|
|
| Message/Author |
|
|
|
|
Hi,
I have a problem concerning the trend analysis and I hope you can help me with that. Here is my syntax.
VARIABLE:
NAMES = id_new sitzung Inter Intra HSCL
Inter_L1 Intra_L1 WW_L1 HSCL_L1
Inter_L2 Intra_L2 HSCL_L2;
MISSING = all (-99);
USEVARIABLES = Inter Intra time;
CLUSTER = id_new;
LAGGED = inter(1);
TINTERVAL = sitzung(1);
WITHIN = Inter Intra time;
BETWEEN = ;
ANALYSIS:
TYPE = twolevel random;
ESTIMATOR = bayes;
DEFINE:
time = sitzung;
MODEL:
%WITHIN%
p_11 | Inter ON Inter&1;
p_22 | Intra ON Intra&1;
p_12 | Intra ON Inter&1;
p_21 | Inter ON Intra;
t1 | Inter ON time;
t2 | Intra ON time;
if I leave the time variable completely out, the model runs fine. On the between level all variables and random effects are correlated. I hope you can help me with the problem. Thanks in advance. |
|
|
|
|
| Looks correct. Please send the output (and data if possible) to Support. |
|
|
|
|
| Also send the gh5 file so we can look at the traceplot to see if a parameter is going awry. |
|
|
| Back to top |
|
|
