Declining Racial Stratification in Marriage Choices? Trends in Black/White Status Exchange in the United States,1980 to 2010

Abstract

The status exchange hypothesis suggests that partners in black/white marriages in the United States trade racial for educational status, indicating strong hierarchical barriers between racial groups. The authors examine trends in status exchange in black/white marriages and cohabitations between 1980 and 2010, a period during which these unions increased from 0.3 percent to 1.5 percent of all young couples. The authors find that status exchange between black men and white women did not decline among either marriages or cohabitations, even as interracial unions became more prevalent. The authors also distinguish two factors driving exchange: (1) the growing probability of marrying a white person as educational attainment increases for both blacks and whites (educational boundaries) and (2) a direct trade of race-by-education between partners (dyadic exchange). Although the theoretical interpretation of exchange has focused on the latter factor, the authors show that status exchange largely emerges from the former.

Keywords

marriage racial stratification African Americans whites cohabitation

Patterns of union formation provide information about social boundaries. If a boundary is socially relevant, most marriages will occur within groups defined by it, and growing intermarriage across a boundary will signal its declining relevance. Attributes such as religion and family background were important boundaries to marriage in the past in the United States, but they have been replaced by education, in a trend from ascription to achievement as the basis for spousal selection (Kalmijn 1991a, 1991b). However, one ascribed attribute continues to have enormous relevance in the American marriage market: the boundary between blacks and whites. Although the number of racial and ethnic intermarriages has grown since the 1970s, and black/white marriages have experienced the largest increase in relative terms (Qian and Lichter 2011; Wang 2012), the odds of a white/black union are still much lower than intermarriage between whites and any other ethnoracial group (Qian and Lichter 2007, 2011; Rosenfeld 2008).

The low prevalence of black/white couples may signal not just homophily—a nonhierarchical preference for fellow group members—but rather a persistent racial status hierarchy. The status exchange hypothesis is predicated on such a hierarchy. The hypothesis proposes that in black/white marriages, partners trade racial status (whiteness, in the U.S. context) for socioeconomic status such that the union is beneficial for both partners (Davis 1941; Merton 1941). Status exchange has been described as a “second-order” kind of homogamy in which “partners with equivalent resources need not be similar with respect to all characteristics relevant to marriage choice; their overall equivalence could result from a balance of pluses and minuses in different areas” (Schoen and Wooldredge 1989). Given a gender-based division of labor and gendered expectations, the most common form of exchange in the United States would be one in which a black man exchanges his educational or economic status for the racial status of a white woman, and interracial marriages formed by a black man and a white woman comprise the majority of black/white couples.

Status exchange will exist only if whiteness carries a hierarchical advantage in a particular society. As put by Kalmijn (2010), status exchange in racially mixed couples “can be regarded as evidence that in the marriage market—and presumably also in society at large—race is treated as a hierarchical variable” (p. 1252, italics in the original). The relevance of status exchange does not emerge from its prevalence, as black/white marriages constitute only a small fraction of all marriages in the United States. Rather, status exchange provides social scientists with a unique window into the persistence of racial stratification in union formation.

The analysis of status exchange offers important insights into trends of racial stratification and intermarriage in the United States, especially over the past few decades. Black/white unions were banned in most U.S. states throughout the first half of the twentieth century, and not surprisingly, the odds of intermarriage were extremely low. Between 1948 and 1965, 14 northern and western states repealed antimiscegenation laws, but only after the Supreme Court’s 1967 decision in Loving v. Virginia did 16 remaining southern states follow suit (Fryer 2007:74). Black/white unions have increased dramatically in the post–civil rights era. According to our analysis of the decennial census and the American Community Survey (ACS), the proportion of all married “young couples” (in which the male partner is 25–34 years old) that included a black and white spouse was only 0.34 percent in 1980. This proportion had increased to 0.56 percent by 1990, 1.09 percent by 2000, and 1.51 percent by 2010. Although proportion is still small, this trend reflects a massive change in the number and visibility of black/white marriages. Even if the racial boundary between whites and blacks is still the most powerful division in the American marriage market (Rosenfeld 2008), intermarriage is much more prevalent now than in past decades. Growing racial intermarriage has not, however, been accompanied by a decline in socioeconomic gaps between blacks and whites. The black/white gap in economic advantage declined in the 1960s but was halted in the 1970s and even reversed thereafter (Morris and Western 1999), resulting in persistent racial disparities in schooling and earnings (DeNavas-Walt, Proctor, and Smith 2012, Figure 1; Fischer and Hout 2006; McDaniel et al. 2011).

Figure 1.

Percentage of black and white individuals 25 to 34 years of age who were married, cohabiting, or not in a union, 1980 to 2010.

There are pronounced gender differences in black/white unions. Since the 1960s, about two thirds of black/white marriages have been formed by a black man and a white woman. Furthermore, black women are particularly isolated in the marriage market (Lichter and Qian 2004) and perceived as less desirable than other groups by those looking for romantic partners (Feliciano, Robnett, and Komaie 2009; Lin and Lundquist 2010). Theories explaining gender asymmetry in black/white intermarriage include the traditional gender role of men in initiating courtship, the suggestion that white men may suffer a higher labor market penalty for interracial marriage than black men, and, particularly, the possibility that education and economic potential may be a less valuable asset for women than men in the marriage market, given traditional gender-based norms and expectations (Gullickson 2006a). However, the value of education among potential female partners has changed in the recent past as the gender gap in educational attainment increasingly favors women. Among blacks, the gender gap has favored women for a long time and has grown in the recent past. Among whites, women’s college completion surpassed men’s in the 1980s, and female advantage has widened ever since (McDaniel et al. 2011). The gender gap favoring men in employment and earnings has declined markedly over the past four decades, even if gender convergence has slowed recently (Cotter, Hermsen, and Vanneman 2015; Goldin 2006). These trends may alter the gender composition of black/white unions toward a more balanced gender ratio and may extend status exchange to interracial couples formed by a black woman and a white man.

The increase in black/white intermarriage was accompanied by a substantial attitudinal change. In 1963, 62 percent of white Americans supported a legal ban on intermarriage (Schuman et al. 1997). By 2002, this percentage had dropped to 11 percent (Smith et al. 2013). Furthermore, according to the General Social Survey, opposition to a close relative marrying a black person declined from 65 percent of all white respondents in 1990 to about 20 percent in 2010. Notably, the reverse opposition to marrying a white person has been consistently low among blacks, hovering around 9 percent over the period of observation. This trend suggests that the resistance of intermarriage emerges largely from whites. Although white resistance has decreased in recent decades, whites still reject blacks as potential spouses for family members at higher rates than they reject any other ethnoracial group (Wang 2012).

