An Evaluation of Gay/Lesbian and Heterosexual Adoption

Abstract

Many experts in the helping professions have agreed that there is no scientific credence to support a gay and lesbian adoption ban. Nevertheless, there continues to be persistent mythology pertaining to outcomes for children adopted by gay and lesbian parents. This position may be somewhat due to the dearth of research that compares heterosexual and homosexual parenting outcomes with adopted children. To respond to this gap in the literature, this study explored the extent of emotional and behavioral problems among children aged 1.5 to 5 years (n = 380) and 6 to 18 years (n = 1,004) with gay and lesbian or heterosexual adoptive parents. A multiple regression analysis was used to assess the association between the dependent variables (child internalizing and externalizing behavior) on adoptive parent sexual orientation (gay and lesbian or heterosexual) while controlling for child age, child sex, pre-adoptive maltreatment, co-sibling adoption, adoption preparation, family income, and family functioning. As hypothesized, results indicted that child internalizing and externalizing behavior was not contingent upon adoptive parent sexual orientation. Rather, regardless of sexual orientation, adoptive parents are likely to encounter similar challenges in terms of risk factors for child behavioral problems and mitigating factors of such behavior. Recommendations for practice, policy, and future research are highlighted.

KEYWORDS:

• gay
• lesbian
• heterosexual
• adoption
• family functioning
• parenting
• child
• behavioral issues

INTRODUCTION

As of 2007, there were approximately 130,000 children in the child welfare system waiting to be adopted (U.S. Department of Health and Human Services, 2007). Yet congress noted in 2007 that there were “serious shortages” of qualified adoptive parents (Library of Congress, 2007). The American Civil Liberties Union (1999) stated that many gay and lesbian families are interested in adopting and willing to adopt children and are often open to accepting the harder to place children such as those who are older. Gates, Badgett, Macomber, and Chambers (2007) stated that there are approximately 2 million gay and lesbian people interested in adopting. Yet policies of adoption agencies, social stigma, and state laws have created barriers for gay and lesbian couples in the adoption process (Shelley-Sireci & Ciano-Boyce, 2002).

Last year, the Florida law that for over 3 decades banned adoptions by gays and lesbians was ruled unconstitutional (Ruggeri, 2008). Florida has been the only state with an outright ban on adoption by gays and lesbians, although in both Arkansas and Utah there are bans any unmarried persons adopting or fostering children. Mississippi bans homosexual couples, but not single lesbian, gay, bisexual, and transgender (LGBT) persons, from adopting (Anderson, 2008). In 2007, New Hampshire repealed a 15-year ban on gay and lesbian adoption, after hearing extensive testimony from children's advocates (Barillas, 2007).

Many experts in the fields of child psychology and social work have agreed that there is no scientific evidence to support a gay and lesbian adoption ban (Anderson, 2008). Among others, Bigner and Bozett (1989) found in their review of research on homosexual parents that there is “no evidence of any kind” that validates beliefs that children living with homosexual parents experience significant negative outcomes. More recently, research conducted by Gartrell, Peyser, and Bos, Paterson and Wainright, and Stacey and Biblarz (as cited in Ryan, in press) all offer that there is no reliable scientific evidence that proves that lesbian or gay parents arrange their homes differently or are inept parents or that their children develop differently from those in heterosexual homes. Organizations such as the American Academy of Pediatrics, the American Medical Association, the American Psychiatric Association, and the National Association of Social Workers all support permitting same-sex couples to adopt (American Civil Liberties Union [ACLU], 1999).

Yet there continues to be persistent mythology pertaining to outcomes for children adopted by gay and lesbian parents (ACLU, 1999). For example, in the Florida case experts called by the state claimed there was a higher incidence of drug and alcohol abuse among gay couples, that they were more unstable than heterosexual unions, and that the children of gay couples suffer from social stigma (Ruggeri, 2008). Many of the persistent and pervasive myths are grounded in homophobia and a heterosexual bias that views heterosexual couples as the gold standard of parenting practices. It is argued that heterosexual parents have better results in raising children. This position may be based in some part on the lack of research that compares heterosexual and homosexual parenting outcomes with adopted children.

Toward this end, this study will present findings from a sample of gay and lesbian adoptive parents and heterosexual adoptive parents. This study compared the extent of emotional and behavioral problems of adopted children given the sexual orientation of their adoptive parents. Specifically, controlling for various confounding variables, we hypothesized that parent sexual orientation would not have a significant impact upon the emotional and behavioral functioning of their children.

LITERATURE REVIEW

Research on gay and lesbian adoptive families has generally been scarce (Van Voorhis & Wagner, 2001). However, the literature is growing (Ryan, 2007). There are a few studies that have looked particularly at the functioning level of gay and lesbian adoptive families. In 2004, Ryan and Cash surveyed 183 lesbian and gay families on a variety of topics that included family dynamics. Participants reported that those in partnerships had been so for a significant amount of time, many had a spiritual affiliation, and many reported a high level of social support. Although findings were based on a nonrandom sample and dependent upon self-report, it was the first large-scale, national adoption study of its kind. Then in 2005, Erich, Leung, Kindle, and Carter examined the functioning of lesbian and gay adoptive families. They found in their convenience sample of 47 parents and their 68 adopted children that all of the families functioned in the average/typical or strength range. Similar to Ryan and Cash's (2004) findings, the parents had developed solid support networks. However, the 2005 pilot study also had the inherent methodological weaknesses similar to many adoption studies, such as homogeneity in the sample, convenience sampling, and self-reporting. In 2007, Ryan used the same sample used earlier by Ryan and Cash (2004) to report specifically on the parenting styles of gay and lesbian adoptive parents and the strengths of the children. Ryan found a high level of parenting skill in the 183 families. As well, the parents assessed the children as having several high-level strengths. Ryan concluded that the children in the sample were “growing up in healthy families with strong, capable parents which has resulted in the children themselves showing many areas of strengths” (p. 128). Although confined by the limitations described about the Ryan and Cash (2004) study, it should be noted that the results of both the Ryan and the Ryan and Cash studies were based on information gained from valid, reliable, and widely used instruments.

