Household rosters, marital and fertility histories
Though it is not currently possible to map the full variety of relationships, nor the pathways to each, there are limited indicators that may give clues to both. There are two primary sources. First, abundant information about household relationships, marital, and fertility histories now exists in most large national surveys. These allow us to determine whether an individual is living with a romantic partner in a cohabiting or marital relationship. Related fertility histories permit an estimate of the relationship between mother and father in longitudinal surveys. Basic demographic household rosters and fertility histories, in short, have furthered our understanding of the role cohabitation and unmarried partnerships play as a pathways or alternative to marriage.
The availability of such information has made it possible to conduct extensive work on the pathways out of cohabiting unions, (into marriage or breaking up) and on fertility in cohabiting unions. Pathways into or among cohabitating unions are poorly understood, and information on this is typically gained from static comparisons of cohabiting and non-cohabiting individuals on basic demographic characteristics (education, race, age, etc.).(29)
The obvious limitation of any enumeration of household members, or fertility histories, is that they will almost always miss any relationship or person that is not co-residential.(30) Only those relationships that exist as cohabiting or married couple situations will be available for study based on this strategy. In the absence of new longitudinal evidence, this is a serious limitation because it prevents us from knowing the precursors to cohabitation or marriage, as they currently exist. In short, we cannot adequately study the pathways to co-residential unions in this fashion because there is little, if any, information about the other person until he or she is co-residing with the respondent.
The most valuable resource for studying union formation, at present, is probably the National Survey of Families and Households.(31) This longitudinal, national survey is now in its third wave, having been launched in 1987. The NSFH includes detailed information on union transitions (into and out of cohabiting and marital unions) as well as some attitudinal information about the advantages or disadvantages of cohabitation or marriage. The` longitudinal design permits an analysis of transitions into and out of relationships, and considerable work has already been done on transitions into and out of cohabiting unions.(32) Moreover, the broad range of respondents included would permit an analysis of union formation among middle aged and elderly individuals. NSFF, as with most surveys, gathers scant information about romantic partners who are not yet co-resident.
Studies of high school students, college females, and unmarried mothers
The other primary source of information comes from surveys and qualitative studies specifically designed to consider youthful attitudes and behaviors as they relate to relationships and marriage. Studies of high-school age students, for example, have shown how teenagers feel about marriage and divorce every year since 1975.(33) At least one national study of female college students has been conducted to assess the current dating customs on American college campuses.(34) And intensive research has focused on unmarried women who have just given birth.(35) Other national surveys of high school students exist, though they are limited in the information gathered about romantic and intimate relationships or the dynamics of such relationships with one obvious exception.(36)
The National Longitudinal Study of Adolescent Health (Add Health) focuses on an in-school cohort of 7th to 12th graders in 1994 studied through 2001-02.(37) Data were collected from adolescent respondents on best friends, romantic partners, and sexual partners. The sampling design generates many pairings for which both participants are respondents. This allows for the analysis of the process of pair formation and dissolution. Especially valuable is a list of relationship events or characteristics (e.g., where you met, age and race of partner, whether exchanged gifts, whether told others you were a couple, types of sexual expression, etc.) that can be sequenced (allowing one to determine the order of events in relationships). A sample of 1,507 partners of respondents were interviewed at Wave III. The sample consists of one-third married, one-third cohabiting, and one-third dating partners. The available data are for respondents seven or eight years after the first interview - age 18 to 26. As such, a growing number are now approaching the typical ages for transitions into serious romantic (cohabiting or married) relationships. The third wave of Add Health focuses primarily on the pattern of, attitudes about, and influences on transitions. Future waves will be especially valuable for understanding pathways to union formation among young adults.
The indicators included in Add Health serve as a model for how trajectories to union formation might be studied among youths and adults. To do so would require a different sampling design (to capture adults in all stages of the life course) and other significant modifications of administration and follow-up. Still, in my opinion, this is the best model we currently have for how to study pathways to union formation because it explicitly includes romantic and non-residential partners.