|
|
Parallel Process and Multi-level Mode... |
|
Message/Author |
|
|
Hello, I am new to MPLUS and was interested in using parallel process latent growth curve analyses simultaneously with multi-level modeling (2-levels: time, individual and neighborhood). First, is this possible to do since I understand that the parallel process requires a wide-formatted data, whereas multi-level modeling requires a long-formatted data? If so, what would you recommend for a data structure? Also, is it correct to say that when looking at our WITHIN results, the estimates/SEs account for the correlation of individuals living within the same neighborhood? Any guidance you could provide would be much appreciated. Thank you. |
|
|
Yes, this can be done. You have as columns in your data the two processes for all time points, say: y1, y2, ..., yT, z1, z2, ..., zT plus covariates and repeat that over all people in all clusters (data don't have to be given cluster by cluster). And, yes, the SEs then account for the intra-neighborhood correlations because you model this correlation by allowing Between (Level2) level variances. |
|
Back to top |
|
|