Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

In the last 10 years, there has been increased attention given to the demographic analysis of same-sex couples and their families. This attention may be attributed to multiple factors, including current debates and social movements regarding issues of sexual orientation and legislative rights. Additionally, one paramount factor contributing to the increased attention is the availability of better data on same-sex households. While these data are limited, the inclusion of various measures and indicators of sexual orientation and access to large-scale nationally representative data gives demographers a starting point to examine issues of sexual orientation. Currently, there is growing support for the argument that sexual orientation does have an effect on demographic processes and life outcomes. This chapter presents an overview of the current demographic research on same-sex families. It describes how same-sex families are understood by demographers, limitations to this demographic research, and demonstrates how demography can illuminate issues of same-sex families. This chapter also draws on research outside of demography, primarily from family studies, to further highlight substantive concerns and future directions for demographic research.

Overall, there has been very little academic research conducted on same-sex couples and families. While some empirical data, most of which is qualitative, has been collected on gay and lesbian families (primarily within psychological and sociological family studies), this research is quite limited, resulting in gay men and lesbians being effectively ignored within family studies (Allen and Demo 1995; Demo and Allen 1996). Some undergraduate family studies textbooks suggest that the primary reason for this is due to the relatively small population of sexual minorities (Starbuck 2002; Benokraitis 2011). However, one might also speculate that the social stigma of homosexuality, stereotypes, and issues of social tolerance could play a part in the types of questions considered and data collected on gay and lesbian families.

The population size of gay and lesbian families is still in question. Sociological family studies texts cite conservative estimates on gay and lesbian families, suggesting that they comprise at least 5% of families in the United States (Baca Zinn and Eitzen 1999). Demographic literature has suggested there are just over 600,000 same-sex unmarried partner households in the U.S. as of the 2000 Census (Smith and Gates 2001), the 2008 American Community Survey estimates 565,000 households (Gates 2009), and most recently the 2010 Census has released estimated counts of 646,464 same-sex partner households (O’Connell and Feliz 2011). These estimates, as will be discussed later, are conservative, given that they do not include single-parent families headed by gay men or lesbians, and do not include couples who choose not to identify their relationship as a “same-sex partnership” on census surveys. These data nonetheless indicate that there is a notable subpopulation living in gay and lesbian families, warranting an assessment of their demographic characteristics and outcomes. This chapter explores many of these issues, including prevalence, demographic characteristics, and outcomes, relative to heterosexual families.

2 Background on Same-Sex Families

The family plays an important role in society and for individuals. It is in the family that we spend the most time and where we learn how to participate in society. Often, our strongest social and emotional attachments are to our families. However, there is great diversity among families just as with the individuals that comprise them. Macro factors such as urbanization, the presence of market economies, education structures, and the move towards individualism, have all had an effect on how families are defined and organized (Waite 2005). Additionally, social tolerance, legislation and public policy, and social movements have a strong influence over who is allowed to marry, when individuals are likely to marry, and who is expected to make decisions in the family.

As with heterosexual families, same-sex families come in many shapes and sizes. They are influenced by similar contextual factors, such as education and workplace expectations. However, same-sex families may also face issues that are not always applicable to heterosexual families. For example, since most jurisdictions deny same-sex couples the right to legally marry, they do not share the same legal rights and protections as married heterosexual couples and, because of this, nor do their children (Demo et al. 2000; Cahill et al. 2002). Moreover, there are many laws, policies, and practices in the U.S. and internationally concerning the regulation of adoption, foster care, child custody, and visitation issues which are biased in favor of heterosexual people, their relationships, and their families—also limiting constructions of families (Cahill et al. 2002).

Same-sex relationships and families often face more social biases; they are frequently taken less seriously and are less accepted than their heterosexual counterparts. Same-sex partners may feel less free to show affection towards one another in public or to talk about their home life while at work. Also, compared to heterosexual couples, they are less likely to be extended “couple” privileges with respect to invitations, occupational benefits, and so forth, or to be viewed as authentic couples (Baca Zinn and Eitzen 1999, 2008). Such factors influence the organization and health of families and relationships.

A discussion of issues related to sexual orientation leads to many important questions for social scientists and demographers wishing to have a greater understanding of same-sex families. For example, many who object to the marriage of same-sex individuals argue that marriage is a necessary environment for the raising of children and that since gay men and lesbians cannot procreate naturally, they should not be permitted to marry. Nonetheless, many same-sex partners are raising children despite these “biological” limitations. The presence of children in same-sex households then raises questions such as: How do children come to be in the household? Are they adopted or are same-sex partners employing other means to have children biologically? How are children and families affected by the lack of legal and social recognitions afforded their heterosexual counterparts? While demographic research may not yet be able to answer all of these questions, the discussion in this chapter lays the foundation to present a more complete picture of same-sex families and enhance what has already been learned from family studies research.

