Measuring Attitudes toward Open Adoption: Spanish Validation of the Open Adoption Scale

Introduction

Open adoption is a form of adoption characterized by the arrangement of some degree of communication or contact between the biological family and the adopted child or the adoptive family (Berástegui, 2016; Grotevant, 2000). It has also been labeled inclusive adoption or contact adoption. Open adoptions are very heterogeneous regarding the type, level, and frequency of contact, the people involved, or the degree and mode of formalization. The degree of communication varies from a simple exchange of non-anonymous information (semi-open adoptions) to the regular schedule of meetings (fully open adoptions) (Hass, 2015; ISS/CIR, 2015). In addition, this contact can be maintained with different members of the family of origin (biological parents, grandparents, aunts/uncles, or siblings).

Open adoption has been used since the 90s in countries, such as the USA, Great Britain, Germany, Austria, Canada, or New Zealand. In Spain, open adoption has been possible since Ley 26/2015 [Law 26/2015] on the Protection of Children and Adolescents entered into force. Under this law, open adoption is formalized judicially (Diez, 2018) and can be structured as the initial provision for a child or as a way to ensure permanency in previous family foster arrangements. Moreover, it primarily aims to maintain sibling connections when a permanent family placement in the same household has not been secured (Adroher et al., 2023). However, its initial implementation has been slow and uneven and, to date, has not been evaluated (Adroher et al., 2023; Diez, 2018; Martin, 2020).

The development of open adoption in Spain has clashed with the traditional vision of the family and a culture of closed adoption, which considers essential strict respect for confidentiality in the adoptive triad and the rupture of all ties with the family of origin (Rosser & Berástegui, 2017). Closed adoption is based on the belief that secrecy is necessary to protect the child in adoption and helps triad members to heal and move on with their lives. In contrast, openness is presumed to force the child to have dual loyalties, leads to confusion, and will inevitably create competition between the adoptive and birth families, interfering with attachment in the adoptive family (Ryan et al., 2011).

This traditional view is part of cultural beliefs about open adoption and can interact with the professional discourses and adoptive projects of families offering to adopt. Although Spanish professionals have expressed positive attitudes toward open adoption, they warn of some barriers for the families and the professionals themselves (Rosser & Berástegui, 2017). In a recent study, Corral et al. (in press) interviewed 22 adoption professionals who highlighted their crucial role in supporting adoptive families in open adoption. Although the professionals acknowledge that open adoption is beneficial, they stated that its implementation is challenging and face significant cultural, technical, and administrative difficulties. They concluded that, although open adoption is increasingly accepted, its implementation requires a mindset change.

In contrast with these beliefs, international literature consistently highlights the positive impact of open adoptions on all members of the adoptive triad, especially children, whose best interests should be the primary consideration of decision-making in adoption (Smith et al., 2020). Research has found a positive effect of open adoption on children’s well-being and psychological adjustment (del Pozo de Bolger et al., 2021; Grotevant et al., 2017; Siegel, 2013; Wolfgram, 2008). In a recent systematic review, Smith et al. (2020) report that open adoption promotes more open communication about adoption and origins in the adoptive family, and a better development of children’s adoptive identity, fostering self-esteem and decreasing their possible feelings of guilt, sadness, and abandonment in the adoption process. It also fosters a better parent-child attachment (Lo & Grotevant, 2020), allowing the children to expand their circle of support, maintaining their link to two families, and conflict of loyalties in the children (Smith et al., 2020). On the other hand, contact with members of their biological families is not related to greater difficulties of adaptation or to the desire to return to their biological family (Smith et al., 2020). In this sense, research has found some benefits for families and the relationship between them, such as the reduction of anxiety and insecurity, a more empathetic and realistic view of the biological parents, and an increase in communication about adoption in the adoptive parents with their children, or the reduction of anxiety, fears, and the grieving process of the biological mothers (Brodzinsky, 2005; McSherry et al., 2008; Neil, 2004; Smith & Logan, 2004).

Some risks are also indicated, especially in cases of previous family abuse or when the family of origin is reluctant to adoption. It should be noted that most of the research on open adoption has been developed in private adoptions and not in adoptions from care (Faulkner & Madden, 2012; Frasch et al., 2000; Ward et al., 2022), which are more similar to adoptions in European countries and, more specifically, in the Spanish system (Adroher et al., 2023; Berástegui, 2016; Gómez, 2018).