In sum, both the acceptance and prevalence of black/white intermarriage has increased in American society, suggesting a decline in the strong hierarchical racial barriers that give rise to status exchange (Kalmijn 2010). This leads to our first hypothesis:

Hypothesis 1: Status exchange in black/white unions has declined over time as these unions have become more prevalent and accepted.

What Does Status Exchange Really Capture?

The theory of status exchange suggests that black/white unions would often involve an exchange of status in which the white partner trades a higher racial status for a higher status on another attribute, generally measured as education. Although other measures of status such as earnings and wages have been used, education is our preferred variable because it strongly predicts economic well-being and other important outcomes such as health, psychological well-being, and political participation (Pallas 2000), and because most people complete their education before establishing a union. Alternative measures of status, such as earnings or wages, are much more likely to endogenously change after marriage if individuals adjust their labor market involvement precisely as a result of union formation.

The status exchange hypothesis posits that interracial unions diverge from educational homogamy and include a white partner who is less educated than the black partner, a pattern known as “white educational hypergamy.” Early attempts to test the status exchange theory for black/white intermarriage assumed that the educationally asymmetric couples predicted by the theory—those characterized by white hypergamy—would be the most common form of interracial union. These studies found little support for the theory because educationally homogamous marriages were found to be the most common form of union in black/white intermarriage, just as they are for other types of marriage.

As argued by Gullickson (2006b), this test is excessively strict because most interracial marriages are, similar to their intraracial counterparts, educationally homogamous. A more appropriate test of exchange compares interracial and intraracial couples. Specifically, white partners in interracial marriages are expected to be more educationally hypergamous (marrying up) and less educationally hypogamous (marrying down) than their respective sexes in intraracial unions. Conversely, black partners in interracial unions should be more educationally hypogamous and less educationally hypergamous than their respective sexes in racially endogamous unions. Using such a criterion, several studies have found evidence of status exchange in marriages between black men and white women in the United States but weaker or no indication of exchange in marriages between black women and white men (Gullickson 2006b; Hou and Myles 2013; Kalmijn 1993, 2010; Qian 1997; Schoen and Wooldredge 1980), although one study did not find evidence of exchange in black/white marriages (Rosenfeld 2005) and elicited an interesting debate (Gullickson and Fu 2010; Kalmijn 2010; Rosenfeld 2010).

What the status exchange literature has been slow to realize is that educational asymmetry, in the form of white hypergamy, conflates two distinct factors contributing to exchange. First, in racially stratified societies, both educational status and racial status (whiteness) operate as sources of advantage. If racial status and educational status are at least partially fungible, then individuals with higher education will have better access to white partners, and white individuals will have better access to more educated partners (Fu 2001). As a consequence, both whites and blacks will be more likely to marry whites as their education increases, resulting in race-specific effects of education on interracial marriage: education will increase the likelihood of interracial marriage among blacks but will decrease the likelihood of interracial marriage among whites, a component referred to as “educational boundaries” by Gullickson (2006b) and “market exchange” by Gullickson and Torche (2014).

In addition to “educational boundaries,” there may be a direct trading of racial for educational status in interracial couples, resulting in an asymmetric difference in educational attainment between spouses. This is the notion to which status exchange theory has traditionally referred. White partners in interracial marriages are expected to be more educationally hypergamous and less educationally hypogamous than white partners in intraracial marriages, and black partners in interracial marriages should be more educationally hypogamous and less educationally hypergamous than black partners who marry blacks. This educational asymmetry in interracial marriages was called “dyadic exchange” by Gullickson and Torche (2014) because it refers to a direct trading of resources between partners.

Educational boundary parameters capture the hypothesis that highly educated black men are more likely to marry white women than fellow blacks with lower levels of education and the hypothesis that highly educated white women are less likely to marry black men than fellow white women with lower levels of education. Note that these hypotheses say nothing about the level of education of the potential partner chosen by black men and white women, respectively. If these hypotheses are true, then white wives married to black husbands will be more likely to be educationally hypergamous than white women in racially endogamous marriages. The reason is that highly educated black men are selected into racial intermarriage, and highly educated white women are selected out of intermarriage, increasing the odds of white hypergamy even if no direct exchange of racial for educational status exists.

As in other forms of assortative mating, educational boundaries emerge from a combination of individual preferences and marriage market exposure to potential partners of different races and educational levels (Kalmijn 1998). For example, given persistent racial stratification in U.S. society, white individuals may request more education in a black partner than in a white partner (preferences), and they may also be more likely to be exposed to potential black partners with higher levels education in settings such as neighborhoods and educational institutions (exposure).

It is important to stress that both “educational boundaries” and “dyadic exchange” emerge from the same premise, namely, that racial stratification shapes spousal selection and that whiteness is treated as a source of status advantage in contemporary American society. However, educational boundaries do not require nor expect a direct trading between partners (Gullickson and Torche 2014). They assume only that each person in the marriage market seeks the “best” potential partner in terms of both educational and racial status and that these attributes are to some extent exchangeable, such that a white person will be more likely to marry a black person with high rather than low level of educational attainment if he or she decides to cross the racial boundary in intermarriage. Because the parameter estimates capturing educational boundaries and dyadic exchange partially overlap, if educational boundaries are not properly modeled, they will appear as dyadic exchange, a direct trade between partners. This potential limitation has affected most analyses of exchange in the United States to date (exceptions are Gullickson 2006b and Hou and Myles 2013).

Distinguishing the two forces driving exchange is important because educational boundaries signal a more general phenomenon that applies to all interracial couples, not only to the small proportion of educationally asymmetric couples. Whereas most prior research conflates these dimensions, we separate them in a log-linear analysis of trends to assess their relative relevance in accounting for the phenomenon that the exchange literature has reduced to a direct trading of resources between partners.

Marriages and Cohabitations

To date, almost all studies of status exchange in black/white unions have focused on married couples. However, the proportion of individuals who are married in early adulthood has declined over time, and cohabitations have increased. The “retreat from marriage” among young adults varies across race and education. Figure 1 shows the proportion of white and black individuals 25 to 34 years of age who are married, cohabiting, or not in a union (labeled “single” for simplicity) between 1980 and 2010.¹ Among blacks, marriage declined from 59 percent to 34 percent of those 25 to 34 years old, while singlehood increased from 35 percent to 51 percent and cohabitation grew from 6 percent to 15 percent. The changes were less dramatic but still substantial among whites, with a decline in marriage from 80 percent to 61 percent and increases of singlehood from 16 percent to 24 percent and cohabitations from 4 percent to 15 percent.

Marriages are not a random sample of young adults. Among individuals 25 to 34 years of age, cohabiters have lower levels of schooling than either married or single individuals. Educational differences between married and single individuals are less pronounced. Among whites and black women 25 to 34 years old, single young adults have consistently higher levels of schooling than married ones, suggesting marriage postponement among the highly educated. Among black men, single and married individuals have similar levels of schooling. These patterns remained quite constant between 1980 and 2010 despite generalized educational expansion among all groups (Appendix Table 1 in the online supplemental material gives educational attainment by marital category for blacks and whites between 1980 and 2010).