2.1 Family Studies Research on Same-Sex Families

Family studies research on same-sex families tends to concentrate primarily on individuals within same-sex families—their attributes, feelings, and behaviors. The social and emotional effects on children growing up in Gay, Lesbian, Bisexual, and Transgendered (GLBT) families have also been examined (Allen and Demo 1995; Stacey and Biblarz 2001; Biblarz and Savci 2010). Although there are many reasons to study same-sex families and relationships, it could be argued that the literature has been concentrated in areas that have also been the center of the “family values” debate in mainstream America. Many of the research questions posed tend to focus on proving or disproving elements of this debate, such as examining the effects of these relationships on children or assessing how same-sex families affect or may affect heterosexual society.

Most findings on same-sex couples and families have come from a few large-scale family studies (see Blumstein and Schwartz [1983] or the Lawrence Kurdek series [1987, 1992]), and various other small-scale studies, all of which used convenience samples consisting of respondents who were overwhelmingly white, middle-class, young adults with above-average levels of education (Blumstein and Schwartz 1983; Kurdek 1987, 1992). They have found that, like most heterosexual individuals, gay men and lesbians seek and desire secure, intimate relationships. Gay men and lesbians also look for the same relationship qualities as their heterosexual counterparts, such as spending time together, sharing intimate feelings, having equal power in their relationships, and being monogamous (Starbuck 2002; Kurdek 1992). The majority of current research indicates that there is little difference between same-sex partners and heterosexual cohabiting couples when it comes to issues of stability, conflict, problem-solving, decision-making, interpersonal violence, and the division of household labor (Demo et al. 2000; Sarantakos 1996; Carrington 1999). Likewise, rates of relationship dissolution are about the same for gay men, lesbians, and cohabiting heterosexuals, all of which are higher than those for married heterosexual couples (Starbuck 2002). These findings support the notion that same-sex households and families are similar to unmarried heterosexual households and families. The findings could also lend support to the idea that a marriage contract does provide an increased barrier to dissolution, something to which most gay men and lesbians do not have access.

It is also argued that same-sex partners and their families may be more stable than has been suggested in the family literature (Gottman et al. 2002). The National Gay and Lesbian Task Force has asserted that same-sex partners exist in large numbers, and they are “stable, productive households and have many of the same needs as do opposite sex couples” (Bradford et al. 2002: iv). There have been no clear explanations provided in the current literature to account for the lower levels of stability of same-sex unmarried partners in comparison to married partners. For instance, dissimilarities in rates of stability between heterosexual married couples and same-sex couples cannot be definitively attributed directly to sexual orientation. Instead, additional factors, such as external stresses stemming from heterosexist norms, lack of social privileges, legal rights, and other issues may also be contributing to the observed instability.

Methodological issues in family research may also be contributing to overstated or inaccurate instability measures for gay male and lesbian families. Findings within family studies rely heavily, or in some cases solely, on data that are often well over 15 years old and are not nationally representative. Moreover, research has been based on the ideal of the traditional marriage construct. This may well affect the manner in which researchers have undertaken analyses (Brines and Joyner 1999). Frequently, they tend to view cohabitation as a form of “trial marriage” rather than as its own form of relationship, with unique attributes and characteristics (Brines and Joyner 1999). These views are likely to affect how surveys are designed and how behaviors are interpreted. For example, it is often assumed that couples have a higher level of commitment to one another and to their relationship by the mere fact that they are married. In contrast, unmarried cohabiters are viewed as less committed and their relationships as less serious compared to married couples. This has implications for same-sex couples who cannot legally marry and whose relationships go unrecognized or are viewed as illegitimate.

However, in a recent review of new scholarship on GLBT families, Biblarz and Savci (2010) note the rapid growth of work in the past decade within family studies. They assert that this is due to an increase in available data that include questions and measures of sexual orientation (including both nationally representative data collection projects and more focused qualitative studies), and better research designs being employed by social scientists (Biblarz and Savci 2010). This increase in research and its diversity regarding researchers, questions of interests, and methodologies (longitudinal, surveys, interviews, etc.) has led to a more cumulative knowledge regarding GLBT families.

Moreover, it should be noted that findings and conclusions from prior family studies research has overwhelmingly indicated that gay male and lesbian parents are just as capable of raising children as are their heterosexual counterparts (Stacey and Biblarz 2001; Cianciotto and Cahill 2003; Biblarz and Savci 2010). Likewise, most of the major child advocacy organizations recognize “gay and lesbian parents as good parents, and assert that children can and do thrive in gay and lesbian families” (Cahill et al. 2002: 69). These organizations include: the American Academy of Pediatrics, the American Psychological Association (APA), the National Association of Social Workers in conjunction with the APA, the American Psychoanalytic Association, and the American Academy of Family Physicians (Cahill et al. 2002).

3 Demographic Research on Same-Sex Families: Substantive Concerns

As a fledgling field, the social demography of same-sex families is inundated with many issues including the lack of nationally representative data sets, concerns of measurement and conceptualization, and the social stigma and assumptions surrounding sexual minorities. In addition, these issues directly affect one another and what is demographically known about same-sex families.