Articles on other salient gay and lesbian adoption topics have also been published in the recent past. Most frequent in the gay and lesbian adoption literature are nonempirical articles that provide guidelines for practice and important policy implications. For example, in 2000 Mallon drew on the research literature, case examples, and interviews to elucidate the issues surrounding gay and lesbian adoptive parenting and to provide child welfare policy implications. In the same manner, Mallon (2007) used the research literature, case examples, and 31 years of child welfare practice wisdom to provide competent practice and assessment guidelines for child welfare workers in the consideration of lesbian and gay people as prospective foster/adoptive parents. While stating that just as not every heterosexual should adopt, not all gay and lesbian people should foster parent or adopt, the author does point to the need for the same “fair process and open opportunity” as heterosexuals are afforded (p. 84). Matthews and Cramer (2006) also provided policy and practice knowledge for adoption professionals in the consideration of gay and lesbian headed families. Ryan, Pearlmutter, and Groza (2004) used systems theory to provide a model for adoption workers when making decisions about gay and lesbian persons as adoptive parents. Although greatly limited in terms of research impact, these articles provided practice and policy considerations that are particularly useful and timely. The potential of the articles are specifically relevant when considering the heterosexist opinions of some child welfare workers (Ryan, 2000).

In the same topic vein but empirically based, Ryan and Whitlock (2007) completed a cross sectional study of 96 lesbian adoptive parents to explore their perception of the adoption experience. Findings indicated that the adoptive mothers found the process generally positive and were somewhat/very satisfied with the experience, although all of the women perceived bias on the part of the adoption professionals. Ryan and Whitlock's study had the oft-found limitations of a convenience sample and self-report measures. However, the importance of allowing the lesbian adoptive mothers to provide their own perceptions of the adoption process is considered important in and of itself. In a critical examination of research methods and processes, Jacobson (1995) noted that marginalized groups must be given the opportunity to share their perceptions of potentially disempowering and discriminatory experiences.

There are a few studies in the literature that specifically compare gay/lesbian parents to heterosexual parents. Shelley-Sireci and Ciano-Boyce (2002) compared the process of becoming parents for lesbian adoptive, lesbian birth, and heterosexual adoptive parents. They surveyed 18 lesbian adoptive mothers from 10 different households, 49 lesbian birth mothers from 27 households, and 44 heterosexual adoptive families from 26 different households. Due to high levels of variance between co-parents, results were based on individual rather than family responses. There was no significant difference in how long lesbian or heterosexual adoptive parents were together before they adopted or in how long the adoption process took. Lesbian adoptive parents were more likely to adopt a girl for their first child, although this could be explained by their higher rate of adoption from China. Both lesbian and heterosexual parents in the sample rated their adoption experience as “neutral.” However, lesbian parents perceived more discrimination as a result of their being lesbian and unmarried. As well, they also believed at a much greater rate than the heterosexual parents did that they needed to omit information in the home study in order to have a successful outcome. Shelley-Sireci and Ciano-Boyce's exploratory study faced the same challenges described above of those who study gay and lesbian parenting, such as a small, homogeneous sample. Kindle and Erich (2005) also faced these same sampling limitations in their study of the perceptions of social support by heterosexual and homosexual adopters. Using the Family Support Scale, they obtained data from 47 gay/lesbian and 25 heterosexual respondents. They found no significant difference in the overall levels of support. However, the heterosexual sample had higher levels of support from relatives and biological children, while the gay and lesbian sample reported higher levels of support from their partner and day care center. It is important to note that some differences between the sample sets, such as the heterosexual sample having a higher number of biological children and the homosexual sample having a high rate of partnership, could attribute for the respective higher rates of support.

The most current empirical study available on the adoption experience is a comparison of 82 gay or lesbian adoptive families to 1,071 heterosexual families (Brooks, Kim, & Wind, in press). The study examined family characteristics, child characteristics, preparation for adoption, and the adoption experience. Brooks et al. (in press) found that the gay and lesbian adopters received similar adoption preparation and had similar availability to services in comparison to the heterosexual adopters. Differences between the groups were found in that gay and lesbian adoptive families utilized educational and informational services, while heterosexual families utilizes clinical services. As well, gay and lesbian adoptive parents rated legal advice, child counseling, and communication classes more helpful that their heterosexual counterparts. Overall, the findings suggested that the roles of agency services were similar for gay and lesbian adoptive families and heterosexual adoptive families. As well, the gay and lesbian families were found to “… adopt similar kinds of children, receive similar kinds of support, and find the services they receive as helpful as straight adoptive families” (Brooks et al., in press, conclusion section, ¶1).

Leung, Erich, and Kanenberg (2005) compared family functioning in gay/lesbian, heterosexual, and special needs adoptions. Using three data sets of 86 special needs adoptive parent respondents, 47 gay/lesbian adoptive parent respondents, and 25 heterosexual adoptive parent respondents, data were collected via five highly regarded self-report measures. Findings included “no negative effects upon family functioning associated with gay/lesbian sexual orientation of adoptive parents” (p. 1042). In fact, the results suggested that functioning was higher when homosexual families adopted older children. The authors noted the threats to internal and external validity presented by the use of convenience sampling and cross-sectional studies. Most recently, Erich, Kanenberg, Case, Allen, and Bogdanes (2009) analyzed factors affecting adolescent attachment in both homosexual and heterosexual adoptive parents. Using self-report measures, convenience sampling, and small sample sizes, they did not find significant differences based upon parent sexual orientation.

