Skip to main content
. 2019 Jan 17;14(1):e0210877. doi: 10.1371/journal.pone.0210877

Table 4. Model comparisons for 5 latent class estimations using 100 parametric bootstrapped likelihood ratio tests.

Parametric boostrapped likelihood ratio test (2 times the Loglikelihood difference) p
Number of latent classes
2 versus 1 257.60 <0.001
3 versus 2 144.49 <0.001
4 versus 3 84.36 <0.001
5 versus 4 60.27 <0.001