3.1 Defining Family

In studying same-sex families, one large issue derives from the diverse definitions and conceptualizations of families, and more specifically same-sex families. Academically, family has been defined as individuals having either biological or marital associations that are culturally recognized (Waite 2005; Baca Zinn and Eitzen 1999). Families have also been described as being responsible for the bearing and the raising of children, for comprising the structure within which individuals reside, and for being the means by which property is shared and passed down (Waite 2005; Baca Zinn and Eitzen 1999). However, others have asserted that these sorts of definitions, often based on legal or structural terms, are quite problematic in reference to families outside of the ideal nuclear heterosexual family type. They are often too narrow or are not inclusive of some of the increasingly more common “alternative” family types, such as cohabiting couples without children, GLBT families, and “families of choice” (Weston 1991; Cahill et al. 2002; Waite 2005). Most typically, GLBT families refer to families that consist of at least one gay male, lesbian, bisexual, or transgendered parent with one or more child, or to a gay or lesbian couple irrespective of whether children are present (Cahill et al. 2002). Also popular in family studies, is the term “gay families”—where gay is employed as an umbrella term for GLBT and holds to the same definition as GLBT families (Baca Zinn and Eitzen 2008). However, these terms have been criticized because families do not have sexual orientations, rather it is individuals who make up families that have sexual orientations (Baca Zinn and Eitzen 2008). Nevertheless, these terms persist in family literature.

Conversely, “families of choice” are defined more by emotional ties rather than legal terms or blood relations; they can include friends, lovers, co-parents, children and/or relatives from prior relationships—most anyone, who provides emotional and/or material support (Weston 1991; Cahill et al. 2002). Weston (1991) notes that “families of choice” have become a very real option for many gay, lesbian, bisexual, and transgendered people who have been shunned from their own biological families. Most notably, “family of choice” does not include an actual parameter for sexual orientation, such as with gay or GLBT family terminology. It could be argued, however, that the notion of choice delegitimizes this conceptualization of a family as a “real” family further making this term problematic and controversial too.

In the 1970s researchers began to focus more on the household form and its variations, due in part to social changes in the life course, such as increases in age at first marriage and in the percentage of single-headed households (Seidman 1993; Weston 1991). This created a shift in the understanding of definitions and roles associated with the family and, consequently, affected how scholars studied the family. This new way of looking at the family resulted in changing conceptions of family and, thus brought a greater visibility of gay male and lesbian families. However, the major challenge for gay male and lesbian families is still confronting the ideology of dominant American culture. As Weston (1991) notes, homosexuality is associated with deviance, singleness, and unnaturalness, all of which directly counter the traditional image of the family that encompasses heterosexuality, morality, and nature (Weston 1991). Moreover, in application to topics of the family, this ideology has been especially resilient to change. This resilience also speaks, at least in part, to the lack of available data on GLBT families, as there is a significant cultural lag between these understandings and empirical research.

In addition to the aforementioned terms and issues, demographic research has also drawn on same-sex families, along with same-sex couples, same-sex partners, and same-sex unmarried partners to reference gay and lesbian families (Baumle et al. 2009; Gates 2009; Black et al. 2000). However, the conceptualization of these terms differs based on the survey parameters under which the data are collected (as will be discussed in more detail in the next section). Given that the majority of demographic research on same-sex families has come from census data that do not specifically address sexual orientation, definitions are limited to households which encompass “same-sex unmarried partners.” As such, this definition is a subset of gay and GLBT families, since most definitions of a family would also include single parents. In accordance with this demographic work drawing on census data, this chapter utilizes same-sex families and gay and lesbian families to speak to households that are organized around two same-sex unmarried partners, regardless of the presence of children. This phrasing is in line with most of the economic and social demographic research that employs the census data (Black et al. 2000).

3.2 Data and Measures

To date, there are only five nationally representative datasets in the U.S. from which we are able to demographically examine issues of homosexuality and same-sex families in the United States. There is the General Social Survey (GSS), the National Health and Social Life Survey (NHSLS),Footnote 1 National Survey of Family Growth (NSFG),Footnote 2 and the Census and American Community Survey (ACS). Depending on the survey, the manner in which homosexuality and families are measured varies due to how sexual orientation has been defined and conceptualized by researchers. Sexual orientation can be defined in terms of sexual behavior, sexual desire, and self-identification, or any combination of the three (Laumann et al. 1994). Most commonly, analyses draw on self-identification and behavioral measures of sexual orientation (Baumle et al. 2009). For example, the GSS includes only behavioral measures of homosexuality, while the NHSLS and NSFG has measures of all three conceptualizations. The U.S. Census and ACS only allow for the analysis of self-identified same-sex unmarried partners who lived in the same household.

Regarding the demographic analysis of same-sex families, the census and ACS data are arguably the best, and certainly the largest, datasets on same-sex partners (Black et al. 2000; Baumle et al. 2009; Gates 2009). The U.S. Census seeks to enumerate all American households decennially, while the ACS samples the population yearly. Both surveys are conducted by the Census Bureau and provide wide-ranging information concerning demographics, economics, and the spatial distribution of the U.S. population (Gates 2009). Gates (2009) asserts that the census and ACS data “provide vital and widely trusted information about same-sex couples and their families that cannot be acquired from any other data sources” (Gates 2009: 1).

