|
|
| Message/Author |
|
|
|
|
Dear. Muthen.
I have an EMA data, which combines several observations to produce one latent variable Y.
A total of 65 people respond four times a day for a week and come up with the value each time they respond.
1. I'm not sure which of the two level time series or cross classified time series (N> 1) is more appropriate (see the presentation Johns Hopkins Workshops, 08-17-2017, Part 3 (Muthen)).
2. In the case of the two level time series, smoking example used two variables: urge and negative affect. is it the same as in my case?
3. In addition, my variable is a latent variable, but are the above two methods applicable? I considered latent growth model before.
4. If I were to intervene and divide it into two groups, would I add this to the 'between' command?
5. Can you recommend a paper that analyzes EMA data with DSEM?
Thank you very much for your time. |
|
|
|
|
1. Sounds like you should use two-level modeling.
2. You mention a latent variable so perhaps you have several factor indicators measured at each time point.
3. You can do factor analysis DSEM.
4. See all the applied DSEM papers at
http://www.statmodel.com/TimeSeries.shtml |
|
|
| Back to top |
|
|
