Great video and very helpful explanations. Can I ask you something? I would like to use STATA for a multilevel model with a categorical dependent variable (therefore not a dummy variable). My dependent variable has 6 categories. Is it possible in STATA 17.0? I don't know if there is a specific command
Great video. It was very helpful. Does the multilevel regression have assumptions like linear regression? How can they be tested in Stata? Do you have any literature recommendations?
Thank you very much for your kindly feedback, Mr. Mike Crowson Thank you very much for your kindly feedback, Ph.D. Mike Crowson, How can we check influential and outlier observations in the multilevel binary logistic regression model in Stata?
Great video, very clear explanations. But I'd like to note that the use of ICC as an indicator of whether the researcher should use multilevel modeling is disputed. ICC, after all, is a function of between-group variation in means. Even when there is little variation in group means, the relationships of interest (e.g. the one between s.e.s. and math score) may vary across the groups.
Thanks Mike for such useful illustrations. I'm interested in the same application but on survey data where survey weights should be incorporated , and weight re-scaling might be needed. Did you do any videos about this type of analysis?
Hi Mike I have a doubt. Which model can we run if, for example we are studying only one school? In that case, we have school level characteristics as well as student level variables.
If you are only studying students at a single school you can't do multilevel analysis. You can only do a standard regression. To use multilevel, you would need data on students from multiple schools
@@mikecrowson2462 Thank you for the reply. So, can we take school-level variables and student-level variables in the same model? Because school level characteristics will be the same for all students. Which model will be preferred in such case?
Thanks Dr. could you explain the basic concepts like random effect, fixed effect, linear mixed effect model, generalized linear mixed effect, Generalized estimated equation...their difference in concept .just one video it will help us a lot especially for beginners
Thank you for this useful video I have a question. I run several mixed effect models. In my results, ICC for the level 1 model is higher than the null model. And ICC for the level 2 model is less that ICC for the level 1. What does this mean?
Hello Mike, first of all thank you so much for your videos! What I'm finding is that I can't run the "lrtest" command to compare my models because the number of observations is different for each of them, depending on the missing data they are affected by, Do you have any video on how to deal with the missing data in order to properly carry out multi-level modeling?
Hi there. Unfortunately, I have yet to find a way around this problem, apart from ensuring there is no difference in the number of missing cases between the models. I suppose you could listwise delete cases that are likely to create this problem, or use some type of imputation. Believe me, if I find something that works, I'll it up! Thanks for your comment and question!
@@mikecrowson2462 Thanks for your reply. I have been trying multiple imputation as a means to solve the problem, but the knowledge required to carry it out in such a large dataset (I am using TIMSS 2015) is far beyond me. Listwise deletion would greatly decrease the size of my sample. Do you think doing "lrtest m1 m2, force" is acceptable? Or could I just not do a likelihood-ratio test and directly compare my models' with the null model in terms of explained variance using Snijders and Bosker (2012)'s formula? R2 = 1 - (("model-n σ2" + "model-n τ2") / ("null model σ2" + "null model τ2)) Where : σ2 = level-two random error variance τ2 = level-one random error variance I am really sorry to bother you, and again I want to thank you very much for your help. I am a phd student and unfortunately I am not getting much support in my faculty, statistics-wise.
thank you for your video! I am already subscribe to your channel. I am wondering that how to use cross-classified multilevel model. What is the difference between cross-classified and normal multilevel model? Can i find manuel and tutorial about cross-classified multilevel model?
Thank you very much for your kindly feedback, Mr. Mike Crowson Thank you very much for your kindly feedback, Ph.D. Mike Crowson, How can we check influential and outlier observations in the multilevel binary logistic regression model in Stata?
Thank you! You cannot imagine how helpful this is when writing my thesis during this corona lockdown!
Great video and very helpful explanations. Can I ask you something? I would like to use STATA for a multilevel model with a categorical dependent variable (therefore not a dummy variable). My dependent variable has 6 categories. Is it possible in STATA 17.0? I don't know if there is a specific command
Great video. It was very helpful. Does the multilevel regression have assumptions like linear regression? How can they be tested in Stata? Do you have any literature recommendations?
Thanks for this great video. I have a 2 level sem modetating mediation model. Can I do the the cfa and hypotheses test with stata?
Do you need to allow the slope for SES to randomly vary (in this case varying independently from intercept) in order to use it in an interaction term?
Thank you very much for your kindly feedback, Mr.
