GLM-Multiple Logistic Regression for ordinal variables

[ainda612]

Hello,

I am trying to find out if some demographic characteristics influence in the value given to some geographic points marked in a map. Those values are given in a scale of 1 to 10 (ordinal variables) and each subject may have marked more than one point (therefore I have repeated measures).

To analyze this I was advised to perform the Generalised Linear Model -Fixed Effects - Multiple logistic regression, but when I tried to do it a dialog box appeared saying "Repeated measures analysis is not supported for the multinomial probability distribution". I've been looking on the web for an explanation of this message but haven't found anything that could apply for this case.

I'm not an expert on statistics so could someone please tell me what does this mean? Is this a problem with the data I have or should I perform a different analysis?

Thanks in advance for your help
Paola

[Maguin, Eugene]

Look at the documentation for GenLinMixed. Repeated measures is a sub-sub command. I can't comment on what shows in the menu box; I don't use them.
Gene Maguin

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of ainda612
Sent: Tuesday, October 27, 2015 8:32 PM
To: [hidden email]
Subject: GLM-Multiple Logistic Regression for ordinal variables

Hello,

I am trying to find out if some demographic characteristics influence in the value given to some geographic points marked in a map. Those values are given in a scale of 1 to 10 (ordinal variables) and each subject may have marked more than one point (therefore I have repeated measures).

To analyze this I was advised to perform the Generalised Linear Model -Fixed Effects - Multiple logistic regression, but when I tried to do it a dialog box appeared saying "Repeated measures analysis is not supported for the multinomial probability distribution". I've been looking on the web for an explanation of this message but haven't found anything that could apply for this case.

I'm not an expert on statistics so could someone please tell me what does this mean? Is this a problem with the data I have or should I perform a different analysis?

Thanks in advance for your help
Paola



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/GLM-Multiple-Logistic-Regression-for-ordinal-variables-tp5730887.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

[Bruce Weaver]

It's actually a key word on the DATA_STRUCTURE sub-command.  From the FM under GENLINMIXED > DATA_STRUCTURE Subcommand:

REPEATED_MEASURES Keyword
The REPEATED_MEASURES keyword gives the repeated (or within-subject) effect. This effect defines the ordering of repeated measurements within subjects. If some measurements do not appear in the data for some subjects, then the existing measurements are ordered and the omitted measurements are treated as missing values. If REPEATED_MEASURES is not specified, then no repeated measures model is built.
v Specify a single field or a list of fields connected by asterisks (*) or the keyword BY.
v Each distinct combination of the values of the fields defines a separate repeated measure.
v Any field specified on the REPEATED_MEASURES keyword cannot be used as a SUBJECTS keyword or on the FIELDS subcommand.
v Each distinct repeated measures value must occur only once within a subject. Generally speaking, you should not use a WEIGHT field if REPEATED MEASURES is specified.


MIXED, on the other hand, has a REPEATED sub-command.  But in that case, the name is rather misleading, as it "specifies the residual covariance matrix in the mixed-effects model".  So for example, it can be used to allow for heterogeneous variances, etc, even in the absence of any repeated measures factors in the model.  

HTH.

Maguin, Eugene wrote

Look at the documentation for GenLinMixed. Repeated measures is a sub-sub command. I can't comment on what shows in the menu box; I don't use them.
Gene Maguin

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of ainda612
Sent: Tuesday, October 27, 2015 8:32 PM
To: [hidden email]
Subject: GLM-Multiple Logistic Regression for ordinal variables

Hello,

I am trying to find out if some demographic characteristics influence in the value given to some geographic points marked in a map. Those values are given in a scale of 1 to 10 (ordinal variables) and each subject may have marked more than one point (therefore I have repeated measures).

To analyze this I was advised to perform the Generalised Linear Model -Fixed Effects - Multiple logistic regression, but when I tried to do it a dialog box appeared saying "Repeated measures analysis is not supported for the multinomial probability distribution". I've been looking on the web for an explanation of this message but haven't found anything that could apply for this case.

I'm not an expert on statistics so could someone please tell me what does this mean? Is this a problem with the data I have or should I perform a different analysis?

Thanks in advance for your help
Paola



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/GLM-Multiple-Logistic-Regression-for-ordinal-variables-tp5730887.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

[Art Kendall]

What are the value labels of the 10 point scale?  Are these pure rankings of 10 stimulus points on the map?

Often rating scales are not severely discrepant form interval level of measurement.

Please explain your design in more detail.  It is possible that some of you analyses can be done using CATREG.
The virtue of that is to check whether you have one of the unusual situations where you get meaningfully better fit assuming ordinal levels of measurement than you would assuming interval level.