It is via the “unmarried partner” response that individuals are able to identify as being unrelated and living in a household in a “marriage-like” relationship with one another. It is assumed that these data represent same-sex households (male-male or female-female) occupied by partnered individuals in a gay relationship (Baumle et al. 2009; Black et al. 2000, 2003; Simmons and O’Connell 2003; Walther and Poston 2004; Gates and Ost 2004). While they do have limitations, the census and ACS data have given researchers the ability to examine under-explored issues regarding sexual orientation. These data are not generalizable to the entire GLBT population. For one they do not include single individuals. They do not encompass those who do not identify as living in a “marriage-like” relationship, nor do they directly address sexual orientation or transsexuality. To draw on these data, one must understand and employ the same clear-cut definition of what a same-sex partner is, as dictated by the Census Bureau (Black et al. 2000; Smith and Gates 2001; Gates and Ost 2004; Walther and Poston 2004). These data, however, are very useful for the study of same-sex families as long as researchers are clear about to whom the data refer. The importance of this cannot be overstated.

There are obvious methodological limitations regarding the conceptualization and measurement of sexual orientation, and its application to demographically studying same-sex families. First, there is a lack of a common, consistent definition in surveys, as noted in the “Defining Families” section of this chapter. Second, there are problems obtaining sufficiently representative sample sizes, as the GLBT population is considered to be relatively small and hidden. Third, and most notably, there is the absence of questions that address sexual orientation in large-scale data collections. These limitations are related to the social stigma attached to homosexuality that affects survey design and the manner in which individuals respond to survey questions about sexual orientation and behaviors (Laumann et al. 1994; Baumle et al. 2009).

These methodological problems are the same problems that are inherent in gathering and analyzing data on most stigmatized groups. While data and analyses have been highly criticized due to these methodological limitations, the research derived from these data, in combination with the fact that the American public has become more socially tolerant of homosexuality, has led to increased discourse and greater visibility of GLBT populations and issues. All of these factors have resulted in a cycle that creates more conversations, controversies, questions, and research that culminates in better data and better understanding of the GLBT population and sub-populations.

4 Empirical Findings

4.1 Same-Sex Parents and Their Children

Thus far, most demographic research analyzing same-sex partners and their children has been largely descriptive—limited to summations of various parenting rates and general demographic breakdowns (Smith and Gates 2001; Cahill et al. 2002; Simmons and O’Connell 2003). As with the family literature, there have been very few demographic studies specifically addressing same-sex families, or the children of gay male and lesbian parents. Once again, this is largely due to the lack of quality data addressing these subjects. There are relatively few quality surveys from which to estimate the number of children with gay male and lesbian parents and it is difficult to draw reliable conclusions from these studies, as most of them are not based on representative samples. Table 14.1 contains a list of data sources for demographic research on same-sex families.

Table 14.1 Data sources of demographic same-sex family research
Full size table

Regarding children in same-sex households, Badgett (2001) notes that according to the Voter Research Surveys and the Yankelovich Monitor, the proportion of children in lesbian households is roughly equal to that in heterosexual women’s households, whereas gay male households are about half as likely as heterosexual male households to have children. However, data from the General Social Survey/National Health and Social Science Life Survey (GSS/NHSLS) and the 1990 Census suggest that there are somewhat lower numbers of children in same-sex households (Badgett 2001). According to work done by Black and his associates (2000), the combined GSS-NHSLS data indicate that 28% of lesbians and 14% of gay men have children in their households (Black et al. 2000). Data from the 1990 U.S. census indicate lower percentages of children in same-sex, as compared to heterosexual, households. According to the 1990 census, only 20% of female same-sex households and 5% of male same-sex households have children, compared to 57% of married households (Black et al. 2000; Badgett 2001). Badgett (2001) asserts that this sizable difference in the census data between same-sex and heterosexual families is perhaps a reflection of the exclusion of single-parent households, or a bias in reporting patterns for same-sex households where couples with children may be less likely to disclose their relationship on the census questionnaire.

In a report on educational policy and issues affecting GLBT youth, Cianciotto and Cahill (2003) state that estimates range between two and eight million gay and lesbian parents in the U.S. With respect to the number of children with one or more gay, lesbian, or bisexual parent, a range of from one to fourteen million has been estimated (Cianciotto and Cahill 2003). A poll conducted in 2000 by the Kaiser Family Foundation indicates that 8% of the 405 self-identified gay, lesbian, and bisexual respondents had children under 18 in their households (Cianciotto and Cahill 2003). The Black Pride Survey 2000 indicates that 21% of the black gay, lesbian, bisexual, or transgendered participants reported being biological parents, while 2.2% reported being adoptive or foster parents (Cahill et al. 2002). Moreover, 12% reported currently living with children, while 25% had at least one child (Cahill et al. 2002).

Drawing on the 600,000 same-sex unmarried partners enumerated in the 2000 U.S. Census, Simmons and O’Connell (2003) find that 34% of the female same-sex unmarried partner households and 22% of the male same-sex unmarried partner households contained at least one child under 18. Cianciotto and Cahill (2003: 1) assert that for female same-sex unmarried partner households, this rate “is not that much lower than the percentage of married opposite-sex households with children (46%) or the percentage of unmarried opposite-sex households with children (43%).” They observe, however, that male same-sex partner households “parent at about half the rate of married couples (22% vs. 46%)” (Cianciotto and Cahill 2003: 1).

