Duke University Press

Table 6

Goodness-of-fit statistics for multiracial self-identification models

	Gender + Controls	+ Regime	+ Generation	+ Gender × Generation	+ Gender × Generation × Regime
Log-Likelihood	−2,417	−2,282	−2,268	−2,263	−2,250
Akaike Information Criterion	4,852	4,590	4,564	4,556	4,553
Bayesian Information Criterion	4,910	4,673	4,653	4,651	4,726
Number of Variables	8	12	13	14	26

	Gender + Controls	+ Regime	+ Generation	+ Gender × Generation	+ Gender × Generation × Regime
Log-Likelihood	−2,417	−2,282	−2,268	−2,263	−2,250
Akaike Information Criterion	4,852	4,590	4,564	4,556	4,553
Bayesian Information Criterion	4,910	4,673	4,653	4,651	4,726
Number of Variables	8	12	13	14	26

Note: N (observations) = 4,330. See Table 4 (Models 2a and 2b) and Table 5 for detailed model estimates.