Message/Author |
|
Kiki Yang posted on Saturday, August 20, 2005 - 4:15 am
|
|
|
Hello! This is my output. I would like to know does the "Pearson Chi-Square Value" of "Chi-Square Test of Model Fit for the Binary and Ordered Categorical (Ordinal) Outcomes" mean overall or other? My English is poor. I hope you can understand my question. TESTS OF MODEL FIT Loglikelihood H0 Value -5736.266 Information Criteria Number of Free Parameters 27 Akaike (AIC) 11526.531 Bayesian (BIC) 11642.102 Sample-Size Adjusted BIC 11556.395 (n* = (n + 2) / 24) Entropy 0.883 Chi-Square Test of Model Fit for the Binary and Ordered Categorical (Ordinal) Outcomes Pearson Chi-Square Value 0.001 Degrees of freedom cannot be computed for this model part. Likelihood Ratio Chi-Square Value 0.001 Degrees of freedom cannot be computed for this model part. |
|
|
These two tests are for the latent class indicators only. They are not for the full model. |
|
Kiki Yang posted on Saturday, August 20, 2005 - 4:55 pm
|
|
|
Thank you for your respond. I want to be sure again. Is the P value test for the latent class indicators or categorical observed variables? |
|
bmuthen posted on Sunday, August 21, 2005 - 1:37 pm
|
|
|
The p value test is for the latent class indicators, which are the categorical observed variables. |
|
Kiki Yang posted on Sunday, August 21, 2005 - 4:23 pm
|
|
|
Thank you very much for your help. |
|
|
Hello everyone, I'm trying to run a latent class analysis with four continuous variables and two categorical variables. I keep getting an error stating that the degrees of freedom cannot be calculated. As a result, I do not have enough information to assess the model fit. Do you have any suggestions? Here is a sample of the relevant output: MODEL FIT INFORMATION Number of Free Parameters 24 Loglikelihood H0 Value -985.151 H0 Scaling Correction Factor 1.123 for MLR Information Criteria Akaike (AIC) 2018.301 Bayesian (BIC) 2098.053 Sample-Size Adjusted BIC 2022.013 (n* = (n + 2) / 24) Chi-Square Test of Model Fit for the Binary and Ordered Categorical (Ordinal) Outcomes Pearson Chi-Square Value 1.455 Degrees of freedom cannot be computed for this model part. Likelihood Ratio Chi-Square Value 1.322 Degrees of freedom cannot be computed for this model part. |
|
|
These tests are not overall tests of model fit. They test only the observed versus expected frequency tables for the categorical latent class indicators. Absolute fit statistics are not available with categorical variables and maximum likelihood estimation. Nested models can be compared using -2 times the loglikelihood difference which is distributed as chi-square. |
|
CB posted on Thursday, March 05, 2015 - 11:52 am
|
|
|
Hello, I'm running a latent class analysis with 4 categorical indicators and I too keep getting an error stating that degrees of freedom cannot be computed for Pearson chi-square and Likelihood Ratio chi-square. If these aren't overall tests of model fit, is it an issue if degrees of freedom cannot be computed? Is it because I'm using categorical indicators that the df cannot be computed? Any insight on this is welcome. Thanks!! |
|
|
Perhaps you have other parts to the model than the 4 indicators. |
|
Chenqq posted on Wednesday, February 28, 2018 - 6:06 am
|
|
|
Dear friends, Who can help me explain this: "Statistical analyses should be performed independently on each of these five plausible values and results should be aggregated to obtain the final estimates of the statistics and their respective standard errors. " In other words, could you povide me with detailed procedures in Mplus to deal with the five pvs of PISA? Thanks in advance! |
|
|
Use the Type=Imputation option in the Data command. See the V8 UG on our website, pp. 572-573. |
|
Chenqq posted on Friday, March 02, 2018 - 4:41 pm
|
|
|
Thank you so much for your response,Bengt O. Muthen. |
|
Gizem Samdan posted on Wednesday, November 06, 2019 - 6:43 am
|
|
|
I am running a LCA. 1.What does it mean if the p Value of the Pearson Chi-Square and/or the likelihood ratio square is 1.000?They are not significant so it means that the model is doing a right classification but 1 cannot be correct either, is it? 2. And what if the values differ like for example Pearson Chi Square Value is 786 and Likelihood Ratio Chi-Square is 308? How can I interpret these issues? Thank you in advance! |
|
|
1. That's usually when Pearson and Likelihood ratio chi-2's disagree which is due to too many zero cells - so that neither is dependable and both should be ignored. 2. Ignore both because they disagree too much (for the same reason as above). |
|
|
Thank you for your answer. I have one more question : So since I cannot interpret the chi square and the likelihood ratio chi square for the model fit of my LCA, I have been looking for other fit indices of relative model fit like: BIC, AIC, Entropy,Lo-Mendell-Rubin adjusted likelihood ratio test and the bootstrap likelihood ratio. Furthermore, the classes of the output can be interpreted very well and I have also checked the standardized bivariate and univariate residuals in the tech10 output- none of these are over 1.96. So taking all together I think these are enough indices to say that the model shows a good fit. Would you agree to this assumption ? |
|
|
Yes, I would agree with that. |
|
Back to top |