Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

What is the difference between covariates or factors, in a ordinal regression?

2,762 views
Skip to first unread message

Sushma Dhital

unread,
Feb 27, 2017, 9:07:45 AM2/27/17
to
I read that independent categorical variable should be treated as factors, is it true?

Bruce Weaver

unread,
Feb 27, 2017, 4:56:13 PM2/27/17
to
On Monday, February 27, 2017 at 9:07:45 AM UTC-5, Sushma Dhital wrote:
> I read that independent categorical variable should be treated as factors, is it true?

Yes, Factors are categorical explanatory variables (e.g., sex, marital status), and Covariates are continuous (or at least quantitative) explanatory variables (e.g., age, body weight).

Rich Ulrich

unread,
Feb 27, 2017, 10:13:40 PM2/27/17
to
Of course -- you can code up the dummy variables for
the factors and do the whole analysis of ANOVA with a
regression program. That is what all modern stat-packages
do, internally.

If you want to think of one factor as being in the model as
a logical "covariate" (with 1 or more d.f.), that is fine.


--
Rich Ulrich

Lijun Zhao

unread,
Sep 12, 2023, 7:11:06 AM9/12/23
to
hi Rich,
but, ideally, the sex should be entered in the factor section. right? The only thing is not to make contrasts. correct?


Lijun

Rich Ulrich

unread,
Sep 18, 2023, 1:12:10 PM9/18/23
to
On Tue, 12 Sep 2023 04:11:04 -0700 (PDT), Lijun Zhao
<lijunz...@gmail.com> wrote:
I don't know why you say 'ideally' -- If you have NO interest in
interactions or means, your output is easier to read if it is a
covariate.

> The only thing is not to make contrasts. correct?

? If you don't want contrasts ... I don't follow what you ask.

--
Rich Ulrich
0 new messages