Assessing Attitudes toward Open Adoption

Although the myths underpinning closed adoption have been refuted by international research, the attitude toward open adoption has a major impact on its development and outcomes (Brown et al., 2007). Technicians’ negative attitudes are related to a lower tendency to propose and promote this type of adoption, considering that they are not usually in the child’s best interests or showing a greater tendency to suspend contacts (Robinson, 2017; Sobol et al., 2004). Similarly, families’ negative attitudes can lead to a closer adoption project, lower willingness toward this type of adoption, less suitability for open adoption, and a negative interpretation of the challenges that contact may involve (Sorek et al., 2020). A study shows greater resistance among families than professionals when promoting openness and more open attitudes in more experienced and trained technicians (Ryan et al., 2011). Hence, assessing cultural beliefs or adherence to myths about openness is essential in designing and supporting this measure (Berástegui, 2016; Brown et al., 2007).

The Open Adoption Scale (OAS; Brown et al., 2007) evaluates the myths or negative beliefs about open adoption concerning the three vertices or protagonists of the adoptive triad (biological family, adoptive family, and adopted minor), such that a low score would indicate a more positive attitude toward open adoption. These myths were originally drawn from a literature review and followed a content validity process. This questionnaire differs from other questionnaires aimed at assessing open communication within the adoptive family, such as the Origins and Differences Communication Scale (Berástegui, 2005; Berástegui & Jódar, 2013) or the Adoption Communication Scale (ACS; Aramburu et al., 2015; Brodzinsky, 2006; Grotevant et al., 2009), as it addresses the attitude toward open adoption structurally. It was initially designed as a training and research tool to inform child welfare personnel and policymakers of the influence of adoption-related myths on child welfare adoption practice (Brown et al., 2007).

In the Spanish-speaking context, validating this measure can be very useful. First, previous research has underlined the importance of considering the cultural dimension when planning and implementing open adoption (Ryan et al., 2011). Having a tool in Spanish to investigate social attitudes toward this form of adoption can help accompany this measure’s implementation in Spain and be a starting point to evaluate its reception in other Spanish-speaking contexts. In addition, it can be used to evaluate future families’ specific attitudes in their selection, recruitment, and evaluation processes. Not least, such a measure can help assess the change in the adoptive family’s pre-preparation processes (Ryan et al., 2011) and the training of professionals and teams (Siegel, 2013), which are success factors of open adoption. Therefore, this research aims to validate the Spanish version of the Open Adoption Scale (OAS).

Method

Participants

Convenience and snowball sampling were used to recruit participants. To carry out a factor analysis, Nunnally (1994), and Bentler and Chou (1987) recommended using a sample 10 times larger than the number of the items of the scale, so an attempt was made to get a large enough sample to meet this requirement.

The total sample consisted of 440 participants, 75.7% women (n = 333) and 24.3% men (n = 107). Ages ranged from 18 to 71 years, with a mean age of 30.6 (SD = 11.79). Most of the participants reported having university studies (96.8%), and not living as a couple (69.4%) or having children (75.2%). Table 1 shows these and other socio-demographic characteristics of the participants.

Using the SPSS procedure of creating random samples of 50% of cases, participants were divided into two groups to perform an exploratory factor analysis (EFA) with Group 1 (n = 220) and a confirmatory factor analysis (CFA) with Group 2 (n = 220).

Instruments

Sociodemographic Questionnaire

This questionnaire, prepared for the present study, included sociodemographic characteristics (sex, age, educational level, couple status and years of cohabitation, presence or absence of children and their number). In addition, to examine convergent validity, as no other scales have been found to assess myths about open adoption, four items of negative attitudes toward open adoption were included within the adoptive triad. They were as follows: 1) No good will come from biological parents keeping in touch with the adoptive family; 2) No good will come from biological parents keeping in touch with their children when they are adopted by other families; 3) No good will come from the adopted child having contact with their biological family; 4) No good will come from adoptive parents having contact with the biological family of the adopted child. Participants had to rate their level of agreement on these four items from 1 (very low) to 10 (very high).

