|
|
A question about SK tests |
|
Message/Author |
|
milan lee posted on Saturday, December 07, 2013 - 4:23 pm
|
|
|
Skewness and Kurtosis tests are recommended in Dr.Muthen's chpt published in 2004. However, as mentioned in the section of "Statistical checking", the SK tests cannot be applied to longitudinal data with missing data or with outliers. So, if I want to keep the missing data using ML, I cannot use the SK tests? On the other hand, supposedly that the dataset contains no missing data but a high non-normality in the outcomes. Can I still apply SK tests? Thank you! |
|
|
Yes and yes. But I've found that the SK testing often isn't that useful so I wouldn't rely very heavily on it. |
|
Back to top |
|
|