Message/Author |
|
Anonymous posted on Wednesday, April 27, 2005 - 6:09 am
|
|
|
Just a quick question. I am working on developing a monte carlo CFA simulation that requires Likert type data (e.g., 1=strongly disagree, 2=disagree, 3=agree, 4=strongly agree). This is not a situation where I have the data before hand, so I will have to simulate the data from scratch. In other words, I want to restrict the data generation to numbers between 1-4. How can I tell Mplus that I want 6 independent variables that have the Likert type data? I understand the code for the 6 independent variables, but not the restriction of the data between 1-4. Please, would you give me an example of this restriction using the Mplus code? Thank you, in advance, so much for the help. |
|
bmuthen posted on Wednesday, April 27, 2005 - 6:27 am
|
|
|
You use a setup in line with ex 5.10, which exists in a Monte Carlo version on the Mplus CD. Here, you specify that the outcome is categorical with a certain number of thresholds (#thresholds = #categories minus 1). The theshold values give the probabilities of being in each category. |
|
hamy temkit posted on Monday, September 13, 2010 - 3:38 pm
|
|
|
Hi, I appreciate any input you could have on this. I would like to run a monte carlo power multigroup analysis, using the cutpoints option to define my grouping variable and would like to, not recourse to generating external files to do so but rather run basically the simulation in one step. Here is a snippet of my code. I am getting an error message regarding grouping option not being recognized. I am not surprised since we are not using the DATA option. Is there a way to tweek the code in order to do this. I have an Mplus License and can send you my code if needed. Thanks. TITLE: MONTE CARLO POWER MONTECARLO: names are x y1 y2 y3 y4; cutpoints = x(0) ; nobs = 2000; nreps = 1000; seed = 2224; ngroups = 2; grouping is x (0=G1 1=G2); Model population: Model population-G1: Model population-G2 Model : Model-G1: Model-G2: OUTPUT: TECH1 TECH9; |
|
|
See mcex5.15.inp which comes on the Mplus CD and is available on the website. A grouping variable is not needed. |
|
hamy temkit posted on Tuesday, September 14, 2010 - 11:29 am
|
|
|
Hi, Thank you so much for the prompt reply. I think this might work for my example. Thanks. |
|
Eric Teman posted on Monday, January 16, 2012 - 12:59 pm
|
|
|
When using the Monte Carlo command, is there a way to save the population covariance matrix as a data file? |
|
|
No, that is not an option for the MONTECARLO command. |
|
Eric Teman posted on Monday, January 16, 2012 - 2:37 pm
|
|
|
How would I go about generating a population matrix based on certain specifications without having to use an input data file? |
|
|
See Slides 114-116 of the Topic 4 course handout which can be found on the website. |
|
Hyahn posted on Thursday, October 29, 2015 - 1:45 am
|
|
|
Dear, Muthen. I am trying to generate MAR missing mechanism with MONTECARLO simulation. Here's the part of the commands of ex12.4 and ex12.9. MODEL MISSING: [y1-y4@-1]; y1 on x*.4; .. MODEL MISSING: [y1-y4@15]; y1 on u1@-30; I'd like to ask about the meaning of these two lines, y1 on x*.4; y1 on u1@-30; What does the value(.4, -30) mean? How can I set the values for generating MAR pattern? |
|
|
The line y1 on x*.4; says that as x increases the logit for missing on y1 increases 0.4. So just like regular logistic regression. The line y1 on u1@-30; means that when the binary predictor u1 = 1 the logit for missing on y1 is -30, that is, the probability is very small. In both cases, the logit and the probability are influenced also by the intercept of the logistic regression. |
|
Back to top |