Message/Author |
|
jonas helao posted on Tuesday, September 10, 2019 - 3:21 pm
|
|
|
In chapter 2 in your book on mediation modeling, it is said that the standardized m and y use their estimated standard deviation instead of their sample standard deviation. What is meant here? |
|
|
Two things. The model-estimated value can be different than the sample value. And, using the parameter label for the estimated standard deviation takes into account the sampling variation in the SD which is not the case if you just plug in the sample value. |
|
jonas helao posted on Wednesday, September 11, 2019 - 4:23 pm
|
|
|
Thank you. Does Mplus give the estimated standard deviation in the output? |
|
|
If you want the values, you find them in the Residual output. Well, you find variances so you have to take the square root to get the SD. If you want formulas to express standardized results, you have to express this in Model Constraint command using labels for parameters in the Model command. |
|
Back to top |