redundant parameters in mutlinomial logistic regression

[Greg]

Hi everyone,

This forum has been very helpful in the past and hopefully it will also be with this question I have.

My understanding is that the categorical predictors are entered in the 'factors' box, and the non-categorical in the 'covariates' box. However, when entering all the predictors in the 'covariates' box, I do not get the message "parameters set to zero becasue they are redundant". In addition, I also noticed that when all predictors are entered as 'covariates', the results, from a quick glance, "seem" to make sense.

Does this make sense? I apologize if this is a trivial question, but I haven't had a lot of hands-on experience with multi log regression. Any suggestions/ideas would be appreciated!

Regards,
Grigoris

[Swank, Paul R]

Variables entered into the factors box are assumed categorical and are reparameterized because a regression assumes ordered variables. In coding a categorical variable with indicator codes (0,1), one category is redundant because an intercept is included in each model by default. A variable in the covariate box is assumed ordered and is not reparametrized.

Dr. Paul R. Swank,
Professor and Director of Research
Children's Learning Institute
University of Texas Health Science Center-Houston


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of grigoris
Sent: Friday, February 27, 2009 12:55 PM
To: [hidden email]
Subject: redundant parameters in mutlinomial logistic regression

Hi everyone,

This forum has been very helpful in the past and hopefully it will also be
with this question I have.

My understanding is that the categorical predictors are entered in the
'factors' box, and the non-categorical in the 'covariates' box. However,
when entering all the predictors in the 'covariates' box, I do not get the
message "parameters set to zero becasue they are redundant". In addition, I
also noticed that when all predictors are entered as 'covariates', the
results, from a quick glance, "seem" to make sense.

Does this make sense? I apologize if this is a trivial question, but I
haven't had a lot of hands-on experience with multi log regression. Any
suggestions/ideas would be appreciated!

Regards,
Grigoris

--
View this message in context: http://www.nabble.com/redundant-parameters-in-mutlinomial-logistic-regression-tp22252171p22252171.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
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