Table 2.

Results from latent class analysis for establishing three racialized classes among African Americans in two urban midwestern cities (N=306).

Model (Number of latent classes)*	Loglikelihood	Best H0 replicated (Yes/No)	# of parameters	AIC		BIC	SSABIC	LMR-LRT (p)	Entropy
1	−792.4	yes	5	1594.7		1613.4	1597.5	--	--
2	−655.8	yes	11	1353.5		1384.5	1359.6	<.0001	.87
3	−627.7	yes	17	1289.4		1352.7	1298.8	<.0001	.94
4	−617.6	yes	23	1281.2		1366.9	1293.9	.0012	.95

Note: SSABIC = sample size adjusted BIC; LMR-LRT = Vuong-Lo-Mendal-Rubin Likelihood Ratio Test comparing current model with a model with one less latent profile.

^*Even though the measures of model fit suggest that a 4-class solution was better than a 3-class solution, we selected the 3-class model as our best model because in the 4-class model two of the classes contained less than 10% of the sample, thus making that model a less desirable solution.