Skip to main content
. Author manuscript; available in PMC: 2024 Apr 1.
Published in final edited form as: Prev Sci. 2021 Jul 7;24(3):505–516. doi: 10.1007/s11121-021-01262-3

Table 1.

Comparison of fit measures for models with a different number of classes

Full data (N = 90)
Outliers removed (N = 85)
Measure 1 class 2 classes 3 classes 1 class 2 classes 3 classes
Log likelihood −403.57 −341.51 −312.67 −333.26 −280.93 −254.98
Relative entropy .732 .756 .786 .795
BIC 866 805 810 724 682 692
SABIC 825 719 680 683 597 563
DBIC 842 755 734 700 632 617
HQ-AIC 846 764 749 699 642 633
HT-AIC 839 764 784 705 645 678
CLC 716 674 587 548

Lower values of information criteria indicate better fit. Relative entropy and CLC require multiple classes to be computed and are undefined for the 1-class model

BIC Bayesian Information Criteria, SABIC Sample Size Adjusted BIC, DBIC Draper BIC, HQ-AIC Hannan-Quinn Akaike Information Criteria, HT-AIC Hurvich-Tsai Akaike Information Criteria, CLC Classification Likelihood Criteria