Skip to main content

Table 2 Model Selection Criteria of the Growth Mixture Model (GMM) analysis

From: Physical activity and sedentary behavior trajectories and their associations with quality of life, disability, and all-cause mortality

 

Growth mixture model (GMM)

Likelihood Ratio Test p-value

Free parameters

LL

AIC

BIC

aBIC

Classes-size (%)

Entropy

Bootstrapped

Vuong-Lo-Mendell-Rubin

Lo-Mendell-Rubin

Physical activity

1-class

14

−23,840.52

47,709.05

47,791.9

47,747.4

100

    

2-classes

13

− 2758.87

5543.74

5620.63

5579.33

63,37

0.98

< 0.01

< 0.01

< 0.01

3-classes

21

− 1682.76

3407.52

3531.72

3464.99

37,20,43

0.99

< 0.01

1.00

1.00

4-classes

29

− 1909.13

3876.26

4047.78

3955.64

0,59,28,13

0.99

1.00

0.34

0.34

Sedentary behavior

1-class

16

−31,918.59

63,869.17

63,963.01

63,912.17

100

    

2-classes

14

−13,259.06

26,546.12

26,628.21

26,583.73

91,9

0.92

< 0.01

0.14

0.14

3-classes

22

−13,113.08

26,250.16

26,320.54

26,282.41

90,2,8

0.93

< 0.01

0.01

0.01

4-classes

30

−13,063.23

26,186.45

26,362.36

26,267.04

0,86,3,11

0.91

1.00

0.53

0.53

  1. Notes: LL Log Likelihood; AIC Akaike information criteria; BIC Bayes Information criteria; aBIC sample-size adjusted BIC