GLM

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

GLM

Anthony Santella
Dear Listserv,

I have 2 models, and would like your scholarly input on which SPSS
procedures you would use to find the predictive value of the independent
variables on the dependent variable.  My data comes from a large hospital
database of close to 9000 observations over 6 years (about 1500
observations a year, not the same people from year to year).

1. 1 continuous dependent variable, 9 categorical indendepent variables

2. 1 categorical dependent variable (variable is dichotomous), 9
categorical indendepent variables

I was under the impression that I should use a factorial ANOVA, but am not
sure anymore.

Thank you in advance,
Anthony
Reply | Threaded
Open this post in threaded view
|

Re: GLM

statisticsdoc
Anthony,

Briefly, for the second of these, you will want to use Logistic Regression.
A couple of questions I would first ask are:

Are any individuals represented more than once in the 9,000 observations?
You mentioned that the dataset does not contain the same people from year to
year, but are there any repeat observations within years?

With so many categorical independent variables (presumably with many
levels), does your data contain any empty cells?

Best,

Steve

For personalized and professional consultation in statistics and research
design, visit
www.statisticsdoc.com


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]]On Behalf Of
Anthony Santella
Sent: Sunday, January 14, 2007 3:52 PM
To: [hidden email]
Subject: GLM


Dear Listserv,

I have 2 models, and would like your scholarly input on which SPSS
procedures you would use to find the predictive value of the independent
variables on the dependent variable.  My data comes from a large hospital
database of close to 9000 observations over 6 years (about 1500
observations a year, not the same people from year to year).

1. 1 continuous dependent variable, 9 categorical indendepent variables

2. 1 categorical dependent variable (variable is dichotomous), 9
categorical indendepent variables

I was under the impression that I should use a factorial ANOVA, but am not
sure anymore.

Thank you in advance,
Anthony