Question about difference between GEE and Cox regression

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Question about difference between GEE and Cox regression

chantelanuit
Hi,

I have to analyze data related to a prospective cohort that involves
following a sample of respondents over 6 years. Measurement happens 5 times
and outcome is smoking onset. There are several predictors to be included
(smoking environment, school rules about smoking, peers smoking, etc.).

What is the difference between the question asked by :

a) GEE with repeated measurement and students nested in schools
and
b) Cox regression ?

I understand that Cox regression is about time to outcome modelling, but I
am unsure why one technique might be more relevant than the other.

Thanking you in advance for any suggestion,

Michael

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Reply | Threaded
Open this post in threaded view
|

Re: Question about difference between GEE and Cox regression

Maguin, Eugene
Michael,

Let's ignore the 'GEE' and 'students nested in schools' for the moment. Repeated measurements is about variation in sample means, for continuous variables, or sample proportions, for dichotomous variables, over time. Cox regression is about the hazard rate, the rate of events per unit time, in a continuous time model or the likelihood of the event per time block in a discrete time model. The nested in schools business applies to cox regression as well as repeated measures. I think a very useful book for you to look at would be Judith Singer and John Willett's book on multilevel models, which you are describing, and cox regression/survival analysis/event history analysis.

Gene Maguin


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Michael
Sent: Tuesday, April 03, 2012 10:28 PM
To: [hidden email]
Subject: Question about difference between GEE and Cox regression

Hi,

I have to analyze data related to a prospective cohort that involves
following a sample of respondents over 6 years. Measurement happens 5 times
and outcome is smoking onset. There are several predictors to be included
(smoking environment, school rules about smoking, peers smoking, etc.).

What is the difference between the question asked by :

a) GEE with repeated measurement and students nested in schools
and
b) Cox regression ?

I understand that Cox regression is about time to outcome modelling, but I
am unsure why one technique might be more relevant than the other.

Thanking you in advance for any suggestion,

Michael

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD