This post concerns the situation where you have a dependent variable with three or more unordered categories.

Typical options for analysis include:

- Multinomial logistic regression
- Discriminant Function Analysis
- Categorical Regression (Optimal Scaling)

**Resources for Multinomial Logistic Regression**

- David Garson's Logistic Regression Resources: The material focuses on binary logistic regression in SPSS. It does however make some mention of multinomial logistic regression. Also, many of the ideas of multinomial logistic regression can be seen as a straightforward extension of binary logistic regression.
- UCLA provide annotated SPSS output for a multinomial logistic regression
- Chan (PDF 2mb) provides an example of a multinomial logistic regression with SPSS tips.
- I previously posted some resources on binary logistic regression.

hi, i am very new to SPSS, i want to ask

ReplyDeletemy spss version 16

The step by step to do binary logistic is

1. Analyze

2. Regression

3. Binary Logistic...

But under the regression, my spss on display 3 choice which is

Linear

Curve estimation

partial least square

ordinal

I read many book and search it on google, asking me to click

1. Analyze

2. Regression

3. Binary Logistic...

mengzhen89@gmail.com

Hope you can help me

Mengzhen

@Meng Zhe

ReplyDeleteMy guess is that you have the student version of SPSS. You might need to install an add-on module.

I'm not sure about pricing these days, but there used to be a Grad Pack that might do what you want.

See Regression Models

http://www.spss.com/software/statistics/regression/

thanks you , can I have you email ? I am very new in this, i cant find any help within my city to assist me in my research that need to use Multinomial Logistic Regression. I really need help. If you are able to help me, can email me at mengzhen89@gmail.com ?

ReplyDelete@mengzhen89 I suggest you post any questions on stats.stackexchange.com

ReplyDeleteHello! Do you know whether I can do a repeated measures multinomial logistic regression using SPSS?

ReplyDeleteThank you!

Caroline

Hi Caroline,

ReplyDeleteI'm not sure.

Perhaps try asking on stats.stackexchange.com

Hi Im Hiroshi. Can I borrow data to best perform the Multinomial Logistic Regression? pls send it to my email trish_perez2006@yahoo.com your help would mean a lot to me.. Thank you so much..

ReplyDeleteHere's an example

Deletehttp://www.ats.ucla.edu/stat/R/dae/mlogit.htm with data here:

http://www.ats.ucla.edu/stat/r/dae/mlogit.csv

Thank you Jeromy Anglim for giving me an example for the Multinomial Logistic Regression. Is it possible that i can use ordinal data with my dependent variable? As what i've read, nominal data is applicable with this test. And as I explored it, one of the assumptions is the Proper Model Specification. I don't know how to satisfy this assumption using SPSS 18. In short I'm still novice with this test. Hope you could help me all through out the test. And wish to formulate a model with my study. I have my data here but no idea how to finish this.

DeleteHello Jeromy,

ReplyDeleteAny suggestions for how to switch the reference group for the independent variable in multinomial logistic regression? It appears that the default is the last, and I can't seem to change it. I figured how to set the reference group for the DV, but not for the IVs.

In addition, how do you get the n for each subgroup of the IVs? MLR does not seem to report this, but it is being requested?

Thanks, Jamie

If you want to change the reference group on the DV, just recode the DV. E.g., If your DV was coded 0 and 1 then you could switch the coding to change the reference group like so:

ReplyDeleteRECODE y (0=1) (1=0) INTO y_new.

EXECUTE.

i am doing project on thyroid affected people in Guwahati city, Assam, but cannot find any way to analyse it or to convert it in SPSS. will you help me out. and what the results of SPSS shows us?

ReplyDeleteSee here: http://jeromyanglim.blogspot.com.au/2011/03/how-to-ask-me-statistics-question.html

Delete