Dear Statalist users,
I am new in SEM models, and trying to fit my model.
I have questioner with several waves and 3 levels of analysis. In the data set, I have information in individual level (first level of analysis), team level (second level of analysis), and waves of study (third level of analysis). Also, I have 2 latent variable (construct) with 4 observed variables defining each of them.
Considering the fact that each individual's response is correlated during different waves, also the two constructs are correlated, what is the suitable model?
1. Should I define panel at first and then ignore the correlations among the responses of each individual? how I can make it? panel of each individual based on team or wave? or what?
Although, if I define panel I can not repeat the name of the variable in different waves.
2. Another option is to define a fixe effect, but is it good enough for such a model? and how I should use it?
3. Can I use the group option to define the fixed effect?
I tried the first version by defining all the potential relations as correlation which obviously failed.
The code was super long which dose not worth to mention here
I am new in SEM models, and trying to fit my model.
I have questioner with several waves and 3 levels of analysis. In the data set, I have information in individual level (first level of analysis), team level (second level of analysis), and waves of study (third level of analysis). Also, I have 2 latent variable (construct) with 4 observed variables defining each of them.
Considering the fact that each individual's response is correlated during different waves, also the two constructs are correlated, what is the suitable model?
1. Should I define panel at first and then ignore the correlations among the responses of each individual? how I can make it? panel of each individual based on team or wave? or what?
Although, if I define panel I can not repeat the name of the variable in different waves.
2. Another option is to define a fixe effect, but is it good enough for such a model? and how I should use it?
3. Can I use the group option to define the fixed effect?
I tried the first version by defining all the potential relations as correlation which obviously failed.
The code was super long which dose not worth to mention here