In sum, decline in marriage rates is evident for all groups among young adults, but individuals with higher levels of education appear to be more likely to postpone marriage, while individuals with lower levels of education are more likely to establish cohabitations. Cohabiters are consistently the most educationally disadvantaged category among both genders and racial groups.²

As cohabitations become more prevalent and offer a “principle of cohesion” possibly different than marriage (Brines and Joyner 1999), the question about status exchange should be extended to this kind of union. If cohabitation is simply an “informal marriage” (Macklin 1987:322, cited in Schoen and Weinick 1993), then trends of status exchange between cohabiting and married unions should compare similarly. Two theoretical perspectives, however, suggest that status exchange may differ between marriages and cohabitations. First, the “specialization and trading” perspective suggests that couples maximize gains to marriage by specializing in realms in which they have a comparative advantage and by trading areas of advantage with their spouses, with men specializing in market work and women specializing in domestic work (Becker 1991). Because cohabitation lacks the legal protections, expectations, and long-term commitment of marriages, the risks associated with specialization are too high among cohabiters, and partners operate on the principle of equality in terms of their economic resources (Brines and Joyner 1999). This would result in more similarity in human capital and earning power among cohabiters compared with married couples. Second, the “winnowing” or “sequential selection” approach sees cohabitation as a stage in the mate selection process, with the transition from cohabitation to marriage marked by increasing selectivity of partners (Blackwell and Lichter 2000, 2004). As a result, cohabiters will be less similar than spouses in every domain.

If, as highlighted by the literature, status exchange is a “second-order” form of homogamy in which minority partners “compensate” for their low racial status offering their higher social status (Schoen and Wooldredge 1989), then the two theoretical approaches predicting differences in assortative mating between married and cohabiting couples have different implications for status exchange. The “specialization and trading” hypothesis predicting stronger homogamy in economic resources among cohabiting couples would also predict a higher degree of status exchange. In contrast, the “winnowing” approach, predicting less homogamy among cohabitations than marriages, would suggest lower levels of exchange.

To the best of our knowledge, only one study exists to date comparing status exchange between cohabitations and marriages (Blackwell and Lichter 2000). That analysis found weaker exchange among cohabiters, consistent with the “winnowing” hypothesis. Following that study, we predict as follows:

Hypothesis 2: Status exchange is weaker among cohabitations than marriages.

Data and Analytical Strategy

We examine trends in status exchange between 1980 and 2010 to address three questions: Has status exchange weakened as black/white marriages have become more prevalent? Is the phenomenon traditionally called exchange a dyadic trading of status that makes the union beneficial for both partners, or is it better described as emerging from the differential association between educational attainment and the probability of racial intermarriage (educational boundaries)? Is status exchange equally strong among married and cohabiting couples, or are there differences signaling dissimilar “principles of cohesion”?

We analyze representative samples of married and cohabiting couples in the United States in 1980, 1990, 2000, and 2008 to 2012 (which we call 2010 for simplicity). We use “long-form” responses from the 1980, 1990, and 2000 censuses from the University of Minnesota Population Center (Integrated Public Use Microdata Series). The long-form respondents represent a 5 percent cross-sectional sample of U.S. households at each time of collection. After 2000, the Census Bureau discontinued the long form and replaced it with the ACS, which samples 1 percent of the U.S. population annually. To approximate a 2010 sample comparable with previous long-form samples, we use a pooled data set that combines and weights the respondents from five consecutive years of ACS collection between 2008 and 2012 (Ruggles et al. 2010). We identify heterosexual marriages and cohabitations and exclude any records in which the responses on age, sex, marital status, or householder relationship by either husband or wife were reallocated by the Census Bureau in postcollection procedures (approximately 3 percent of sampled couples). We identify marriages in our sample by the “spouse” category in the relationship-to-householder question. Cohabitations are identified by matching household heads with their “unmarried partners” in the 1990, 2000, and 2010 data sets. The 1980 data set does not include the unmarried partner category in the relationship-to-householder question. Rather, it classifies individuals as “partner/roommate.” This induces some misclassification error in 1980, as some dyads coded as cohabitations may be roommates.³ We include native-born and foreign-born respondents in the analysis. Some studies restrict their samples to native-born respondents to ensure that unions have been established in the domestic marriage market, but this strategy misses respondents who migrated before adolescence and thus found their partners in the United States. As a robustness check, we restrict the sample to native-born respondents and replicate all analyses; results are unaltered.

Interracial coupling may vary by age at first union, union dissolution, and remarriage. For example, unions that are established later in life or after prior unions have ended may be more likely to cross racial lines (Fu 2010; Porterfield 1982) or racially exogamous unions may be more likely to be dissolved (Bratter and King 2008; Fu and Wolfinger 2011; Heaton 2002). To build a comparable series, we would ideally examine first unions over time. However, no time-series information on union date or parity is currently available. As an alternative, we select couples in which the male partner is 25 to 34 years of age for each year of observation. This choice increases the probability of capturing first unions while ensuring that the large majority of individuals would have completed their educational careers.

For each year, we produce contingency tables that cross-classify male partner’s educational attainment by male partner’s race by female partner’s educational attainment by female partner’s race. Educational attainment is categorized into three levels: (1) a high school degree or less, (2) some college, and (3) a bachelor’s or an advanced degree. Race has two categories: (1) non-Hispanic white and (2) non-Hispanic black. All other racial and ethnic identities are excluded from the analytical sample. Beginning in 2000, the Census Bureau allowed individuals to indicate multiple racial identities. The Integrated Public Use Microdata Series data set includes a constructed single-race variable that preserves comparability before and after this change, which we use in this analysis (Ruggles et al. 2010). Each contingency table contains 36 cells. Data are weighted to represent population distributions. Appendix Table 2 in the online supplemental material offers the weighted frequencies for each of the 36 cells among married and cohabiting couples for all years of observation. These frequencies can be used to replicate all analyses offered here.

Measuring Status Exchange

Following past research, we use log-linear models in the assessment of status exchange. The advantage of log-linear models is that they account for changes in marginal distributions of the variables examined; for example, interracial unions will be a larger proportion of all unions for blacks than for whites simply because whites are a larger proportion of the American population. By the same token, an increase in educational attainment among women over time will result in growing educational hypergamy among men driven simply by a widening educational gap favoring women. By accounting for these differences in marginal distributions, we can examine net associations between race and education (for an intuitive explanation of the advantages of log-linear models to examine assortative mating, see Kalmijn 2010).