Using the 5% Public Use Microdata Sample of the 2000 U.S. Census, my coauthors and I (Baumle et al. 2009) explored demographics related to same-sex families and their households. The next three sections draw on, and expand, this work.

4.2 Prevalence and Composition of Same-Sex Families

The census data permit an exploration of the prevalence and characteristics of same-sex families. Drawing on the 5% PUMS sample for the 2000 Census (at current the 2010 data is not fully available for analysis), we found that same-sex unmarried partner households contain 31,972 male and 32,756 female partners, for a total of 64,728 same-sex unmarried partners. There are 30,973 other members in these households including 21,111 individuals under the age of 18. Of those under the age of 18, 20,868 can be identified as children in same-sex unmarried partner households.Footnote 3 Overall, 15% of male same-sex partners and 21% of female same-sex partners have children present in their household.

The same-sex partners in this sample are predominantly white and have attended at least some college. On average the male partners are 45 years old with a household income of $79,000 and the female partners are 43 years old with an income of $67,000.

Over 57% of the male unmarried partners and 50% of the female unmarried partners are categorized as “never married.” With regard to children in the household, 85% of the male households and 78% of the female households, report having no children. A comparison of some of the demographic characteristics of same-sex households to different-sex households appears in a later section.

With regard to other household members, children are the primary co-residents with same-sex partners (as compared to other familial or non-familial household members). Table 14.2 shows the relationship of all individuals in the household to the householder. Excluding the unmarried partners, the next five largest categories of people in the households are: “children” (21%), “other non-relatives” (1.6%), “stepchildren” (1.6%), “grandchildren” (1.5%), and “housemates/roommates” (just over 1%).

Table 14.2 Relationship to head of household of same-sex households (2000)
Full size table

One challenge faced when drawing on the census data is that the census question about children is not phrased in a manner that permits a distinction between biological or adopted children. Moreover, it cannot be determined whether the children belong to another member in the household (Badgett 2001). These data limitations restrict a complete understanding of the relationships within the families of gay men and lesbians. However, children can be identified in relation to the householder and children have been categorized as being an adopted child or a natural-born child of the householder. This, of course, does not reflect how the child might or might not be related to the unmarried partner who is not listed as the householder on the census form.

In Table 14.3, we take a closer look at the children in same-sex households and their relationship to the householder. As previously mentioned, there are 20,868 children in the sample. There are 8,381 children in gay male partnered households and 12,487 in lesbian partnered households. Although the 5% PUMS data include 21,111 individuals under the age of 18, 243 individuals whose indicated relationships were inconsistent with that of a parent/child relationship were dropped. These individuals appeared to fall outside the “child” category, either because (1) they were living as adults, as indicated by their assignment to the “head/householder” or “unmarried partner” relationship categories, or (2) their relationship to the householder was indicated as “housemates/roommates” or “roomers/boarders/lodgers,” suggesting a non-parental relationship. However, the “other non-relatives” category was included due to its size and the ambiguity of the category in association with a parental relationship. For example, this may be a logical choice for categorizing children who have been informally adopted by the householder.

Table 14.3 Relationship of children to head of household of same-sex households (2000)
Full size table

The majority of the children in the sample are white; however, the racial and ethnic breakdown of children is more diverse when compared to the racial and ethnic breakdown of the same-sex unmarried partners in the sample. On average, children are 8 years old, with an education level between the first and fourth grades. As shown in Table 14.3, “children,” “adopted children,” “stepchildren,” “grandchildren,” and “other non-relatives” comprise the top five relationships for children to householder; these categories account for just over 96% of all the children in these households. Children identified as the “children” of the householder likely include children who are the biological offspring of the householder. They may also be children who were products of artificial reproductive technologies and/or surrogacy. In such a case, even if the householder did not contribute biologically to the birth of the child, he or she still might consider the child his or her “natural child.”

The “adopted child” category is most likely used by an individual who has engaged in the formal legal adoption of a child (in the past or during the current relationship), or who has adopted the child of their partner (Baumle et al. 2009). Children in this category could also be the natural born child from a previous heterosexual relationship or a child resulting from artificial reproductive technologies (i.e. in the case of a female same-sex couple, one woman might bear the child and the other might formally adopt the child). It is further assumed that the “stepchild” category refers to children of one partner who are from prior relationships—irrespective of type, whether heterosexual or same-sex.

The above-mentioned categories are the more easily reasoned and recognized categories, whereas the “other non-relatives” category poses a greater challenge because we are unable to ascertain the actual relationship between the children and same-sex partners. Past work has suggested that this is a reasonable category for children who have been informally adopted by the head of household (Baumle et al. 2009).

