Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Binary and ologit

From   "Shittu, Aminu" <[email protected]>
To   "[email protected]" <[email protected]>
Subject   Re: st: Binary and ologit
Date   Tue, 31 Jul 2012 17:37:13 -0700 (PDT)

 Hi Maarten,

Apologies my initial posting might be vague. I am suppose to add that none of the over 12,000 animals in my 30 years dataset is healthy. We selected animals that were all diagnosed with our particular disease of interest, whose 2 clinical stages (acute and chronic) were recorded in sheep, but it only said 'diseased' in cattle - same disease though. Could this diagnosis with different levels (acute, chronic and diseased) be treated as ordinal even though cattle has only 1 level in the dataset, and 2 levels for sheep?


From: Maarten Buis <[email protected]>
To: [email protected] 
Sent: Tuesday, July 31, 2012 5:32:52 PM
Subject: Re: st: Binary and ologit

The easiest way is to combine the chronic an acute into diseased and
use a binary logit. So, say you have one variable --- lets call it
status --- with 4 categories: 0 healty, 1 diseased (cattle only), 2
acute (sheap only), 3 chronic (sheap only). Than you can first create
a new variable:

recode status (2=1) (3=1) , gen(newstatus)

and use that as response in your logit command.

Hope this helps,

On Tue, Jul 31, 2012 at 6:08 PM, Shittu, Aminu wrote:
> The type of response in this multivariable data looks pretty strange to me. In the data, there is only one diagnosis of disease in cattle, but in sheep it was categorised as either acute or chronic based on laboratory confirmation. Binary logistic regression will work well for sheep, if we are to drop the cattle specie. It will be ordered logit (acute, chronic or disease) if we are to work with both (cattle and sheep), but cattle belongs to one category of response so I was wondering on the kind of modelling approach should I use if both species of animals are to be included in the model?
> Yours,
> Aminu.
> *
> *   For searches and help try:
> *
> *
> *

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index