Our initial assessment of status exchange uses the hypergamy ratio (Kalmijn 1993, 2010). The hypergamy ratio captures the excess white hypergamy (white spouses “marrying up” in education) in interracial marriages compared with what would be predicted if the pattern of educational assortative mating (EAM) were symmetric between spouses independent of their race. To generate hypergamy ratios, we follow two steps. First, we fit a log-linear model of quasi-symmetry that adjusts for any differences in marginal distribution of education across gender and racial group and that assumes that the remaining association between spouses’ education is symmetric; that is, it does not depend on the race of the spouses. Following Kalmijn (2010), we write the quasi-symmetry model for a table cross-classifying husband’s education (i), wife’s education (j), husband’s race (k), and wife’s race (l) as follows:

\begin{array}{l} \log (F_{i j k l}) = H E_{i} + W E_{j} + H R_{k} + W R_{l} + H E_{i} * \\ H R_{k} + W E_{j} * W R_{l} + (D_{H E} * D_{W E}) + \\ (D_{H E} * D_{W E}) * (H R * W R) + E N + \\ E N * 1 / 2 (H E + W E) \end{array}

The first four terms (HE_i + WE_j + HR_k + WR_l) adjust for the marginal distribution of the variables, and the next two terms (HE_i * HR_k + WE_j * WR_l) account for educational differences across races separately for men and women. The (D_HE * D_WE) terms capture EAM. Each of these parameter estimates captures an odds ratio for the association of each combination of partner’s education levels.⁴ Higher odds ratios capture stronger homogamy (i.e., a lower probability of crossing the barrier between the two educational categories in intermarriage). Note that these EAM terms apply to all couples, regardless of which spouse has each particular level of education, capturing symmetric educational associations. The (D_HE * D_WE) * (HR * WR) interaction terms allow EAM to vary depending on the racial composition of the couple, distinguishing three types of couples: white/white, black/black, and interracial. To address the tendency toward racial endogamy, the model also includes EN, a dummy variable distinguishing interracial from racially endogamous marriages. Because prior literature has shown an association between educational attainment and racial endogamy (Kalmijn 1993; Qian 1997), we include an interaction term between racial endogamy and the “average” educational level of the couple, EN * 1/2(HE + WE). By averaging education across both spouses, this term captures variation in endogamy across levels of schooling while retaining symmetry within the model.

The essential characteristic of this model is that it assumes that EAM is symmetric regardless of the racial composition of the couple, with no excess white educational hypergamy in interracial couples. We then predict cell counts expected from this symmetric model. Using these predicted counts, we divide the predicted number of individuals in each racial group marrying up in education (hypergamy) by the predicted number of individuals marrying down in education (hypogamy) for each racial combination of the couple (white/white, white/black, black/white, and black/black). In the case of interracial marriages, this quotient is obtained for the white spouse. For racially endogamous couples, the quotient is obtained for the male partner. We repeat these steps for the observed cells and then divide the quotient based on observed frequencies by the quotient based on frequencies predicted by the symmetric model to produce the hypergamy ratio. The hypergamy ratio therefore captures any excess observed white educational hypergamy in interracial couples (and excess male partner’s hypergamy in intraracial couples) compared with what would be predicted by a model assuming symmetric EAM (Kalmijn 2010). A value of 1 for the hypergamy ratio indicates that EAM is symmetric, and no exchange is present. Values larger than 1 in interracial unions indicate that white individuals married to black spouses will be more educationally hypergamous than expected by a symmetric pattern of EAM, supporting the status exchange hypothesis. Given the traditional gender division of labor, white hypergamy is expected to be more pronounced when the black spouse is male.

Distinguishing Components of Exchange: Educational Boundaries and Dyadic Exchange

The hypergamy ratio provides a parsimonious account of status exchange using a single coefficient and facilitates evaluating change over time. However, it does not yield tests of statistical significance, and it conflates two distinct forces potentially contributed to exchange: educational boundaries and dyadic exchange. Following Gullickson (2006b) and Gullickson and Torche (2014), we distinguish these components using a log-linear formulation. For a table cross-classifying male partner’s education (i), female partner’s education (j), male partner’s race (k), and female partner’s race (l), the log-linear model capturing educational boundaries and dyadic exchange is expressed as

\begin{array}{l} \log (F_{i j k l}) = H E_{i} + W E_{j} + H R_{k} + W R_{l} + \\ (H E_{i} * H R_{k}) + (W E_{j} * W R_{l}) + (H E_{i} * W E_{j}) + \\ (H R_{k} * W R_{l}) \\ + \sum_{p = 1}^{z} η_{p} u_{p i j} + \sum_{q = 1}^{z} δ_{q} w_{q i j} + τ (x_{i j k l}) + γ (y_{i j k l}) \end{array}

The model accounts for the marginal distribution of either spouse’s race and education, different marginal distributions of education by husband’s race (HE_i * HR_k) and wife’s race (WE_j * WR_l), a general pattern of EAM (HE_i * WE_j), and a general pattern of racial endogamy (HR_k * WR_l) across all interracial and intraracial couples. All these terms are symmetric (i.e., they assume that there is nothing particular about interracial couples that distinguish them from racially endogamous couples). The educational boundary components of exchange are captured by the $\sum_{p = 1}^{z} η_{p} u_{p i j} + \sum_{q = 1}^{z} δ_{q} w_{q i j}$ terms, where p and q index levels of education, and u_pij and w_qij are defined as

\begin{array}{l} u_{p i j} = {\begin{matrix} \begin{array}{l} 1 if (i > p and BM / WF) or \\ (j > p and WM / BF) \end{array} \\ 0 otherwise \end{matrix} \\ w_{q i j} = {\begin{matrix} \begin{array}{l} 1 if (i > q and BM / WF) or \\ (j > q and WM / BF) \end{array} \\ 0 otherwise \end{matrix} \end{array}

Each of the terms η_p and δ_q indicates how the log-odds of interracial marriage change for blacks and whites, respectively, when moving up one level of education from either p or q (i.e., from high school graduate to some college and from some college to college graduate, for z = 2). As Gullickson and Torche (2014), we use difference coding, in which the parameter estimates capture the difference between adjacent pairs in an ordinal scale (e.g., some college or more compared with high school or less).

When we add these terms, separately by gender, they exactly fit the three-way interactions of husband’s race by wife’s race by husband’s education (HE * HR * WR), and husband’s race by wife’s race by wife’s education (WE * HR * WR), as well as the lower order two-way interactions involved. For male partners, these three-way interaction terms capture how his education (HE) shapes the chances of interracial pairing (HR * WR). For female partners, they capture how her education (WE) shapes the chances of interracial pairing (HR * WR).

By expressing these three-way interactions as educational boundaries, we are able to assess which educational levels have a stronger influence on the probability of racial intermarriage separately for whites and blacks. The status exchange theory indicates that educational boundaries should be positive for blacks, signaling increasing chances of racial intermarriage as their education increases, but negative for whites, signaling lower chances of intermarriage as their education increases.

