Skip to main content

Table 2 Information criteria for models of k classes and Lo-Mendell-Rubin test for improvement between models with k and k-1 classes

From: Uncovering steroidopathy in women with autism: a latent class analysis

k classes df -log likelihood AIC BIC LMR test
P value
1 8 -6,424.359 12,878.718 12,949.54  
2 21 -2,906.206 5,862.413 5,977.781 0.0000
3 30 -2,869.814 5,817.628 5,997.602 0.6174
  1. The lowest values on the information criteria indicate the best model (shown in bold), weighing absolute model fit and number of estimated parameters. A description of the two and three class models can be found in Additional file1: Table S1. AIC, Akaike Information Criterion; BIC, Bayesian Information Criterion; df, degrees of freedom; LMR, Lo-Mendell-Rubin. P value in bold indicates significance P < 0.05).