Journal of Marriage and Family ncfr YUE QIAN University of British Columbia Gender Asymmetry in Educational and Income Assortative Marriage The reversal of the gender gap in education in educational attainment has coincided with has reshaped the U.S.marriage market.Draw- substantial changes in marriage patterns.In ing on data from the 1980 U.S.Census and the 2012,21%of married women had spouses who 2008-2012 American Community Surveys,the were less educated than they were-a twofold author used log-linear models to examine gen- increase from 1980 (Wang,2014).During der asymmetry in educational and income assor- the period when the gender gap in education tative mating among newlyweds.Between 1980 narrowed and eventually reversed,has the nor- and 2008-2012,educational assortative mating mative practice of women marrying men of reversed from a tendency for women to marry higher socioeconomic status (i.e.,hypergamy) up to a tendency for women to marry down in been eroded?When they marry less-educated education,whereas the tendency for women to men,do women also marry men with lower marry men with higher incomes than themselves incomes,thereby challenging the traditional persisted.Moreover,in both time periods,the breadwinning role of men in the family?In tendency for women to marry up in income was answering these questions,this article offers generally greater among couples in which the a critical empirical investigation of gender wife's education level equaled or surpassed that asymmetry in assortative mating and presents a of the husband than among couples in which detailed picture of the state of gender equality the wife was less educated than the husband. in heterosexual marriages. The author discusses the implications of the ris- Women married to men with less edu- ing female advantage in education for gender cation than themselves are often thought to change in heterosexual marriages. challenge the traditional,male-dominant sta- tus in marriage (Kaukinen,2004;Schwartz Han,2014).The previously nonnormative Women have made greater gains in educational arrangement-educational hypogamy (i.e.,mar- attainment than men during the past few decades riages in which the wife has more education in the United States.Currently,women earn than the husband)-has become more common about 60%of bachelor's and master's degrees (Schwartz Mare,2005).Does increasing edu- and half of all doctoral degrees (DiPrete cational hypogamy indicate a shift away from Buchmann,2013).The rising female advantage the convention of mate selection that embodies male dominance?In this article,I argue that an exclusive focus on educational assortative Department of Sociology,University of British Columbia, mating provides an incomplete understand- 6303 NW Marine Drive,Vancouver,BC Canada V6T 1ZI ing of mate selection patterns and overstates (yue.qian@ubc.ca). gender change in heterosexual marriages.This This article was edited by Kelly Raley. study advances prior work toward a more Key Words:demography.education,gender.marriage,mate comprehensive understanding of the gendered selection.U.S.population. and multidimensional nature of mate selection Journal of Marriage and Family(2016) 1 D0L:10.1111/jomf.12372
Yue Qian University of British Columbia Gender Asymmetry in Educational and Income Assortative Marriage The reversal of the gender gap in education has reshaped the U.S. marriage market. Drawing on data from the 1980 U.S. Census and the 2008–2012 American Community Surveys, the author used log-linear models to examine gender asymmetry in educational and income assortative mating among newlyweds. Between 1980 and 2008–2012, educational assortative mating reversed from a tendency for women to marry up to a tendency for women to marry down in education, whereas the tendency for women to marry men with higher incomes than themselves persisted. Moreover, in both time periods, the tendency for women to marry up in income was generally greater among couples in which the wife’s education level equaled or surpassed that of the husband than among couples in which the wife was less educated than the husband. The author discusses the implications of the rising female advantage in education for gender change in heterosexual marriages. Women have made greater gains in educational attainment than men during the past few decades in the United States. Currently, women earn about 60% of bachelor’s and master’s degrees and half of all doctoral degrees (DiPrete & Buchmann, 2013). The rising female advantage Department of Sociology, University of British Columbia, 6303 NW Marine Drive, Vancouver, BC Canada V6T 1Z1 (yue.qian@ubc.ca). This article was edited by Kelly Raley. Key Words: demography, education, gender, marriage, mate selection, U.S. population. in educational attainment has coincided with substantial changes in marriage patterns. In 2012, 21% of married women had spouses who were less educated than they were—a twofold increase from 1980 (Wang, 2014). During the period when the gender gap in education narrowed and eventually reversed, has the normative practice of women marrying men of higher socioeconomic status (i.e., hypergamy) been eroded? When they marry less-educated men, do women also marry men with lower incomes, thereby challenging the traditional breadwinning role of men in the family? In answering these questions, this article offers a critical empirical investigation of gender asymmetry in assortative mating and presents a detailed picture of the state of gender equality in heterosexual marriages. Women married to men with less education than themselves are often thought to challenge the traditional, male-dominant status in marriage (Kaukinen, 2004; Schwartz & Han, 2014). The previously nonnormative arrangement—educational hypogamy (i.e., marriages in which the wife has more education than the husband)—has become more common (Schwartz & Mare, 2005). Does increasing educational hypogamy indicate a shift away from the convention of mate selection that embodies male dominance? In this article, I argue that an exclusive focus on educational assortative mating provides an incomplete understanding of mate selection patterns and overstates gender change in heterosexual marriages. This study advances prior work toward a more comprehensive understanding of the gendered and multidimensional nature of mate selection Journal of Marriage and Family (2016) 1 DOI:10.1111/jomf.12372
2 Journal of Marriage and Family by examining how income and education jointly prior to the gender reversal in educational attain- shape assortative mating patterns. ment.Then,pooling data from the 2008-2012 Compared with educational differentials American Community Surveys,I examined between two spouses,men's income advan- recent marriage cohorts for couples who got tage over their wives is more central to their married during the period when the gender gap identity as breadwinners and household heads in education favored women.Overall,this study (Tichenor,2005).The smaller amount of money reveals whether and how gendered features of women bring into the household relative to their partner choice in heterosexual marriages have husbands contributes to women's subordinate changed in recent decades. status and lower bargaining power in the fam- ily (England,2003).Despite the importance of spouses'relative income in shaping power THEORETICAL FRAMEWORKS AND HYPOTHESES dynamics between them (Bittman et al.,2003; Assortative mating has received substantial Tichenor,2005),research on income assortative scholarly interest (for reviews,see Blossfeld, mating has been scant. 2009:Kalmijn,1998;Schwartz,2013).Indi- Despite the decline and eventual reversal of viduals have preferences for spouses but face the gender gap in education (DiPrete Buch- constrained opportunities in the marriage mar- mann,2013),a gender gap in pay persists:In kets (Oppenheimer,1988;Schwartz,2013). 2011,the median hourly pay of women was 84% Different theories have been developed to that of men (Economic Policy Institute,2012). conceptualize the underlying processes deter- On one hand,women's advantage in education mining partner choice.In her career-entry may enable them to be more economically inde- theory,Oppenheimer(1988)applied job-search pendent and thus put less emphasis on economic theory to mate selection and maintained that traits when evaluating potential spouses(Press, as women's employment and labor force 2004).On the other hand,evidence suggests attachment increase,women are increasingly that men may still feel uncomfortable forming evaluated as potential spouses on the basis relationships in which they have lower status of their socioeconomic traits.Accordingly, than their female partners(Bertrand,Kamenica, men begin competing for highly educated or Pan,2015;Fisman,Iyengar,Kamenica, high-earning women.At the same time,the Simonson,2006).This persistent gendered deteriorating economic position of less-skilled norm as well as the gender pay gap mean that men and the declining economic security of men-especially men who do not have an edu- strict gender specialization where the husband cational advantage over their wives-may marry is the single earner in the family would also lead women with lower incomes than themselves to to changes in men's attitudes and behaviors in preserve their status and gender role expecta- mate selection(Oppenheimer,1994,1997). tions in marriage.Hence,I