The x_ijkl and y_ijkl terms capture “dyadic exchange.” They are given by

\begin{array}{l} x_{i j k l} = {\begin{matrix} \begin{array}{l} 1 if (i > j and BM / WF) or \\ (j > i and WM / BF) \end{array} \\ 0 otherwise \end{matrix} \\ y_{i j k l} = {\begin{matrix} \begin{array}{l} 1 if (i < j and BM / WF) or \\ (j < i and WM / BF) \end{array} \\ 0 otherwise \end{matrix} \end{array},

where x_ijkl identifies white hypergamy and y_ijkl identifies white hypogamy. Because the parameters apply only to interracial couples, racially endogamous couples are used as a baseline for comparison. These parameters involve a four-way interaction (HE * WE * HR * WR). For interracial marriages (defined by HR * WR), the black partner’s education is expected to be larger than the white partner’s education (captured by HE * WE) when compared with intraracial unions, such that white hypergamy should be positive and white hypogamy should be negative.

Our analysis proceeds as follows. We first offer some descriptive evidence on interracial union trends between 1980 and 2010. Then, we conduct a parsimonious assessment of status exchange trends using the hypergamy ratio. Finally, we examine the two components of exchange to determine whether it is driven by educational boundaries or dyadic exchange. All analyses are conducted separately for married and cohabiting couples.

Descriptive Findings: Trends in Interracial Unions

Figure 2A presents rates of racial exogamy among married individuals 25 to 34 years old from 1980 to 2010. Given that the probability of intermarriage depends on the relative size of each racial group, we present rates of racial intermarriage separately for blacks and whites. Among white men, the probability of marrying a black woman rises from 0.08 percent among those born around 1950 to 0.7 percent for those born around 1980. Among white women, there is an increase from 0.3 percent to 1.5 percent. As expected given group size, the proportion of blacks marrying whites is higher. In 1980, 4 percent of black men and 1 percent of black women married a white partner; by 2010, as many as 15 percent of black men and 8 percent of black women married interracially. Translated to the proportion of all couples, only 0.34 percent of young married couples included a black spouse and a white spouse in 1980. For the cohorts born three decades later, 1.5 percent of young marriages were black/white. Even if interracial couples were still a small proportion of all marriages, they were much more visible in 2010 than in prior decades.⁵ Trends of racial exogamy are very similar among cohabiting couples (Figure 2B). Exogamy is somewhat more pronounced among cohabiters—in particular, black men have consistently been more likely to cross racial barriers in establishing cohabitations—but there is convergence across union type over time.

Figure 2.

Proportion of 25- to 34-year-old white individuals married or cohabiting with blacks and proportion of 25- to 34-year-old black individuals married or cohabiting with whites, 1980 to 2010.

The gender composition of black/white couples changed as interracial unions became more prevalent. Figure 3 shows a decline in the proportion of interracial marriages formed by a black man and a white woman, from 81 percent to 69 percent among marriages and from 87 percent to 79 percent among cohabitations. Although still infrequent, white male/black female couples are a growing proportion of interracial couples, a trend possibly associated to women’s gains in educational attainment and labor force participation.

Figure 3.

Gender asymmetry in black/white marriages and cohabitations: proportion of white/black young couples (male partner 25–34 years old) formed by a black man and a white woman, 1980 to 2010.

Trends in Status Exchange

We now move to the core of the analysis. Appendix Table 3 in the online supplemental material includes the full set of parameter estimates and standard errors from the quasi-symmetry model used to generate hypergamy ratios across years. Figure 4A summarizes the findings from this model, showing the hypergamy ratios for interracial marriages from 1980 to 2010. Hypergamy ratios of 1 indicate educational symmetry between partners, and values greater than 1 reveal status exchange. As expected, racially endogamous couples display hypergamy ratios very close to 1, indicating no exchange.

Figure 4.

Hypergamy ratio among young marriages and cohabitations (male partner 25–34 years old), 1980 to 2010.

The story is different for interracial couples. White wives in interracial marriages are much more likely to be educationally hypergamous (“marrying up” in education) than their counterparts in intraracial marriages. More important, white hypergamy did not decrease over time. The hypergamy ratio among black husbands/white wives was 1.5 in 1980, rose to almost 1.7 in 2000, and then returned to 1.5 in 2010. This means that in 1980, white women in interracial marriages were 50 percent more likely than white women in intraracial marriages to have husbands with more education than themselves; strikingly, the hypergamy ratio was the same in 2010, even after 30 years of increasing prevalence and acceptance of interracial marriage.

The finding is nearly identical among black male/white female cohabiters. The hypergamy ratio lingers around 1.6 for the entire period of observation, without indication of a decline in status exchange as interracial couples become more prevalent.

We now move to the unions formed by a white man and a black woman, also presented in Figures 4A and 4B. Consistent with prior research (Qian 1997), exchange is weaker among white male/black female unions, with the hypergamy ratio varying between 1 and 1.2 over time. Again, there is virtually no difference between married and cohabiting couples, except in the recent past, when exchange declines among marriages and increases among cohabitations. This increase suggests a possible convergence in exchange across gender combinations for cohabiting couples, although it is too early to tell if this change constitutes trend.

The main conclusion from Figure 4 is that status exchange among black male/white female couples did not decline over time even though these unions are increasing in number. This refutes hypothesis 1. Furthermore, exchange is equally strong for both marriages and cohabitations, refuting hypothesis 2 and suggesting that the “principles of cohesion” are similar across these union types among these young couples, at least with respect to the exchange of educational for racial status.

What Drives Exchange and What Does It Tell Us about Racial Distances?

The hypergamy ratio is a useful summary measure, but it conflates two factors driving exchange: the higher probability of marrying a white partner for both blacks and whites as their education increases (“educational boundaries”) and a direct trading of resources in interracial marriage resulting in excess white educational hypergamy and the avoidance of white hypogamy (“dyadic exchange”). The final part of the analysis distinguishes these two components. Given that white male/black female couples show very weak evidence of exchange, the analysis focuses on black male/white female unions. We estimate equation 2 separately for married and cohabiting couples. Appendix Tables 4A and 4B in the online supplemental material offer the full set of parameter estimates, standard errors, and significance tests.

Figure 5A displays the educational boundary parameters for married couples. These parameters capture the odds of racial intermarriage for black men and white women as their education increases. Given that three educational categories are used, two contrasts are estimated: the difference in odds of intermarriage between those with some college or more versus those with a high school diploma or less, and the difference between college graduates versus those with less than a college degree.

Figure 5A.

Educational boundary parameters, black male/white female young married couples (husband 25–34 years old), 1980 to 2010.

