only marginally significantp = 0.021
However, the Lo–Mendell–Rubin likelihood ratio test (LMR-LRT) showed that the 2-class latent profile model significantly improved over the 1-class model ( p < 0.001), the 3-class model significantly improved over the 2-class model ( p < 0.001), but the 4-class model was only marginally significant over the 3-class model ( p = 0.021), and the 5-class model was not significant over the 4-class model ( p = 0.131).