametric test on data I extracted from maps.
In practice I have a column that represents the landscape Dynamics of a
certain time period (there are 3 dynamics, each of them marked by the
number 1, 2 or 3) and the other column with the values of a topographic
variable (for example the slope) . In all, there are more than 90,000 pairs
of values.
Going to do the test in R, for all the dynamics and for all the variables,
I get out of the values of chi-square elevated (even in the order of
thousands) and a p-value always <2.2e-16 .... why? Where can the error be? in
the script or in the test approach?
Thanks in advance

        [[alternative HTML version deleted]]

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

Fwd: Problem with Kruskal–Wallis test

mngon
Dear all,
I am a beginner with R (and also with the statistics) for which I hope to
be clear.
I should do this non-parametric test on data I extracted from maps.
In practice I have a column that represents the landscape Dynamics of a
certain time period (there are 3 dynamics, each of them marked by the
number 1, 2 or 3) and the other column with the values of a topographic
variable (for example the slope) . In all, there are more than 90,000 pairs
of values.
Going to do the test in R, for all the dynamics and for all the variables,
I get out of the values of chi-square elevated (even in the order of
thousands) and a p-value always <2.2e-16 .... why? Where can the error be? in
the script or in the test approach?
Thanks in advance

        [[alternative HTML version deleted]]

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

Re: Problem with Kruskal–Wallis test

Michael Dewey-3
In reply to this post by mngon
Dear Giuseppe

If I understand you correctly you have a very large sample size so it is
not surprising that you get very small p-values. Eevn a scientifically
uninteresting difference can become statistically significant with large
samples. You probably need to define a metric for meaningful differences
between groups and calculate a confidence interval for it.

Michael

On 21/12/2018 15:37, Giuseppe Cillis wrote:

> Dear all,
> I am a beginner with R (and also with the statistics) for which I hope to
> be clear.
> I should do this non-parametric test on data I extracted from maps.
> In practice I have a column that represents the landscape Dynamics of a
> certain time period (there are 3 dynamics, each of them marked by the
> number 1, 2 or 3) and the other column with the values of a topographic
> variable (for example the slope) . In all, there are more than 90,000 pairs
> of values.
> Going to do the test in R, for all the dynamics and for all the variables,
> I get out of the values of chi-square elevated (even in the order of
> thousands) and a p-value always <2.2e-16 .... why? Where can the error be? in
> the script or in the test approach?
> Thanks in advance
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> [hidden email] 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.
>

--
Michael
http://www.dewey.myzen.co.uk/home.html

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

Re: Problem with Kruskal–Wallis test

mngon
Dear Michael,
Thanks for your answer.
So, I'm not an expert in R and statistics, how can I create this interval
of confidence of groups?
Thanks
Gc

Il giorno sab 22 dic 2018, 13:34 Michael Dewey <[hidden email]> ha
scritto:

> Dear Giuseppe
>
> If I understand you correctly you have a very large sample size so it is
> not surprising that you get very small p-values. Eevn a scientifically
> uninteresting difference can become statistically significant with large
> samples. You probably need to define a metric for meaningful differences
> between groups and calculate a confidence interval for it.
>
> Michael
>
> On 21/12/2018 15:37, Giuseppe Cillis wrote:
> > Dear all,
> > I am a beginner with R (and also with the statistics) for which I hope to
> > be clear.
> > I should do this non-parametric test on data I extracted from maps.
> > In practice I have a column that represents the landscape Dynamics of a
> > certain time period (there are 3 dynamics, each of them marked by the
> > number 1, 2 or 3) and the other column with the values of a topographic
> > variable (for example the slope) . In all, there are more than 90,000
> pairs
> > of values.
> > Going to do the test in R, for all the dynamics and for all the
> variables,
> > I get out of the values of chi-square elevated (even in the order of
> > thousands) and a p-value always <2.2e-16 .... why? Where can the error
> be? in
> > the script or in the test approach?
> > Thanks in advance
> >
> >       [[alternative HTML version deleted]]
> >
> > ______________________________________________
> > [hidden email] 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.
> >
>
> --
> Michael
> http://www.dewey.myzen.co.uk/home.html
>

        [[alternative HTML version deleted]]

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

Re: Problem with Kruskal–Wallis test

Michael Dewey-3
I think you need to talk to someone who uses your sort of geographic
data to find out what an appropriate metric for comparing your variables
is. Only then will you know what might be a suitable way forward.