As predicted by theory, educational boundaries are largely significant and positive for black men. These parameters indicate that as the education of black men increases, their chances of having a white wife substantially increase. For example, in 1980 black men with at least some college had 63 percent higher odds of marrying a white woman than those with high school diplomas, and black male college graduates were 47 percent more likely to marry a white woman compared with black men without college degrees. Educational boundaries for black men declined in the most recent two decades, though they remained statistically significant. In 2000, the only significant educational boundary captures the higher likelihood for black men with some college to marry a white woman compared with men with lower education. In 2010, a college degree is the main distinguishing boundary: black male college graduates have 20 percent higher odds of marrying a white woman than black men with lower levels of schooling.

Interestingly, white wives’ educational boundaries became stronger as black husbands’ were weakening. Educational boundaries are largely insignificant at the conventional p < .05 level among white women in 1980 and 1990. But white female educational boundaries increase in (negative) magnitude and become statistically significant in 2000 and 2010, signaling a growing avoidance of black men as the education of white women increases in these later years. In 2000, the odds of interracial marriage among college-educated white women were 0.75 as large as women with lower levels of education; in 2010, college-educated women have odds of marrying a black man only 0.68 as large as white women with lower levels of education. Thus, as black men’s educational boundaries declined across cohorts, these boundaries increased across cohorts among white women. This change probably emerges from the growing relevance of education as an asset for women in the marriage market and the decline in traditional gender patterns in racial intermarriage. The result of these trends for men and women, when considered jointly, is that the educational boundary component of exchange remained relatively constant across cohorts.

Figure 5B shows dyadic exchange, distinguishing white hypergamy (expected to be positive) and white hypogamy (expected to be negative). Two sets of dyadic exchange parameters are shown: the total effect before controlling for educational boundaries, and the effect net of educational boundaries.

Figure 5B.

Dyadic exchange parameters: total effect and net of educational boundaries, black male/white female young married couples (husband 25–34 years old), 1980 to 2010.

The comparison provides information on the extent of confounding that is prevalent in the status exchange literature, which has thus far focused on the dyadic, direct trading of race for educational resources among partners. As shown in Figure 5B, the excess white hypergamy disappears or substantially declines after accounting for educational boundaries. In 1980, shown in the leftmost panel, the odds of white hypergamy were 90 percent higher than educational homogamy, while the odds of white hypogamy were only 36 percent higher. This result would signal an “excess” white hypergamy in black male/white female marriages, consistent with dyadic exchange. However, after accounting for educational boundaries, white hypergamy and white hypogamy are very similar and statistically indistinguishable, reaching 25 percent and 43 percent, respectively. This indicates that mixed couples are more likely to be educationally heterogamous than racially endogamous couples but not in a particular direction that favors white hypergamy. This result is consistent over time, with the exception of 2000, when there is some indication of white hypergamy. Besides this exception, there is no indication of excess white hypergamy after the educational boundary component of exchange has been accounted for (in contrast, total dyadic exchange parameters without controls for educational boundaries are comparable, and lead to similar conclusions, as the hypergamy ratios).

The distinction of two components of exchange yields an important finding. Exchange is better described as emerging from educational boundaries rather than as a direct trade between spouses. Indeed, exchange appears to be largely a by-product of growing chances of racial intermarriage for blacks as their education increases and declining chances of intermarriage for whites as their education increases.

Figure 6 replicates the analysis of status exchange components for cohabiters. According to Figure 6A, the centrality of educational boundaries is even more pronounced among cohabiters than among married couples. Educational boundaries are extremely large in 1980 and less pronounced but still substantial from 1990 to 2010. Black men with some college education have odds 30 percent higher than their counterparts with high school or less to cohabit with a white woman in 1990, 59% higher in 2000 and 21% higher in 2010. Furthermore, educational boundaries tend to be more gender symmetric than for married couples. Consistently over time, black men with higher education have higher chances of marrying white women, while white women with higher education have lower chances of marrying black men. The single exception to this pattern is 1990, when educational boundaries for white women go in the direction opposite to theory, but the magnitude of these effects is rather small.

Figure 6A.

Educational boundary parameters, black male/white female young cohabiting couples (male partner 25–34 years old), 1980 to 2010.

Figure 6B displays dyadic exchange parameters for cohabiting unions. Before controlling for educational boundaries, white hypergamy tends to be positive and white hypogamy tends to be negative across cohorts, signaling again a dyadic race-by-education exchange between partners. These parameter estimates change once educational boundaries are accounted for, with white hypergamy becoming statistically insignificant or even reversing signs. As with educational boundaries, a partial exception is 1990, in which dyadic exchange plays a more important role. For all other periods, educational boundaries appear to account for white educational hypergamy, rather than a direct trading of racial-for-educational status.

Figure 6B.

Dyadic exchange parameters: total effect and net of educational boundaries, black male/white female young cohabiting couples (male partner 25–34 years old), 1980 to 2010.

These findings suggest that partners in interracial marriages and cohabitations do not directly exchange sources of status to achieve a mutually beneficial arrangement. Instead, individuals appear to enact a preference for whiteness even if they are unaware of it (Quillian 2006). Such preference is not just nonhierarchical in-group favoritism. Rather, in a racially stratified society in which whiteness is a source of advantage, the odds that a black person partners with a white person increase with education, regardless of the education of the potential white partner, and conversely, the odds that a white individual has a black partner decrease with education, regardless of the education of the potential black partner.

Discussion

Even if the boundary between blacks and whites is still the most powerful division in the American marriage market (Rosenfeld 2008), the prevalence and stated acceptance of white/black unions has increased dramatically. These trends have occurred in context of a substantial change of race relations in which blatant racial prejudice and discrimination have given way to strong norms against overt racism, while socioeconomic gaps between blacks and whites remain persistently high.

In this context, it is relevant to examine trends in status exchange in black/white unions. Status exchange provides a strong test of racial stratification in the marriage market. Simply put, if status exchange persists over time, there is indication that race is still treated as a hierarchical attribute in the marriage market (Kalmijn 2010). We find compelling evidence that status exchange remained strong among black male/white female couples from 1980 through 2010. Moreover, although some have claimed that changes in patterns of intermarriage may be offset by diverging trends among the growing number of cohabitations, we find that status exchange trends among black male/white female unions are similar across type of union. In contrast, status exchange is slight among the small but growing number of white male/black female couples.

Overall, our findings align with recent scholarship arguing that racial hierarchical distinctions persist in American society despite the decline in overt racism (Bobo 2011; Bobo and Smith 1998; Bonilla-Silva 2010; Jackman 1994; McConahay 1986; Sears 1988). The evidence presented here adds further caution to the claim that the United States has moved or is moving toward a colorblind society over the past three decades.

