TABLE 2.
Model fit indices for final latent class models for frequency/conversation subsamplesa
Number of | Log Likelihood | Information Criteria | LMR Likelihood Ratio Test | Entropy | ||||||
---|---|---|---|---|---|---|---|---|---|---|
Caller Type | Conversation | Classes | Free Parameters | AIC | BIC | aBIC | Statistic | p-value | ||
One-Time | Single | 3 | 86 | −508313 | 1016798 | 1017609 | 1017336 | 6258.49 | <0.001 | 0.758 |
Two-Time | First | 3 | 86 | −89028 | 178228 | 178888 | 178615 | 1340.22 | <0.001 | 0.776 |
Second | 3 | 86 | −80983 | 162137 | 162798 | 162524 | 901.11 | <0.001 | 0.753 | |
Three+ Time | First | 3 | 86 | −82756 | 165683 | 166336 | 166063 | 1436.56 | <0.001 | 0.816 |
Last | 3 | 86 | −70912 | 141996 | 142649 | 142375 | 841.43 | 0.038 | 0.732 |
Data were from n=153,514 messages to Crisis Text Line 2013–2017 (n=122,909 individuals); AIC, Akaike information criterion; BIC, Bayesian information criterion; aBIC, adjusted Bayesian information criterion; LMR, Lo-Mendell-Rubin