Thus, the current literature on gay and lesbian adoption and comparison of gay and lesbian adoption to heterosexual adoption supports New Hampshire's and Florida's recent decisions to repeal bans on gay and lesbian adoption. The current, albeit limited, literature continues to demonstrate that gay and lesbian parents are viable, strong, and an untapped source for children in need of adoption. Building on the need for continued research on gay and lesbian family functioning and specifically in comparison to heterosexual functioning, the current study was completed.

METHODOLOGY

Design and Data Collection Process

This study utilized a cross-sectional survey design and consisted of two data sets. The first data set consisted of families participating in the Florida Adoption Project (FAP). The goal of the FAP is to examine key indicators of success among the population of parents who had adopted a child through Florida's public child welfare system and were receiving a special needs subsidy payment. Adoptive parents in Florida receiving an adoption subsidy for at least one adopted child in their care were eligible for the study. Parents electing to be removed from the survey mailing, those with incorrect addresses, and children ineligible for the study (i.e., older than 18) were excluded from the frame. Parents were sent a coded packet that included a cover letter, informed consent form, and survey instrument. The completion of the survey indicated the parent's consent. This study is based on the first wave of data that were collected annually at three points via a mail survey of adopting parents. In some cases, adoptive parents completed surveys on siblings whom they adopted. Thus, 1,694 families returned surveys for 2,382 children.

The second data set consisted of gay and lesbian individuals and couples throughout the United States. The sample gay and lesbian adoptive parents were obtained through a variety of media in order to ensure an adequate sample size for this comparative study. Newspaper ads were placed in several metropolitan gay and lesbian weekly newspapers, adoption magazines, and gay parenting magazines as well as a designated Web site. Once contact was initiated by the participants via e-mail or phone, a research team member would contact parents and describe the purpose of the study and the procedures, and if they wished to participate, the researcher obtained their address. A cover letter, consent form, and survey were sent to all participants who were interested in participating in the survey. Of the 281 surveys distributed, 183 surveys were returned.

Sample

To ensure comparability between the two data sets, three sampling criteria measures were followed. First, only gay and lesbian adoptive couples and heterosexual adoptive couples met inclusion criteria for this study. Regardless of sexual orientation, single adoptive parents were excluded. Second, although some FAP adoptive parents completed surveys on siblings whom they adopted, only the oldest adopted sibling was included in this study. This is consistent with the gay and lesbian data set where the child of focus was the oldest adopted child when they had adopted siblings. Thus, the data set included adoptive parents who had adopted one child or the oldest child in sibling adoptions. Finally, the subsample for the present analysis included parents who adopted children aged 1.5 through 18 years.

The subsample of children were classified as having a gay and lesbian adoptive parent (n = 155) or heterosexual adoptive parent (n = 1,229). For the analyses that follows, we divided the sample into two age groups of children, 1.5 to 5 years (n = 380) and 6 to 18 years (n = 1,004). Of the 1.5- to 5-year-old children, 86 (22.6%) were children of gay or lesbian parents and 294 (77.4%) were children of heterosexual parents. In terms of the 6- to 18-year-old children, 69 (6.9%) were children of gay or lesbian parents and 935 (93.1%) were children of heterosexual parents.

Measures

Although the gay and lesbian survey contained information specific to gay and lesbian parents, both surveys contained common domains that are reported on in this study, including information about parent and child characteristics, family composition and dynamics, the child's pre-adoptive history (maltreatment history), and current emotional and behavioral functioning. Measures were treated as independent, dependent, and control variables. The main independent variable in this study was parent's sexual orientation (1 = heterosexual; 0 = gay and lesbian). The main dependent variable was child internalizing and externalizing behavior. In order to rule out alternative explanations for child behavior, we include several covariates in the models.

Dependent Variables

The two dependent variables were measured by the raw scores on the internalizing problems and externalizing problems scale of the Child Behavior Checklist (CBCL) 1.5 to 5 and 6 to 18 versions (Achenbach & Rescorla, 2001). The CBCL is one of the most widely used measures of child behavioral functioning and has sound psychometric properties (Achenbach & Rescorla, 2001). The items are rated by caregivers on a 3-point Likert-type response format (0 = not true, 1 = somewhat or sometimes true, 2 = very true or often true). A high score indicates more child behavior problems. In the present sample, Cronbach's alpha coefficients for internalizing and externalizing were more than .88 for both versions.

Control Variables

Based on their empirical associations with adoption outcomes, control factors were included in the multivariate models. These include sex (1 = female, 0 = male) (Brooks, Simmel, Wind, & Barth, 2005), child age (Sharma, McGue, & Benson, 1996), pre-adoptive history of maltreatment (1 = yes, 0 = no) (Nalavany & Ryan, 2008; Simmel, 2007), co-sibling adoption (1 = yes, 0 = no) (Erich & Leung, 2002), adoption preparation (Nalavany, 2006; Simmel, 2007), family income (Ryan, Glover, & Cash, 2006), and family functioning (Leung & Erich, 2002).

