LCA syntax problem PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
Message/Author
 Hee-Kyung Kwon posted on Monday, December 02, 2002 - 9:04 pm
The followings are the input and output I have. The program keeps saying that "CINDICATORS" option is unknown.... What's wrong with my syntax?


INPUT INSTRUCTIONS

Title: ALCOHOL
Data: FILE IS C:\HKTEMP\ALC.DAT;
TYPE IS INDIVIDUAL;
FORMAT IS FREE;
NOBSERVATIONS ARE 350;
VARIABLE:NAMES ARE AFFDIS AUSE BUSE CUSE DUSE dis sex par neo paralc con;
USEVARIABLES ARE DIS AUSE BUSE CUSE DUSE;
MISSING=all(999);
CLASSES=C(3);
CATEGORICAL = DIS;
ANALYSIS:TYPE=MIXTURE;
MITERATIONS=500;
MODEL:
%OVERALL%
I BY AUSE@1 BUSE@1 CUSE@1 DUSE@1;
S BY AUSE@-2 BUSE@-1 CUSE@0 DUSE@1;
[AUSE@0 BUSE@0 CUSE@0 DUSE@0];
i*7 s*.7;
i with s*.5;

C#1 BY DIS*-8;
C#2 BY DIS*4;
C#3 BY DIS*6;

%C#1%

[I*1.3 S*.3];

%C#2%


[I*2.3 S*.5];

%C#3%

[I*3.3 S*.5];


Mplus VERSION 2.01 PAGE 2


Mplus VERSION 2.01 PAGE 3


Mplus VERSION 2.01 PAGE 4


*** ERROR in Model command
C BY U statements are no longer supported. Please refer to Chapter 9 of the
Mplus User's Guide on how to change these statements using the new language.


MUTHEN & MUTHEN
11965 Venice Blvd., Suite 407
Los Angeles, CA 90066

Tel: (310) 391-9971
Fax: (310) 391-8971
Web: www.StatModel.com
Support: Support@StatModel.com

Copyright (c) 1998-2001 Muthen & Muthen
 Linda K. Muthen posted on Tuesday, December 03, 2002 - 6:11 am
The message from the output you sent,

C BY U statements are no longer supported. Please refer to Chapter 9 of the
Mplus User's Guide on how to change these statements using the new language.

is telling you that the c BY u syntax has been changed in Version 2. If you refer to pages 132-134 in Chapter 9 of the Mplus User's Guide, you will find the new syntax.

You can also look at the mixture modeling examples on the website which show the new syntax.
 Robert Urban posted on Monday, December 20, 2010 - 10:55 am
Dear Dr. Muthén,

With a colleague we wanted to perform a LCA with a complex sample (teachers' ratings).
We wrote the syntax:

Variable: names are gender class teacher u1-u25 x1-x25 y1-y25;

usevariables = x2 x10 x15 x21 x25;
categorical = x2 x10 x15 x21 x25;
missing is blank;
stratification is teacher;
classes = c (2);
Analysis: Type = Mixture Complex ;
Starts = 500 10;

Output: tech1 tech8 tech10 tech11;

However the output is empty. The message is the following:

INPUT READING TERMINATED NORMALLY

LCA teacher SDQ complex sample

Could you help us to clarify the problem.

Thank you very much.
Best regards,
Robert
 Linda K. Muthen posted on Monday, December 20, 2010 - 12:16 pm
Please send the input, data, output, and your license number to support@statmodel.com.
 socrates posted on Tuesday, November 22, 2011 - 11:26 pm
I built a LCA with 2 latent classes and ran this model in Mplus Version 3.11. So far, everything went well.

However, when running exactly the same syntax with exactly the same dataset in Version 6.1, I get the following error message:

THE MODEL ESTIMATION DID NOT TERMINATE NORMALLY DUE TO AN INSUFFICIENT
NUMBER OF E STEPS. INCREASE THE NUMBER OF MITERATIONS. ESTIMATES
CANNOT BE TRUSTED.

Why does this happen?

P.S. The error message still appears after increasing the miterations up to 500,000.
 Linda K. Muthen posted on Wednesday, November 23, 2011 - 1:38 pm
Please send the output and your license number to support@statmodel.com.
 Oxnard Montalvo posted on Thursday, May 17, 2012 - 6:25 pm
Hi,
I have an latent class model with 7 indicators, a continuous covariate and binary group. I am unsure how to construct the syntax to hold item response probabilties invariant over the groups, but leave the other parameters (latent statuses and covariate effects) to vary between the genders. Can you advise as to what the syntax should look like?
Cheers,
Oxnard
 Linda K. Muthen posted on Friday, May 18, 2012 - 1:52 pm
It should look like Example 7.12 but without the direct effect from x to u4.
 Bengt O. Muthen posted on Friday, May 18, 2012 - 6:08 pm
A week or so ago, somebody asked about using TECH11 (LMR) and TECH14 (BLRT) to determine the number of latent classes. They were concerned about needing to have their largest class last. An Mplus Web Note No. 14 is now available and will shortly be posted that gives a detailed description of how to use these methods. The note describes the K-1STARTS and LRTSTARTS options that go with TECH11 and TECH14. It points out that re-ordering the classes to have the largest class last is not necessary. Avoiding local optima by replication of the best loglikelihood value is discussed. The note also shows the value of the OPTSEED option. A 3-step strategy is presented for doing the analyses successfully and with as little computational time as possible.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: