MARK REGNERUS RESEARCH: How different are the grown children of parents of same-sex unions?

How different are the grown children of parents of same-sex unions? Results of a New Family Composition Study

Mark Regnerus

Department of Sociology and Population Research Center, University of Texas at Austin, USA

Article Information

Information about the article:

Received February 1, 2012

Redesigned February 29, 2012

Adopted March 12

2012 year

Keywords :

Parenting in same-sex marriage

Family composition

Early adulthood

Sampling problems


The New Family Structures Study (NFSS) is a sociological information project where a survey was conducted among a wide range of randomly selected young Americans (ages 18-39) raised in different types of families. In this NFSS debut article, I compare how adult children of parents with same-sex romantic relationships succeed in 40 different social, emotional, and relational output variables when compared to six other types of families. The results demonstrate numerous systematic differences, especially between the children of women in lesbian relationships and married (heterosexual) biological parents. These results are also usually consistent in multivariate contexts, assuming a much greater heterogeneity of situations in lesbian families than the conclusions drawn from studies of a non-representative sample of lesbian families. The NFSS has proven to be an eloquent, multilateral database that can greatly assist family professionals in understanding the depth of influence of family composition and related elements.

© 2012 Elsevier Inc. All rights reserved.

The text has been translated from the original:


1. Introduction

Child well-being has long been at the center of a public debate on marriage and family policies in the United States. The importance of this issue has not diminished even now, when the legislative, judicial authorities of the states and voters are trying to determine the legal boundaries of marriage. Sociology data remains one of the few sources of information useful in legal debates regarding marriage and adoption rights, highly valued by both proponents and opponents of same-sex marriage. Behind the marriage and child development policy are questions about the possible impact of family composition on children: the number of parents present and actively involved in the lives of children, their genetic relationship with children, the marital status of parents, their gender differences or similarities, and the number of changes in family structure . In this introduction to the New Family Composition Survey (NFSS), I compare how young people of different family backgrounds succeed in 40 different social, emotional, and relational categories. In particular, I focus on how respondents who claim that their mother is in a same-sex relationship with another woman or their father - with another man are comparable to their original, holistic heterosexual married families using nationally representative data from large-scale random sampling among American youth. Sociologists specializing in the frequency of changes in family composition have until recently noted a pattern of higher stability and the social advantage of complete (heterosexual) families compared to single mothers, cohabiting couples, adoptive parents and divorced, custody of children (Brown, 2004; Manning et al., 2004; McLanahan and Sandefur, 1994). In 2002, Child Trends, a respected non-governmental research organization, provided detailed information on how important it is for a child to grow up in the “presence of two biological parents ” (emphasized by them; Moore et al., 2002, p. 2). Out of wedlock motherhood, divorce, cohabitation, families with a step-parent were widely perceived as something less successful in important areas related to development (such as education, behavior problems, emotional well-being), largely due to the comparative fragility and instability of such relations.


However, when the American Sociological Survey article was published in 2001, which discussed the results of studies on sexual orientation and parenting, sociologists Judith Stacey and Tim Biblarz began to note that although between children from same-sex and heterosexual unions and there are some differences, there are not so many of them as sociologists expected, and these differences should not necessarily be interpreted as shortcomings. Since then, public opinion, formed by comparative studies of same-sex parenting, claims that there are very few noticeable differences in children whose parents are gay or lesbian (Tasker, 2005; Wainright and Patterson, 2006; Rosenfeld, 2010). Moreover, recent studies have demonstrated a number of possible benefits of parenting by lesbian couples (Crowl et al., 2008; Biblarz and Stacey, 2010; Gartrell and Bos, 2010; MacCallum and Golombok, 2004). Thus, the scientific opinion on parenting by gays and lesbians began to challenge the pre-existing beliefs about the perceived benefits of parenting in biologically holistic, complete heterosexual families.