Specifically addressing adopted children, Gates et al. (2007) find that 4% of adopted children in the United States are being raised by gay or lesbian parents—3% of which are in single parent households and 1% in coupled same-sex households. They, further, estimate that 14,100 children are fostered by a gay or lesbian parent (Gates et al. 2007). This number represents close to 3% of children in all forms of foster care. In considering just non-kin foster care, the percentage doubles to 6% where 5% are estimated to be in a single gay or lesbian parent home and 1% in same-sex coupled homes (Gates et al. 2007).

4.3 Factors Affecting the Presence of Children in Same-Sex Households

In examining factors affecting the presence of children in same-sex households, it is important to consider both individual and contextual characteristics of same-sex partners. As such, demographic characteristics of same-sex parents and their households (race, ethnicity, household income, age, previous marital status) could be important predictors of children in same-sex households. In addition, contextual characteristics such as region of residence and whether households were located in a state with restrictive state-level family laws could play a role.

Baumle and colleagues (2009) drew on the 2000 Census Public Use Microdata sample and found that the odds of having a child present in the household are 40% higher for lesbians compared to gay men, controlling for other demographic characteristics. Results also indicate that racial or ethnic minorities are more likely to have children present in their households. With regard to household income, partners in households where children are present earn slightly less than those who live in households where there are no children. Regional location did not appear to have a significant effect on the presence of children in same-sex households.

The largest predictor of children being present in a same-sex household is whether individuals had indicated a previous marital relationship on the census. According to the census data, approximately half of same-sex unmarried partners may be categorized as having been previously married, which could indicate a heterosexual relationship. Individuals who indicated a previous marital relationship were almost three times more likely to have a child in their household than those who marked the “never married” or “not applicable” category. While this result does not speak to how the children come to be in same-sex households, it does lend support to the notion that many children present in same-sex households may be from previous heterosexual marriages and relationships.

Continuing this line of research, we sought to further investigate the effects of state-level variables on the odds of children being present in same-sex households (Baumle and Compton 2011). Specifically, we examined whether formal law plays a central role in family formation outcomes for gay men and lesbians when considering the effect of both positive and negative family laws (such as whether gay men and lesbians are able to adopt and foster, and matters of surrogacy and second parent adoption irrespective of sexual orientation), as well as “pro-gay” and “anti-gay” legislation outside of family laws (i.e. the presence of sodomy and antidiscrimination laws).

Employing a multilevel analysis, we found that negative formal laws appeared to play little or no role in family formation outcomes. Laws prohibiting same-sex couples from adopting, fostering, or surrogacy had no statistically significant effect on the presence of children in households. However, laws prohibiting second parent adoption did result in lower odds of children being present in a household. Further, positive laws – measured as a combination of adoption and second parent adoption lawsFootnote 4 – increased the odds of children being present. Overall, these results are compatible with prior sociolegal research finding that individuals are less likely to consult formal law in decisions regarding their everyday lives – particularly with regard to family matters – but are more likely to do so with regard to family issues concerning “business” matters, such as wills, estates, guardianship, and transfers of property (Baumle et al. 2009).

These results further indicate that, overall, living in a state with antidiscrimination legislation increases the odds of a child being present, and that higher concentrations of same-sex partners in the state increases the odds of children being present. The presence of an anti-sodomy law did not have a statistically significant effect on the presence of children in same-sex households. These findings lend further support to the notion that a “friendly” environment might increase the prevalence of children in same-sex households, but that negative laws are less powerful predictors.

Overall, these analyses show the importance of considering both individual and contextual characteristics when examining outcomes for same-sex families, especially considering the current legal and political climate and controversies surrounding the GLBT population.

4.4 Demographic Comparisons Across Couple Types

Another goal in demographically studying same-sex families and households is to assess the degree to which same-sex partnerships are comparable to heterosexual married and unmarried partnerships. Politically, it has been argued that providing legally sanctioned marriages to gay men and lesbians is unnecessary because they provide no real benefit that cannot be gleaned through contractual agreements. However, research indicates that married individuals are healthier, live longer, and tend to have more assets and accumulate more wealth than individuals who are not married (Waite 1995; Waite and Gallagher 1999; Blumstein and Schwartz 1983). In addition, and as previously discussed, the stability of same-sex relationships has been questioned and same-sex relationships are generally considered by family studies literature and society as less stable with lower levels of commitment than relationships where individuals are married.

In work with Baumle et al. (2009), we found that same-sex couples largely fall between unmarried heterosexual couples and married couples with regards to their demographic characteristics and their standard of life and relationships commitment variables. These include social and economic indicators of relationship commitment and stability, most notably home ownership, the presence of children, and living in a dual-income household. While the rates for the presence of children in their same-sex homes are considerably lower compared to heterosexual couples and lesbian partners appear to be slightly more educated than the all other partner types, these data suggest that same-sex households may be more similar to married households than previously suggested by family studies. For example, it appears that same-sex couples have greater financial commitments and dependence on one another than do heterosexual unmarried partners, although these do appear to be less than those of heterosexual married households. Sixty-seven percent of same-sex partners own their homes, compared to 46% of unmarried heterosexual partners and 82% of married households. On average, same-sex unmarried partner households report $77,708 in household income, whereas unmarried heterosexual partner households report $55,798 and married households report $77,669. On the surface it may appear that same-sex households make more than cross-sex households. However, compared to married heterosexual men, partnered gay men earn significantly less, but slightly more than heterosexual unmarried partnered men. Conversely, partnered lesbians earn more than both their married and unmarried counterparts (Baumle et al. 2009).