Family functioning was measured by the Family Function Style Scale (FFSS) (Trivette, Dunst, Deal, Hamby, & Sexton, 1994). The FFSS is based on the strengths perspective and measures the degree to which the respondent believes one's family is represented by different strengths and capabilities. The FFSS consists of 26 items (e.g., “We believe that something good always comes out of even the worst situations”) rated on a 5-point scale ranging from (1) “not at all like my family” to (3) “sometimes like my family” to (5) “almost always like my family.” We used the total score as a global assessment of family functioning. Higher scores reflect more positive family functioning, whereas low scores indicate more negative family functioning. The coefficient alpha for the scale with this sample was .95.

To assess adoption preparation, we used an 8-item scale that asked parents to rate their satisfaction with the services provided when adopting their child (Nalavany, 2006). Questions dealt with parents’ perception of satisfaction with the social workers responsiveness to questions and concerns, accuracy of timeline given for the adoption, the quality and quantity of medical information provided, social worker's knowledge of adoption-related issues, handling of problems, and concern for the adopted child and family. Responses to each item were based on a 5-point Likert type scale, with higher scores representing more satisfaction with services provided at adoption. In the present sample, Cronbach's alpha was .95.

RESULTS

Table 1 summarizes child and parent characteristics according to age group and parent sexual orientation. In addition to descriptive statistics, t-tests for continuous variables and chi-square for dichotomous and categorical variables were conducted to determine group differences between the children with gay and lesbian parents and heterosexual parents (see Table 1). To be consistent with the multivariate analyses to follow, we present the results according to the two different age groups of children, 1.5 to 5 years and 6 to 18 years.

TABLE 1 Characteristics of Adopted Parents and Children

	1.5 to 5 Years (n = 380)		6 to 18 Years (n = 1,004)
	Gay/Lesbian	Heterosexual	Gay/Lesbian	Heterosexual
	(n = 86)	(n = 294)	(n = 69)	(n = 935)
Parent race/ethnicity
Caucasian	100%∗	78.8%	94.2%∗	76.4%
African American	0%	12.7%	0%	18.9%
Latino(a)	0%	7.1%	1.4%	3.4%
Biracial/other	0%	1.4%	4.3%	1.3%
Parent education
High school/GED/associate's degree	14.0%∗	71.2%	24.6%∗	72.7%
Bachelor's degree	27.9%	19.6%	26.1%	15.9%
Master's/doctoral degree	58.1%	9.3%	49.3%	11.4%
Transracial adoption (yes)	11.6%∗	38.7%	18.8%	26.1%
Parent age, years	43.69 (6.52)	42.37 (7.81)	46.06 (6.19)	48.02 (8.88)
Family income	$118,619∗	$62,798	$111,207∗	$58,222
Years partnered/married	10.45 (5.31)∗	14.19 (8.46)	12.79 (7.19)∗	19.62 (10.16)
Child race/ethnicity
Caucasian	25.6%∗	46.7%∗	39.1%∗	56.9%
African American	11.6%	22.8%	14.5%	23.3%
Latino(a)	18.6%	9.0%	14.5%	7.9%
Biracial/other	44.2%	21.5%	30.4%	12.0%
Child placement age	1.06 (1.33)	1.01 (1.13)	3.72 (3.75)	4.07 (3.67)
Child current age	3.69 (1.35)∗	4.28 (1.30)	10.47 (3.20)∗	12.1 (3.23)
Sibling adoption (yes)	7.0%∗	16.7%	18.8%∗	38.8%
Child sex (male)	62.8%	54.4%	58.0%	51.6%
FFSS∗∗ total score	110.16 (11.19)	111.98 (13.85)	105.49 (11.52)∗	109.37 (14.38)
Adoption preparation	33.56 (6.00)	29.45 (9.67)∗	31.01 (7.68)	29.29 (10.03)
Type of abuse prior to adoption
Neglect/abandonment (yes)	31.4%∗	49.3%	53.6%∗	67.0%
Physical abuse (yes)	4.7%∗	17.7%	33.3%	34.4%
Sexual Abuse (yes)	2.3%	1.4%	23.2%	17.6%
CBCL internalizing problems	5.53 (4.26)	6.79 (7.30)	6.94 (7.18)	6.81 (7.06)
CBCL externalizing problems	7.88 (5.06)∗	10.28 (9.28)	8.57 (9.78)∗	11.45 (11.70)
Note. CBCL = Child Behavior Checklist; FFSS = Family Functioning Style Scale.
∗p < .01.

Sample Characteristics: 1.5 to 5 Years

The ethnic makeup of the parents was not diverse, in that 100% of the gay and lesbian and 78.8% of the heterosexual parents were Caucasian. A significant majority of gay and lesbian parents (58.1%) had obtained master's or doctoral degrees, whereas 71.2% of the heterosexual parents had completed high school, GED, or associate's degrees. As such, the annual income of gay and lesbian parents (M = $118,619) was significantly more than that of heterosexual parents (M = $62,798). Significantly more heterosexual parents (38.7%) had adopted a child of different ethnicity than did gay and lesbian parents (11.6%). The average age of gay and lesbian parents (M = 43.69 years) and heterosexual parents (M = 42.37 years) was approximately equal. Gay and lesbian parents were partnered for significantly less years (M = 10.45 years) than heterosexual parents were married (M = 14.19 years). The majority of gay and lesbian parents resided in the West (34.5%), followed by the Northeast (25.9%), Midwest (20.1%), Southeast (9.8%), Southwest (8.6%), and Canada (1.1%).

