Toward a More Nuanced Understanding of Intercoupling

Abstract

Distinguishing between interethnic and interracial coupling offers new understandings about family diversity, as well as the assimilation of immigrants and their children. An analysis of the Children of Immigrants Longitudinal Study, a survey (N = 1,480) of the children of Asian and Latin American immigrants who live in Southern California, finds that a more accurate conceptualization of mixed couples should distinguish between interethnic and interracial couples. Although race remains a significant boundary in relationship formation, ethnicity is also important, especially in terms of language and religion. This study has demonstrated that race and ethnicity are both important social boundaries for understanding mixed couples and denote different types of mixed couples—interethnic versus interracial couples—and different types of lived experiences. At the very least, this analysis of mixed couples calls for the inclusion of interethnic couples within studies of intermarriage.

Keywords

family diversity intermarriage intercoupling assimilation acculturation language religion

Introduction

Two major trends contribute to greater diversity: immigration and intermarriage (Hidalgo & Bankston, 2010; Weaver, Umana-Taylor, Hans, & Malia, 2001). Partially because of immigration, existing research on intermarriage does not capture the diversity of couples who cross racial and ethnic boundaries. Immigrants migrating to the United States and forming relationships with natives “may disrupt traditional family processes” (Clark, Glick, & Bures, 2009, p. 853), especially in terms of assimilation patterns. Intermarriage is a litmus test often used by scholars for determining whether social groups are assimilating and reducing the social distance that separates them. Is it important to distinguish between racial (perceived physical differences) and ethnic (ancestry, history, and culture) social groups? What constitutes intermarriage has varied depending on the questions asked and the data sets available. Given the social changes of the last 40 years, now is a good time to reconsider how intermarriage is conceptualized and studied. Focusing on immigration and intermarriage simultaneously yields a richer understanding of family diversity.

Contemporary studies on intermarriage generally focus on interracial marriages. Interethnic relationships are only just beginning to be studied in a post-1960 era of mass migration. Studies typically focus on interracial relationships and, by default, lump interethnic relationships with co-ethnic relationships (couples who share the same race and ethnicity). This study challenges that assumption. This is especially important given that immigration is changing the racial and ethnic composition of the United States. This influx is likely to erode racial and ethnic boundaries, thus leading to larger numbers of couples that are mixed, either racially or ethnically or both.

This study uses a more nuanced approach to studying intermarriage. I introduce intercoupling as a concept for understanding key differences between interethnic and interracial couples, as well as differences between married, cohabiting, and dating couples. This study is somewhat exploratory in the sense that we do not know a lot about the specific differences between these various types of mixed relationships. Are interethnic and interracial couples similar, or two different types of mixed couples? This overarching question guides this study.

More specifically, I analyze the Children of Immigrants Longitudinal Study (CILS), which is a survey (N = 1480) of the children of Asian and Latin American immigrants (or de facto second generation¹) who live in Southern California. This survey addresses how the respondents think about mixed relationships, with regard to both race and ethnicity, and more specifically, with regard to language and religion. Language and religion both play an important role in how the respondents think about ethnic differences. Overall, the survey focuses on a variety of factors to address two questions: first, how are interethnic and interracial couples different from each other; and second, how are interethnic and interracial couples different from co-ethnic couples.

This article is divided into four sections. The first section highlights how this study will differ from typical intermarriage studies. This discussion includes key differences between race and ethnicity as it relates to intermarriage. The second section illuminates how the CILS data set is uniquely qualified to offer new insights into mixed relationships. The third section outlines the key findings from this study. The survey data reveals how religious differences affect interethnic and interracial relationships, as well as linguistic differences that can be thought of as a form of acculturation to the host society for these children of immigrants. In the concluding section, I explore how these findings help us understand how immigrants and their children are assimilating into our society, specifically through the formation of mixed relationships and what that means in terms of family diversity.

Relationships, Ethnicity, and Race

This study adds to the current literature on intermarriage in three important ways: First, rather than limit this study to married relationships, attention is refocused on intercoupling. Intercoupling draws attention to three types of modern relationships: marriage (and engagement), cohabitation, and dating. Second, the study examines intercoupling among the children of Asian and Latin American immigrants. Third, I make a key distinction between interethnic and interracial relationships. Intercoupling is synonymous with mixed coupling. Thus, co-ethnic couples share a relationship with someone of the same racial and ethnic identity, whereas mixed couples are either racially or ethnically mixed.

Intercoupling

The focus on couples rather than marriages is timely given increasing cohabitation rates since the 1990s (Smock, 2000). Studies point to important differences between marriage and cohabitation (Blackwell & Lichter, 2000; Guzzo, 2006; Lamanna & Riedmann, 2008); the qualitatively different mate selection processes of cohabiting couples cannot be ignored (Blackwell & Lichter, 2000). Patterns of cohabitation differ by culture and by national-origin (Landale, Oropesa, & Bradatan, 2006; Manning & Landale, 1996; Oropesa, 1996), and cohabiting couples are twice as likely as married couples to be in a relationship with someone of a different racial or ethnic background (Fields & Casper, 2001; Simmons & O’Connell, 2003). Dating patterns are also likely to be qualitatively different than marriage and cohabiting patterns (Joyner & Kao, 2005; McClintock, 2010). To the extent that dating leads to cohabitation and marriage, dating relationships are an important but relatively understudied aspect of intercoupling.

Asian and Latin American Immigrants

The largest immigrant flows in the United States over the past several decades have been from Latin America and Asia, comprising 78% of immigrants (Portes & Rumbaut, 2006). With increasing immigration from Asia and Latin America, the predominance of Black–White interracial marriage is changing (Vermeulen, 1999). According to the U.S. Census, the immigrant-origin population (first and second generations) is close to 69 million—almost one fourth of the national population (Current Population Survey [CPS], 2003-2006). One third of this population resides in California (Foner, Rumbaut, & Gold, 2000), whereas California makes up only 12% of the total U.S. population. In 2000, San Diego was among the top 10 metropolitan areas in the United States in numbers and percentage of first- and second-generation Asians and Latinos. Specifically, San Diego is also one of the most common destinations for immigrants from the Philippines, Laos, Cambodia, Vietnam, and Mexico (and to a lesser extent China, Taiwan, Japan, Korea, and India; CPS, 2003-2006).

Mixed Couples

Using Cornell and Hartmann’s (2006) definition of race and Schermerhorn’s (1979) definition of ethnicity, I treat race and ethnicity as two separate, but sometimes overlapping, group characteristics. Racial groups are defined by the group or by others based on perceived common physical characteristics held to be inherent (e.g., White, Asian, Black). Ethnic groups are collectivities within the larger society defined by common ancestry, language, religion, and cultural focus.

This study compares co-ethnic couples (couples who share the same race and ethnicity) with two distinctive types of mixed couples—interracial couples (e.g., Asian–White or Latino–Black) and interethnic couples (e.g., Filipino–Chinese or Mexican–Peruvian). To my knowledge, empirical studies comparing the differences between interethnic and interracial couples are unavailable. Most studies on racial intermarriage ignore marriages that I would term interethnic (e.g., Japanese–Korean or Cuban–Mexican).

The conflation of racial intermarriage and ethnic intermarriage

Most intermarriage studies focus on racial intermarriage (Alba & Nee, 2003; Bean & Stevens, 2003; Blumberg & Roye, 1979; Larsson, 1965; McNamara, Tempenis, & Walton, 1999; Qian & Lichter, 2007; Rosenblatt, Karis, & Powell, 1995). Consider two of these intermarriage studies and the implications their findings have for making predictions about assimilation.

In a study of children of Japanese ancestry born in the 1980s, Alba and Nee (2003) reported that 63% had “mixed ethnic ancestry, suggesting the erosion of a racial boundary” (p. 94). This statement may be accurate if race is understood as a subcategory of ethnicity. If, however, ethnicity is a subcategory of race, then predicting the erosion of a racial boundary requires a closer examination of ethnic groups. Individuals of Japanese descent may choose from among different coupling patterns. They may couple within ethnic and racial boundaries and thus form relationships with other Japanese. Or, they may cross interethnic boundaries and marry other Asians. Or, they may cross interracial boundaries and marry Whites, Blacks, or people of other races.

In fact, there is some evidence that Asian men and women are outmarrying² across ethnic boundaries. A dramatic increase in interethnic marriages has been accompanied by a relative decrease in interracial marriages. Significant increases in the size of the foreign-born population and their greater propensity to marry within their own or other Asian groups may account for the rise in interethnic marriages (Shinagawa & Pang, 1996). And yet only a limited number of studies have examined interethnic marriages (Gilbertson, Fitzpatrick, & Yang, 1996; Jiobu, 1988; Kalmijn & Van Tubergen, 2010; Perlmann, 1997; Qian, Blair, & Ruf, 2001; Qian & Cobas, 2004; Qian & Lichter, 2007; Shinagawa & Pang, 1988, 1996).

To their credit, Alba and Nee (2003) point out,

The out-marriage rate of [Japanese Americans] is such that without sizable new immigration from Japan, which is unlikely, this ethnic group appears to be on the road to amalgamation with Whites and, to a lesser but growing extent because of pan-Asian marriages, other Asian groups. (p. 93)

But they assume that race and ethnicity are interchangeable. However, if an increasingly larger percentage of Japanese Americans marry other Asians, then the Japanese may not be “on the road to amalgamation” at all. The eroding ethnic distinctions may, in fact, reinforce racial distinctions. This is an example of what Yinger (1994) refers to as dissimilation. If Asians are less willing to cross racial lines than to cross ethnic lines, the outcome may be a form of dissimilation or segmented assimilation rather than a linear assimilation path. Crossing racial lines clearly denotes a type of assimilation and declining social distance, but if Japanese Americans cross ethnic lines more readily than racial lines, the result is not full assimilation.

Bean and Stevens (2003) report that 21% of “Asian wives” and 18% of “Latino wives” have husbands of different “race or ancestry” and that 10% of “Asian husbands” and 15% of “Latino husbands” have wives of different race or ancestry. The use of panethnic labels places the focus on race and diminishes the importance of ethnic distinctions between the various groups. What does “Asian” or “Latino” mean? Common race or ancestry among Latinos ignores the variety of ethnicities represented: Mexicans, Cubans, and a growing number of immigrants from Central and South America as well as the Dominican Republic. The Asian panethnic label ignores differences between Chinese, Koreans, Japanese, and Filipinos. Intermarriage is not fully explored when scholars use panethnic terms such as Asian or Latino.

Less is known about Latin American–origin groups, but there is evidence that they marry other Latin American–origin groups at high rates (Gilbertson et al., 1996; Qian & Cobas, 2004). These findings do not discredit Alba and Nee (2003) or Bean and Stevens (2003) but instead make measuring social and cultural assimilation using data on intermarriage problematic. The increase in interethnic marriages among Asians and Latinos could simply be “viewed as part of a larger process of assimilation in which social barriers created by ethnic boundaries attenuate as social relations expand across ethnic groups” (Alba & Nee, 2003, p. 263-264), or it could mean that crossing ethnic lines is easier than crossing racial lines and thus signals a more complex assimilation path over time.

Few studies examine ethnic intermarriages (Fu, 2007; Gilbertson et al., 1996; Jiobu, 1988; Perlmann, 1997; Qian et al., 2001; Qian & Cobas, 2004; Qian & Lichter, 2007; Shinagawa & Pang, 1988, 1996). For example, Gilbertson et al. (1996) examined intermarriage among Latinos in New York, finding a high rate of intermarriage between Dominicans and Puerto Ricans. They concluded that poverty among both Dominicans and Puerto Ricans and a distinct intermarriage pattern pointed “to a cleavage within the broader Hispanic population of New York City and argue against an undifferentiated Hispanic melting pot interpretation of intermarriage patterns in New York City” (Gilbertson et al., 1996, p. 457). Qian et al. (2001) and Qian and Cobas (2004) used the 1990 census to confirm that large numbers of immigrants from Asia and Latin America and their descendants were marrying across ethnic lines within panethnic racial categories. In a comparison of 1990 and 2000 census data, Qian and Lichter (2007) reported a decline in intermarriage with Whites and substantial increases in marriages between native- and foreign-born co-ethnics. The changing patterns of intermarriage were associated with growing numbers of immigrants and with ethnic native populations marrying foreign-born ethnics. Unfortunately, given the focus on race, it is difficult to know whether the changes are the result of co-ethnic marriages or interethnic marriages.

Research Questions

Understanding intermarriage dynamics offers a glimpse into the future of American society, but predictions about assimilation processes requires a more specific understanding of interethnic and interracial coupling among the children of recent immigrants. In what follows, I use data from the CILS to answer the bigger question of whether or not interethnic and interracial couples are more similar, or two different types of mixed couples. More specifically, the survey data are used to examine (a) How are interethnic and interracial couples different from each other? and (b) How are interethnic and interracial couples different from co-ethnic couples?

Methodology

CILS Survey Data

The CILS³ is the largest and longest study of its kind conducted in the United States (Portes & Rumbaut, 2001). The CILS baseline sample was drawn from eighth and ninth graders enrolled in all San Diego City schools in 1991 and 1992, with respondents ranging from 13 to 17 years of age. Eligible respondents were either foreign-born youth who had come to the United States before age 12 or U.S.-born children of immigrants (with at least one foreign-born parent). The respondents were subsequently surveyed in 1995 and again in 2001-2003. CILS, which focuses on the children of immigrants, includes respondents representing many Asian countries (the Philippines, Vietnam, Cambodia, Laos, China, Taiwan, and other Asian countries) as well as many Latin American countries (predominantly Mexico, with smaller samples from other Latin American countries).

CILS measures race and ethnicity by asking separate questions related to racial and ethnic identities. The race question is self-reported, but the respondents choose from among structured answers. The ethnicity question, however, is open-ended, allowing respondents to write in their ethnicity. What follows is the exact wording of the race and ethnicity questions from the CILS questionnaires⁴:

How do you identify, that is, what do you call yourself? ________________________

(Examples: Asian, Hispanic, Latino, American, Mexican, Mexican-American, Filipino, Filipino-American, Vietnamese, Vietnamese-American, Lao, Lao-American, Hmong, Hmong-American, Cambodian, Cambodian-American, Chinese, Chinese-American, Black, African-American, etc.)

Which of the following race categories listed do you consider yourself to be?

(Circle number or fill in)

1. White 2. Black 3. Asian 4. Multiracial 5. Other (specify): __________________

The most recent questionnaire, completed when the respondents were in their mid-20s, included the above questions about the ethnicity and race of their “spouse or partner.” This item allows us to make a distinction between co-ethnic, interethnic, and interracial coupling practices of the respondents.

The CILS data set offers a number of advantages and disadvantages. As mentioned above, one of the key advantages is the measure of both a racial and ethnic identity—especially with a write-in option that allows the respondent to self-identify their ethnicity as opposed to limited forced options. To complement this, it also has the partner’s racial (close ended) and ethnic (open ended) identification, which allows us to ascertain whether or not they might be considered an interracial and/or interethnic couple. In addition, respondents in CILS can be either married, cohabiting, dating, or not in a relationship. This allows us to analyze the different stages of relationships they are in, as well as to compare these various groups with each other. Aside from ethnic identity and relationship questions, the CILS questionnaire asks dozens of questions about sociocultural and economic characteristics, living situation, identity issues, and opinions about various other issues, including key questions on generation, language, and religion. On the other hand, one of the key disadvantages is that CILS does not ask the generation, language, and religion of the respondent’s spouse or partner, although I am not aware of any such data sets that ask such information about the respondent’s partner.

Finally, CILS is not a nationally representative sample and thus we find different immigrant groups than in other parts of the country. Moreover, the percentage of mixed couples in other parts of the country is significantly lower than in Southern California. In a previous analysis using the CPS from 2003 to 2006 (see Chapters 4 and 5 of Morgan, 2009), I found that 7% of total married couples (14% of Asian and 15% of Latino couples) and 15% of cohabiting couples (47% of Asian and 26% of Latino couples) in the United States are in interracial relationships. More specifically, among the second generation, 23% of Asian couples and 20% of Latino couples are in interracial marriages. A comparison of mixed couples in Southern California (via CILS) and California (via CPS) show that Asians and Latinos are more likely to be in mixed relationships in Southern California. Surprisingly, the percentage of mixed couples in Southern California compares closely with the rest of the United States (for Asians, 20% in Southern California compared with 16% in the rest of the United States, and for Latinos, 22% in Southern California compared with 17% in the rest of the United States). Even so, the population in Southern California is much more diverse than the rest of the United States and thus mixed relationships may not look the same in other parts of the country. It is indicative of possible future U.S. trends in intermarriage, however, as we have seen continual increases in intermarriages in the past several decades.

An Analytical Map of Relationships in CILS

Figure 1 gives a detailed breakdown of the types of relationships the CILS respondents reported in the third wave of the study (2001-2003). Of the 1,480 respondents, 60% have a partner and thus are in a relationship. Each respondent with a partner identified their spouse or partner’s racial and ethnic identity. Using these data, respondents are categorized as within a co-ethnic or a mixed relationship. If the race or ethnicity they report for their partner is different from their own, they are classified as being in a mixed relationship (either ethnically or racially). Notice that the 843⁵ people in relationships were split evenly between co-ethnic and mixed couples. Mixed couples can then be broken down into interethnic and interracial couples. Any respondent who identifies the ethnic identity of their partner differently from their own is classified as being in an interethnic relationship, whereas any respondent who identifies the racial identification of their partner differently from their own is classified as being in an interracial relationship. There are of course couples who are in both an interracial and an interethnic relationship, in which case preference is given to the interracial category.⁶

Figure 1.

Mixed relationships and the Children of Immigrants Longitudinal Study (CILS), Wave III (N = 1,480)

Note that 38% of the mixed couples are interethnic relationships but not interracial relationships. These are couples where both are racially Asian or Latino, but are from different ethnic groups: for example, Filipino–Chinese, Mexican–Salvadoran, or Mexican–Latino. If a respondent identifies his or her ethnic identity as Mexican and then identifies his or her partner as Latino, I make the presumption that their relationship is interethnic solely on the fact that the respondent identified a different ethnic identity for their partner compared with their own ethnic identity. These classifications are highly subjective; if the respondents consider their partners’ ethnic identity to be different from their own—no matter how someone from the outside might view them—they are classified as interethnic. The bold circle in Figure 1 represents the respondents included in most studies of intermarriage. In other words, when only interracial couples are included in a study, scholars ignore all the interethnic couples (in this case 38% of the mixed couples) and generally treat interethnic couples as co-ethnic couples (e.g., see Bean & Stevens, 2003). For this study, I analyze the 829 respondents who are in a relationship.

Findings

This study addresses two specific questions using the CILS survey data: (a) How are interethnic and interracial couples different from each other? and (b) How are interethnic and interracial couples different from co-ethnic couples?

Variables

As noted before, interethnic relationships are those in which the individuals report a different ethnic category, whereas interracial relationships are those in which the individuals report a different racial category. A co-ethnic relationship is, by default, a relationship in which the individuals report the same ethnic and racial identities. Individuals not currently in a relationship are not included in the analysis.

Demographic characteristic include sex, age, presence of children, education, and socioeconomic status (SES). Type of relationship is also considered: dating, engaged, cohabiting, and married. Ethnicity factors include ethnic identity, generation in the United States, language, and religion. Race is measured using the panethnic categories. To establish which social factors are most likely to account for interethnic/co-ethnic difference and interracial/co-ethnic difference, I conduct a multinomial logistic regression. Multinomial logistic regression allows for the comparison of three nonordered categories and is superior to a set of binary logistic regression models (Hoffmann, 2003).

Bivariate Analysis of CILS

Table 1 presents descriptive statistics for all the independent variables tested. Given the vast literature on intermarriage, we know that interracial relationships are very different from co-ethnic relationships. What is more problematic is how distinctive interethnic relationships may be. Traditional approaches lump these couples with co-ethnic couples and compare them to interracial relationships. If the percentages of interethnic and interracial relationships are the same across the characteristics listed, then there is no point in making a distinction, and we are therefore justified in labeling these mixed relationships as “interracial couples.” If we do see important differences, then we should either distinguish between interethnic and interracial couples, or at least include interethnic relationships into a broader mixed relationship category. These differences could manifest themselves in a number of ways: the size of the difference, whether the differences are significant (both substantively and statistically), and based on the patterns that exist in the data. In other words, we could see important differences, either by the fact that one category—interethnic or interracial relationships—is consistently larger than the other or by the different patterns that emerge between the two categories. As can be seen in Table 1, the distributions across coupling relationships are distinct and statistically significant for most of the variables examined.

Table 1.

Descriptive Statistics for Types of Couples in Southern California

				Total
Characteristics	Co-ethnic Couples^a, %	Interethnic Couples^b, %	Interracial Couples^c, %	%	N
Total	50	19	31	100	843
Sex**
Male	56	20	24	100	328
Female	46	19	35	100	515
Parent SES index***
Lowest SES quartile	60	21	20	100	276
Second quartile	48	18	34	100	188
Third quartile	45	19	36	100	205
Highest SES quartile	43	17	40	100	174
Education
Some high school	57	19	24	100	54
High school graduate	54	20	26	100	178
Some college	50	18	32	100	438
College graduate	46	20	35	100	142
Age
23 years old	46	24	29	100	209
24 years old	51	17	32	100	339
25 years old	49	19	32	100	199
26 years old	55	16	29	100	76
27 years old	60	15	25	100	20
Relationship status***
Married	61	16	23	100	267
Engaged	41	20	39	100	69
Cohabiting	34	27	39	100	88
Single, dating	48	19	33	100	419
Children***
Have children	59	15	26	100	321
No children	45	21	34	100	522
Race (panethnicity)***
Other (Latino)	59	13	28	100	346
Asian	51	23	26	100	372
Multiracial	18	21	61	100	104
White	50	20	30	100	10
Black	27	46	27	100	11
Ethnic identity***
Mexican	60	15	25	100	282
Filipino	46	15	39	100	330
Cambodian–Lao–Hmong	59	24	17	100	97
Vietnamese	41	33	26	100	81
Chinese	30	20	50	100	10
Latin, other	23	42	35	100	26
Asian, other	0	29	71	100	17
Religion***
Catholic	55	14	31	100	498
Buddhist	53	30	17	100	104
Other religion	53	15	32	100	78
Protestant	37	19	44	100	27
No religion	31	30	39	100	132
Acculturation variables
Language***
Foreign-language dominant	74	12	14	100	77
Fluent bilingual	62	13	25	100	224
Limited bilingual	50	26	24	100	119
English dominant	40	21	39	100	423
Generation***
1.5 (foreign-born, 6+)	57	15	28	100	203
1.75 (foreign-born, under 6)	53	24	22	100	217
2 (U.S.-born, 2 foreign born parents)	54	15	31	100	296
2.5 (U.S.-born, 1 foreign born parent)	25	25	50	100	127

Note: SES = socioeconomic status. The row percentages add up to 100% and thus represent all those respondents in relationships.

Source: Children of Immigrants Longitudinal Study (CILS), Wave III, San Diego Sample, 2001-2003.

Coethnic = Same racial and ethnic identity for both partners.

Interethnic = Different ethnic identity from partner; “How do you identify, that is what do you call yourself?”

Interracial = Different race category from partner; racial categories—White, Black, Asian, Multiracial, and Other.

Chi-square significance: ***p < .001. **p < .01. *p < .05.

The literature on intermarriage has established some important demographic differences (such as sex, SES, education, and age) between individuals in co-ethnic and interracial relationships (Kalmijn, 1998). In Table 1, males are more likely than females to report a co-ethnic relationship. Gender differences in interethnic coupling are not apparent, but women are more likely to report interracial coupling. The higher the parental socioeconomic level, the more likely respondents will report interracial coupling (40% among the highest SES quartile). Although SES does not influence interethnic coupling, co-ethnic coupling is most common among the poor (60%). A similar pattern is found among more educated respondents. The most educated are less likely to be in co-ethnic relationships (only 24% among those with graduate school education, although the N is relatively small in this category). Even so, among the college educated, one third report an interracial relationship compared with only one fourth among high school graduates. Again, socioeconomic differences are less a factor in interethnic coupling. Considering only the demographic factors, we find more variation among respondents in co-ethnic and interracial relationships than among respondents in interethnic relationships.

The focus on intercoupling instead of intermarriage reveals interesting differences across respondents. Married respondents are much more likely to be in co-ethnic relationships (61%), whereas cohabiting respondents are more likely to be in interethnic relationships (27%) and interracial (39%). Moreover, the distribution across dating relationships appears to reflect the overall pattern of relationships: 48% co-ethnic, 33% interracial, and 19% interethnic. Respondents with children are more likely to be in co-ethnic relationships. Respondents without children are slightly less likely to be in co-ethnic relationships and only slightly more likely to be in interracial relationships. Perhaps more formalized relationships are more likely to be co-ethnic (Kalmijn, 1998).

Racial (or panethnic) categories and ethnic identity highlight some important differences between respondents in interethnic and interracial relationships. Latinos and Mexicans are much more likely to report a co-ethnic relationship than are Asians, who are more likely to report an interethnic relationship. Six out of 10 multiracial respondents reported being in an interracial relationship. Most of these multiracial respondents are also in interethnic relationships, but precedence was given to interracial relationships. Few children of immigrants reported a “White” or “Black” racial identity.

Variation by ethnic identity is more prevalent. For example, while 60% of Mexicans report a co-ethnic relationship, only 46% of Filipinos do so. While 59% of Cambodians, Laotians, and Hmong report co-ethnic relationships, only 41% of Vietnamese do so. Filipinos are more likely than Mexicans to report interracial coupling and Vietnamese are more likely to report interethnic coupling. Many of these differences can be explained by macrostructural factors such as group size (Blau & Schwartz, 1997). Nonetheless, all this highlights the important racial and ethnic differences between interracial and interethnic couples.

Information about religious affiliation is somewhat limited in the CILS survey because only the respondent’s affiliation is reported. Nonetheless, more than half of all Catholic (55%) and Buddhist (53%) respondents reported co-ethnic coupling. Among those that report no religion, 30% are in interethnic relationships and 39% are in interracial relationships. Protestants are few in numbers and, therefore, it is difficult to assess differences across relationship type for this group.

The last two variables in Table 1, language and generation, are measures of acculturation (Gordon, 1964). The foreign-language dominant respondents were much more likely to be in co-ethnic relationships (74%) and much less likely to in interethnic relationships (12%). Comparing across respondents who report foreign-language dominance, bilingualism, and English dominance, the bilingual respondents are slightly less likely to report interracial relationships (24%) and the English dominant are more likely (39%). One might expect that more language acculturated respondents might meet and form relationships with interracial partners. A similar pattern is found when examining generation differences. Those respondents with foreign-born parents (1.5, 1.75, and 2nd) are more likely to report co-ethnic relationships. However, respondents with a U.S.-born parent (2.5) are much less likely to report a co-ethnic relationship (only 25%) and are just as likely to report an interethnic relationship (25%). In comparison, 50% report interracial relationships. The percentages presented in Table 1 point to possible justification for at least the inclusion of interethnic relationships within a “mixed relationship” category, if not the separation of mixed couples into interethnic and interracial categories.

Multivariate Analysis of CILS

Table 2 presents the results of a multinomial regression analysis with the type of relationship (co-ethnic, interethnic, and interracial relationships) as the dependent variable. A multinomial regression allows for the comparison of respondents in co-ethnic relationships with respondents in interethnic as well as interracial relationships. Because of small N categories and potential collinearity problems across independent variables, I collapsed some categories.⁷ In addition, three sets of variables were entered in stepwise fashion to form three models: first, demographic variables, including relationship status and children; second, panethnic categories and ethnic groups; and third, religious preference and assessments of acculturation—English dominance and a U.S.-born parent (2.5 generation).

Table 2.

Multinomial Logistic Regression Model Predicting the Type of Mixed Relationship by Coethnic Relationships: Children of Immigrants in Southern California

	Interracial Relationships (vs. Coethnic)
Predictors	Wald	Sig.	Odds	Wald	Sig.	Odds	Wald	Sig.	Odds
Demographic variables
Gender (1 = female)	1.2	NS	1.25	0.6	NS	1.17	0.7	NS	1.19
Parental SES index	0.1	NS	1.09	0.0	NS	1.01	0.2	NS	0.84
Years of education completed	0.4	NS	0.96	1.6	NS	0.92	1.5	NS	0.92
Age (years)	1.0	NS	0.90	1.5	NS	0.88	0.7	NS	0.92
Relationship status
Engaged	1.3	NS	1.54	1.2	NS	1.54	1.0	NS	1.50
Cohabiting	9.7	**	2.80	5.9	*	2.30	4.8	*	2.17
Single, dating	0.0	NS	1.05	0.1	NS	0.92	0.1	NS	0.92
(Married)
Children (1 = yes)	7.9	**	0.51	5.0	*	0.58	5.8	*	0.55
Race (panethnicity)
Latino panethnicity				7.3	**	0.36	2.8	NS	0.51
Asian panethnicity				11.1	**	0.28	4.3	*	0.42
(All others)
Ethnic identity
Mexican				16.3	***	0.13	8.5	**	0.21
Filipino				10.0	**	0.22	5.3	*	0.30
Cambodian– Lao–Hmong				4.1	*	0.33	4.4	*	0.28
Vietnamese				1.1	NS	0.57	1.3	NS	0.51
(Asian and Latin other)
Religion
Catholic							10.6	**	0.37
Other religion							6.8	**	0.33
Buddhist							0.8	NS	0.70
Protestant							0.7	NS	0.60
(No religion)
Acculturation:
English dominant							4.5	*	1.64
U.S.-born parent (2.5 generation)							6.5	*	2.36
Intercept	0.5	NS		4.1	*		2.6	NS
R ²	0.042^a	0.095^b	χ²(1) = 72.1	0.097^a	0.207^b	χ²(1) = 165.5	0.122^a	0.254^b	χ²(1) = 207.4
Demographic variables
Gender (1 = female)	15.4	***	2.01	12.3	***	1.92	11.8	***	1.92
Parental SES index	13.9	***	2.50	2.9	NS	1.72	1.3	NS	1.45
Years of education completed	0.0	NS	1.00	0.0	NS	1.00	0.0	NS	0.99
Age (years)	0.8	NS	1.08	0.5	NS	1.07	1.1	NS	1.10
Relationship status
Engaged	5.2	*	2.11	4.7	*	2.10	4.2	*	2.06
Cohabiting	11.7	***	2.83	8.4	**	2.54	7.4	**	2.48
Single, dating	3.0	NS	1.45	2.5	NS	1.43	2.5	NS	1.44
(Married)
Children (1 = yes)	3.4	NS	0.69	3.1	NS	0.68	3.5	NS	0.66
Race (panethnicity)
Latino panethnicity				18.8	***	0.25	11.6	***	0.32
Asian panethnicity				38.7	***	0.14	24.8	***	0.17
(All others)
Ethnic identity
Mexican				10.8	**	0.21	5.0	*	0.33
Filipino				2.3	NS	0.50	1.3	NS	0.58
Cambodian– Lao–Hmong				3.8	NS	0.35	2.2	NS	0.41
Vietnamese				1.8	NS	0.49	0.5	NS	0.66
(Asian and Latin other)
Religion
Catholic							6.2	*	0.50
Other religion							2.7	NS	0.54
Buddhist							4.6	*	0.40
Protestant							0.0	NS	1.02
(No religion)
Acculturation
English dominant							13.5	***	2.16
U.S.-born parent (2.5 generation)							2.7	NS	1.63
Intercept	2.2	NS		0.0	NS		0.3	NS
R ²	0.042^a	0.095^b	χ²(1) = 72.1	0.097^a	0.207^b	χ²(1) = 165.5	0.122^a	0.254^b	χ²(1) = 207.4

Note: SES = socioeconomic status. Reference categories for independent variables are in parentheses.

