Dear all,
I have a very quick question.
In a binary logistic regression model, say I have a 3 category independent variable - categories red, blue, green. Say my dependent variable is whether someone develops a disease (1) or not (0). I code the categorical variable using 2 dummy binary variables (indicator coding) using green as the reference category i.e.
X1 X2
Red 1 0
Blue 0 1
Green 0 0
When I generate my model, I have, hypothetically, the coefficients: for x1 : beta 1 = 1.2 ( i.e. Odds Ratio=exp(1.2)=3.3) and for x2: beta 2 = 2.8 (i.e. odds ratio=exp(2.8)=16.4).
Now my p value for beta 1 is < 0.0005 and my p value for beta 2 is also < 0.0005 .
Now does the significance of the coefficients mean that:
a) the *odds* of developing the disease for someone in the red category is significantly different from the *odds* of developing the disease for someone in the green category.
And
b) the *odds* of developing the disease for someone in the blue category is significantly different from the *odds* of developing the disease for someone in the green category.
Thanks for this help - texts usually provides lots of detail on how to interpret the categorical variables in a model but fail to elaborate on how to express the meaning of a significant coefficient in a categorical scenario.
Kindest Regards,
Kim
Dr Kim Pearce PhD, CStat
Senior Statistician
Haematological Sciences
Institute of Cellular Medicine
William Leech Building
Medical School
Newcastle University
Framlington Place
Newcastle upon Tyne
NE2 4HH
Tel: (0044) (0)191 282 0451
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|