Ordinal regression with some categories combined for some data

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

Ordinal regression with some categories combined for some data

Bob O'Hara-2
Hi!

I've been asked about a problem where I think I can see how to write
the model, but don't know if it's been implemented in R. It's not
something I work on a lot, so I'm hoping someone else can point me to
an answer straight away.

The researcher has been carrying out germination experiments: lost of
seeds are put in several conditions (temperature humidity etc.), and
every few days they are checked to see if they have germinated.
Because the days are discrete I think it makes sense to view this as
an ordinal regression problem (rather than as an interval censored
survival analysis). But what makes this tricky is that there are days
when the researcher only checked some seeds. So for some seeds the
germination might fall into more than one category.

Is there a package in R that can handle this, i.e. do an ordinal
regression where for some observations the categories are interval
censored? Or is it easier to go straight to a full interval-censored
survival analysis?

Bob

--
Bob O'Hara

Biodiversity and Climate Research Centre
Senckenberganlage 25
D-60325 Frankfurt am Main,
Germany

Tel: +49 69 798 40226
Mobile: +49 1515 888 5440
WWW:   http://www.bik-f.de/root/index.php?page_id=219
Blog: http://occamstypewriter.org/boboh/
Journal of Negative Results - EEB: www.jnr-eeb.org

______________________________________________
[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: Ordinal regression with some categories combined for some data

Thierry Onkelinx
Dear Bob,

I don't know any package that handles ordinal data the way you are looking
for. I'd just would comment on the ordinal regression. Would the time of
loss be the ordinal response? That seems inefficient to me when you have a
lot of time points (= lots of ordinal classes). IMHO the survival analysis
would make more sense.

Best regards,

Thierry

ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Research Institute for Nature and
Forest
team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance
Kliniekstraat 25
1070 Anderlecht
Belgium

To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey

2016-01-15 16:48 GMT+01:00 Bob O'Hara <[hidden email]>:

> Hi!
>
> I've been asked about a problem where I think I can see how to write
> the model, but don't know if it's been implemented in R. It's not
> something I work on a lot, so I'm hoping someone else can point me to
> an answer straight away.
>
> The researcher has been carrying out germination experiments: lost of
> seeds are put in several conditions (temperature humidity etc.), and
> every few days they are checked to see if they have germinated.
> Because the days are discrete I think it makes sense to view this as
> an ordinal regression problem (rather than as an interval censored
> survival analysis). But what makes this tricky is that there are days
> when the researcher only checked some seeds. So for some seeds the
> germination might fall into more than one category.
>
> Is there a package in R that can handle this, i.e. do an ordinal
> regression where for some observations the categories are interval
> censored? Or is it easier to go straight to a full interval-censored
> survival analysis?
>
> Bob
>
> --
> Bob O'Hara
>
> Biodiversity and Climate Research Centre
> Senckenberganlage 25
> D-60325 Frankfurt am Main,
> Germany
>
> Tel: +49 69 798 40226
> Mobile: +49 1515 888 5440
> WWW:   http://www.bik-f.de/root/index.php?page_id=219
> Blog: http://occamstypewriter.org/boboh/
> Journal of Negative Results - EEB: www.jnr-eeb.org
>
> ______________________________________________
> [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: Ordinal regression with some categories combined for some data

Duncan Mackay-4
In reply to this post by Bob O'Hara-2
Hi

I have never seen germination experiments carried out as ordinal regression.
Most germination tests are done using a nls model.
For some species germination may only start a week after planting and then
germinate over 2 or 3 days.
If all germinated over the experimental period and  there is no change in
germination between readings then read them accordingly.
If there are ungerminated seeds at the end; have these then had germinable
tests been applied to them?.

I agree with Thierry about survival analysis.
If however you are wanting to get into latency then ordinal models may be OK
Another package to do ordinal regression is VGAM

Regards

Duncan

Duncan Mackay
Department of Agronomy and Soil Science
University of New England
Armidale NSW 2351
Email: home: [hidden email]



-----Original Message-----
From: R-help [mailto:[hidden email]] On Behalf Of Bob O'Hara
Sent: Saturday, 16 January 2016 01:49
To: r-help
Subject: [R] Ordinal regression with some categories combined for some data

Hi!

I've been asked about a problem where I think I can see how to write
the model, but don't know if it's been implemented in R. It's not
something I work on a lot, so I'm hoping someone else can point me to
an answer straight away.

The researcher has been carrying out germination experiments: lost of
seeds are put in several conditions (temperature humidity etc.), and
every few days they are checked to see if they have germinated.
Because the days are discrete I think it makes sense to view this as
an ordinal regression problem (rather than as an interval censored
survival analysis). But what makes this tricky is that there are days
when the researcher only checked some seeds. So for some seeds the
germination might fall into more than one category.

