Using co-variates in a logistic regression analysis

[Moshe Marko]

Hi,

 

I have two groups of older adults of "disable" and "non-disable". The goal is to predict the presence or absence of "disability" based on values of age, gender, body mass index, and leg strength. Comparing the two groups, only the age was found to be significantly different between the groups. Do I need to input all the four variables into the logistic model? If so, can I use step wise regression?

 

Moshe  

 

 

Moshe Marko, PT, DPT, MHS, OCS, CSCS
Assistant Professor
Department of Physical Therapy Education
College of Health Professions
SUNY Upstate Medical University
Room 2232  Silverman Hall
750 Adams Street
Syracuse, NY 13210-1834
315 464 6577
FAX [phone]
[hidden email]

[Swank, Paul R]

Do you mean that individually, only age was related? If so, it would be possible that all the variables together would predict better than age. First, the power of the test depends upon the strength of relation for the combined variables. Since age is significant, you may find that another variable added to the model becomes significant when age is in the model also because the power has increased. Also, there is the possibility of suppression if the predictors are correlated. One variable that does not relate to the outcome may “suppress” superfluous variance in another and therefore add significantly to prediction. I would not use any type of selection routine however. They do not what people think and can be very problematic.

 

Dr. Paul R. Swank,

Children's Learning Institute

Professor, Department of Pediatrics, Medical School

Adjunct Professor, School of Public Health

University of Texas Health Science Center-Houston

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Moshe Marko
Sent: Wednesday, September 21, 2011 2:07 PM
To: [hidden email]
Subject: Using co-variates in a logistic regression analysis

 

Hi,

 

I have two groups of older adults of "disable" and "non-disable". The goal is to predict the presence or absence of "disability" based on values of age, gender, body mass index, and leg strength. Comparing the two groups, only the age was found to be significantly different between the groups. Do I need to input all the four variables into the logistic model? If so, can I use step wise regression?

 

Moshe  

 

 

Moshe Marko, PT, DPT, MHS, OCS, CSCS
Assistant Professor
Department of Physical Therapy Education
College of Health Professions
SUNY Upstate Medical University
Room 2232  Silverman Hall
750 Adams Street
Syracuse, NY 13210-1834
315 464 6577
FAX [phone]
[hidden email]