Mike Crowson
Thank you very much for your kindly feedback, Ph.D. Mike Crowson, How can we check influential and outlier observations in the multilevel binary logistic regression model in Stata?
Very helpful, thank you! One question: Do I include control variables like age or gender already in the Nullmodel?
Great video, very clear explanations. But I'd like to note that the use of ICC as an indicator of whether the researcher should use multilevel modeling is disputed. ICC, after all, is a function of between-group variation in means. Even when there is little variation in group means, the relationships of interest (e.g. the one between s.e.s. and math score) may vary across the groups.
Thanks Mike for such useful illustrations. I'm interested in the same application but on survey data where survey weights should be incorporated , and weight re-scaling might be needed. Did you do any videos about this type of analysis?
thank you very much for your video, I just want to know how can I get the fixed effect test (F-TEST IN SPSS) in stata
Hi Mike
I have a doubt. Which model can we run if, for example we are studying only one school? In that case, we have school level characteristics as well as student level variables.
If you are only studying students at a single school you can't do multilevel analysis. You can only do a standard regression. To use multilevel, you would need data on students from multiple schools
@@mikecrowson2462 Thank you for the reply.
So, can we take school-level variables and student-level variables in the same model? Because school level characteristics will be the same for all students.
Which model will be preferred in such case?
Sir, can we have a random slope of a 2nd level variable in a 2 level model.
how can I get standardized coefficients (beta) in multilevel regression ?
Thank you for your work! But I have to say the google drive links is failed (error 404). I don't know if that is my problem.
Hi there, the links should be fixed now. I cannot understand how links can randomly go bad! Thanks for your patience!
Thanks Dr. could you explain the basic concepts like random effect, fixed effect, linear mixed effect model, generalized linear mixed effect, Generalized estimated equation...their difference in concept .just one video it will help us a lot especially for beginners
Thank you!! It's very helpful!
Thank you! Very useful video. Samurais function rules!
You are very welcome, Rodrigo!
Another question
Is the mixed effects model usefull when I have a daily time serise variable (level 1) and country cultural effects ( level 2)?
Thank you for this useful video
I have a question. I run several mixed effect models. In my results, ICC for the level 1 model is higher than the null model. And ICC for the level 2 model is less that ICC for the level 1. What does this mean?
Hello, the google drive seems to be not working
Hi there, the links should be fixed now. I cannot understand how links can randomly go bad! Thanks for your patience!
Hello Mike, first of all thank you so much for your videos!
What I'm finding is that I can't run the "lrtest" command to compare my models because the number of observations is different for each of them, depending on the missing data they are affected by,
Do you have any video on how to deal with the missing data in order to properly carry out multi-level modeling?
Hi there. Unfortunately, I have yet to find a way around this problem, apart from ensuring there is no difference in the number of missing cases between the models. I suppose you could listwise delete cases that are likely to create this problem, or use some type of imputation. Believe me, if I find something that works, I'll it up! Thanks for your comment and question!
@@mikecrowson2462 Thanks for your reply.
I have been trying multiple imputation as a means to solve the problem, but the knowledge required to carry it out in such a large dataset (I am using TIMSS 2015) is far beyond me. Listwise deletion would greatly decrease the size of my sample.
Do you think doing "lrtest m1 m2, force" is acceptable?
Or could I just not do a likelihood-ratio test and directly compare my models' with the null model in terms of explained variance using Snijders and Bosker (2012)'s formula?
R2 = 1 - (("model-n σ2" + "model-n τ2") / ("null model σ2" + "null model τ2))
Where :
σ2 = level-two random error variance
τ2 = level-one random error variance
I am really sorry to bother you, and again I want to thank you very much for your help. I am a phd student and unfortunately I am not getting much support in my faculty, statistics-wise.
thank you for your video! I am already subscribe to your channel. I am wondering that how to use cross-classified multilevel model. What is the difference between cross-classified and normal multilevel model? Can i find manuel and tutorial about cross-classified multilevel model?
Thank you! it´s fantastic!
p.s. thanks for this awesome video
Increíbleee, te quiero besarrrrrr
2 advertisements in 7:33 mins ... your choice but you definitely lost my there.
I think you should be thankful to Mr. Crowson for his effort and getting this help for free rather than complaining about the advertisement.
@@johanneshuttner530 If smth is free, you are the product.
Thank you very much for your kindly feedback, Mr.
Mike Crowson
Thank you very much for your kindly feedback, Ph.D. Mike Crowson, How can we check influential and outlier observations in the multilevel binary logistic regression model in Stata?