Skip to main content

Table 1 Model fit statistics for cubic latent class growth analysis one to six class solutions

From: Associations between developmental timing of child abuse and conduct problem trajectories in a UK birth cohort

Fit statistics

1 class

2 classes

3 classes

4 classes

5 classes

6 classes

LL (No. of Para.)

−32,500.791 (6)

−29,037.350 (11)

−28,581.588 (16)

− 28,484.751 (21)

−28,388.525 (26)

− 28,354.757 (31)

BIC

65,057.220

58,176.705

57,311.546

57,164.237

57,018.152

56,996.981

SSABIC

65,038.153

58,141.749

57,260.700

57,097.502

56,935.527

56,898.467

Entropy

 

0.763

0.725

0.709

0.688

0.697

LMR-LRT p-value

 

0.0000

0.0000

0.0403

0.0016

0.1998

Adj. LMR-LRT p-value

 

0.0000

0.0000

0.0426

0.0018

0.2061

BLRT p-value

 

0.0000

0.0000

0.0000

0.0000

0.0000

Group size (%)a

 C1

10,648

8598 (80.7%)

447 (4.2%)

480 (4.5%)

135 (1.3%)

1566 (14.7%)

 C2

 

2050 (19.3%)

2343 (22.0%)

1643 (15.4%)

465 (4.3%)

7691 (72.2%)

 C3

  

7858 (73.8%)

8019 (75.3%)

1616 (15.2%)

84 (0.8%)

 C4

   

506 (4.8%)

7720 (72.5%)

167 (1.6%)

 C5

    

712 (6.7%)

597 (5.6%)

 C6

     

543 (5.1%)

  1. Note. Based on N = 10,648, LL Log-Likelihood value, No. of Para. Number of estimated (freed) parameters, BIC Bayesian Information Criterion, SSABIC Sample Size Adjusted BIC, LMR-LRT Lo-Mendell-Rubin Likelihood Ratio Test, Adj. LMR-LRT Adjusted LMRT-LRT, BLRT Bootstrapped Likelihood Ratio Test, C Class. a Based on most likely latent class membership