Prev: Top R packages for SAS programmers
Next: reading raw file
From: RKM on 11 Apr 2010 19:15
I'm doing an ordinal logistic regression, and I'm new to them (and
somewhat new to SAS), so I'm a bit confused about how to interpret the
output.

The variable I'm interested in is a covariate, and it's significant,
with a positive parameter estimate. However, my understanding is that
the sign of the odds ratio (and thus the sign of the parameter
estimate?) doesn't always indicate the direction of the effect in this
type of model. I was expecting a negative relationship based on a
graph of the data.

Can anyone tell me how I determine the direction of the relationship
from the output?

Thanks so much.
From: Adam on 14 Apr 2010 09:25
Is it possible you're getting your parameter estimates and odds ratios mixed up?

Odds ratios are always positive, but if they are less than one, that's equivalent to a negative parameter estimate.

---
frmsrcurl: http://compgroups.net/comp.soft-sys.sas/Interpreting-Glimmix-parameter-estimates
From: RKM on 20 Apr 2010 10:22
Which would mean that a negative paramater estimate always means an
inverse relationship between the dependent and independent variable,
and a positive estimate means a positive relationship? I'm getting a
positive estimate but the graph definitely looks like a negative
relationship.

On Apr 14, 9:25 am, Adam <u...(a)compgroups.net/> wrote:
> Is it possible you're getting your parameter estimates and odds ratios mixed up?
>
> Odds ratios are always positive, but if they are less than one, that's equivalent to a negative parameter estimate.
>
> ---
> frmsrcurl:http://compgroups.net/comp.soft-sys.sas/Interpreting-Glimmix-paramete...

 | 
Pages: 1
Prev: Top R packages for SAS programmers
Next: reading raw file