go beyond prior stud- The United States has experienced changes ies by incorporating both education and income in women's economic roles since the 1980s: into the analysis of gender asymmetry in assorta- Women's employment has increased(England, tive mating to provide a better understanding of 2010),occupational gender segregation and the changes in gender role expectations and spouses' gender pay gap have declined(Blau,Brummund, relative socioeconomic status in marriage. Liu,2013;Blau Kahn,2007;England, This study investigates how men's and 2010),and strikingly,the gender gap in educa- women's education and income jointly shape tion has reversed from favoring males to favoring mate selection patterns in the United States.I females (DiPrete Buchmann,2013).In light ask(a)whether the patterns of educational and of women's progress in education and employ- income assortative marriage are symmetrical ment in recent decades,the career-entry theory with respect to gender and (b)whether they have suggests the following hypotheses: changed in recent decades when gender inequal- Hypothesis 1:The tendency for women to marry ity in education and employment has changed men whose education is lower than their own substantially.To this end,I used data from increased between 1980 and 2008-2012. the 1980 U.S.Census to examine educational Hypothesis 2:The tendency for women to marry and income assortative mating patterns among men whose income is higher than their own couples who married in 1979-1980,a period decreased between 1980 and 2008-2012
2 Journal of Marriage and Family by examining how income and education jointly shape assortative mating patterns. Compared with educational differentials between two spouses, men’s income advantage over their wives is more central to their identity as breadwinners and household heads (Tichenor, 2005). The smaller amount of money women bring into the household relative to their husbands contributes to women’s subordinate status and lower bargaining power in the family (England, 2003). Despite the importance of spouses’ relative income in shaping power dynamics between them (Bittman et al., 2003; Tichenor, 2005), research on income assortative mating has been scant. Despite the decline and eventual reversal of the gender gap in education (DiPrete & Buchmann, 2013), a gender gap in pay persists: In 2011, the median hourly pay of women was 84% that of men (Economic Policy Institute, 2012). On one hand, women’s advantage in education may enable them to be more economically independent and thus put less emphasis on economic traits when evaluating potential spouses (Press, 2004). On the other hand, evidence suggests that men may still feel uncomfortable forming relationships in which they have lower status than their female partners (Bertrand, Kamenica, & Pan, 2015; Fisman, Iyengar, Kamenica, & Simonson, 2006). This persistent gendered norm as well as the gender pay gap mean that men—especially men who do not have an educational advantage over their wives—may marry women with lower incomes than themselves to preserve their status and gender role expectations in marriage. Hence, I go beyond prior studies by incorporating both education and income into the analysis of gender asymmetry in assortative mating to provide a better understanding of changes in gender role expectations and spouses’ relative socioeconomic status in marriage. This study investigates how men’s and women’s education and income jointly shape mate selection patterns in the United States. I ask (a) whether the patterns of educational and income assortative marriage are symmetrical with respect to gender and (b) whether they have changed in recent decades when gender inequality in education and employment has changed substantially. To this end, I used data from the 1980 U.S. Census to examine educational and income assortative mating patterns among couples who married in 1979–1980, a period prior to the gender reversal in educational attainment. Then, pooling data from the 2008–2012 American Community Surveys, I examined recent marriage cohorts for couples who got married during the period when the gender gap in education favored women. Overall, this study reveals whether and how gendered features of partner choice in heterosexual marriages have changed in recent decades. Theoretical Frameworks and Hypotheses Assortative mating has received substantial scholarly interest (for reviews, see Blossfeld, 2009; Kalmijn, 1998; Schwartz, 2013). Individuals have preferences for spouses but face constrained opportunities in the marriage markets (Oppenheimer, 1988; Schwartz, 2013). Different theories have been developed to conceptualize the underlying processes determining partner choice. In her career-entry theory, Oppenheimer (1988) applied job-search theory to mate selection and maintained that as women’s employment and labor force attachment increase, women are increasingly evaluated as potential spouses on the basis of their socioeconomic traits. Accordingly, men begin competing for highly educated or high-earning women. At the same time, the deteriorating economic position of less-skilled men and the declining economic security of strict gender specialization where the husband is the single earner in the family would also lead to changes in men’s attitudes and behaviors in mate selection (Oppenheimer, 1994, 1997). The United States has experienced changes in women’s economic roles since the 1980s: Women’s employment has increased (England, 2010), occupational gender segregation and the gender pay gap have declined (Blau, Brummund, & Liu, 2013; Blau & Kahn, 2007; England, 2010), and strikingly, the gender gap in education has reversed from favoring males to favoring females (DiPrete & Buchmann, 2013). In light of women’s progress in education and employment in recent decades, the career-entry theory suggests the following hypotheses: Hypothesis 1: The tendency for women to marry men whose education is lower than their own increased between 1980 and 2008–2012. Hypothesis 2: The tendency for women to marry men whose income is higher than their own decreased between 1980 and 2008–2012
Educational and Income Assortative Marriage Previous studies on assortative mating exam- underlying the economic model of marriage ined education and income separately.It is (Becker,1981)that individuals make rational unclear how these two variables work together marriage decisions and marry only if the utility As I elaborate later,drawing on an"uneven and of marriage exceeds the utility of remaining stalled gender revolution"perspective(England, single.It posits that individuals balance unequal 2010,p.149)and status exchange theory (Davis, traits through exchange to maximize their gains 1941;Merton,1941),I hypothesize that hus- from marriage (Davis,1941;Merton,1941). bands'income advantage over their wives would Hence,when women marry down in education be more apparent among couples in which the (i.e.,marry men whose education is lower than wife has an education level surpassing or equal their own),they may marry up in income (i.e., to that of the husband. marry men whose income is higher than their An "uneven and stalled gender revolution" own).Such balance between more-educated perspective(England,2010)posits that progress wives and higher earning husbands should be toward gender equality has been uneven welcomed by both men and women choosing Change in heterosexual relationships has been marital partners.From women's perspective, much more limited when compared with gender when searching for a spouse in a pool of change in the world of paid work and education. less-educated men,women have more to gain For example,men are still expected to propose if they choose men with higher incomes.After marriage (Sassler Miller,2011),and the all,men do not need equivalent levels of edu- vast majority of women take their husbands' cation to have higher incomes than women. surnames (Goldin Shim,2004).Mate pref- Even women who work full-time tend to earn erences are still gendered such that women less than men of comparable or less educa- have a stronger preference for income than men tion (Institute for Women's Policy Research, (Hitsch,Hortacsu,Ariely,2010).In addition, 2015).From men's perspective,although men although men value potential wives'economic have placed more importance on the financial roles more today than they did in the past(Buss. prospects of a potential spouse over time(Buss Shackelford,Kirkpatrick,Larsen,2001),they et al.,2001),they may value women's high still appear to hesitate to choose women whose status only up to the point when women's status status exceeds their own status (England,2011; exceeds their own status (Bertrand et al.,2015; Graf Schwartz.2011). England,2011;Graf Schwartz,2011).For The "uneven and stalled gender revolution" example,a speed dating study found that men perspective also points out that the progress did not value women's intelligence or ambition toward gender equality in paid work has recently when it exceeded their own (Fisman et al., stalled,as measured by changes in women's 2006).Psychology experiments showed that labor force participation,gender pay gap,and men's self-esteem was lower when their part- occupational sex segregation since the 1990s ners succeeded than when their partners failed, (Blau et al.,2013;Blau Kahn,2007;England, whereas women's self-esteem was not