|
Message/Author |
|
In-Hee Choi posted on Wednesday, April 21, 2010 - 10:05 pm
|
|
|
I am doing a simulation study about the effects of DIF patterns on detecting latent classes in Mixture Rasch model. There are 30 items and two latent classes of subjects. Among 30 items, last three items (28-30) have the same difficulties in two classes, in other words, they are the anchor items. And the difficulty parameters of the first 28 items are set to be freely estimated in two classes. These are the model statements I have used. %OVERALL% f BY y1-y30@1; [y28$1 y29$1 y30$1]; %c#1% [f@0]; f; [y1$1-y27$1]; %c#2% [f]; f; [y1$1-y27$1]; And I specified number of classes as two, and "TYPE = MIXTURE", and "ALGORITHM=INTEGRATION". However, in output, the thresholds for y28$1-y30$1 were not the same. Is there any something wrong in input statement? Thanks in advance! |
|
In-Hee Choi posted on Wednesday, April 21, 2010 - 10:20 pm
|
|
|
I want to clarify some command. In model statement, for each class, I used this command for every 27 items. [y1$1 y2$1 y3$1 ... y25$1 y26$1 y27$1]; |
|
|
Please send the full output and your license number to support@statmodel.com. |
|
Back to top |
|
|