Skip to main content
. 2023 Jul;223(1):309–318. doi: 10.1192/bjp.2022.192

Table 2.

Latent class mixed model fit of the three SANS dimension trajectories

Number of classes Number of parameters Fit statistics % of the sample in each class
AIC BIC aBIC Entropy Class 1 Class 2 Class 3 Class 4
Expressivity
1 4 8476.66 8492.15 8479.46 1.00 100.00%
2 8 8261.06 8292.04 8266.66 0.90 89.30% 10.70%
3a 12 8027.25 8073.72 8035.65 0.90 79.44% 10.70% 9.86%
4 16 8035.25 8097.20 8046.45 0.63 78.31% 10.99% 10.70% 0.01%
Experiential
1 4 9711.55 9727.04 9714.35 1 100%
2 8 9630.94 9661.92 9636.54 0.56 80.28% 19.71%
3a 12 9550.71 9597.18 9559.11 0.75 72.11% 15.77% 12.11%
4 16 9558.71 9620.67 9569.91 0.48 70.74% 16.62% 12.68% 0.01%
Attention
1 4 7144.43 7159.94 7147.25 1 100%
2 8 6977.83 7008.85 6983.47 0.72 78.15% 21.85%
3a 12 6912.45 6958.99 6920.92 0.85 77.46% 19.89% 3.08%
4 16 6920.45 6982.50 6931.74 0.80 76.50% 19.89% 3.36% 0.01%

SANS, Scale for the Assessment of Negative Symptoms; AIC, Akaike information criterion; BIC, Bayesian information criterion; aBIC, sample size-adjusted Bayesian information criterion.

a.The preferred model.