I am looking for Multilevel Modeling Papers that fits the general description below;
Outcome is student course grade. And a student can take multiple courses in a semester. For example, one student might be taking
three college level courses from three different faculty and another student might be taking only two courses from a different set
of two faculty, a third student might be taking two different courses from the same faculty, or a student might be taking only one course.
- I have predictors at the student, classroom, and faculty level.
- Initially, this looks like a multiple membership design, but my understanding of multiple membership is students have one outcome variable. In this scenario, students can have more than one outcome.
- I don't think this is repeated measures design, students are taking the courses during the same semester and receive their course grade the end of the semester.
- In this design, courses are uniquely nested within faculty, but students are not necessarily uniquely nested within courses.
As I mentioned above, students might be taking multiple courses from different and sometimes same faculty.
Any help is greatly appreciated.
Thank you
Take a look at the book in the link. It might help. Best wishes, David
https://www.amazon.com/Analysis-Regression-Multilevel-Hierarchical-Models/dp/052168689X/ref=sr_1_3?s=books&ie=UTF8&qid=1526930356&sr=1-3&keywords=Andrew+Gelman
This is a great free online course for multilevel modeling:
http://www.bristol.ac.uk/cmm/learning/online-course/
They have great hands-on examples with MLwiN, R and Stata.
Also, you can handle the multiple outcomes issue in a Structural Equation Modeling (SEM) environment with a crossed random or fixed effects model. I cannot recall any particular paper right now, but I am pretty sure you can get some answers from structural equation modeling listserv located at http://www2.gsu.edu/~mkteer/semnet.html
hi
book:
Andy Field-2017. Discovering Statistics Using IBM SPSS Statistics-SAGE Publ
It is possible to have a multivariate multilevel model with more than one response nested within individuals who may then be cross-classified by some higher level structure. But I would think I would treat Faculty as a fixed effect rather than as a random classification. The number of responses an individual takes could be imbalanced eg some take 1 ,some take 8.
Here is some guidance on multivariate multilevel models
http://www.bristol.ac.uk/cmm/software/mlwin/mlwin-resources.html#multipleresp
If you can wait few weeks, a paper I wrote using multivariate multilevel models is going to be published on Social Compass. Despite the different topic, I use the same structure: outcomes nested into individuals nested into countries with some higher livel units (religious denominations in my case, faculty in yours) as fixed effects.
And, of course, Kelvyn Jones' paper can be an amazing guidance as it was for me
Article Investigating the Macro Determinants of Self-Rated Health an...
Nedim that is good to hear.
Have a look at
https://www.researchgate.net/post/Can_you_share_any_examples_of_the_use_of_multivariate_regression
Nedim, I will send how far I have got to you privately; and have now done so.
Kelvyn
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Powders