affected 2010).Similarly,U.S.attitudes toward egali- by their partners'performance(Ratliff Oishi, tarian gender roles and the share of women 2013).Although these studies did not directly who keep their surnames after marriage have test men's reaction to their partners'education changed little since the 1990s(Cotter,Hermsen, or income,they suggest that men may avoid a Vanneman,2011;Goldin Shim,2004).This potential spouse who has both higher education stalling of progress toward gender equality sug- and higher income than themselves. gests that the norm against marriages in which Overall,the "uneven and stalled gender women have higher status than their husbands revolution"perspective (England,2010)and has changed little in recent decades. status exchange theory suggest the following In light of the rising female advantage in hypotheses: education,status exchange theory in the inter- Hypothesis 3a:The tendency for women to marry marriage literature provides a theoretical basis to up in income is greater among couples in which understand how individuals avoid status reversal the wife has more education than the husband in marriage via assortative mating (e.g.,Davis, than among couples in which the wife has less 1941;Gullickson,2006;Gullickson Fu,2010: education than the husband. Gullickson Torche.2014:Merton.1941).Sta- Hypothesis 3b:The greater tendency for women tus exchange theory shares the assumption to marry up in income when they marry down
Educational and Income Assortative Marriage 3 Previous studies on assortative mating examined education and income separately. It is unclear how these two variables work together. As I elaborate later, drawing on an “uneven and stalled gender revolution” perspective (England, 2010, p. 149) and status exchange theory (Davis, 1941; Merton, 1941), I hypothesize that husbands’ income advantage over their wives would be more apparent among couples in which the wife has an education level surpassing or equal to that of the husband. An “uneven and stalled gender revolution” perspective (England, 2010) posits that progress toward gender equality has been uneven. Change in heterosexual relationships has been much more limited when compared with gender change in the world of paid work and education. For example, men are still expected to propose marriage (Sassler & Miller, 2011), and the vast majority of women take their husbands’ surnames (Goldin & Shim, 2004). Mate preferences are still gendered such that women have a stronger preference for income than men (Hitsch, Hortaçsu, & Ariely, 2010). In addition, although men value potential wives’ economic roles more today than they did in the past (Buss, Shackelford, Kirkpatrick, & Larsen, 2001), they still appear to hesitate to choose women whose status exceeds their own status (England, 2011; Graf & Schwartz, 2011). The “uneven and stalled gender revolution” perspective also points out that the progress toward gender equality in paid work has recently stalled, as measured by changes in women’s labor force participation, gender pay gap, and occupational sex segregation since the 1990s (Blau et al., 2013; Blau & Kahn, 2007; England, 2010). Similarly, U.S. attitudes toward egalitarian gender roles and the share of women who keep their surnames after marriage have changed little since the 1990s (Cotter, Hermsen, & Vanneman, 2011; Goldin & Shim, 2004). This stalling of progress toward gender equality suggests that the norm against marriages in which women have higher status than their husbands has changed little in recent decades. In light of the rising female advantage in education, status exchange theory in the intermarriage literature provides a theoretical basis to understand how individuals avoid status reversal in marriage via assortative mating (e.g., Davis, 1941; Gullickson, 2006; Gullickson & Fu, 2010; Gullickson & Torche, 2014; Merton, 1941). Status exchange theory shares the assumption underlying the economic model of marriage (Becker, 1981) that individuals make rational marriage decisions and marry only if the utility of marriage exceeds the utility of remaining single. It posits that individuals balance unequal traits through exchange to maximize their gains from marriage (Davis, 1941; Merton, 1941). Hence, when women marry down in education (i.e., marry men whose education is lower than their own), they may marry up in income (i.e., marry men whose income is higher than their own). Such balance between more-educated wives and higher earning husbands should be welcomed by both men and women choosing marital partners. From women’s perspective, when searching for a spouse in a pool of less-educated men, women have more to gain if they choose men with higher incomes. After all, men do not need equivalent levels of education to have higher incomes than women. Even women who work full-time tend to earn less than men of comparable or less education (Institute for Women’s Policy Research, 2015). From men’s perspective, although men have placed more importance on the financial prospects of a potential spouse over time (Buss et al., 2001), they may value women’s high status only up to the point when women’s status exceeds their own status (Bertrand et al., 2015; England, 2011; Graf & Schwartz, 2011). For example, a speed dating study found that men did not value women’s intelligence or ambition when it exceeded their own (Fisman et al., 2006). Psychology experiments showed that men’s self-esteem was lower when their partners succeeded than when their partners failed, whereas women’s self-esteem was not affected by their partners’ performance (Ratliff & Oishi, 2013). Although these studies did not directly test men’s reaction to their partners’ education or income, they suggest that men may avoid a potential spouse who has both higher education and higher income than themselves. Overall, the “uneven and stalled gender revolution” perspective (England, 2010) and status exchange theory suggest the following hypotheses: Hypothesis 3a: The tendency for women to marry up in income is greater among couples in which the wife has more education than the husband than among couples in which the wife has less education than the husband. Hypothesis 3b: The greater tendency for women to marry up in income when they marry down
Journal of Marriage and Family in education did not change between 1980 and same education level did not change between 1980 2008-2012 and2008-2012. In addition,among couples in which hus- bands and wives are equals with respect to edu- cation,husbands'income advantage would help PREVIOUS RESEARCH AND THE PRESENT STUDY attain their higher status relative to their wives. Changing patterns of assortative mating may Moreover,I hypothesize that among couples reflect changes in marginal distributions of who share the same education level,the ten- husbands'and wives'traits as well as changes dency for women to marry up in income may in associations between spouses'traits once be less pronounced at both ends of the edu- marginal distributions have been taken into cational distribution.Despite recent declines in account (Hou Myles,2008;Mare,1991; marriage rates especially among less-educated Qian Lichter,2007).For example,the pro- individuals(Cherlin,2010),the symbolic impor- portion of marriages in which the wife had tance of marriage remains high in that marriage more education than the husband may have has become“"a marker of prestige”(Cherlin, increased between 1980 and 2008-2012 mainly 2004,p.855).Individuals from lower social classes require high economic security before because of disproportionate gains in women's educational attainment rather than changes in entering marriage(Cherlin,2004;Smock,Man- ning,Porter,2005).Because less-educated the association between spouses'education. The theoretical frameworks reviewed previ- women tend to have poor economic prospects (Blau,1998),they cannot afford to marry a man ously,however,suggest continuity and change in the association between spouses'traits such who also has poor economic prospects.In con- as education and income,net of changes in trast,higher education provides women with marginal distributions of these traits,during higher earning power.Thus,highly educated women have greater economic independence the 1980 to 2008-2012 period.In addition to and may consider a husband's breadwinning the theoretical relevance,changes in the asso- ciation between spouses'traits are thought of capability as less important than other criteria when seeking a future spouse (Press,2004).In as indicators of changes in social boundaries addition,highly educated men may be more will- between groups and changes in the function of ing than less-educated men to marry women who certain traits in mate selection (Hou Myles, earn more than they do (South,1991)because 2008;Qian Lichter,2011).Given the theo- they hold more egalitarian gender ideologies retical and substantive importance,the current (Davis Greenstein,2009).Yet,at the same study follows prior highly regarded research on time,men with no high school diploma,but not assortative mating (e.g.,Mare,1991;Qian Lichter,2007,2011;Schwartz Mare,2005) their female counterparts,have experienced a dramatic decline in income in recent decades and uses log-linear models to estimate asso- (White Rogers,2000).The economic vulner- ciations between spouses'traits (specifically, ability of men with very low levels of education education and income),independent of the suggests that a pronounced tendency for women changing marginal distributions of these traits. to marry up in income may not be particularly Indeed,among many traits that play a role in evident among couples in which both spouses the choice of a spouse,sociologists have most have less than a high school education.In light often examined assortative marriage patterns of these considerations,I propose the following with respect to education (e.g.,Blossfeld,2009; hypotheses: Mare,1991,2016;Schwartz Mare,2005). Education is multifaceted,reflecting cultural Hypothesis 4a:The tendency for women to marry resources such as values,beliefs,and life styles up in income is generally greater among cou- as well as earnings potential (Bruze,2011; ples who share the same education level than Kalmijn,1994;Sweeney,2002).In addition, among couples in which the wife has less educa- schools provide an institutional setting wherein tion than the husband.This tendency is likely less individuals can interact and build romantic pronounced at both ends of the educational distri- relationships with potential spouses (Kalmijn, bution. 1998).Prior research on educational assortative Hypothesis 4b:The greater tendency for women to marriage in the United States found that from marry up in income among couples who share the 1960 to the early 2000s,men and women