The analysis also takes issue with the very notion of exchange, as traditionally understood in the literature. Exchange has been treated as a direct trade of resources between black and white partners. Merton (1941) defined “a reciprocal compensatory situation in which the black male ‘exchanges’ his higher economic position for the white female’s higher caste status” (p. 372), and Davis (1941) indicated that “the class achievements of certain Negro males enable them to bargain for females of the white caste who stand low in the class hierarchy” (p. 389, cited in Gullickson and Torche 2014). This analysis has delved into the components of status exchange, distinguishing educational boundaries from dyadic exchange. Educational boundaries capture the growing probability of marrying a white person as educational attainment increases for both blacks and whites, which results in opposing probabilities of intermarriage for these racial groups. Dyadic exchange, in turn, captures the excess white educational hypergamy in interracial marriages compared with intraracial marriages, net of educational boundaries. If educational boundaries are not properly modeled, they will appear as dyadic exchange, a limitation affecting most analyses of exchange to date.

The analysis of status exchange’s components yields clear findings. For both married and cohabiting couples, educational boundaries largely account for status exchange. Education increases the chances that black men intermarry (regardless of the educational level of the potential white spouse) and/or decreases the chances that white women intermarry (again, regardless of the educational level of the potential black spouse). These phenomena result in educational asymmetry among the interracial couples that are formed. Critically, net of these educational boundaries, white hypergamy resulting from dyadic exchange is not more likely than white hypogamy among interracial couples. In other words, interracial couples are more educationally heterogamous than intraracial couples, but this is driven by the increasing propensity to marry a white spouse for both blacks and whites as their education increases rather than by a higher prevalence of individual dyads in which one partner directly trades racial status for the other partner’s educational status.

The distinction between educational boundaries and dyadic exchange is important because educational boundaries apply to all interracial couples and not only those that are educationally heterogamous. That is, educational boundaries capture a more general phenomenon than originally thought by theorists of status exchange, signaling that whiteness is treated as a source of status in the U.S. marriage market. Strikingly, the growing prevalence of racial intermarriage has not resulted in the decline of status exchange. In contrast, as suggested by similar patterns of racial intermarriage in Brazil, the growth of racial intermarriage may even paradoxically facilitate the expression of a black/white hierarchy by making interracial unions less exceptional, the disadvantage associated with blackness more acceptable, and the gains associated with marrying a white person more valuable (Gullickson and Torche 2014; Telles 2004).

This analysis has focused on black/white intermarriage given the strong barriers between these groups historically formed in the United States through centuries of slavery, segregation, and discrimination. Other work has extended the question about status exchange to intermarriage between whites and both Latino and Asian minorities (Fu 2001; Hwang, Saenz, and Aguirre 1995; Qian 1997), but no research to date has ascertained the specific factors—educational barriers or dyadic exchange—driving what the literature has identified as exchange patterns among these ethnoracial groups. Extending the question about the sources of exchange for these groups is an important task, but one that is beyond the scope of this article. Such analysis would require an assessment of the changing prevalence of different ethnoracial marital combination and the historically shaped sources of stratification between groups, as well as an adequate treatment of the gender-specific patterns of partner selection for different ethnoracial groups (Feliciano et al. 2009; Lin and Lundquist 2010). We trust future research will undertake this task, as it will provide needed information about the changing contours and patterns of ethnoracial stratification in the changing U.S. marriage market.

Footnotes

Acknowledgements

This study uses data from the Integrated Public Use Microdata Series. We thank the Minnesota Population Center for making the data available. We also thank Cynthia Feliciano, Nicole Marwell, Ann Morning, and Lawrence Wu for helpful comments and suggestions.

Supplemental Material

The online supplemental material is available at .

Notes

Author Biographies

Florencia Torche is Professor of Sociology at New York University. Her scholarship examines stratification and mobility, educational inequality, assortative mating, and the early emergence of disadvantage.

Peter Rich is Assistant Professor of Policy Analysis and Management at Cornell University. His work evaluates how individual decisions in housing, school, and marriage markets reflect and potentially reinforce racial and economic inequalities, and how such decisions are impacted by policy and local context.

References

Becker

1991. A Treatise on the Family. Exp. ed. Cambridge, MA: Harvard University Press.

Black

Gates

Sanders

Taylor

2006. “The Measurement of Same-sex Unmarried Couples in the 2000 U.S. Census.” California Center for Population Research, On-Line Working Paper CCPR-023-07.

Blackwell

Lichter

2000. “Mate Selection among Married and Cohabiting Couples.” Journal of Family Issues 21(3):275–302.

Blackwell

Lichter

2004. “Homogamy among Dating, Cohabiting, and Married Couples.” Sociological Quarterly 45(4):719–37.

Bobo

2011. “Somewhere between Jim Crow & Post-racialism: Reflections on the Racial Divide in America Today.” Daedalus 140(1):11–36.

Bobo

Smith

1998. “From Jim Crow Racism to Laissez-faire Racism: The Transformation of Racial Attitudes in America.” Pp. 182–220 in Beyond Pluralism: Essays on the Conception of Groups and Group Identities in America, edited by Katkin

Landsman

Tyree

Urbana: University of Illinois Press.

Bonilla-Silva

2010. Racism without Racists: Color-blind Racism and the Persistence of Racial Inequality in the United States (3rd ed.). Lanham, MD: Rowman & Littlefield.

Bratter

King

2008. “‘But Will It Last?’: Marital Instability among Interracial and Same-race Couples.” Family Relations 57(2):160–71.

Brines

Joyner

1999. “The Ties That Bind: Principles of Cohesion in Cohabitation and Marriage.” American Sociological Review 64(3):333–55.

10.

Cotter

Hermsen

Vanneman

2009. “The Stalled Gender Revolution.”Retrieved October 2015 (http://www.bsos.umd.edu/socy/vanneman/endofgr/default.html).

11.

Davis

1941. “Intermarriage in Caste Societies.” American Anthropologist 43(3):376–95.

12.

DeNavas-Walt

Proctor

Smith

2012. “Income, Poverty, and Health Insurance Coverage in the United States: 2011.” Current Population Reports P60-243. Washington, DC: U.S. Census Bureau.

13.

Feliciano

Robnett

Komaie

2009. “Gendered Racial Exclusion among White Internet Daters.” Social Science Research 38(1):39–54.

14.

Fischer

Hout

2006. Century of Difference. New York: Russell Sage.

15.

Fryer

2007. “Guess Who’s Been Coming to Dinner? Trends in Interracial Marriage over the 20th Century.” Journal of Economic Perspectives 21(2):71–90.

16.

V. K.

2001. “Racial Intermarriage Pairings.” Demography 38(2):147–59.

17.

V. K.

2010. “Remarriage, Delayed Marriage, and Black/White Intermarriage, 1968–1995.” Population Research and Policy Review 29(5):687–713.

18.