Michael

On 22/12/2018 17:27, Giuseppe Cillis wrote:

> Dear Michael,
> Thanks for your answer.
> So, I'm not an expert in R and statistics, how can I create this
> interval of confidence of groups?
> Thanks
> Gc
>
> Il giorno sab 22 dic 2018, 13:34 Michael Dewey <[hidden email]
> <mailto:[hidden email]>> ha scritto:
>
>     Dear Giuseppe
>
>     If I understand you correctly you have a very large sample size so
>     it is
>     not surprising that you get very small p-values. Eevn a scientifically
>     uninteresting difference can become statistically significant with
>     large
>     samples. You probably need to define a metric for meaningful
>     differences
>     between groups and calculate a confidence interval for it.
>
>     Michael
>
>     On 21/12/2018 15:37, Giuseppe Cillis wrote:
>      > Dear all,
>      > I am a beginner with R (and also with the statistics) for which I
>     hope to
>      > be clear.
>      > I should do this non-parametric test on data I extracted from maps.
>      > In practice I have a column that represents the landscape
>     Dynamics of a
>      > certain time period (there are 3 dynamics, each of them marked by the
>      > number 1, 2 or 3) and the other column with the values of a
>     topographic
>      > variable (for example the slope) . In all, there are more than
>     90,000 pairs
>      > of values.
>      > Going to do the test in R, for all the dynamics and for all the
>     variables,
>      > I get out of the values of chi-square elevated (even in the order of
>      > thousands) and a p-value always <2.2e-16 .... why? Where can the
>     error be? in
>      > the script or in the test approach?
>      > Thanks in advance
>      >
>      >       [[alternative HTML version deleted]]
>      >
>      > ______________________________________________
>      > [hidden email] <mailto:[hidden email]> 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.
>      >
>
>     --
>     Michael
>     http://www.dewey.myzen.co.uk/home.html
>

--
Michael
http://www.dewey.myzen.co.uk/home.html

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

Re: Problem with Kruskal–Wallis test

Bert Gunter-2
In reply to this post by mngon
"So, I'm not an expert in R and statistics" ....

So you need to seek local help from someone who is. Statistics is usually
off-topic for this list -- it is about R programming primarily. And online
is probably not a good venue for the sort of discussion you need anyway.

Cheers,
Bert

Bert Gunter

"The trouble with having an open mind is that people keep coming along and
sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )


On Tue, Dec 25, 2018 at 5:37 AM Giuseppe Cillis <[hidden email]> wrote:

> Dear Michael,
> Thanks for your answer.
> So, I'm not an expert in R and statistics, how can I create this interval
> of confidence of groups?
> Thanks
> Gc
>
> Il giorno sab 22 dic 2018, 13:34 Michael Dewey <[hidden email]>
> ha
> scritto:
>
> > Dear Giuseppe
> >
> > If I understand you correctly you have a very large sample size so it is
> > not surprising that you get very small p-values. Eevn a scientifically
> > uninteresting difference can become statistically significant with large
> > samples. You probably need to define a metric for meaningful differences
> > between groups and calculate a confidence interval for it.
> >
> > Michael
> >
> > On 21/12/2018 15:37, Giuseppe Cillis wrote:
> > > Dear all,
> > > I am a beginner with R (and also with the statistics) for which I hope
> to
> > > be clear.
> > > I should do this non-parametric test on data I extracted from maps.
> > > In practice I have a column that represents the landscape Dynamics of a
> > > certain time period (there are 3 dynamics, each of them marked by the
> > > number 1, 2 or 3) and the other column with the values of a topographic
> > > variable (for example the slope) . In all, there are more than 90,000
> > pairs
> > > of values.
> > > Going to do the test in R, for all the dynamics and for all the
> > variables,
> > > I get out of the values of chi-square elevated (even in the order of
> > > thousands) and a p-value always <2.2e-16 .... why? Where can the error
> > be? in
> > > the script or in the test approach?
> > > Thanks in advance
> > >
> > >       [[alternative HTML version deleted]]
> > >
> > > ______________________________________________
> > > [hidden email] 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.
> > >
> >
> > --
> > Michael
> > http://www.dewey.myzen.co.uk/home.html
> >
>
>         [[alternative HTML version deleted]]
>
> ______________________________________________
> [hidden email] 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]]

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

Re: Problem with Kruskal–Wallis test

JohnDee
In reply to this post by mngon
On Fri, 21 Dec 2018 16:37:54 +0100
Giuseppe Cillis <[hidden email]> wrote:

> Dear all,
> I am a beginner with R (and also with the statistics) for which I
> hope to be clear.
> I should do this non-parametric test on data I extracted from maps.
> In practice I have a column that represents the landscape Dynamics of
> a certain time period (there are 3 dynamics, each of them marked by
> the number 1, 2 or 3) and the other column with the values of a
> topographic variable (for example the slope) . In all, there are more
> than 90,000 pairs of values.
> Going to do the test in R, for all the dynamics and for all the
> variables, I get out of the values of chi-square elevated (even in
> the order of thousands) and a p-value always <2.2e-16 .... why? Where
> can the error be? in the script or in the test approach?
> Thanks in advance
>
Your question and your problem is not about R, but rather about why your
results do not match your expectations.  There are a lot of questions
you should be considering like, why use Kruskal-Wallis, before you even
start wondering about R?  

JWDougherty

______________________________________________
[hidden email] 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.