|
|
Calculating within-cluster variance |
|
Message/Author |
|
|
Hello, In addition to the cluster mean, I would like to calculate the (within-)cluster variance or standard deviation for some of my level 1-IVs. My aim is to use the (within-)cluster variance as IV on level 2, thereby representing the diversity/heterogeneity of the variable on the group level (similar to Ahearne/Lam/Hayati/Kraus 2013, p. 44). Is this possible with Mplus? Thank you! Eva |
|
|
Not automatically in a single step, but you can do TYPE=TWOLEVEL BASIC to get the variance and then enter it as a level-2 IV. |
|
|
Is it possible to compute a separate variance for each cluster? This would help in computing rwg. |
|
|
UG ex 9.28 shows how a random variance can be handled in Mplus. You can also get the estimated values in all clusters for any random effect by using the FSCORES option (see the UG index). |
|
Back to top |
|
|