Dining table step 1 contains the BIC and G 2 values, and the main coefficients of interest, namely those estimates from Models 1 and 2 involving Whites, Japanese, or Japanese Americans only. All BIC values were much less than 0, which indicated that both models fit better than the saturated model and the basic model with ? r , ? R , and ? rR . The improvement in fit highlights the substantial role that compositional and educational pairing patterns play in explaining the racial and ethnic pairing patterns. For the prewar cohort, Model 1 fit better than Model 2, but for the resettlement cohort, as expected, Model 2 fit better than Model 1. This result suggests that a major gender asymmetry entered the racial pairing patterns of the Japanese or Japanese Us citizens and you may Whites after World War II. eleven) from the design matrix of Model 2, the BIC value increased (not shown). This result indicates that veteran-husband and Japanese-war-bride intermarriages were only partly responsible for the emergence of the gender asymmetry in the level of intermarriage after World War II.
Table 1
Note: SE = important mistake. Coefficient estimates for pairings connected with most other fraction racial teams (Blacks, Latinos, or any other Asians) and other quotes appear regarding the article authors on the demand. To have covariates maybe not found in Table 1, look for Formula step 1.
Because Model 2 fit better than Model 1 for the resettlement cohort, we used Model 2 to calculate all adjusted odds ratios for comparability across cohorts. Table 1 also includes the coefficients of Model 2 obtained from the sample eliminating White veteran-husband and Japanese-wife intermarriages for the resettlement cohort. Additional tests (not shown but available on request) indicated statistically significant between-cohort changes in the patterns of spousal pairing on race or ethnicity involving Japanese or Japanese Americans and Whites (?1, ?step three, ?cuatro, ?eight, ?8, ?11, and ?twelve) in each regional subsample.
Modified Opportunity Ratios
Data 2A (Japanese Americans) and you can 2B (Japanese) expose the fresh new adjusted odds ratios toward a couple of cohorts. Brand new rates was projected by detatching the fresh has an effect on off sociodemographic composition and you may instructional combining. To possess Japanese pairings with Whites about resettlement cohort, the brand new reestimated opportunity rates regarding the attempt excluding the intermarriages from Light veteran husbands and you may Japanese wives are showed ( Shape 2.B ).
Japanese Americans and Whites
New modified chances rates ( Contour 2A ) have been significantly more similar to the heightened exclusion theory than simply with the absorption theory. They indicated that, in the place of a great y, there’s a boost in this new tendency to your homogamy between the prewar cohort in addition to postwar resettlement cohort. A lot more analyses, hence broke up the new resettlement cohort into one or two subcohorts, showed that the new adjusted chance rates for the early (1946–1955) while the late (1956–1966) resettlement cohorts (21,658 and 15,434, respectively) had been greater than chances proportion 
of one’s prewar cohort (10,509). Ergo, even though the premier escalation in the chances proportion is present in the sooner 50 % of brand new resettlement cohort, proof of heightened exception to this rule continued towards latter element of one to cohort. As the achievement don’t transform, we establish right here the outcome from the folded resettlement cohort try to profit from big test size.
Additional analyses including showed that the odds ratios denied and you may changed in the historical models, primarily for the removal of racial compositional has an effect on (? r and you can ? Roentgen ). The team systems out-of Japanese Americans in accordance with Whites when you look at the fraction (computed off Dining table step 1 ) doubled for females (0.15% prewar and you can 0.33% resettlement) and you may tripled for males (0.11% prewar and 0.34% resettlement) amongst the cohorts. That it development led to so much more intermarriages. But fundamental which racial compositional transform, the newest desire into homogamy in fact enhanced. Brand new prewar cohort’s chances ratio into Japanese Western–White pairings became approximately ten,100000 which have adjustments having racial composition and you can remained at that top if together with otherwise excluding the dimensions of area, training, otherwise seasoned condition. Regarding the resettlement cohort, racial structure accounted for about half this new homogamous inclination. Part, rather than knowledge and you will seasoned condition, taken into account much of the rest half. Whenever removing (the brand new aspect off) region out of Model dos, chances ratio (thirty-five,810) substantially increased. When deleting degree as an alternative, the odds ratio (15,413) denied some. For this reason, in the event degree just weren’t managed, the end would have stayed undamaged. When deleting experienced updates regarding Model 2, the odds proportion (17,781) in addition to changed very little.