SPSS logistic multilevel regression (generalised linear model)

[Laila]

Hi I have a few questions: 1. Do I need to restructure my data before I can begin analysis (what is the syntax) 2. How do I do the analysis (what is the syntax please) 3. How do I interpret the findings (I have found some pricey books but was hoping there was a youtube video or annotated interpretation of spss output somewhere that I can draw on). So here it goes. I am predicting if students went on to remain in college in 2014. I have the same students who completed the same attitudinal surveys in 2011 and 2013. They indicated in 2013 if they had enrolled in college for 2014 (yes/no). The independent variables from the survey are on a likert scale. I will also include other independent variables such as gender, ethnicity and prior attainment. I would also like to partition out the variance at the student and school level (and therefore conduct a multi-level logistic regression). I am using spss v 22. I have looked it up and so far what I can gather is I need to go into the command help and use generalised liner models and go from there. My spss data is currently structured as each student has a single record so all their 2011 and 2013 data are on the same line ( I matched on student name, school and gender between the 2011 and 2013 data). Any help would be greatly appreciated before I start pulling out my hair! Thank you!!!!

[Maguin, Eugene]

First of all, may I suggest that if you are at a university that you try very hard to find someone to work/consult with. I think it will make the process, which may be your first time through this analysis, easier and more defensibly rigorous.

 

That said,

References. Look at Joop Hox’s website. Also the UCLA IDRE website. Have you googled this topic? I suggest that you subscribe to the multilevel listserv (http://www.jiscmail.ac.uk/lists/multilevel.html) and look through the archives on this topic.

Q1: Right now your dataset apparently looks like this:

Schoolid, studentcasenumber, school_level variables, student_level variables

Your data structure is students nested within schools. It looks like you now have this data structure. You  don’t need to restructure your data.

 

Q2: Let’s see what you know. Write out your level 1 and level 2 models and the combined model. My opinion is that you need to know those to put together the fixed effects and random effects subcommands.

 

You’ve read the documentation for genlinmixed. Which of the examples is most similar to your planned analysis?

 

Q3: Roughly, you interpret these results as if you had done a normal distribution multilevel model. Suppose your current DV was had a normal distribution. Could you set-up, analyze and interpret such a model?

 

Gene Maguin

 

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Laila
Sent: Friday, September 18, 2015 7:20 PM
To: [hidden email]
Subject: SPSS logistic multilevel regression (generalised linear model)

 

Hi I have a few questions: 1. Do I need to restructure my data before I can begin analysis (what is the syntax) 2. How do I do the analysis (what is the syntax please) 3. How do I interpret the findings (I have found some pricey books but was hoping there was a youtube video or annotated interpretation of spss output somewhere that I can draw on). So here it goes. I am predicting if students went on to remain in college in 2014. I have the same students who completed the same attitudinal surveys in 2011 and 2013. They indicated in 2013 if they had enrolled in college for 2014 (yes/no). The independent variables from the survey are on a likert scale. I will also include other independent variables such as gender, ethnicity and prior attainment. I would also like to partition out the variance at the student and school level (and therefore conduct a multi-level logistic regression). I am using spss v 22. I have looked it up and so far what I can gather is I need to go into the command help and use generalised liner models and go from there. My spss data is currently structured as each student has a single record so all their 2011 and 2013 data are on the same line ( I matched on student name, school and gender between the 2011 and 2013 data). Any help would be greatly appreciated before I start pulling out my hair! Thank you!!!!

---

View this message in context: SPSS logistic multilevel regression (generalised linear model)
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