Skip to main content

Table 2 Best fit and diagnostic criteria for latent class models of one to six class solutions

From: How are different clusters of physical activity, sedentary, sleep, smoking, alcohol, and dietary behaviors associated with cardiometabolic health in older adults? A cross-sectional latent class analysis

Class

LL

AIC

BIC

Adjusted BIC

Entropy

LMR (p-value)

BLRT (p-value)

Smallest class count (n)

Smallest class size (%)

One-class

-2401.0

4832.1

4890.2

4842.6

 N/A

N/A

N/A

355

100

Two-class

-2325.7

4713.5

4833.5

4735.2

0.61

148.9 (< 0.01)

-2401.0 (< 0.01)

167/188

47.0/53.0

Three-class

-2307.6

4709.2

4891.2

4742.1

0.69

35.9 (0.91)

-2325.7 (0.09)

56/199/100

15.8/56.1/28.2

Four-class

-2294.1

4714.2

4958.1

4758.3

0.82

27.1 (1.00)

-2307.8 (0.67)

89/67/143/56

25.1/18.9/40.3/15.8

Five-class

-2279.5

4717.0

5022.9

4772.2

0.76

29.6 (0.80)

-2294.4 (1.00)

135/74/69/41/36

38.0/20.8/19.4/11.5/10.1

Six-class

-2265.6

4721.3

5089.1

4787.7

0.76

40.3 (0.78)

-2286.0 (0.33)

46/81/40/30/77/81

13.0/22.8/11.3/8.4/21.7/22.8

  1. Bold values indicate the value corresponding to the “best” model according to each fit indicator.
  2. LL: Log Likelihood; AIC: Akaike’s Information Criterion; BIC: Bayesian Information Criterion; LMR: Lo-Mendell-Rubin likelihood ratio test; BLRT: Bootstrapped likelihood ratio test.