The ethnic makeup of the children of gay and lesbian adoptive parents was predominately biracial (44.2%), whereas the majority of children of heterosexual parents were Caucasian (46.7%). The age of placement in the adoptive home was approximately equal for children of gay and lesbian parents (M = 1.06 years) and heterosexual parents (M = 1.01 years). In terms of current age, children of heterosexual parents were significantly older (M = 4.28 years) than children of gay and lesbian parents (M = 3.69 years). Sex was approximately equally distributed, with 62.5% and 54.4% of children of gay and lesbian parents and heterosexual parents, respectively, being male. A significant majority of children of heterosexual parents were adopted with their sibling(s) (16.7%) as compared to children of gay and lesbian parents (7.0%). A higher percentage of children of heterosexual parents had pre-adoptive histories of neglect (49.3%) and physical abuse (17.7%) as compared to children of gay and lesbian parents (31.4% and 4.7%, respectively).

Sample Characteristics: 6 to 18 Years

The ethnic makeup of the parents was not diverse, in that 94.2% of the gay and lesbian and 76.4% of the heterosexual parents were Caucasian. A significant majority of gay and lesbian parents (49.3%) had obtained master's or doctoral degrees, whereas 72.7% of the heterosexual parents had completed high school, GED, or associate's degrees. Accordingly, the annual income of gay and lesbian parents (M = $111,207) was significantly more than that of heterosexual parents (M = $58,222). There were no differences in transracial adoptions between heterosexual parents (26.1%) and gay and lesbian parents (18.8%). The average age of gay and lesbian parents (M = 46.06 years) and heterosexual parents (M = 48.02 years) was approximately equal. Gay and lesbian parents were partnered for significantly less years (M = 12.79) than heterosexual parents were married (M = 19.62).

The ethnic makeup of the children of gay and lesbian adoptive parents was predominately Caucasian (39.1%) and biracial (44.2%), whereas the majority of children of heterosexual parents were Caucasian (56.9%). The age of placement in the adoptive home was approximately equal for children of gay and lesbian parents (M = 3.72 years) and heterosexual parents (M = 4.07 years). In terms of current age, children of heterosexual parents were significantly older (M = 12.1 years) than children of gay and lesbian parents (M = 10.47 years). Sex was approximately equally distributed, with 58.0% and 51.6% of children of gay and lesbian parents and heterosexual parents, respectively, being male. A significant majority of children of heterosexual parents were adopted with their sibling(s) (38.8%) as compared to children of gay and lesbian parents (18.8%). Over half of the children of gay and lesbian parents (53.6%) and heterosexual parents (67%) had pre-adoptive histories of neglect, whereas approximately equal proportions had pre-adoptive histories of physical abuse (33.3% and 34.4%) and sexual abuse (23.2% and 17.6%), respectively.

Bivariate Analyses

t-tests (see Table 1) were conducted to determine the impact of parent sexual orientation on family functioning, adoption preparation, internalizing problems, and externalizing problems. In the 1.5- to 5-year-old group, the results indicated that gay and lesbian parents reported significantly more satisfaction with the adoption preparation process (t = –3.72, df = 378, p < .001) and significantly less externalizing problems (t = 2.29, df = 378, p < .05) of their adopted children than heterosexual parents reported. However, no statistically significant associations were found between family functioning and internalizing problems.

The results for the 6- to 18-year-old group indicated that heterosexual parents reported significantly higher levels of family functioning (t = 2.18, df = 1002, p < .05) and more externalizing problems among their children (t = 1.99, df = 1002, p < .05) than reported by gay and lesbian parents. However, no statistically significant associations were found between adoption preparation services and internalizing problems.

To identify potential multicollinearity problems for the multivariate analyses, Pearson's correlations were utilized to assess the associations among continuous analysis variables. The findings indicated a strong positive association between child placement age and child age at survey for the 1.5- to 5-year-old children (r = .462, p < .01) and for the 6- to 18-year-old children (r = .356, p < .01). Since child age at survey was shown to be statistically different for the bivariate analyses (see Table 1), we elected to control for child age at survey and omit child placement age for the multivariate analyses due to the shared variance between these variables.

Multivariate Analyses

Based on the literature, several ordinary least squares multiple regression models were performed and are presented in this section. These examined how, when controlling for, child age, child sex, pre-adoptive maltreatment, co-sibling adoption, adoption preparation, family income, and family functioning, and whether parent sexual orientation would predict internalizing and externalizing behaviors among 1.5- to 5- and 6- to 18-year-old adopted children. Specifically, we hypothesized that sexual orientation of the parents would not be predictive of internalizing and externalizing behaviors. Multiple regression assumptions were checked. Sensitivity analyses did not reveal any outliers or observations that adversely altered the regression coefficients. Various transformations were attempted to improve the distribution of the family functioning variable (FFSS), but none of these transformations made any appreciable difference. Therefore, the family functioning variable was used in its original form for the analyses.

Internalizing and Externalizing Problems: 1.5 to 5 Years

As shown in Table 2, only one variable emerged as a predictor of internalizing problems. An adopted child who was adopted as part of a sibling group would have significantly higher levels of internalizing problems (b = 2.26, p < .05) compared to children not adopted as part of a sibling group, holding all other variables constant. The variables included in the model accounted for only 4.2% of the variance in internalizing behavior.