Additionally, 71% of same-sex unmarried partner households have dual labor force participation, compared to 74% of unmarried partner households and 64% of married partner households. These results also could support the notion that there may be an income or wealth advantage for those who are married compared to those who are not or cannot marry.Footnote 5

Findings from more recent data sources, like the 2007 and 2008 American Community Survey (ACS), further support the notion that same-sex couples differ from heterosexual cohabiting partners. Drawing on the 2007 ACS data, O’Connell and Lofquist (2009) found that heterosexual unmarried partners tend to be younger, less educated, and have lower household incomes than same-sex partners and heterosexual married partners. However, heterosexual married couples are least likely to be in an inter-racial relationship, least likely to both be employed, and most likely to own their homes according to the 2007 ACS data (O’Connell and Lofquist 2009).

In addition to assessing how same-sex and opposite-sex couples compare, demographic analyses using census data have also been concerned with examining how “married” same-sex partners might differ from “cohabiting” same-sex partners. These questions arise particularly in relation to the self-selection of the “unmarried partner” versus “spouse” categories of the “relationship to the head of household” question on the U.S. Census. Before 2004, same-sex couples did not have access to legal marriage. This does not mean that same-sex couples did not endeavor to organize their relationships as if legally married; indeed, some couples identified on surveys as “married” even in the absence of a legal marriage. In 2004, Massachusetts became the first state to legalize same-sex marriage. Connecticut and California (which briefly allowed same-sex marriage, but currently does not) followed suit in 2008. As of January 2012, Iowa, Vermont, New Hampshire, the District of Columbia, and New York have also legalized same-sex marriage, while Rhode Island, Maryland, New Mexico, and Illinois recognize same-sex marriages granted from other states (NGTLF 2010).Footnote 6

Currently, the U.S. Census Bureau does not allow same-sex partners to indicate a spousal relationship on the census due to guidelines set forth by the 1996 Defense of Marriage Act (DOMA) which defines “marriage” and “spouse” for federal purposes. As such, same-sex partners that do identify a spousal relationship are edited into the “unmarried partner” category in public use files. However, the internal data files do contain the edited responses, allowing data to be teased out via imputation flags in the more recent ACS data and in the upcoming 2010 Census (O’Connell and Lofquist 2009).

So, are households that indicate “unmarried partner” different from households that indicate “spouse”? According to the most recent ACS data, they are in fact different. Drawing on the 2007 ACS data, partners that indicate a spousal relationship are slightly older, more likely to have children, and more likely to own their homes compared to partners that indicate an unmarried partner relationship (O’Connell and Lofquist 2009). Moreover, same-sex spouses are less likely to have an inter-racial partner and less likely to have both partners employed (O’Connell and Lofquist 2009). The 2008 ACS data also indicate that same-sex spouses are different from same-sex unmarried partners (Gates 2009). Consistent with the 2007 data, same-sex spouses are older, twice as likely to be raising children, more likely to be homeowners, and have lower employment rates than their unmarried partners. Additionally, same-sex spouses are most likely to be female, to have lower education levels, and lower incomes than their unmarried counterparts (Gates 2009).

Compared to heterosexual married spouses, same-sex spouses are similar in age, education levels, income levels, homeownership rates, and whether they were in an inter-racial relationship. However, they differed in that they were less likely to be raising children and they are slightly less likely to have both spouses working compared to the heterosexual spouses (Gates 2009).

Lastly, it is important to note the theoretical and methodological concerns about comparisons among these couple types. For example, it is possible that a small proportion of same-sex couples may be different-sex couples that miscoded their or their partner’s sex. Likewise, these numbers do not capture couples who do not self-identify as unmarried partners, due to the stigma or concerns of confidentiality, nor are we able to infer relationship of children in the household to anyone other than to “person #1.” This last issue is perhaps the most problematic for studies of same-sex families drawing on the census data. As such, caution should be used when drawing conclusions about same-sex couples and their families. Nevertheless, these analyses do shed light on same-sex families, their children, and the manner in which their relationships compare to those of heterosexual relationships. They also draw attention to issues of same-sex marriage and how marital status (whether couples are legally married or perceive themselves to be married) may be an especially important variable of interest for future demographic summaries of same-sex couples and their households.

5 Theoretical Issues and Research Directions

Kinsey argued that it was impossible to enumerate how many gay men and lesbians are in the population (Baca Zinn and Eitzen 2008). This was primarily due to the lack of data and the difficulty in counting a hidden and stigmatized population. Although strides have been made, to date, we still have no count of the gay male and lesbian population and are still grappling with the same theoretical and methodological issues. Available data are limited and are prone to criticism. However, with an understanding of the data limitations and conceptual assumptions, reliable counts at the household level are increasingly possible, allowing better access to same-sex partners and their families.