4 Journal of Marriage and Family in education did not change between 1980 and 2008–2012. In addition, among couples in which husbands and wives are equals with respect to education, husbands’ income advantage would help attain their higher status relative to their wives. Moreover, I hypothesize that among couples who share the same education level, the tendency for women to marry up in income may be less pronounced at both ends of the educational distribution. Despite recent declines in marriage rates especially among less-educated individuals (Cherlin, 2010), the symbolic importance of marriage remains high in that marriage has become “a marker of prestige” (Cherlin, 2004, p. 855). Individuals from lower social classes require high economic security before entering marriage (Cherlin, 2004; Smock, Manning, & Porter, 2005). Because less-educated women tend to have poor economic prospects (Blau, 1998), they cannot afford to marry a man who also has poor economic prospects. In contrast, higher education provides women with higher earning power. Thus, highly educated women have greater economic independence and may consider a husband’s breadwinning capability as less important than other criteria when seeking a future spouse (Press, 2004). In addition, highly educated men may be more willing than less-educated men to marry women who earn more than they do (South, 1991) because they hold more egalitarian gender ideologies (Davis & Greenstein, 2009). Yet, at the same time, men with no high school diploma, but not their female counterparts, have experienced a dramatic decline in income in recent decades (White & Rogers, 2000). The economic vulnerability of men with very low levels of education suggests that a pronounced tendency for women to marry up in income may not be particularly evident among couples in which both spouses have less than a high school education. In light of these considerations, I propose the following hypotheses: Hypothesis 4a: The tendency for women to marry up in income is generally greater among couples who share the same education level than among couples in which the wife has less education than the husband. This tendency is likely less pronounced at both ends of the educational distribution. Hypothesis 4b: The greater tendency for women to marry up in income among couples who share the same education level did not change between 1980 and 2008–2012. Previous Research and the Present Study Changing patterns of assortative mating may reflect changes in marginal distributions of husbands’ and wives’ traits as well as changes in associations between spouses’ traits once marginal distributions have been taken into account (Hou & Myles, 2008; Mare, 1991; Qian & Lichter, 2007). For example, the proportion of marriages in which the wife had more education than the husband may have increased between 1980 and 2008–2012 mainly because of disproportionate gains in women’s educational attainment rather than changes in the association between spouses’ education. The theoretical frameworks reviewed previously, however, suggest continuity and change in the association between spouses’ traits such as education and income, net of changes in marginal distributions of these traits, during the 1980 to 2008–2012 period. In addition to the theoretical relevance, changes in the association between spouses’ traits are thought of as indicators of changes in social boundaries between groups and changes in the function of certain traits in mate selection (Hou & Myles, 2008; Qian & Lichter, 2011). Given the theoretical and substantive importance, the current study follows prior highly regarded research on assortative mating (e.g., Mare, 1991; Qian & Lichter, 2007, 2011; Schwartz & Mare, 2005) and uses log-linear models to estimate associations between spouses’ traits (specifically, education and income), independent of the changing marginal distributions of these traits. Indeed, among many traits that play a role in the choice of a spouse, sociologists have most often examined assortative marriage patterns with respect to education (e.g., Blossfeld, 2009; Mare, 1991, 2016; Schwartz & Mare, 2005). Education is multifaceted, reflecting cultural resources such as values, beliefs, and life styles as well as earnings potential (Bruze, 2011; Kalmijn, 1994; Sweeney, 2002). In addition, schools provide an institutional setting wherein individuals can interact and build romantic relationships with potential spouses (Kalmijn, 1998). Prior research on educational assortative marriage in the United States found that from 1960 to the early 2000s, men and women
Educational and Income Assortative Marriage 5 especially those at the top and bottom of the degree of marital sorting on income and the educational distribution,increasingly married share of couples in which the wife had more spouses with similar education (Schwartz education or income than the husband were Mare,2005). severely underestimated if prevailing marriages, Income assortative mating has received far as opposed to newlyweds,were examined less attention than educational assortative mat- To investigate the role of women's wages ing.Yet in recent decades,income may have in assortative mating,Sweeney and Cancian become increasingly important in the selection (2004)took an individual-level approach to of marriage partners.The median age at first examine the association between women's pre- marriage has risen substantially:Between 1980 marital wages and the economic standing of and 2011,it increased from 24.7 to 28.7 for men the men they married.They found an increase and from 22.0 to 26.5 for women (U.S.Cen- between two cohorts of women in the posi- sus Bureau,2011).Individuals wait until they tive correlation between women's premarital attain stable employment and income and even wages and wages of their husbands,and there- wealth (such as savings,a car,or a home)before fore argued that women's economic prospects they get married (Edin Kefalas,2005;Schnei- became more important in determining their der,2011).As individuals marry later,often after marriage prospects.Unfortunately,Sweeney they have established their economic roles,they and Cancian's (2004)study,along with other are more likely to evaluate potential spouses on prior studies that indicated a positive correlation the basis of current incomes rather than future between spouses'earnings (Cancian Reed, economic prospects as proxied by educational 1999;Schwartz,2010),failed to reveal the attainment (Oppenheimer,1988). within-couple difference in earnings.A positive Prior studies examined associations between correlation between spouses'income implies spouses'earnings in prevailing marriages,that that high-earning individuals (both men and is,all existing marriages at the time of a survey, women)marry higher earning spouses than and found a growing resemblance in husbands' low-earning individuals.In fact,a correlation and wives'earnings (Cancian Reed,1999; between spouses'incomes can be strongly posi- Schwartz,2010).These studies did not,how- tive,even if most women marry men with higher ever,adequately assess the role of income in levels of income than themselves (Kalmijn, assortative entry into marriage.Earnings change 1998).Hence,we do not know whether mate for both spouses after marriage (Cooke,Boyle, selection has changed to the point where men Couch,Feijten,2009).The associations and women no longer tend to form marriages between spouses'earnings in prevailing mar- in which the husband has higher income than riages are based on couples at various durations the wife.The present study departs from prior of marriage,and thus are the combined results research by using log-linear models to investi- of spousal resemblance at the time of marriage gate the prevalence of and changes in income formation and postmarriage divisions of house- hypergamy in which women marry men with hold and market labor between the spouses higher incomes than themselves,net of gender (Schwartz,2010).In addition,the associations differences and shifts in income distributions between spouses'earnings in prevailing mar- This study thus evaluates how marriage is gen- riages suffer from bias due to divorce,because dered from the very start by examining gender spouses'relative earnings influence the risk of asymmetry in assortative mating. divorce (Teachman,2010). In contrast,this study offers a precise account of the trends in assortative mating by examining METHOD education and income at the time of marriage among newlyweds in two time periods.Examin- Data and Measurement ing newlyweds is most suitable for investigating I used data from the 5%sample of the 1980 trends in assortative mating and the role of census and the American Community Survey marriage markets in shaping who marries whom (ACS)2008-2012 5-year sample.The data (Kalmijn,1998:Schwartz Mare,2012). came from the Integrated Public Use Microdata Analyzing newlyweds also avoids bias arising Series project at the University of Minnesota from marital dissolution and changes in traits (https://usa.ipums.org/usa/).The 1980 census after marriage.The appendix shows that the and the ACS from 2008 to 2012 are well suited