TABLE 2 Regression Models Predicting CBCL Internalizing and Externalizing Problems

	CBCL Internalizing		CBCL Externalizing
	Problems		Problems
	B	β	B	β
1.5- to 5-Year-Old Children
Constant	9.83		17.39
Child age, years	.432	.085	.147	.023
Sibling adoption (1 = yes, 0 = no)	2.26∗	.118	2.63∗	.108
Adoption preparation	−.058	−.078	−.117∗	−.124
FFSS total score	−.031	−.062	−.052	−.080
Annual income	−9.43E-6	−.078	−7.55E-6	−.049
Child sex (1 = female, 0 = male)	.267	.020	.204	.012
Physical abuse (1 = yes, 0 = no)	1.61	.087	1.73	.072
Neglect/abandonment (1 = yes, 0 = no)	.897	.066	.361	.021
Sexual abuse (1 = yes, 0 = no)	.499	.009	.057	.001
Parent sexuality (1 = heterosexual, 0 = gay and lesbian)	−.312	−.019	.945	.046
Adjusted R ^2	.042		.035
6- to 18-Year-Old Children
Constant	16.41		30.44
Child age, years	.238∗∗	.110	.227∗	.064
Sibling adoption (1 = yes, 0 = no)	−.599	−.041	1.17	.049
Adoption preparation	−.113∗∗∗	−.158	−.247∗∗∗	−.211
FFSS total score	−.072∗∗∗	−.144	−.139∗∗∗	−.170
Annual income	−1.12E-5∗	−.070	−2.44E-5∗∗	−.093
Child sex (1 = female, 0 = male)	−.478	−.034	−2.56∗∗∗	−.110
Physical abuse (1 = yes, 0 = no)	.721	.048	2.36∗∗	.097
Neglect/abandonment (1 = yes, 0 = no)	−.431	−.029	−.420	−.017
Sexual abuse (1 = yes, 0 = no)	2.76∗∗∗	.150	4.44∗∗∗	.147
Parent sexuality (1 = heterosexual, 0 = gay and lesbian)	−.682	−.024	1.53	.034
Adjusted R ^2	.101		.160
Note. CBCL = Child Behavior Checklist; FFSS = Family Functioning Style Scale.
∗p < .05; ∗∗p < .01; ∗∗∗p < .001.

Two variables emerged as predictors of externalizing problems. Table 2 shows that an adopted child who was adopted as part of a sibling group would have significantly higher levels of internalizing problems (b = 2.63, p <.05) as compared to children not adopted as part of a sibling group, holding all other variables constant. Adopted children's externalizing problems significantly decreased as their adoptive parents reported more satisfaction with the adoption preparation process (b = –.117, p < .05). The variables included in the model accounted for only 3.5% of the variance in internalizing behavior. As hypothesized, the sexual orientation of adoptive parents was not a significant predictor of internalizing problems or externalizing problems.

Internalizing and Externalizing Problems: 6 to 18 Years

As shown in Table 2, several variables emerged as significant predictors of internalizing behaviors. A 1-year increase in child age at survey (b = .238, p < .01) and pre-adoptive sexual abuse (b = 2.76, p < .001) was predictive of more internalizing problems. On the other hand, an increase in annual income (b = –.001, p < .001), family functioning (b = –.072, p < .001), and parental satisfaction with adoption preparation services (b = –.113, p < .001) was predictive of significantly fewer internalizing problems. The variables included in the model accounted for 10.1% of the variance in internalizing behavior.

Several variables also emerged as significant predictors of externalizing problems. As illustrated in Table 2, a 1-year increase in child age at survey (b = .227, p < .05), pre-adoptive history of physical abuse (b = 2.36, p < .01), and pre-adoptive history of sexual abuse (b = 4.46, p < .001) accounted for more externalizing problems. Overall, girls were significantly less at risk for externalizing problems as compared with boys (b = 2.56, p < .001). The control variables of annual income (b = –.002, p < .01), family functioning (b = –.139, p < .001), and parental satisfaction with adoption preparation services (b = –.247, p < .001) were predictive of fewer externalizing problems. The variables included in the model accounted for 16.0% of the variance in externalizing behavior. As hypothesized, the sexual orientation of adoptive parents was not a significant predictor of internalizing problems or externalizing problems.

DISCUSSION

Limitations and Strengths

As with many adoption studies, the present study possesses some limitations that need to be considered in order to understand the influence on the findings and their implications. First, as this study utilized a nonrandom sample and cross-sectional design, the findings cannot be generalized to the broader population of adoptive parents and children nor can causal inferences be made. Longitudinal studies in adopted children are needed to examine the developmental trajectories and causal pathways of adjustment (Simmel, 2007). Second, secondary confirmation of variables was absent. Future research could incorporate other informants, such as caseworkers, teachers, or other professionals, to garner a more reliable indication of child behavior. Third and perhaps most significantly, the comparability of the gay and lesbian and heterosexual parents may be suspect since both samples were collected for different study purposes. There were also considerably more heterosexual parents than gay and lesbian parents in both analyses, with ratios approximately 3 to 1 and 13 to 1, respectively. When the number of participants in each group is highly disproportionate, the power to detect statistically significant results is compromised (Cohen, Cohen, West, & Aiken, 2002). That no statistically significant differences were found between gay and lesbian and heterosexual parents may possibly be the result of decreased power. Similarly, the finding that pre-adoptive sexual abuse was nonsignificant in the 1.5- to 5-year-old internalizing and externalizing behavior analyses may be an artifact of decreased power because only 6 children had such histories. We elected to include pre-adoptive sexual abuse in the analyses to be consistent with the analyses in the 6- to 18-year-old adopted children.

Notwithstanding theses limitations, the study contained several strengths. The study adds to our understanding of the factors that may contribute to adopted child internalizing problems and externalizing problems in gay and lesbian families and heterosexual families. In doing so, this study used measures with sound psychometric properties and variables that were measured similarly between the two samples, and to the best of our knowledge, it is the largest comparative study on gay and lesbian and heterosexual adoptive parents. The findings and implications of this study are described below.