Source: Children of Immigrants Longitudinal Study (CILS), Wave III Survey, San Diego Sample, 2001-2003 (N = 829).

Pseudo R²: ^a = (McFadden); ^b = (Nagelkerke).

***

p < .001. **p < .01. *p < .05. NS = not significant.

The analysis demonstrates that a number of factors differentiate between co-ethnic coupling and interracial coupling but do not differentiate between co-ethnic coupling and interethnic coupling. Among the most interesting effects, women are almost twice as likely (1.92) to report interracial coupling compared with men, but no such gender differences discriminate between interethnic and co-ethnic coupling. Cohabiting respondents are 2.17 times more likely to be in interethnic relationships and 2.48 times more likely to be in interracial relationships compared with married respondents. This points to the stronger racial and ethnic boundary that a formal link between two people, such as marriage, poses as opposed to a more informal relationship such as cohabitation. Surprisingly, once other factors are controlled, we find only a slight difference in the presence of children among interethnic relationships and no significant differences in the presence of children among interracial relationships. The fact that just 39% of married people in CILS are in mixed relationships compared with 66% of cohabiting respondents probably explains why those with children are less likely to be in mixed relationships, since married people are more likely to have children than their cohabiting counterparts.

In terms of racial panethnic categories, only Asians are less likely to be in interethnic relationships compared with “all others” (a category that includes multiracials, Whites, and Blacks), whereas Latino and Asian respondents are much less likely to form interracial relationships compared with “all others.” It appears that Latino panethnicity becomes insignificant for interethnic relationships once we account for religion and acculturation, and in this case religion may be a proxy for panethnic identities. In a similar fashion, it appears that parental SES becomes insignificant for interracial relationships once we account for racial and ethnic identities. For ethnic identities, Mexicans are less likely to be in interethnic and interracial relationships compared with “Asian and Latin other” (a category that includes dozens of other countries from these regions), whereas Cambodians, Laotians, Hmong, and Vietnamese are less likely to be in interethnic relationships, but just as likely to be in interracial relationships compared with “Asian and Latin other.” We see that the strength of association for all the panethnic and ethnic group variables (as evidenced by the Wald statistics), both for interethnic and interracial relationships, accounts for a significant portion of the variance—although in different patterns for each type of intercoupling.

The third panel in Table 2 adds the religious preference and acculturation variables and completes the regression for predicting mixed relationships. Religion is a strong predictor of intercoupling. Catholics are the least likely to report both interethnic and interracial coupling compared with those with no religion, and Buddhists are also less likely to report being in an interracial relationship. When English is the dominant language, respondents are 1.64 times more likely to be in an interethnic relationship and 2.16 times more likely to be in an interracial relationship. Finally, once other factors are controlled, being born of U.S.-born parents is not associated with being in an interracial relationship. However, respondents are 2.37 times more likely to be in an interethnic relationship. It is not clear why there is a difference in having a U.S.-born parent between interethnic and interracial coupling. A future analysis of in-depth interviews could shed light on this interesting finding.

The addition of religious preference and acculturation has a large effect on which variables are important in explaining mixed relationships (both interethnic and interracial). The panethnic and ethnic variables weaken considerably in the third models (as evidenced by the Wald statistic). It appears that these variables lose some strength in their association because of the strong association between religious preference and acculturation. These findings offer further support for the conclusion that lumping interethnic and co-ethnic couples together, as is often done when scholars compare intermarried couples to homogenous couples, misses some distinctiveness in intercoupling that should not be ignored.

Another way to interpret the multinomial regression model is to calculate predicted probabilities. The predicted probabilities for co-ethnic, interethnic, and interracial couples by religion and acculturation, controlling for all other factors, are reported in Figures 2 and 3. These categories are mutually exclusive and therefore the probabilities add up to 100. The predicted probability that a Catholic, Protestant, Buddhist, those with another religion, and those without a religion will be in a co-ethnic, interracial, and interethnic relationship are reported in Figure 2. Catholics and Buddhists are most likely to be in a co-ethnic relationship. However, Buddhists are less likely to be in an interracial relationship while Catholics are more likely to be in an interracial relationship. In comparison, Protestants are much more likely to report an interethnic relationship. Among those reporting no religious affiliation, the differences in type of relationship is muted. The distinctive differences between Catholic, Buddhist, and Protestant are interesting, especially given that the effects of panethnicity and ethnic identity are already accounted for. These results suggest the importance of religion in understanding mixed relationships (Glenn, 1982; Johnson, 1980; Kalmijn, 1991; Rosenfeld, 2008; Sherkat, 2004).

Figure 2.

Predicted probabilities for type of mixed relationship by religion

Figure 3.

Predicted probabilities for type of mixed relationship by acculturation (language and parental nativity)

The effects of acculturation are demonstrated in Figure 3. The first two columns compare those respondents who still speak a foreign-language (foreign-language dominant, fluent bilingual, or limited bilingual) versus those who speak only English. Those who speak a foreign language are 20% more likely to be in a co-ethnic relationship, 16% less likely to be in an interracial relationship, and slightly less likely to be in an interethnic relationship. The second two columns compare those that have one U.S.-born parent (2.5 generation) versus those that have foreign-born parents. Again, those respondents with foreign-born parents are more likely to be in co-ethnic relationships and less likely to be in interethnic and interracial relationships. Generally, the more acculturated the respondent, the less likely they are to be in a co-ethnic relationship and the more likely they are to be in an interethnic and interracial relationship. We would expect more acculturated respondents to be in interracial relationships, but it is somewhat unexpected to see that more acculturated respondents are also more likely to be in interethnic relationships.