Educational and Income Assortative Marriage 5 especially those at the top and bottom of the educational distribution, increasingly married spouses with similar education (Schwartz & Mare, 2005). Income assortative mating has received far less attention than educational assortative mating. Yet in recent decades, income may have become increasingly important in the selection of marriage partners. The median age at first marriage has risen substantially: Between 1980 and 2011, it increased from 24.7 to 28.7 for men and from 22.0 to 26.5 for women (U.S. Census Bureau, 2011). Individuals wait until they attain stable employment and income and even wealth (such as savings, a car, or a home) before they get married (Edin & Kefalas, 2005; Schneider, 2011). As individuals marry later, often after they have established their economic roles, they are more likely to evaluate potential spouses on the basis of current incomes rather than future economic prospects as proxied by educational attainment (Oppenheimer, 1988). Prior studies examined associations between spouses’ earnings in prevailing marriages, that is, all existing marriages at the time of a survey, and found a growing resemblance in husbands’ and wives’ earnings (Cancian & Reed, 1999; Schwartz, 2010). These studies did not, however, adequately assess the role of income in assortative entry into marriage. Earnings change for both spouses after marriage (Cooke, Boyle, Couch, & Feijten, 2009). The associations between spouses’ earnings in prevailing marriages are based on couples at various durations of marriage, and thus are the combined results of spousal resemblance at the time of marriage formation and postmarriage divisions of household and market labor between the spouses (Schwartz, 2010). In addition, the associations between spouses’ earnings in prevailing marriages suffer from bias due to divorce, because spouses’ relative earnings influence the risk of divorce (Teachman, 2010). In contrast, this study offers a precise account of the trends in assortative mating by examining education and income at the time of marriage among newlyweds in two time periods. Examining newlyweds is most suitable for investigating trends in assortative mating and the role of marriage markets in shaping who marries whom (Kalmijn, 1998; Schwartz & Mare, 2012). Analyzing newlyweds also avoids bias arising from marital dissolution and changes in traits after marriage. The appendix shows that the degree of marital sorting on income and the share of couples in which the wife had more education or income than the husband were severely underestimated if prevailing marriages, as opposed to newlyweds, were examined. To investigate the role of women’s wages in assortative mating, Sweeney and Cancian (2004) took an individual-level approach to examine the association between women’s premarital wages and the economic standing of the men they married. They found an increase between two cohorts of women in the positive correlation between women’s premarital wages and wages of their husbands, and therefore argued that women’s economic prospects became more important in determining their marriage prospects. Unfortunately, Sweeney and Cancian’s (2004) study, along with other prior studies that indicated a positive correlation between spouses’ earnings (Cancian & Reed, 1999; Schwartz, 2010), failed to reveal the within-couple difference in earnings. A positive correlation between spouses’ income implies that high-earning individuals (both men and women) marry higher earning spouses than low-earning individuals. In fact, a correlation between spouses’ incomes can be strongly positive, even if most women marry men with higher levels of income than themselves (Kalmijn, 1998). Hence, we do not know whether mate selection has changed to the point where men and women no longer tend to form marriages in which the husband has higher income than the wife. The present study departs from prior research by using log-linear models to investigate the prevalence of and changes in income hypergamy in which women marry men with higher incomes than themselves, net of gender differences and shifts in income distributions. This study thus evaluates how marriage is gendered from the very start by examining gender asymmetry in assortative mating. Method Data and Measurement I used data from the 5% sample of the 1980 census and the American Community Survey (ACS) 2008–2012 5-year sample. The data came from the Integrated Public Use Microdata Series project at the University of Minnesota (https://usa.ipums.org/usa/). The 1980 census and the ACS from 2008 to 2012 are well suited
6 Journal of Marriage and Family for this research because they collected infor- income in each time period.Moreover,results mation on respondents'age at first marriage of analyses using wages and salaries were (the census)or whether respondents married substantively the same as those using the total within the past 12 months (the ACS),number income. of times married,and total personal income for Incomes in the 2008-2012 5-year ACS data the previous year.Thus,both data sets allowed file were inflated to 2012 dollars.To perform me to examine newly contracted first marriages log-linear analysis,I have to recode the continu- and to obtain information on both spouses ous income measure into a categorical measure. education and income at the time of marriage. To reduce zero cells while preserving adequate Given the focus on income,I limited my detail in spouses'income (Schwartz,2010), sample to working-age adults.In addition, I classified each individual's income by the as marriage patterns may differ between decile he or she occupied in the income dis- native-borns and immigrants,I only included tribution of the 1980 and 2008-2012 analytic couples in which both spouses were U.S.born. samples,respectively.In other words,income In sum,I used a sample of U.S.-born couples in deciles were defined by ranking all people in which both the husband and wife were aged 18 the period-specific analytic samples by their to 55 years and married for the first time within income.Thus,spouses were classified by time approximately 1 year prior to the census or period based on their income relative to other ACS.Sensitivity analysis (available on request) people irrespective of gender. confirmed that the results did not change if I Similar to prior research (e.g.,Qian,1997; included immigrants or used alternative age Qian Lichter,2007;Schwartz Mare,2012), ranges.After excluding 462 couples in which classified each spouse into one of the four either spouse had negative income or both education levels-less than high school,high spouses had zero income,the final sample sizes school,some college,and college degrees and were 38,016 couples in 1980 and 37,686 couples above.As a robustness check,I experimented in2008-2012. with different classifications of educational More than 80%of newlyweds were non- and income levels (e.g.,income quintiles and Hispanic White couples,and supplementary five education levels)and obtained results analyses of White couples only yielded substan- similar to those reported next.Taken together, tively identical results to those reported next. I produced a five-way table with 3,200 cells Assortative mating patterns might be different (10 Income deciles for husbands x 10 Income for racial or ethnic minority couples considering deciles for wives x4 Education levels for hus- the differences across racial or ethnic groups bands x4 Education levels for wives x2 Time in the gender gap in education,the retreat from periods). marriage,the availability of economically suit- able men,and the marriage pool influenced by the large influx of recent immigrants(Cherlin, Analytical Approach 2010;DiPrete Buchmann,2013;Qian I used log-linear models to examine educa- Lichter,2007,2011:Schoen Cheng,2006). tional and income assortative mating.The chief Sample sizes were too small to separately exam- advantage of log-linear models lies in their abil- ine racial or ethnic minority couples,so I leave ity to estimate associations between spouses' this task to future research.The goal of this characteristics (e.g.,education or income)while article is to provide a general account of gender controlling for husband-wife differences in the asymmetry in assortative mating patterns among marginal distributions of these characteristics U.S.newlyweds. as well as shifts in the marginal distributions Following Cancian and Reed (1999),I (Kalmijn,2010:Qian Lichter,2007,2011; defined an individual's income as his or her total Schwartz Mare,2005;but see Rosenfeld, pretax personal income from all sources for 2005,for a critique of log-linear models).The the previous year.I examined spouses'income first set of models included only educational from all sources rather than their annual wage pairing of spouses.The second set of log-linear and salary earnings because the total income models added associations between spouses' reflects individuals'overall economic quality. income.Finally,I examined how education On average,individuals'wages and salaries interacts with income to shape assortative constituted more than 90%of their own total mating patterns
6 Journal of Marriage and Family for this research because they collected information on respondents’ age at first marriage (the census) or whether respondents married within the past 12 months (the ACS), number of times married, and total personal income for the previous year. Thus, both data sets allowed me to examine newly contracted first marriages and to obtain information on both spouses’ education and income at the time of marriage. Given the focus on income, I limited my sample to working-age adults. In addition, as marriage patterns may differ between native-borns and immigrants, I only included couples in which both spouses were U.S. born. In sum, I used a sample of U.S.-born couples in which both the husband and wife were aged 18 to 55 years and married for the first time within approximately 1 year prior to the census or ACS. Sensitivity analysis (available on request) confirmed that the results did not change if I included immigrants or used alternative age ranges. After excluding 462 couples in which either spouse had negative income or both spouses had zero income, the final sample sizes were 38,016 couples in 1980 and 37,686 couples in 2008–2012. More than 80% of newlyweds were nonHispanic White couples, and supplementary analyses of White couples only yielded substantively identical results to those reported next. Assortative mating patterns might be different for racial or ethnic minority couples considering the differences across racial or ethnic groups in the gender gap in education, the retreat from marriage, the availability of economically suitable men, and the marriage pool influenced by the large influx of recent immigrants (Cherlin, 2010; DiPrete & Buchmann, 2013; Qian & Lichter, 2007, 2011; Schoen & Cheng, 2006). Sample sizes were too small to separately examine racial or ethnic minority couples, so I leave this task to future research. The goal of this article is to provide a general account of gender asymmetry in assortative mating patterns among U.S. newlyweds. Following Cancian and Reed (1999), I defined an individual’s income as his or her total pretax personal income from all sources for the previous year. I examined spouses’ income from all sources rather than their annual wage and salary earnings because the total income reflects individuals’ overall economic quality. On average, individuals’ wages and salaries constituted more than 90% of their own total income in each time period. Moreover, results of analyses using wages and salaries were substantively the same as those using the total income. Incomes in the 2008–2012 5-year ACS data file were inflated to 2012 dollars. To perform log-linear analysis, I have to recode the continuous income measure into a categorical measure. To reduce zero cells while preserving adequate detail in spouses’ income (Schwartz, 2010), I classified each individual’s income by the decile he or she occupied in the income distribution of the 1980 and 2008–2012 analytic samples, respectively. In other words, income deciles were defined by ranking all people in the period-specific analytic samples by their income. Thus, spouses were classified by time period based on their income relative to other people irrespective of gender. Similar to prior research (e.g., Qian, 1997; Qian & Lichter, 2007; Schwartz & Mare, 2012), I classified each spouse into one of the four education levels—less than high school, high school, some college, and college degrees and above. As a robustness check, I experimented with different classifications of educational and income levels (e.g., income quintiles and five education levels) and obtained results similar to those reported next. Taken together, I produced a five-way table with 3,200 cells (10 Income deciles for husbands × 10 Income deciles for wives × 4 Education levels for husbands × 4 Education levels for wives × 2 Time periods). Analytical Approach I used log-linear models to examine educational and income assortative mating. The chief advantage of log-linear models lies in their ability to estimate associations between spouses’ characteristics (e.g., education or income) while controlling for husband–wife differences in the marginal distributions of these characteristics as well as shifts in the marginal distributions (Kalmijn, 2010; Qian & Lichter, 2007, 2011; Schwartz & Mare, 2005; but see Rosenfeld, 2005, for a critique of log-linear models). The first set of models included only educational pairing of spouses. The second set of log-linear models added associations between spouses’ income. Finally, I examined how education interacts with income to shape assortative mating patterns
Educational and Income Assortative Marriage 1 To begin,my basic model is as follows: log ()=Model 1 + 1og(/）=+EY+Er +层+， (2) ++增+m "iit where r is a set of parameter estimates for homogamy of each educational group +r, (1) (0=1 when i=k=1,...,0=4 when i=k=4, and =0otherwise),and is the educa- where HE is husband's education (i=1,...4). tion hypogamy parameter (P=1 when iI and 0 oth- (Schwartz Mare,2005).In 14.25%of cells with counts of 0(i.e.,456 of 3,200),I settikr to 1 erwise).and represent changes in the odds of income homogamy and hypergamy. (Schwartz Mare,2005).Empty cells need not respectively,between 1980 and 2008-2012,net be problematic in log-linear analyses (Agresti, of shifts in the marginal distributions of hus- 2002).As a robustness check,I added 0.5 to bands'and wives'income deciles. each cell and obtained substantively the same Finally,to test Hypotheses 3a,3b,4a,and results. 4b,I examined how the gender asymmetry in To test Hypothesis 1,I modeled the associa- income assortative mating differed by the educa- tions between husbands'and wives'education. tional pairing of spouses.I included interaction I modeled the odds of educational homogamy terms between the income assortative mating and hypogamy relative to the odds of educational parameters and the educational homogamy and hypergamy by adding variable diagonal param- hypogamy parameters as well as changes in eters (Qian,1997)and a hypogamy parameter these interaction terms by year.The model is that was coded 1 if the wife had more educa- tion than the husband.Then I added interaction log ()Model3+ terms between the year and the variable diag- onal parameters and between the year and the ++层 hypogamy parameter to model changes in edu- cational assortative mating.The model becomes + (4)
Educational and Income Assortative Marriage 7 To begin, my basic model is as follows: log ( 𝜇ijklt∕tijklt) = 𝜆 + 𝜆HEY it + 𝜆WEY kt + 𝜆HIY jt + 𝜆WIY lt + 𝜆HEHIY ijt + 𝜆WEWIY klt , (1) where HE is husband’s education (i=1, …, 4), WE is wife’s education (k =1, …, 4), HI is husband’s income category (j=1, …, 10), WI is wife’s income category (l=1, …, 10), and Y is period (t =1, 2). Thus, 𝜇ijklt is the expected number of marriages between men with education i in income decile j and women with education k in income decile l in period t. This model includes variations in the distributions of husband’s and wife’s education and income by year (𝜆HEY it , 𝜆WEY kt , 𝜆HIY jt , 𝜆WIY lt ), the associations between education and income for both husbands and wives and their variations by year (𝜆HEHI ij , 𝜆WEWI kl , 𝜆HEHIY ijt , 𝜆WEWIY klt ), and all lower order terms. The 1980 census is self-weighting, whereas the ACS 2008–2012 5-year sample contains weights to ensure that the multiyear sample is representative of the population during the entire 5-year period. I incorporated the weights by an offset tijklt, which is the inverse of the total weighted frequency of the cell divided by the unweighted cell count (Agresti, 2002, p. 391). To preserve the original sample size in the ACS sample, I rescaled the original weights so that the sum of the weights equaled the sample size (Schwartz & Mare, 2005). In 14.25% of cells with counts of 0 (i.e., 456 of 3,200), I set tijklt to 1 (Schwartz & Mare, 2005). Empty cells need not be problematic in log-linear analyses (Agresti, 2002). As a robustness check, I added 0.5 to each cell and obtained substantively the same results. To test Hypothesis 1, I modeled the associations between husbands’ and wives’ education. I modeled the odds of educational homogamy and hypogamy relative to the odds of educational hypergamy by adding variable diagonal parameters (Qian, 1997) and a hypogamy parameter that was coded 1 if the wife had more education than the husband. Then I added interaction terms between the year and the variable diagonal parameters and between the year and the hypogamy parameter to model changes in educational assortative mating. The model becomes log ( 𝜇ijklt∕tijklt) = Model 1 + 𝛾O ik + 𝛾P p + 𝛾O ik𝜆Y t + 𝛾P p 𝜆Y t , (2) where 𝛾O ik is a set of parameter estimates for homogamy of each educational group (O =1 when i=k =1, …, O =4 when i=k =4, and O =0 otherwise), and 𝛾P p is the education hypogamy parameter (P=1 when il and 0 otherwise). 𝛿S jl𝜆Y t and 𝛿Q q 𝜆Y t represent changes in the odds of income homogamy and hypergamy, respectively, between 1980 and 2008–2012, net of shifts in the marginal distributions of husbands’ and wives’ income deciles. Finally, to test Hypotheses 3a, 3b, 4a, and 4b, I examined how the gender asymmetry in income assortative mating differed by the educational pairing of spouses. I included interaction terms between the income assortative mating parameters and the educational homogamy and hypogamy parameters as well as changes in these interaction terms by year. The model is log ( 𝜇ijklt∕tijklt) = Model 3 + 𝛾O ik 𝛿S jl𝜆Y t + 𝛾P p 𝛿S jl𝜆Y t + 𝛾O ik 𝛿Q q 𝜆Y t + 𝛾P p 𝛿Q q 𝜆Y t , (4)