Parent Sexual Orientation

In terms of sexual orientation, there were no significant differences between the heterosexual and homosexual parents. This finding is consistent with prior research that compares these two specific groups of adoptive parents (Erich et al., 2009; Kindle & Erich, 2005; Leung et al., 2005), although it should be noted that Leung et al. (2005) found that gay and lesbian adoptive parents fared better than heterosexual adoptive parents did with older adoptive children.

The sexual orientation of the adoptive parents in this sample had no significant impact on the internalizing or externalizing behaviors of the children. The results support the ACLU's demand that gays and lesbians be given an equal and fair opportunity to pursue adoption. Access to a just and open adoption process requires changes in both practice and policy. First, the findings support the need for specialized training of adoption workers in order to prepare them to make nondiscriminatory assessments and to engage in sensitive practice with gays and lesbians (Mallon, 2007; Ryan, 2000). Logically, this training would begin during the education experience. Logan and Kershaw (1994) found that social work education contains a heterosexist bias. Swank and Raiz (2007) discovered that having a class that explicitly explores homophobia cultivated a greater acceptance of gays and lesbians for social work students. In particular, schools of social work and even more specifically child welfare courses should prepare potential foster child and adoption workers to properly assess and advocate for gay and lesbian adoptive parents. For example, the Evan B. Donaldson Adoption Institute at Case Western Reserve University incorporates knowledge about gay and lesbian adoptive parents in its course work (Groza, n.d.). As well, future adoption workers should explore their own thoughts and feelings in regard to homophobia and increase their self-awareness.

The results also have timely policy implications. It is important to note that while the Florida ban on gay adoption was ruled unconstitutional, it was appealed by the state and is likely to go to the Florida Supreme Court for a final decision (Ruggeri, 2008). And as mentioned earlier, Arkansas, Mississippi, and Utah all still maintain laws that limit the rights of gays and lesbians to adopt. The results of this study support the growing body of knowledge that could aid in overturning the unconstitutional laws currently in effect.

Many of the adoptive parents in this sample faced stressors; however, their sexual orientation was not a predictor of those stressors. The following sections discuss the findings in regard to these stressors, specifically the emotional and behavioral issues of the adoptive children and the factors that might buffer such behavior.

Sex

The adoption analyses in 6- to 18-year-olds revealed that boys were more at risk for poor behavioral outcomes. Commentary reviews (Freeark et al., 2005) and empirical research (Brodzinsky, Smith, & Brodzinsky, 1998; Gibbs, Barth, & Houts, 2005) suggest that adoptive parents are seemingly more concerned about receiving supportive services for their adopted sons as opposed to daughters. This may be partially a result of adopted boys’ inclination for externalizing disorders (Simmel, 2007; Simmel, Barth, & Brooks, 2007). In contrast to internalizing behavior, adoption outcome research suggests that externalizing behavior, such as breaking rules, running away, fighting, and threatening behavior, is most disconcerting to adoptive parents and is associated with placement instability and adoption disruptions (Brodzinsky et al., 1998). Longitudinal research is needed to determine whether unique trajectories exist separately for boys and girls. Qualitative research may add to our understanding of possible dynamics surrounding why adopted boys are seemingly at higher risk for externalizing problems and what supports their families require.

Pre-adoptive History of Sexual Abuse

None of the pre-adoptive abuse variables emerged as significant predictors of internalizing and externalizing problems in the younger child analyses. There are a host of possible explanations for this counterintuitive finding. As mentioned previously, we elected to include pre-adoptive sexual abuse in the analyses to be consistent with the older child analyses despite only 6 children having such histories. Clearly, statistical errors may account for this finding. A far more interesting question is why there were so few children with pre-adoptive sexual abuse histories as compared to children with physical abuse and neglect histories. One possible explanation is that younger children may lack to vocabulary to disclose sexual abuse or may keep it a secret due feelings of betrayal, powerlessness, and shame (Finkelhor & Browne, 1985). It is only when they feel safe with their adoptive parents that they may disclose sexual abuse (Smith & Howard, 1994).

It is possible that the effects of sexual abuse may lay dormant until a later developmental stage, such as in adolescence when symptoms may be triggered by certain stressors related with that developmental stage (Downs, 1993). Lending support to this developmental theory is that pre-adoptive sexual abuse was a significant predictor of more severe internalizing and externalizing behavior among the older children. These findings are consistent with a comprehensive literature review on the adoption adjustment of children with pre-adoptive sexual abuse histories (Nalavany & Ryan, 2008). For adopted children with pre-adoptive sexual abuse, their problematic adjustment at adoption may not abate over time, and they may experience difficulty seeking help and support from their adoptive parents. For example, a recent study found that adopted boys with pre-adoptive sexual abuse were 17.4%, 24.6%, and 17.1% more likely to exhibit clinically significant aggressive behavior across a 3-year period, respectively (Nalavany, Ryan, & Hinterlong, 2009). The small sample of children with pre-adoptive sexual abuse precluded sex-specific analyses. It should be noted, however, that the dynamics (e.g., gender role socialization) and effects (internalizing and externalizing symptoms) of sexual abuse may be vastly different for boys and girls (Spiegel, 2003). Future research should explore the sex-specific experiences of parents who adopt male and female children with pre-adoptive sexual abuse.

Interestingly, pre-adoptive history of physical abuse emerged as a significant predictor of externalizing problems only for the older child analyses, albeit the magnitude of the effect was less pronounced as compared to children with pre-adoptive sexual abuse. On the other hand, pre-adoptive neglect was not significant in any of the analyses. It is unclear whether a similar developmental trajectory exists for physical abuse and neglect. Longitudinal research is needed to understand the developmental trajectory of pre-adoptive sexual abuse, physical abuse, and neglect and the pre- and post-adoption service needs of families.