Is there a package in R that can handle this, i.e. do an ordinal
regression where for some observations the categories are interval
censored? Or is it easier to go straight to a full interval-censored
survival analysis?

Bob

--
Bob O'Hara

Biodiversity and Climate Research Centre
Senckenberganlage 25
D-60325 Frankfurt am Main,
Germany

Tel: +49 69 798 40226
Mobile: +49 1515 888 5440
WWW:   http://www.bik-f.de/root/index.php?page_id=219
Blog: http://occamstypewriter.org/boboh/
Journal of Negative Results - EEB: www.jnr-eeb.org

______________________________________________
[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.

______________________________________________
[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: Ordinal regression with some categories combined for some data

Frank Harrell
In reply to this post by Bob O'Hara-2
This does seem to be a good situation for ordinal regression.  The R rms
package's orm function allows for thousands of categories in Y. But it
doesn't handle censoring.

This discussion would be better for stats.stackexchange.com

Frank
--
------------------------------------------------------------------------
Frank E Harrell Jr Professor and Chairman School of Medicine

        Department of *Biostatistics* *Vanderbilt University*


        [[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.
Frank Harrell
Department of Biostatistics, Vanderbilt University
Reply | Threaded
Open this post in threaded view
|

Re: Ordinal regression with some categories combined for some data

Bob O'Hara-2
In reply to this post by Duncan Mackay-4
Thanks, Thierry & Duncan. I'll go down the survival analysis route.

The data are for central American epiphytes, so not your usual
species. Visually there's definitely differences in the times of
germination, but not in eventual germination, so that's
straightforward.

Bob

On 17 January 2016 at 01:38, Duncan Mackay <[hidden email]> wrote:

> Hi
>
> I have never seen germination experiments carried out as ordinal regression.
> Most germination tests are done using a nls model.
> For some species germination may only start a week after planting and then
> germinate over 2 or 3 days.
> If all germinated over the experimental period and  there is no change in
> germination between readings then read them accordingly.
> If there are ungerminated seeds at the end; have these then had germinable
> tests been applied to them?.
>
> I agree with Thierry about survival analysis.
> If however you are wanting to get into latency then ordinal models may be OK
> Another package to do ordinal regression is VGAM
>
> Regards
>
> Duncan
>
> Duncan Mackay
> Department of Agronomy and Soil Science
> University of New England
> Armidale NSW 2351
> Email: home: [hidden email]
>
>
>
> -----Original Message-----
> From: R-help [mailto:[hidden email]] On Behalf Of Bob O'Hara
> Sent: Saturday, 16 January 2016 01:49
> To: r-help
> Subject: [R] Ordinal regression with some categories combined for some data
>
> Hi!
>
> I've been asked about a problem where I think I can see how to write
> the model, but don't know if it's been implemented in R. It's not
> something I work on a lot, so I'm hoping someone else can point me to
> an answer straight away.
>
> The researcher has been carrying out germination experiments: lost of
> seeds are put in several conditions (temperature humidity etc.), and
> every few days they are checked to see if they have germinated.
> Because the days are discrete I think it makes sense to view this as
> an ordinal regression problem (rather than as an interval censored
> survival analysis). But what makes this tricky is that there are days
> when the researcher only checked some seeds. So for some seeds the
> germination might fall into more than one category.
>
> Is there a package in R that can handle this, i.e. do an ordinal
> regression where for some observations the categories are interval
> censored? Or is it easier to go straight to a full interval-censored
> survival analysis?
>
> Bob
>
> --
> Bob O'Hara
>
> Biodiversity and Climate Research Centre
> Senckenberganlage 25
> D-60325 Frankfurt am Main,
> Germany
>
> Tel: +49 69 798 40226
> Mobile: +49 1515 888 5440
> WWW:   http://www.bik-f.de/root/index.php?page_id=219
> Blog: http://occamstypewriter.org/boboh/
> Journal of Negative Results - EEB: www.jnr-eeb.org
>
> ______________________________________________
> [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.
>
> ______________________________________________
> [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.



--
Bob O'Hara

Biodiversity and Climate Research Centre
Senckenberganlage 25
D-60325 Frankfurt am Main,
Germany

Tel: +49 69 798 40226
Mobile: +49 1515 888 5440
WWW:   http://www.bik-f.de/root/index.php?page_id=219
Blog: http://occamstypewriter.org/boboh/
Journal of Negative Results - EEB: www.jnr-eeb.org

______________________________________________
[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.