Question for chi-square test model fit PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
Message/Author
 Jungeun Lee posted on Thursday, February 14, 2008 - 10:21 am
Hello,

I am running a growth mixture model with count variables. The model ran fine but I got the following warning as to chi-square test model fit.

** Of the 0 cells in the latent class indicator table, 14 were deleted in the calculation of chi-square due to extreme values.

Could you help me understanding ;1)What the warning means; 2) the implication of having such warning; 3) what I can do about it?

Thank you in advance!
 Linda K. Muthen posted on Thursday, February 14, 2008 - 2:32 pm
I would need to see the full output to answer that. Please send it along with your license number to support@statmodel.com.
 Susan E. Collins posted on Tuesday, September 01, 2009 - 12:40 am
Dear Linda,

I had a similar warning:

"** Of the 12210 cells in the latent class indicator table, 78
were deleted in the calculation of chi-square due to extreme values.

** Large values were truncated at 9."

I am also unsure how concerned I should be. I am ok with the truncation...any thoughts?

thanks so much in advance!
Susan
 Linda K. Muthen posted on Tuesday, September 01, 2009 - 8:35 am
When that many cells are deleted, you should ignore the test.
 Susan E. Collins posted on Tuesday, September 01, 2009 - 12:57 pm
Hi Linda,

Thanks for the quick response.

Do you mean "ignore the test" b/c I need to fix my "extreme values" first and then rerun or do you mean "ignore the test" such that i can ignore the chisq but accept the rest of the outcomes? Do I scrap the analysis or keep working on them or accept them as they are?

I guess I just don't understand what this means. could you please clarify?

Thanks much!
Susan
 Linda K. Muthen posted on Tuesday, September 01, 2009 - 1:46 pm
I mean when there are many cells deleted, you should ignore the chi-square test. The rest of the results are fine unless there is another message you are not mentioning. Questions like this are best sent to support with a full output to obtain a full answer.
 Brianna H posted on Friday, March 21, 2014 - 9:44 am
Hello. I am running a Poisson linear growth model with one time-invarying predictor of the slope and three time-invarying predictors of the intercept. The output I requested is "patterns residual standardized TECH1 TECH8," but the Chi-square is not printed in the Model Fit information. The only two fit statistics are the Loglikelihood, AIC, BIC, and adjusted BIC. However, when I specify that one of the predictors of intercept should covary with the other two predictors, I do get a chi-square in the output. Can you please help me understand why this is happening? Thank you.
 Linda K. Muthen posted on Friday, March 21, 2014 - 9:50 am
Please send the two outputs and your license number to support@statmodel.com.
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: