| Message/Author |
|
|
|
|
Hi,
how can I obtain p-values for the (level 1 and level 2) correlations in a two-level-random model?
thanks
Guy |
|
|
|
|
There are two types of correlations: the H1 model (unrestricted mean and variance on each level and no random slopes anywhere) or the H0 model (this is the correlations implied by the random slope model).
To get the H1 model values one can use
ANALYSIS: TYPE = TWOLEVEL;
output: h1se;
or write the H1 model as a model
model:
%within%
y1-y10 with y1-y10;
%between%
y1-y10 with y1-y10;
output:stdY;
To get the H0 model implied correlations ( which are cluster specific) you
need to switch to the Bayes estimator and use the command
output:stdY(cluster); |
|
|
|
|
1. What kinds of correlations are these?
2. Is it possible to get Spearman correlations? |
|
|
|
|
1. They are Pearson product-moment correlations when we deal with continuous variables.
2. Not in Mplus. |
|
|
|
|
I tried this and got an error message that all variables are uncorrelated. However, I have obtained correlations among these variables previously...
MODEL:
%WITHIN%
EVs1 EVN1 EVM1
DM1 OF1
MIE1 MCE1
ECP1
WITH
EVs1 EVN1 EVM1
DM1 OF1
MIE1 MCE1
ECP1;
%BETWEEN%
EVs2 EVN2 EVM2
DM2 OF2
MIE2 MCE2 ECP2
GCP2 JS2
WITH
EVs2 EVN2 EVM2
DM2 OF2
MIE2 MCE2 ECP2
GCP2 JS2; |
|
|
|
|
| Answered on Support. |
|
|
| Back to top |
