Skip to main content

Table 1 Model fit statistics for each of the fitted Latent Profile Analysis (LPA) models

From: Associations between data-driven lifestyle profiles and cognitive function in the AusDiab study

Statistic

1 profile

2 profiles

3 profiles

4 profiles

5 profiles

6 profiles

Log-likelihood

-40613.06

-39871.65

-39498.89

-39316.81

-39158.07

-39082.47

FP

10

18

26

34

42

50

AIC

81246.11

79894.90

79049.78

78701.61

78400.14

78264.93

BIC

81310.34

79894.90

79216.76

78919.98

78669.88

78586.05

adjusted BIC

81278.56

79837.70

79134.14

78811.94

78536.42

78427.17

Entropy

N/A

0.90

0.88

0.86

0.88

0.77

aLMR-LRT

N/A

1461.13, P < 0.0001

734.62, P < 0.0001

358.84, P < 0.0001

312.83, P < 0.0001

149.00, P = 0.139

BLRT

N/A

P < 0.0001

P < 0.0001

P < 0.0001

P < 0.0001

P < 0.0001

Class proportions

100%

88.5%, 11.5%

76.3%, 18.7%, 5%

62.6%, 23.9%, 9.5%, 4.0%

59.1%, 24.1%, 10.4%, 3.8%, 2.6%

33.3%, 25.0%, 25.0%, 10.4%, 3.8%, 2.6%

  1. AIC Akaike Information Criterion, BIC Bayesian Information Criterion, BLRT Bootstrapped-Likelihood Ratio Test, FP Free Parameters, LMR Lo-Mendell-Rubin, LRT likelihood ratio test
  2. aAdjusted Lo-Mendell-Rubin likelihood ratio test for k versus k-1 profiles. Values are two times the loglikelihood difference and corresponding p-value