1.1. Sampling problems in previous studies

However, the question arose about the methodological quality of many studies concerning same-sex parents. In particular, most of them were based on non-random, non-representative data, often obtained using a limited sample that did not allow the results to be transferred to a common set of gay and lesbian families (Nock, 2001; Perrin and Committee on Psychosocial Aspects of Child and Family Health, 2002; Redding, 2008). For example, many published studies on same-sex parent children are based on snowball data (eg, Bos et al., 2007; Brewaeys et al., 1997; Fulcher et al., 2008; Sirota, 2009; Vanfraussen et al., 2003). One noteworthy example is the National Longitudinal Lesbian Family Study, whose analytical articles were especially popular in the media in 2011 (e.g. Huffington Post, 2011 ).

NLLFS uses “convenient sampling”, which attracts candidates entirely by self-selection from ads printed “at lesbian events, in women's stores and lesbian newspapers” in Boston, Washington and San Francisco. Despite the fact that I do not want to downplay the importance of such longitudinal studies - this is considerable skill in itself - such a sampling method is a problem if the goal (or, in this case, the practical result and using the obtained data as generally accepted) is a generalization to level of the entire population. All such samples are biased, often for unknown reasons. One expert claims that as a formal sampling method, “snowball sampling is known to have a number of serious problems” (Snijders, 1992, p. 59). In fact, such samples are easily rejected in favor of “including those who have many relationships, or are associated with a large number of other individuals” (Berg, 1988, p. 531). However, apart from knowing the probability of inclusion of individuals, an objective assessment is not possible.

Now, as Nock (2001) politely pleaded, let's look at a non-representative sample taken from organizations dedicated to promoting gay and lesbian rights — for example, the NLLFS sampling strategy. Suppose, for example, that respondents have a higher level of education than other lesbians who do not often attend such events or bookstores, or live in a different area. If such samples are used for research purposes, then everything related to educational achievements - better health, thoughtful parenting, the availability of public capital and educational opportunities for children - will be biased. Any statements about the population as a whole based on a group that does not represent it as a whole will be distorted, since the selection of lesbian families from this group is less diverse (based on what is known about it) than the selection of a representative audience (Baumle et al., 2009).

To aggravate the problem, the results of nonrandom samples - from which comprehensive statistics cannot be extracted - are regularly compared with nationwide samples of heterosexual families, which undoubtedly consist of a mixture of qualitatively better and worse parents. For example, Gartrell et al. (Gartrell et al., 2011a, b) studied the sexual orientation and behavior of adolescents by comparing the National Survey of Family Growth (NSFG) data with the NLLFS snowball sample. Comparison of the population sampling scale (NSFG) with the selective sampling of youth from same-sex marriage does not provide the statistical reliability required by high-class social science. Until recently, all this was the main way to collect and evaluate data on same-sex parents. This does not mean that snowball sampling is by definition problematic as a data collection technique - it just is not suitable for making useful comparisons with samples that are fundamentally different in terms of sampling characteristics. Snowball sampling and many other types of non-representative sampling are simply not generalizable and not comparable with a wide range of the target population as a whole. Although the researchers themselves usually point out this important limitation, it is often completely lost in the translation and transfer of research results by the media to the public.

1.2. Are there any noticeable differences?

The “lack of differences” paradigm suggests that children from same-sex families show a lack of noticeable negative characteristics when compared with children from other forms of the family. This assumption increasingly even includes comparisons with complete biological, complete families — the form most associated with benefits in the stability and development of children (McLanahan and Sandefur, 1994; Moore et al., 2002).

The answers to questions about significant intergroup differences one way or another usually depend on who they are compared with, what results the researchers studied, whether the evaluated results were considered significant, insignificant, or potentially risky. Some effects — for example, sexual behavior, gender roles, and democratic parenting — have been valued differently in different periods of time in American society.

For the sake of brevity - and to provide enough space for describing NFSS - I will not devote much time to characterizing previous studies, the methodological problems of which have already been considered by NFSS. Several review articles and at least one book have attempted to provide a more thorough assessment of the literature on this subject (Anderssen et al., 2002; Biblarz and Stacey, 2010; Goldberg, 2010; Patterson, 2000; Stacey and Biblarz, 2001a). Suffice it to say that different versions of the phrase “no differences” have been used in a wide variety of studies, reports, testimonies, books and articles since 2000 (e.g., Crowl et al., 2008; Movement Advancement Project, 2011; Rosenfeld, 2010; Tasker , 2005; Stacey and Biblarz, 2001a, b; Veldorale-Brogan and Cooley, 2011; Wainright et al., 2004).