Discussion and Conclusions

Since the 1960s the nature of mixed relationships has been changing with the mass waves of immigrants coming predominantly from Asia and Latin America. Although the number of Black–White relationships has remained fairly constant over the past 50 years, the number of mixed relationships among children of immigrants has increased dramatically. Despite these changes, we still know relatively little about who is entering into mixed relationships, and more important, the differences between co-ethnic, interethnic, and interracial couples. Furthermore, we do not know what meaning intercoupling will have for the larger society in terms of children of immigrants assimilating into the American society.

A bivariate analysis of the CILS survey revealed important differences between interethnic and interracial couples along a number of factors, such as sex, age, presence of children, education, SES, type of relationship, racial identity, ethnic identity, generation, religion, and language. A multivariate analysis then controlled for these factors and compare interethnic and interracial coupling with co-ethnic coupling. For example, women are almost twice as likely to report interracial coupling compared with men, but no such gender differences discriminate between interethnic and co-ethnic coupling. Furthermore, being born of U.S.-born parents is not associated with being in an interracial relationship; however, these respondents are more than twice as likely to be in an interethnic relationship. In terms of religion, Catholics and Buddhists are most likely to be in a co-ethnic relationship. However, Buddhists are less likely to be in an interracial relationship while Catholics are more likely to be in an interracial relationship. In comparison, Protestants are much more likely to report an interethnic relationship. Finally, acculturation plays a key role. Those respondents who speak a foreign language, or with foreign-born parents, are more likely to be in co-ethnic relationships and less likely to be in interethnic and interracial relationships.

We saw from the multivariate analysis of CILS that religious preference will paint a very different picture of mixed couples, depending on whether they are interracial or interethnic couples. In addition, studies have shown that the more acculturated people get, of which language is a good indicator, the more likely they are to become involved in interracial relationships (Gordon, 1964). The findings from this study confirm this, and in addition, show that this is also the case for interethnic coupling. Therefore, even as language may become less of a factor for later generations it appears that interethnic couples will continue to be an important type of intercoupling.

These differences give support to the conjecture that interethnic and interracial couples are two different types of mixed couples. This finding has important implications for the way we look at intercoupling, as well as the theories that are associated with intercoupling. This study has demonstrated that race and ethnicity are both important social boundaries for understanding mixed couples and denote different types of mixed couples—interethnic versus interracial couples—and different types of lived experiences.

These findings also point to important implications for assimilation theory. Although interracial coupling with Whites is a sign of straight-line assimilation into a White majority culture, interethnic coupling seems to indicate another path of assimilation and lends itself more to a segmented process of assimilation. Asian ethnic groups marrying other Asians and Latino ethnic groups marrying other Latinos could be a sign of the formation of Asian and Latino panethnic groups (Qian & Cobas, 2004), if not right away, certainly among the offspring of these couples. As the third generation becomes more prominent in the next several decades, the type of mixed couples their parents formed will have important implications for their own sense of racial and ethnic boundaries, ethnic identity, language use, and religious values.

Just as the analysis of the CILS survey revealed many differences between interethnic and interracial couples, I acknowledge that it also revealed many similarities. Just like interracial couples, interethnic couples are more likely to be in dating and cohabiting relationships than in married relationships. In other words, it is more acceptable to cross racial and ethnic lines in less formal relationships such as cohabitation, but less acceptable to cross the more formal and intimate line of marriage. Marriage is not merely the joining of two individuals, but the union of two families; thus, marriage is a socially challenging action. Those parties involved must withstand the social pressures from friends and family that view them as different.

These types of similarities between interethnic and interracial couples give some support to the conjecture that interethnic and interracial couples could be considered together when talking about intercoupling. Many studies exclude interethnic couples because of legitimate weaknesses in data sets. It is problematic, however, when studies exclude interethnic couples because they see them as conceptually similar to co-ethnic couples. At the very least, this analysis of mixed couples calls for the inclusion of interethnic couples within studies of intermarriage. In this study, we would exclude 38% of the mixed couples if we did not include interethnic couples.

American couples are characterized by great diversity. This study is one of the first of its kind to explicitly compare interethnic relationships with interracial relationships and as such does not offer a definitive statement on the matter. Rather, this is a call for other studies, using various data sets, to explore this issue in more depth. Of particular importance would be a qualitative study that examines in more detail the lived experiences of interethnic couples. Are interethnic couples different from interracial couples, and if so, in what ways? Although this study can only make claims for Southern California, it provides a guide to similar studies in other regions of the United States.

Footnotes

Acknowledgements

I am indebted to Marie Cornwall, Rubén Rumbaut, John Hoffmann, John Lie, Linda Vo, David Brunsma, Tim Heaton, Cardell Jacobson, Ralph Brown, and Ben Gibbs for their helpful feedback. I am equally indebted to the anonymous reviewers who helped me vastly improve this article.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Alba

Nee

(2003). Remaking the American mainstream: Assimilation and contemporary immigration. Cambridge, MA: Harvard University Press.

Bean

F. D.

Stevens

(2003). America’s newcomers and the dynamics of diversity. New York, NY: Russell Sage Foundation.

Blackwell

D. L.

Lichter

D. T.

(2000). Mate selection among married and cohabiting couples. Journal of Family Issues, 21, 275-302.

Blau

Schwartz

(1997). Crosscutting social circles: Testing a macrostructural theory of intergroup relations. Piscataway, NJ: Transaction.

Blumberg

R. G.

Roye

W. J.

(1979). Interracial Bonds. Dix Hills, NY: General Hall.

Clark

R. L.

Glick

J. E.

Bures

R. M.

(2009). Immigrant families over the life course. Journal of Family Issues, 30, 852-872.

Cornell

S. E.

Hartmann

(2006). Ethnicity and race: Making Identities in a changing world (2nd ed.). Thousand Oaks, CA: Pine Forge Press.

Fields

Casper

L. M.

(2001). America’s Families and Living Arrangements (Current Population Report). Washington, DC: U.S. Census Bureau. Retrieved from http://www.census.gov/prod/2001pubs/p20-537.pdf

Foner

Rumbaut

R. G.

Gold

S. J.

(Eds.). (2000). Immigration and immigration research in the United States. Immigration research for a new century. New York, NY: Russell Sage Foundation.

10.

