3) Methods

To obtain initial data, cycle 5 of the National Survey of Family Growth (NSFG95) was used. The survey revealed the number of differentiated female groups aged between 15 and 44. This was followed by collecting couple-level data from respondents to generate cross -- classification of current spouses, cohabiting partners, and dating partners. By means of race/ethnicity and religion. Educational attainment was divided into five categories or levels of education in order to cross reference partners. Racial/ethnic groups were divided into non-Hispanic black, Hispanic, and non-Hispanic white. Asians, American Indians, Eskimos and Aleuts were omitted from the study because of small population sizes. Religion was divided into Catholic, Protestant, Other, and None. Quasi-independence and quasi-crossing parameter models were used to cross-classify partner types.

Racial and religious homogamy for example increased only slightly across the relationship types. In terms of educational attainment, women tended to marry men with a higher education level. In general, however, it was found that substantial homogamy persisted throughout the relationship levels.
5) Conclusions/Implications

The implication is that dating and cohabitation are not used as a filtering process towards greater homogamy in terms of race, religion or education. The data across the three relationship types are remarkably similar.

6) Personal Observations

It is very important to conduct studies in romantic relationships, especially in the light of socially devastating phenomena such as divorce and the resulting effect on children.