Earlier studies of same-sex families usually compared the developmental results of children of divorced lesbian mothers with those of divorced mothers from heterosexual families (Patterson, 1997). The same strategy was followed by psychologist Fiona Tasker (2005), who compared lesbian mothers with divorced heterosexual mothers and found "lack of systematic differences in the quality of family relationships." Wainright et al (Wainright et al. (2004)), using 44 stories from the Add Health nationally representative study database, reported that adolescents living with same-sex female parents showed self-esteem, psychological adaptation, academic achievement, and crime rates, drug use and the quality of family relationships, comparable to 44 demographically “relevant” stories of teenagers of the opposite sex. However, this suggests that in this case too, comparisons were hardly made with respondents from married, stable, biologically holistic families.

The problem is that small-scale samples can help draw conclusions with "no differences." Not surprisingly, statistically significant differences did not appear in studies of only 18, 33, or 44 cases of respondents with same-sex parents, respectively (Fulcher et al., 2008; Golombok et al., 2003; Wainright and Patterson, 2006). Even the analysis of “suitable” samples, carried out in many studies, cannot level the problem of revealing statistically significant differences when the scale of the sample is small. This is a challenge that needs to be resolved in all social science, not to mention doubly important areas where there may be an urge to approve the null hypothesis (that is, there are really no statistically significant differences between the groups). Therefore, one of the most important questions in these studies is simply whether there is enough statistical power to identify significant differences, if any. Rosenfield (Rosenfeld, 2010) was the first scientist to use a large, random sample of the population to compare the effects among children of same-sex parents and married heterosexual parents. He came to the conclusion - after checking the level of education and the income of the parents, and having decided to limit the sample to families who had been living stably together for at least 5 years - there was a lack of statistically significant differences between the two designated groups in a pair of criteria that assess the success of children over the years of primary school.

Sex-related investigations more steadily show clear differences, although the tone of concern over them has declined over time. For example, although it is now increasingly recognized that daughters of lesbian mothers are more likely to be interested in same-sex sexual identity and behavior, concern over these findings has subsided as scientists and the public as a whole have become more open to LGB identity (Goldberg, 2010). Tasker and Golombok (1997) noted that girls raised by lesbian mothers had more sexual partners at the beginning of adulthood than daughters of heterosexual mothers. The sons of lesbian mothers, on the other hand, seem to have the opposite tendency - fewer partners than sons of heterosexual mothers.

More recently, however, the tone about “lack of differences” has shifted somewhat to the advantage that homosexual parents are more competent than heterosexual (Biblarz and Stacey, 2010; Crowl et al., 2008). Their romantic relationship may even be better: a comparative study of Vermont gay civil unions and heterosexual marriages showed that same-sex couples show a higher quality of relationships, compatibility and intimacy, and even fewer conflicts than heterosexual couples (Balsam et al., 2008 ) A review article by Biblarz and Stacey's (2010) on gender and parenting argues that, based on well-known scientific evidence, it can be argued that when parents are two women, it’s on average better than a woman and a man, or at least least a woman and a man with a traditional division of labor. Lesbian co-parents seem to perform better than comparable heterosexual married, biological parents in several ways, although they are denied significant marriage privileges (p. 17).

Even here, however, the authors note that lesbian parents face a “slightly greater risk of separation” caused, in their opinion, by their “asymmetric biological and legal status and high standards of equality” (2010, p. 17).

Another meta-analysis claims that non-heterosexual parents, on average, have significantly better relationships with their children than heterosexual parents, along with the lack of differences in cognitive development, psychological adaptation, gender identity and sexual partner preferences (Crowl et al., 2008 )

However, such a meta-analysis only emphasizes the critical importance of the one who provides the information - almost always it is a small study of a volunteer group whose documented successes in parenting are closely related to recent legislative and judicial debates about rights and legal status. Tasker (Tasker, 2010, p. 36) calls for caution:

Own reports of parenting, of course, may be biased. It is justified to argue that in a biased social climate, lesbian and gay parents may be very interested in presenting some positive picture. Будущим исследованиям следует задуматься о том, чтобы прибегнуть к дополнительным передовым мерам, чтобы устранить потенциальную необъективность…