V. K.

(2007). How many melting pots? Intermarriage, panethnicity, and the Black/non-Black divide in the United States. Journal of Comparative Family Studies, 38, 215-237.

11.

Gilbertson

G. A.

Fitzpatrick

J. P.

Yang

(1996). Hispanic intermarriage in New York City: New evidence from 1991. International Migration Review, 30, 445-459.

12.

Glenn

N. D.

(1982). Interreligious marriage in the United States: Patterns and recent trends. Journal of Marriage and the Family, 44, 555-566.

13.

Gordon

M. M.

(1964). Assimilation in American life: The role of race, religion and national origins. New York, NY: Oxford University Press.

14.

Guzzo

K. B.

(2006). How do marriage market conditions affect entrance into cohabitation vs. marriage? Social Science Research, 35, 332-355.

15.

Hidalgo

D. A.

Bankston

C. L.

(2010). Blurring racial and ethnic boundaries in Asian American families: Asian American family patterns, 1980-2005. Journal of Family Issues, 31, 280-300.

16.

Hoffmann

J. P.

(2003). Generalized linear models. Boston, MA: Allyn & Bacon.

17.

Jiobu

R. M.

(1988). Ethnicity and assimilation. Albany: State University of New York Press.

18.

Johnson

R. A.

(1980). Religious assortative mating the in the United States. New York, NY: Academic Press.

19.

Joyner

Kao

(2005). Interracial relationships and the transition to adulthood. American Sociological Review, 70, 563-581.

20.

Kalmijn

(1991). Shifting boundaries: Trends in religious and educational homogamy. American Journal of Sociology, 56, 786-800.

21.

Kalmijn

(1998). Intermarriage and homogamy: Causes, patterns, trends. Annual Review of Sociology, 24, 395-421.

22.

Kalmijn

Van Tubergen

(2010). A comparative perspective on intermarriage: Explaining differences among national-origin groups in the United States. Demography, 47, 459-479.

23.

Lamanna

M. A.

Riedmann

(2008). Marriages & families: Making choices in a diverse society (10th ed.). Beverly, MA: Wadsworth.

24.

Landale

N. S.

Oropesa

R. S.

Bradatan

(2006). Hispanic families in the United States: Family structure and process in an era of family change. In Tienda

Mitchell

(Eds.), Hispanics and the future of America (pp. 138-178). Washington, DC: National Academies Press.

25.

Larsson

C. M.

(1965). Marriage across the color line (1st ed.). Chicago, IL: Johnson.

26.

Manning

W. D.

Landale

N. S.

(1996). Racial and ethnic differences in the role of cohabitation in premarital childbearing. Journal of Marriage and the Family, 58, 63-77.

27.

McClintock

E. A.

(2010). When does race matter? Race, sex, and dating at an elite university. Journal of Marriage and Family, 72, 45-72.

28.

McNamara

R. P.

Tempenis

Walton

(1999). Crossing the Line. Westport, CT: Praeger Paperback.

29.

Morgan

C. V.

(2009). Intermarriage across race and ethnicity among immigrants: E Pluribus Unions. El Paso, TX: LFB Scholarly.

30.

Oropesa

R. S.

(1996). Normative beliefs about marriage and cohabitation: A comparison of non-Latino Whites, Mexican Americans, and Puerto Ricans. Journal of Marriage and the Family, 58, 49-62.

31.

Perlmann

(1997). Reflecting the changing face of America: Multiracials, racial classification, and American intermarriage (No. 35; Public Policy Brief). Annandale-on-Hudson, NY: Jerome Levy Economics Institution.

32.

Portes

Rumbaut

R. G.

(2001). Legacies: The story of the immigrant second generation. Berkeley: University of California Press.

33.

Portes

Rumbaut

R. G.

(2006). Immigrant America: A portrait (3rd ed.). Berkeley: University of California Press.

34.

Qian

Blair

S. L.

Ruf

S. D.

(2001). Asian American interracial and interethnic marriages: Differences by education and nativity. International Migration Review, 35, 557-586.

35.

Qian

Cobas

J. A.

(2004). Latinos’ mate selection: National origin, racial, and nativity differences. Social Science Research, 33, 225-247.

36.

Qian

Lichter

D. T.

(2007). Social boundaries and marital assimilation: Interpreting trends in racial and ethnic intermarriage. American Sociological Review, 72, 68-94.

37.

Rosenblatt

P. C.

Karis

Powell

D. R. R.

(1995). Multiracial couples: Black & White voices. Thousand Oaks, CA: Sage.

38.

Rosenfeld

M. J.

(2008). Racial, educational and religious endogamy in the United States: A comparative historical perspective. Social Forces, 1, 1-31.

39.

Schermerhorn

R. A.

(1979). Comparative ethnic relations: A framework for theory and research. Chicago, IL: University of Chicago Press.

40.

Sherkat

D. E.

(2004). Religious intermarriage in the United States: Trends, patterns, and predictors. Social Science Research, 33, 606-625.

41.

Shinagawa

L. H.

Pang

G. Y.

(1988). Intraethnic, interethnic, and interracial marriages among Asian Americans in California, 1980. Berkeley Journal of Sociology, 33, 950-1104.

42.

Shinagawa

L. H.

Pang

G. Y.

(1996). Asian American panethnicity and intermarriage. Amerasia Journal, 22, 127-152.

43.

Simmons

O’Connell

(2003). Married-couple and unmarried-partner households: 2000 (Census 2000 Brief). Washington, DC: U.S. Census Bureau. Retrieved from http://www.census.gov/prod/2001pubs/c2kbr01-8.pdf

44.

Smock

P. J.

(2000). Cohabitation in the United States: An appraisal of research themes, findings, and implications. Annual Review of Sociology, 26, 1-20.

45.

Vermeulen

(1999). Essentializing difference? The census, multiculturalism and the multiracials in the USA. In Foblets

M. C.

Lin

P. C.

(Eds.), Cultuur, Etniciteit en Migratie/Culture, ethnicity and migration. Liber Amicorum E. Roosens (pp. 81-98). Leuven, Netherlands: Acco.

46.

Weaver

S. E.

Umana-Taylor

A. J.

Hans

J. D.

Malia

S. E. C.

(2001). Challenges family scholars may face in studying family diversity. Journal of Family Issues, 22, 922-939.

47.

Yinger

J. M.

(1994). Ethnicity: Source of strength? Source of conflict? Albany: State University of New York Press.