8 Journal of Marriage and Family where the key parameters of interest are deciles in both time periods.For example,less and that examine the tendency for women than 5%of husbands but more than 15%of to marry up in income among couples in which wives were in the lowest income decile (=1)in the wife has education equal to and greater both time periods. Descriptive results in Table 1 show that the than the husband,respectively.versus those in which the wife has less education than the percentage of couples in which the husband had husband. more education than the wife declined from 24% in 1980 to 15%in 2008-2012,whereas the share of couples in which the wife had more educa- RESULTS tion than the husband increased from 22%to Descriptive Results 29%during the same period.If two spouses dif- fered in educational attainment,in 1980 the hus- Descriptive results in Figures 1 and 2 show band was more likely to have more education, substantial changes in the distributions of edu- but in 2008-2012 the wife became more likely cation and income among newlywed men and to be the more-educated spouse.From 1980 to women.Figure I shows that educational attain- 2008-2012,the share of couples in which the ment increased for both spouses from 1980 to 2008-2012,but increased more for wives than husband was in a higher income decile than the for husbands.In 1980,about 20%of husbands wife declined,whereas the share of couples in and 17%of wives had a college degree or above, which the wife was in the same or a higher whereas in 2008-2012,about 42%of wives were income decile relative to the husband increased. in this educational group compared with 34%of Notably,for most couples,husbands were in a husbands. higher income decile than their wives regardless Figure 2 shows that income increased more of the time period and the educational pairing of rapidly for wives than for husbands.Percent- spouses.Overall,the descriptive results revealed ages of wives in low income deciles declined. a trend toward wives being more educated than whereas those in high income deciles increased. their husbands coexisting with the traditional The opposite trend was true for husbands. pattern of wives marrying higher income hus- Despite wives'greater gains in income during bands. the decades,wives were still underrepresented These patterns and trends in Table 1 should in higher deciles and overrepresented in lower be interpreted with caution,however,because FIGURE I.PERCENTAGE DISTRIBUTIONS OF EDUCATIONAL ATTAINMENT.BY GENDER AND YEAR. 100% 90% 19.77 16.60 80% 34.38 41.67 70% 28.05 29.53 60% 50% 34.30 40% 35.49 30% 37.61 41.44 20% 24.82 10% 17.93 0% 险0网 498 1980 2008-2012 1980 2008-2012 Husbands Wives Less than hish school ▣High school ☐Some college ■College and above Note.Totals may not sum to 100.00 due to rounding errors.Data were weighted in 2008-2012
8 Journal of Marriage and Family where the key parameters of interest are 𝛾O ik 𝛿Q q and 𝛾P p 𝛿Q q that examine the tendency for women to marry up in income among couples in which the wife has education equal to and greater than the husband, respectively, versus those in which the wife has less education than the husband. Results Descriptive Results Descriptive results in Figures 1 and 2 show substantial changes in the distributions of education and income among newlywed men and women. Figure 1 shows that educational attainment increased for both spouses from 1980 to 2008–2012, but increased more for wives than for husbands. In 1980, about 20% of husbands and 17% of wives had a college degree or above, whereas in 2008–2012, about 42% of wives were in this educational group compared with 34% of husbands. Figure 2 shows that income increased more rapidly for wives than for husbands. Percentages of wives in low income deciles declined, whereas those in high income deciles increased. The opposite trend was true for husbands. Despite wives’ greater gains in income during the decades, wives were still underrepresented in higher deciles and overrepresented in lower deciles in both time periods. For example, less than 5% of husbands but more than 15% of wives were in the lowest income decile (=1) in both time periods. Descriptive results in Table 1 show that the percentage of couples in which the husband had more education than the wife declined from 24% in 1980 to 15% in 2008–2012, whereas the share of couples in which the wife had more education than the husband increased from 22% to 29% during the same period. If two spouses differed in educational attainment, in 1980 the husband was more likely to have more education, but in 2008–2012 the wife became more likely to be the more-educated spouse. From 1980 to 2008–2012, the share of couples in which the husband was in a higher income decile than the wife declined, whereas the share of couples in which the wife was in the same or a higher income decile relative to the husband increased. Notably, for most couples, husbands were in a higher income decile than their wives regardless of the time period and the educational pairing of spouses. Overall, the descriptive results revealed a trend toward wives being more educated than their husbands coexisting with the traditional pattern of wives marrying higher income husbands. These patterns and trends in Table 1 should be interpreted with caution, however, because Figure 1. Percentage Distributions of Educational Attainment, by Gender and Year. Some college College and above 14.56 6.50 12.43 4.92 37.61 24.82 41.44 17.93 28.05 34.30 29.53 19.77 35.49 34.38 16.60 0% 41.67 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1980 19 2008-2012 200 80 8-2012 Husbands Wives Less than hish school High school Note. Totals may not sum to 100.00 due to rounding errors. Data were weighted in 2008–2012
Educational and Income Assortative Marriage 9 FIGURE 2.PERCENTAGE DISTRIBUTIONS OF INCOME DECILE,BY GENDER AND YEAR. Husbands Wives 20 15 10 91012 10 Income Decile ×—1980 —◆—2008.2012 Note.Data were weighted in 2008-2012.Husbands and wives are presented separately for clarity,but their income deciles were calculated based on the pooled sample of both sexes by time period.1=the lowest decile:10=the highest decile. Table 1.Cross-Classification of Educational and Income Assortative Marriage,by Year Income assortative marriage Educational assortative marriage Hypogamy Homogamy Hypergamy Total 1980 Hypogamy 17.42 12.31 70.27 21.91 Homogamy 14.64 12.80 72.56 54.14 Hypergamy 12.76 9.56 77.68 23.95 Total 14.80 11.92 73.28 38,016 2008-2012 Hypogamy 30.97 16.74 52.29 28.69 Homogamy 22.50 1935 58.14 56.13 Hypergamy 15.23 12.95 71.83 15.18 Total 23.83 17.63 58.54 37,686 Note.Hypogamy is marriage in which the wife is more educated or in a higher income decile than the husband.Homogamy is marriage in which two spouses share the same level of education or income.Hypergamy is marriage in which the husband is more educated or in a higher income decile than the wife.Numbers in bold are row percentages,indicating percentages of income assortative marriages by the educational pairing of spouses.Data were weighted in 2008-2012. they may be highly influenced by the educa- in which the wife had more education or income tional and income distributions of husbands and than the husband may have increased between wives as well as changes in these distributions 1980 and 2008-2012 because of the dispro- (Gullickson Fu,2010;Hou Myles,2008; portionate increase in education and income Hout,1983;Kalmijn,2010;Schwartz Mare, for women.Relatedly,as income goes up with 2005).For example,the percentage of couples education for both sexes,women will often
