|
|
| Message/Author |
|
|
|
|
Hello,
I am new to building multi-level models in MPlus and have run into some issues. I want to examine the effect of teaching style (Z and T) on outcome measures of student data (X), as well as the indirect effect on the outcomes via their attitudes (Y and S).
Initial runs indicated I needed to increase Miterations and decrease logcriterion. I have included the code below. When I run this code, the program takes over two weeks to run, and on several attempts the program crashed.
What can I do to lessen the length of the program run?
ANALYSIS:
TYPE = TWOLEVEL RANDOM;
MITERATIONS = 10000;
STARTS = 20;
LOGCRITERION = .0001;
MODEL:
%WITHIN%
YW SW X;
YW by Y1 Y2 Y3;
SW by S1 S2 S3;
X ON YW SW;
%BETWEEN%
Z T YB SB X;
YB by Y1 Y2 Y3;
SB by S1 S2 S3;
Z by Z1 Z2 Z3 Z4 Z5;
T by T1 T2 T3 T4;
Z ON YB(b);
T ON YB(c);
Z ON SB(d);
T ON SB(e);
X ON Z;
X ON T;
X ON YB(f);
X ON YB(g);
MODEL CONSTRAINT:
NEW(indb);
Indb=b*f;
NEW(indc);
Indc=c*f;
NEW(indd);
Indd=d*g;
NEW(inde);
Inde=e*g;
OUTPUT: TECH1 TECH8 CINTERVAL;
Thank you for your help. |
|
|
|
|
| Looks like this should be a straightforward analysis. To help you, we need to see your output/input and data - send to Support@statmodel.com along with your license number. |
|
|
| Back to top |
|
|