Sibling Adoption

Although much of the previous research findings on sibling adoptions in adoptive families and child behavior is mixed (Smith, Garnier, Howard, & Ryan, 2006), sibling adoption was the only significant predictor of both more problematic internalizing and externalizing problems for children 1.5 to 5 years old. Our findings are consistent with previous research regarding the negative effects of sibling adoptions in adoptive families (Boneh, 1979; Erich & Leung, 2002; Leung & Erich, 2002; Smith et al., 2006). In comparison to children not adopted in sibling groups, adopted siblings may be at more risk for behavioral problems because they have at least one common parent and experience similar emotional, behavioral, and social experiences overtime. Our findings suggest the need for sufficient post-adoptive support services early in the adoption life cycle for adoptive families who choose to adopt siblings. However, the findings can not be understood apart from the analyses in 6- to 18-year-olds, where sibling placement was a nonsignificant predictor of internalizing and externalizing problems. It is challenging to reconcile this finding. One explanation may be that parents have more difficulty in accurately differentiating the intensity of emotional and behavioral problems when comparing siblings (Erich & Leung, 2002), which may have the unintended effect of erroneously lowering their total CBCL scores. Additional research is needed to further understand the impact of sibling adoptions on adoption adjustment and clarify the inconsistency of findings across studies.

Adoption Preparation

Parents’ perceived satisfaction with adoption preparation services was a significant factor positively influencing the emotional and behavioral outcomes of the youth in all analyses except internalizing behavior of the younger children. The consistency across the analyses suggests that early in the adoption life cycle, the degree to which adoption workers exemplify concern for the child and family, demonstrate knowledge of adoption-related issues, and respond to adoptive parent problems and questions appears to be of importance to adoptive parents (Simmel, 2007; Smith & Howard, 1999). This suggests that perceived satisfaction with adoption services at adoption may be salient in indirectly helping the child and family adjust to adoption. As potential parents are prepared for their adoptive child's entrance into the family, it is essential that they have a realistic understanding of the emotional and behavioral challenges that they may experience and the understanding of the deleterious impact of a pre-adoptive risk history (e.g., sibling adoption, pre-adoptive child abuse, intellectual disabilities) and its effect on adoption adjustment. Kramer and Houston (1998) found that parents frequently seek support for behavior problems of their children in pre-adoptive placements, with informal, nonagency resources reported to be most helpful. Thus, adoption agencies should ensure that information about these resources is provided to parents prior to placement.

Child Age

The findings suggest that older children (as a proxy for placement age) were at risk for more severe internalizing and externalizing problems, but only for the analyses in older children. Considerable research suggests that older adopted children are at increased risk for emotional and behavioral problems and adoption disruption (Berry & Barth, 1990; Rosenthal, 1993; Sharma et al., 1996; Simmel, Brooks, Barth, & Hinshaw, 2001). According to Pinderhughes (1996), one possible reason for adoption adjustment problems among children adopted at older ages is that there may be profound emotions associated with the loss of their biological families and fears of rejection by their adoptive families. Per Pinderhughes, a possible outcome of these averse life events and worries of abandonment is that many of these children have complex and unpredictable emotional and behavioral problems. Once the children are adopted, these problems may stress the ability of families to successfully incorporate the children in their families.

Family Functioning and Family Income

Family functioning was a nonsignificant predictor in the younger child analyses but emerged as significant for both emotional and behavioral problems in the older child analyses. As predicted, better family functioning was associated with less severe emotional and behavioral problems among the older children. Leung and Erich (2002) found that lower levels of children's behavioral problems predict higher levels of family functioning, which parallels our findings. Brodzinsky's (1990) contention that the most vital environmental factor affecting the adopted child's adjustment is the experiences he or she has with family members is consistent with our results. The finding that family functioning was not a significant predictor of CBCL scores for the younger children may be an artifact of the child's developmental level and shorter length of time in the adoptive home.

According to Groze (1996), success in adoption depends on the balance of resources and stresses assisting/impacting the family. As a proxy for resource availability, our findings demonstrate that higher family income predicted less severe internalizing and externalizing symptoms. An increase in financial supports may buffer some of the barriers and stresses adoptive parents experience in securing medical and behavioral services for their adopted children and enhance family functioning (Ryan et al., 2006).

CONCLUSION

As Mallon (2007) stated, “Not all gay and lesbian people should be foster or adoptive parents. In the same way, not all nongay people would make suitable foster or adoptive parents” (p. 84). Deciding to adopt is an important step and should be meaningfully considered and weighed by potential parents. However, the sexual orientation of the person should not affect that choice or keep that person from pursuing the option if so desired. Gay and lesbian couples should be considered and assessed via an adoption process that is respectful, culturally sensitive, and strengths-based (Matthews & Cramer, 2006). Gay and lesbian adoptive parents, as well as heterosexual adoptive parents, face challenges and issues in their families. Although the emotional and behavioral issues of their children can provide challenges for adoptive parents, their parents’ sexual orientation does not impact their ability to face those challenges.

Social work educators, adoption professionals, and policy makers have a responsibility to face the growing body of knowledge that proves that gay and lesbian parents are as fit as heterosexual parents to adopt. As a result, we must not only educate and practice and create laws that respect the 2 million gay and lesbian potential adoptive parents (Gates, Badgett, Macomber & Chambers, 2007) but we must also serve the 130,000 children in the child welfare system waiting to be adopted (U.S. Department of Health and Human Services, 2007).