|
|
| Message/Author |
|
|
|
|
Hi There,
I am trying to do a LCA with dummy coded variables for the following:
Smoking:
Smoke_try 1=try smoking 1 or twice
Smoke_occ 1=use occasionally, but not every day
smoke_reg 1=use daily
Mairjuana:
MJ_try 1=tried once or twice
Mj_occ 1=use occasionally, but not every week
Mj_reg 1=use weekly
Alcohol:
Alc_occ 1=binge drink 1-3 times a month
Alc_reg 1= binge drink once a week or more
Internalizing disorders and externalizing disorders (binary; 1=>/=1SD above the mean)
I keep getting the following output:
WARNING: THE BEST LOGLIKELIHOOD VALUE WAS NOT REPLICATED. THE SOLUTION MAY NOT BE TRUSTWORTHY DUE TO LOCAL MAXIMA. INCREASE THE NUMBER OF RANDOM STARTS.
IN THE OPTIMIZATION, ONE OR MORE LOGIT THRESHOLDS APPROACHED AND WERE SET AT THE EXTREME VALUES. EXTREME VALUES ARE -15.000 AND 15.000.
Any thoughts as to why I am getting this? How do I interpret the -15/+15? |
|
|
|
|
Here is my input:
! idvariable IS x_stud;
! cluster = x_idsch;
useobservations are (second EQ 1);
usevariables are
smk_occ smk_reg mj_occ mj_reg alc_occ alc_reg int_SD ext_SD;
categorical are
smk_occ smk_reg mj_occ mj_reg alc_occ alc_reg int_SD ext_SD;
classes = c (6);
missing are all (999);
Analysis:
type=mixture;
!type=complex;
output:
modindices tech11 ;
Any help would be great! |
|
|
|
|
+ and - 15 correspond to probabilities of 1 and 0 and are perfectly fine results.
Regarding the Warning, send your full output to Support along with your license number.
We ask that postings be limited to one window - for longer messages, send to Support. |
|
|
| Back to top |
|
|