The 2000 Census gave us the largest-ever nationally representative dataset with which to study same-sex households. It may have even spurred the increased attention given to demographic analyses of same-sex partners and their families in the last 10 years, along with growing social movements related to issues of sexual orientation. It is expected that the 2010 Census will improve upon the 2000 data, giving us better counts and descriptions of same-sex households. As with the 2000 questionnaire, the 2010 questionnaire will not explicitly address sexual orientation of individuals, but it will continue to allow the census to recognize relationships of same-sex partners. It will also mark the first official count of same-sex couples who self-identify as spouses (Conant 2009). Data released to the public will essentially be edited to show only unmarried partners (as with the 2000 data) and same-sex spouses will not be recognized in accordance with Federal DOMA guidelines, however, supplementary data and special reports have been released regarding the numbers of same-sex households—married and unmarried. According to the Census Bureau’s preferred estimates, as of April 2010, there were 131,729 same-sex married households and 514,735 unmarried partner households for a grand total of 646,464 same-sex partner households. This is an 80% increase from the 2000 estimates of total same-sex partner households (O’Connell and Feliz 2011).Footnote 7

The 2010 Census marks the first census since the inception of state-recognized same-sex marriage. Moreover, it will better capture the impact of state-level legislation concerning gay and lesbian families (such as non-discrimination laws, and laws regarding fostering, adopting, and reproductive technologies) that have been in effect for more than 10 years. This is significant because even though legislation may take effect, there is generally a lag between when it is enacted and its impact.

Future work should consider, and continually assess, the changing social and legal landscape and its impact on same-sex families. As demographers, we know that context is important for understanding social behavior and outcomes. For example, it will become more important to separate same-sex unmarried partners and same-sex married partners as more jurisdictions enact same-sex marriage, granting married couples access to certain rights and privileges that are attached to legal marriage. However, caution should also be taken when considering the same-sex married partners and their characteristics as their numbers do greatly exceed the administrative data on same-sex marriages (O’Connell and Lofquist 2009; Leff 2009). In 2000, 30% of same-sex couples indicated a spousal relationship at a time when none of them could have been legally married to one another (Leff 2009). At present, there are approximately 35,000 couples that have been legally married, primarily in California and Massachusetts, and 10 times this rate are identifying spousal relationships (Leff 2009). This suggests that same-sex couples may be applying their own conceptualizations of marriage to represent their relationships (many may have had commitment ceremonies or feel that they live as if they were married), rather than the very formal legally defined definition of marriage that the census employs.

Findings such as the over-estimation of spouses illuminate the complexities of working with surveys that do not consider issues of sexual orientation during their construction. Future population and family surveys should consider this population and related issues of sexual orientation during their construction. Furthermore, future demographic research should be willing to draw on qualitative and multi-method analyses in order to improve our understanding of same-sex families and their understandings of the category meanings related to surveying. Drawing solely on quantitative surveys does not provide a complete picture and we are still unable to ascertain some of the most basic questions about same-sex families. While researchers can be more confident in enumerating how many children are in same-sex households (or those who choose to identify as same-sex households), very little is known about how they come to be in these households and the complexities of their relationships to household members and the state.

Nevertheless, drawing on nationally representative survey data has allowed us to shed light on the presence of same-sex families and their children, and has given us insight to how they demographically compare to heterosexual families. Given the current social, political, and legal climate, this may be especially important because there are many social assumptions that surround same-sex families and sexual orientation. For example, same-sex partners are often associated with privilege—being white, educated, and wealthy. However, income analyses have found that, at the individual level, gay men earn less than married men (Baumle et al. 2009), and approximately 20% of children belonging to gay couples live in poverty compared to 10% of ­children belonging to heterosexual couples (Conant 2009). Likewise, one might assume that external structural (legal and social) barriers and added financial resources may render it challenging for two individuals of the same-sex to have children, absent a prior heterosexual relationship. However, there is not a notable difference in income between same-sex households with children and those without (Baumle et al. 2009). With large-scale nationally representative data, stereotypes can be replaced with factual information.

6 Conclusion

This chapter has reviewed the demographics of the families of gay men and lesbians, including conceptualizations of family, obstacles in studying same-sex families, and some suggestions for future research. It is evident that the environment for demographically studying same-sex families is much better now than compared to 10 years ago. Descriptively, we know a great deal more about same-sex families and their households. We have also begun to make inferences about how same-sex families are organized and how issues related to being a sexual minority affect them. However, within family demography, the study of same-sex families is still a very minor subset of the field. Most demographic research does not consider issues of sexual orientation in relation to greater demographic processes and transitions (largely due to the lack of data). As such, many gaps and questions remain. Nevertheless, there are reasons for continued optimism that our access to this population will continue to open up (especially via the 2010 Census). Reliable counts and portrayals of same-sex families are possible, as is the on-going access to improved measures and data. Hopefully, this will further develop to a point where demographic research on same-sex families contributes to an overall understanding of demographic processes and transitions and, ideally, lead to a point where most, if not all, family demographers will consider issues of same-sex families when they think about survey and research designs, and demographic behaviors.