Educational and Income Assortative Marriage 9 Figure 2. Percentage Distributions of Income Decile, by Gender and Year. 0 5 10 15 20 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Husbands Wives 1980 2008-2012 Percentage Income Decile Note. Data were weighted in 2008–2012. Husbands and wives are presented separately for clarity, but their income deciles were calculated based on the pooled sample of both sexes by time period. 1=the lowest decile; 10=the highest decile. Table 1. Cross-Classification of Educational and Income Assortative Marriage, by Year Income assortative marriage Educational assortative marriage Hypogamy Homogamy Hypergamy Total 1980 Hypogamy 17.42 12.31 70.27 21.91 Homogamy 14.64 12.80 72.56 54.14 Hypergamy 12.76 9.56 77.68 23.95 Total 14.80 11.92 73.28 38,016 2008–2012 Hypogamy 30.97 16.74 52.29 28.69 Homogamy 22.50 19.35 58.14 56.13 Hypergamy 15.23 12.95 71.83 15.18 Total 23.83 17.63 58.54 37,686 Note. Hypogamy is marriage in which the wife is more educated or in a higher income decile than the husband. Homogamy is marriage in which two spouses share the same level of education or income. Hypergamy is marriage in which the husband is more educated or in a higher income decile than the wife. Numbers in bold are row percentages, indicating percentages of income assortative marriages by the educational pairing of spouses. Data were weighted in 2008–2012. they may be highly influenced by the educational and income distributions of husbands and wives as well as changes in these distributions (Gullickson & Fu, 2010; Hou & Myles, 2008; Hout, 1983; Kalmijn, 2010; Schwartz & Mare, 2005). For example, the percentage of couples in which the wife had more education or income than the husband may have increased between 1980 and 2008–2012 because of the disproportionate increase in education and income for women. Relatedly, as income goes up with education for both sexes, women will often
10 Journal of Marriage and Family be less likely to marry a husband with higher I examined educational assortative mating income than themselves when they marry down patterns by fitting Models 2 through 4.To cap- in education when compared with when they ture the tendency for individuals to marry within marry up in education.Thus,gender inequali- their educational groups,I added four param- ties in education and income,income differen- eters along the main educational diagonal in tials across educational and gender groups,and Model 2.The great reduction in the BIC statistic changes in these distributions determine the mar- indicated a strong tendency for individuals to ital "opportunity structure"for men and women marry similarly educated spouses.In Model in the marriage markets (Hou Myles,2008, 3,adding a uniform educational hypogamy p.338).Log-linear models allow me to examine parameter decreased the BIC relative to Model whether the gender asymmetry in educational 2. indicating the gender asymmetric nature and income assortative mating held and whether of educational assortative mating patterns.To there were changes in the asymmetry after con- model changes in educational assortative mat- trolling for the compositional effects such as ing,I added interaction terms between year gender differences and shifts in the marginal dis- and the educational homogamy and hypogamy tributions of education and income. parameters in Model 4,which decreased the BIC.It suggested that the tendency for individ- Results of Log-Linear Models uals to marry within their educational groups (homogamy)and the tendency for women to Table 2 provides the model specifications and marry less-educated husbands (hypogamy) the goodness-of-fit statistics.I present both the changed from 1980 to 2008-2012 even net of deviance and the Bayesian information criterion shifts in the marginal distributions of spouses' (BIC)statistics for model fit,but mainly focus education. on the BIC due to the large sample size (Gul- Next,in Models 5 through 7,I investigated lickson Torche,2014).A smaller value of the patterns of income assortative marriage.In BIC indicates a better-fitting model,and the evi- Model 5,I added 10 dummy variables that dence in favor of the model with more negative indicated the tendency for individuals to marry BIC is considered to be very strong if the BIC within the same income decile.By the BIC, difference between two models is greater than the income homogamy parameters improved 10 (Raftery,1986,1995).Model 1,the baseline the model fit.To examine gender asymmetry in model,is described by Equation 1.This model income assortative mating,I added a uniform assumed no association between the husband's income hypergamy parameter to Model 6.When and the wife's characteristics.Not surprisingly, compared with Model 5,there was a reduction the BIC for Model 1 was much larger than 0, in the BIC in Model 6.Thus,the association indicating a poor model fit. between spouses'income remained gender Table 2.Fit Statistics for Log-Linear Models of Educational and Income Assortative Marriage Model df Deviance BIC 1 Marginals×Y+HExH,×Y+WE×W。×Y 3.042 49.843.63 15.668.09 2 Model 1+EO 3.03820.992.93 -13.137.66 3 Model 2+EHypo 3.037 20.894.95 -13,224.41 4 Model 3+EOxY+EHypoxY 3.03220.432.08 -13.631.11 5 Model 4+IO 3.022 16.931.93 -17.018.91 6 Model 5+IHyper 3.021 15.01437 -18.925.24 7 Model 6+IOxY+IHyper xY 3.01014.849.40 -18.966.63 8 Model7+EO×IO+EHypo×IO+EO×Hyper+EHypo x IHyper 2.95514.177.71 -19.020.42 9 Model8+EOxIO×Y+EHypo×IO×Y+EOx IHyper×Y+EHypo x IHyper x Y 2.90014.039.72 -18,540.50 Note.N=38.016+37.686=75,702;cells =3,200.df =degrees of freedom.Model terms are as follows (dfs are in parentheses):HE=husbands'education (3):WE=wives'education (3):H=husbands'income decile (9):W=wives' income decile(9):Y=year(1):EO=variable diagonal parameters indicating educational homogamy(4):EHypo =educational hypogamy parameter(1):IO=variable diagonal parameters indicating income homogamy (10):IHyper=income hypergamy parameter (1)
10 Journal of Marriage and Family be less likely to marry a husband with higher income than themselves when they marry down in education when compared with when they marry up in education. Thus, gender inequalities in education and income, income differentials across educational and gender groups, and changes in these distributions determine the marital “opportunity structure” for men and women in the marriage markets (Hou & Myles, 2008, p. 338). Log-linear models allow me to examine whether the gender asymmetry in educational and income assortative mating held and whether there were changes in the asymmetry after controlling for the compositional effects such as gender differences and shifts in the marginal distributions of education and income. Results of Log-Linear Models Table 2 provides the model specifications and the goodness-of-fit statistics. I present both the deviance and the Bayesian information criterion (BIC) statistics for model fit, but mainly focus on the BIC due to the large sample size (Gullickson & Torche, 2014). A smaller value of the BIC indicates a better-fitting model, and the evidence in favor of the model with more negative BIC is considered to be very strong if the BIC difference between two models is greater than 10 (Raftery, 1986, 1995). Model 1, the baseline model, is described by Equation 1. This model assumed no association between the husband’s and the wife’s characteristics. Not surprisingly, the BIC for Model 1 was much larger than 0, indicating a poor model fit. I examined educational assortative mating patterns by fitting Models 2 through 4. To capture the tendency for individuals to marry within their educational groups, I added four parameters along the main educational diagonal in Model 2. The great reduction in the BIC statistic indicated a strong tendency for individuals to marry similarly educated spouses. In Model 3, adding a uniform educational hypogamy parameter decreased the BIC relative to Model 2, indicating the gender asymmetric nature of educational assortative mating patterns. To model changes in educational assortative mating, I added interaction terms between year and the educational homogamy and hypogamy parameters in Model 4, which decreased the BIC. It suggested that the tendency for individuals to marry within their educational groups (homogamy) and the tendency for women to marry less-educated husbands (hypogamy) changed from 1980 to 2008–2012 even net of shifts in the marginal distributions of spouses’ education. Next, in Models 5 through 7, I investigated patterns of income assortative marriage. In Model 5, I added 10 dummy variables that indicated the tendency for individuals to marry within the same income decile. By the BIC, the income homogamy parameters improved the model fit. To examine gender asymmetry in income assortative mating, I added a uniform income hypergamy parameter to Model 6. When compared with Model 5, there was a reduction in the BIC in Model 6. Thus, the association between spouses’ income remained gender Table 2. Fit Statistics for Log-Linear Models of Educational and Income Assortative Marriage Model df Deviance BIC 1 Marginals × Y + HE×Hp × Y + WE × Wp × Y 3,042 49,843.63 15,668.09 2 Model 1+EO 3,038 20,992.93 −13,137.66 3 Model 2+EHypo 3,037 20,894.95 −13,224.41 4 Model 3+EO × Y +EHypo × Y 3,032 20,432.08 −13,631.11 5 Model 4+IO 3,022 16,931.93 −17,018.91 6 Model 5+IHyper 3,021 15,014.37 −18,925.24 7 Model 6+IO × Y +IHyper × Y 3,010 14,849.40 −18,966.63 8 Model 7+EO × IO +EHypo × IO + EO × IHyper+EHypo × IHyper 2,955 14,177.71 −19,020.42 9 Model 8+EO × IO × Y +EHypo × IO × Y + EO × IHyper × Y +EHypo × IHyper × Y 2,900 14,039.72 −18,540.50 Note. N =38,016+37,686=75,702; cells=3,200. df =degrees of freedom. Model terms are as follows (dfs are in parentheses): HE=husbands’ education (3); WE= wives’ education (3); Hp =husbands’ income decile (9); Wp = wives’ income decile (9); Y =year (1); EO =variable diagonal parameters indicating educational homogamy (4); EHypo=educational hypogamy parameter (1); IO =variable diagonal parameters indicating income homogamy (10); IHyper=income hypergamy parameter (1)