| Message/Author |
|
|
| Dylan John posted on Monday, June 05, 2017 - 10:24 am
|
|
|
Could someone please look at my syntax and let me know if this is correct. This specific analysis is investigating a 3-latent profile model using 4 continuous indicator variables. Does this seem correct or am I missing anything?
Data:
FILE IS "H:\MPlus Analysis with Weights\SAS_with_weights_not_standardized.txt";
Variable:
NAMES ARE
PERSRUK
T1_ADHD
T2_ADHD
T3_ADHD
T1_INT
T1_CON
T1_IND
T2_INT
T2_CON
T2_IND
T3_INT
T3_CON
T3_IND
weight;
Usevariables are
T1_ADHD
T1_INT
T1_CON
T1_IND;
classes = c(3);
weight is weight;
Analysis:
Type=mixture;
MODEL:
%C#1%
[T1_ADHD T1_INT T1_CON T1_IND];
%C#2%
[T1_ADHD T1_INT T1_CON T1_IND];
%C#3%
[T1_ADHD T1_INT T1_CON T1_IND]; |
|
|
|
|
Correct, but you don't need to include
%C#1%
[T1_ADHD T1_INT T1_CON T1_IND];
%C#2%
[T1_ADHD T1_INT T1_CON T1_IND];
%C#3%
[T1_ADHD T1_INT T1_CON T1_IND];
because this variation in the means is the default. |
|
|
| Dylan John posted on Monday, June 05, 2017 - 4:57 pm
|
|
|
Thank you!
As an aside, could you briefly explain the process of bootstrap likelihood ratio testing to find the best fitting model? |
|
|
|
|
| The Bollen-Stein 1992 paper is a start - SEMNET is good for this kind of topic. |
|
|
| Back to top |
