|
|
Growth Curve Model with Non-normal data |
|
Message/Author |
|
Nathan Smith posted on Tuesday, September 06, 2016 - 2:18 pm
|
|
|
I am working with data that are right skewed. That data are continuous in that they could be any real value between 0 and 100. I am wondering about the most appropriate way to estimate a growth curve model with these data. I imagine I could transform the data so that they are normally distributed. My other hunch is to use a robust estimator. Some guidance would be appreciated. |
|
|
Depending on the substantive setting, you could use MLR log transform use the skew-normal or skew-t distribution. |
|
Back to top |
|
|