Message/Author |
|
MF posted on Thursday, November 05, 2015 - 7:24 pm
|
|
|
Hello, I am conducting a path analysis predicting two continuous variables by three continuous predictors and two interaction terms. My data are not normally distributed and have high kurtosis. The data are independent of one another. N is about 400. I'm trying to decide between using estimator MLR and MLMV. What are the reasons behind using one over the other? Thank you for your help! |
|
|
I would use MLR because it can use all available information. MLMV uses listwise deletion. |
|
MF posted on Friday, November 06, 2015 - 7:05 am
|
|
|
Thank you for your reply! In this particular case, I'm not missing any data. Are there others reason to choose one over the other? |
|
|
They should not be that different. |
|
Back to top |