|
Message/Author |
|
Yue Ma posted on Tuesday, November 07, 2017 - 5:06 pm
|
|
|
I am working on factor analysis (first EFA, then CFA). The observed indicators are continuous with nonnormality. Which estimator should I use for EFA? Which estimator should I use for CFA? Given that there are several estimators that are robust to nonnormality. Thank you! |
|
|
Use MLR. |
|
Yue Ma posted on Tuesday, November 07, 2017 - 7:01 pm
|
|
|
Thank you, Dr. Muthen! There are missing data that are likely missing not at random, such that participants who are lower in the abilities (latent factors) are more likely not be able to complete the performance tests (the observed indicators) and thus are missing on the observed indicators. How do I handle the missing not at random data? Thank you! |
|
|
To learn about NMAR modeling, see our new book as well as Enders'. You may still have MAR if the performance on some of your simpler tasks predict missingness on some of your harder tasks. |
|
Yue Ma posted on Friday, November 10, 2017 - 3:28 pm
|
|
|
Thank you! |
|
Back to top |
|
|