V. K.

Wolfinger

2011. “Broken Boundaries or Broken Marriages? Racial Intermarriage and Divorce in the United States.” Social Science Quarterly 92(4):1096–1117.

19.

Gates

G. J.

2009. “Same-sex Spouses and Unmarried Partners in the American Community Survey, 2008.” Los Angeles: The Williams Institute.

20.

Gates

G. J.

Steinberger

M. D.

2009. “Same-sex Unmarried Partner Couples in the American Community Survey: The Role of Misreporting, Miscoding and Misallocation.” Detroit, MI: Population Association of America Conference.

21.

Goldin

2006. “The Quiet Revolution That Transformed Women’s Employment, Education, and Family.” American Economic Review 96(2):1–21.

22.

Gullickson

2006a. “Black/White Interracial Marriage Trends, 1850–2000.” Journal of Family History 31(3):289–312.

23.

Gullickson

2006b. “Education and Black/White Interracial Marriage.” Demography 43(4):673–89.

24.

Gullickson

V. K.

2010. “Comment: An Endorsement of Exchange Theory in Mate Selection.” American Journal of Sociology 115(4):1243–51.

25.

Gullickson

Torche

2014. “Patterns of Racial and Educational Assortative Mating in Brazil.” Demography 51(3):835–56

26.

Heaton

T. B.

2002. “Factors Contributing to Increasing Marital Stability in the United States.” Journal of Family Issues 23(3):392–409.

27.

Hou

Myles

2013. “Interracial Marriage and Status-caste Exchange in Canada and the United States.” Ethnic and Racial Studies 36(1):75–96.

28.

Hwang

Saenz

Aguirre

1995. “The SES Selectivity of Interracially Married Asians.” International Migration Review 29(2):469–91.

29.

Jackman

1994. The Velvet Glove: Paternalism and Conflict in Gender, Class, and Race Relations. Berkeley: University of California Press.

30.

Kalmijn

1991a. “Shifting Boundaries: Trends in Religious and Educational Homogamy.” American Sociological Review 56(6):786–800.

31.

Kalmijn

1991b. “Status Homogamy in the United States.” American Journal of Sociology 97(2):496–523.

32.

Kalmijn

1993. “Trends in Black/White Intermarriage.” Social Forces 72(1):119–46.

33.

Kalmijn

1998. “Intermarriage and Homogamy: Causes, Patterns, Trends.” Annual Review of Sociology 24:395–421.

34.

Kalmijn

2010. “Educational Inequality, Homogamy, and Status Exchange in Black/White Intermarriage: A Comment on Rosenfeld.” American Journal of Sociology 115(4):1252–63.

35.

Lichter

Qian

2004. “Marriage and Family in a Multiracial Society.” Pp. 169–200 in The American People: Census 2000, edited by Farley

New York: Russell Sage.

36.

Lin

K.-H.

Lundquist

2013. “Mate Selection in Cyberspace: The Intersection of Race, Gender, and Education.” American Journal of Sociology 119(1):183–215.

37.

Macklin

1987. “Nontraditional Family Forms.” Pp. 317–53 in Handbook of Marriage and the Family, edited by Sussman

Steinmetz

New York: Plenum.

38.

McConahay

1986. “Modem Racism, Ambivalence, and the Modern Racism Scale.” Pp. 91–125 in Prejudice, Discrimination and Racism, edited by Dovidio

Gaertner

Orlando, FL: Academic Press.

39.

McDaniel

DiPrete

Buchmann

Shwed

2011. “The Black Gender Gap in Educational Attainment: Historical Trends and Racial Comparisons.” Demography 48(3):889–914.

40.

Merton

R. K.

1941. “Intermarriage and the Social Structure: Fact and Theory.” Psychiatry 4(3):361–74.

41.

Morris

Western

1999. “Inequality in Earnings at the Close of the Twentieth Century.” Annual Review of Sociology 25:623–57.

42.

Pallas

2000. “The Effects of Schooling on Individual Lives.” Pp. 499–525 in Handbook of the Sociology of Education, edited by Hallinan

New York: Kluwer Academic.

43.

Porterfield

1982. “Black-American Intermarriage in the United States.” Marriage and Family Review 5(1):17–34.

44.

Qian

1997. “Breaking the Racial Barriers: Variations in Interracial Marriage between 1980 and 1990.” Demography 34(2):263–76.

45.

Qian

Lichter

2007. “Social Boundaries and Marital Assimilation: Interpreting Trends in Racial and Ethnic Intermarriage.” American Sociological Review 72:68–94.

46.

Qian

Lichter

2011. “Changing Patterns of Interracial Marriage in a Multiracial Society.” Journal of Marriage and the Family 73(5):1065–84.

47.

Quillian

2006. “New Approaches to Understanding Racism and Discrimination.” Annual Review of Sociology 32:299–328.

48.

Rosenfeld

2005. “A Critique of Exchange Theory in Mate Selection.” American Journal of Sociology 110(5):1284–1325.

49.

Rosenfeld

2008. “Racial, Educational and Religious Endogamy in the United States: A Comparative Historical Perspective.” Social Forces 87(1):1–31.

50.

Rosenfeld

2010. “Still Weak Support for Status-caste Exchange: A Reply to Critics.” American Journal of Sociology 115(4):1264–76.

51.

Ruggles

Alexander

J. T.

Genadek

Goeken

Schroeder

M. B.

Sobek

2010. “Integrated Public Use Microdata Series: Version 5.0” [machine-readable database]. Minneapolis: University of Minnesota.

52.

Schoen

Weinick

1993. “Partner Choice in Marriages and Cohabitations.” Journal of Marriage and the Family 55(2):408–14.

53.

Schoen

Wooldredge

1989. “Marriage Choices in North Carolina and Virginia, 1969–71 and 1979–81.” Journal of Marriage and the Family 51(2):465–81.

54.

Schuman

Steeh

Bobo

Krysan

1997. Racial Attitudes in America: Trends and Interpretations. Rev. ed. Cambridge, MA: Harvard University Press.

55.

Sears

1988. “Symbolic Racism.” Pp. 53–84 in Eliminating Racism: Profiles in Controversy, edited by Katz

Taylor

New York: Plenum.

56.

Smith

Mardsen

Hout

Kim

2013. “General Social Surveys, 1972–2012” [machine-readable data file]. Chicago: National Opinion Research Center [producer]; Storrs, CT: Roper Center, University of Connecticut [distributor].

57.

Wang

2012. “The Rise of Intermarriage: Rates, Characteristics Vary by Race and Gender.” Washington, DC: Pew Research Center. Retrieved April 27, 2016 (http://www.pewsocialtrends.org/files/2012/02/SDT-Intermarriage-II.pdf).

58.

Telles

2004. Race in Another America: The Significance of Skin Color in Brazil. Princeton, NJ: Princeton University Press.