Open Adoption Scale (OAS)

This scale, developed by Brown et al. (2007), assesses the level of agreement with various myths about open adoption in adults. It consists of 15 items with 6 response options ranging from 1 (strongly disagree) to 6 (strongly agree). The scale has three subscales of 5 items each, reflecting myths associated with each member of the adoptive triad: the adopted child, the adoptive parents, and the biological parents.

The internal consistency indices (Cronbach alpha) of the original scale and subscales were high: Global OAS = .92, Biological Parents = .85, Adopted Child = .89, and Adoptive Parents = .82.

In the original validation, construct validity was checked by multiple-group CFA (Nunnally & Bernstein, 1994). Regarding the evidence of convergent validity, significant associations were found between the OAS and several items on beliefs about open adoptions. Finally, concurrent validity was verified by observing that social workers had fewer myths about open adoption than students in other disciplines (Brown et al., 2007).

Procedure

We obtained the consent of one of the main authors of the original version, to proceed with the adaptation of this instrument. The current team carried out the first translation from English to Spanish. Two native Spanish speakers with a C2 English level and expertise in child protection, but not research team members, reviewed the translation and back-translated refining the translation with minor changes.

The questionnaire was designed in a physical and an online version through the Sphinx data collection software. The online version of the questionnaire was first completed by two volunteers to guarantee an easy understanding of the instructions, the items, and the requested response method, as well as the correct functioning of the survey tool.

The final version was distributed in its digital version through the database of the research center, which includes participants in other studies and those interested in family childhood issues. It was also sent through digital platforms to reach the largest number of students from various universities through a snowball sampling. In addition, the physical questionnaire was distributed among different groups of students at the University in which the study was carried out.

The Ethics Committee of the Comillas Pontifical University of Madrid approved the study and the data collection followed the principles of research ethics, although it is a voluntary, anonymized study that does not involve intervention or foresee discomfort for the participants.

Data Analysis

First, a maximum likelihood EFA with Oblimin rotation was carried out with Group 1. To determine the number of factors to be extracted, a parallel analysis and a minimum average partial test (Horn, 1965; Velicer, 1976) were carried out. Factors with more than two items and items with loads greater than .40 were considered (Izquierdo et al., 2014; Schmitt, 2011). Secondly, a CFA was conducted with Group 2. The fit indices used were the chi-square statistic, the ratio between chi-square and degrees of freedom (χ²/df), the comparative fit index (CFI), the Tucker-Lewis index (TLI), the root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR). The cut-off criterion for the ratio of χ²/df is ≤ 3; for CFI and TLI, it is ≥ .90; for RMSEA, it is ≤ .06; and for SRMR, it is < .08 (Bentler, 1990; Bentler & Bonett, 1980; Brown, 2015).

Next, the internal consistency of the scale and its subscales was calculated using Cronbach’s alpha and McDonald’s omega. Likewise, after checking that all the quantitative variables were normally distributed using the Kolmogorov-Smirnov test with Lilliefors correction and the analysis of skewness and kurtosis (Curran et al., 1996), several Pearson correlations were calculated to examine the relationship between the scale and its subscales with the four items of negative attitudes toward open adoption. SPSS v.28 was used to perform parallel analysis and EFA, as well as to calculate reliability and correlations; the free software JAMOVI (version 2.3.18; The Jamovi project, 2022), which provides an interface of the R program (Sesé, 2023), was used to perform CFA.

Results

Exploratory Factor Analysis

Firstly, after performing a parallel analysis and a minimum average partial test with Group 1, we determined the adequacy of extracting two factors. Next, EFA was performed with the maximum likelihood method and Oblimin rotation. Nine items loaded on the first factor, related to the myths about the adoptive family (about the adopted child and the adoptive parents). Five items loaded on the second factor, associated with the myths about biological parents. Item AP.4 did not reach a loading of .40 in any of the factors, so it was eliminated, thus retaining 14 of the original 15 items (see Appendix A and B). Table 2 shows the loadings of these 14 items.

Concerning the internal consistency indices, the Open Adoption Scale-Spanish (OAS-S) and its subscales obtained high values of Cronbach’s alpha (OAS-S total = .90, AF = .90, BP = .85) and McDonald’s omega (OAS-S total = .91, AF = .90, BP = .86).

Confirmatory Factor Analysis

Before performing the CFA, the multivariate normality of the fourteen OAS items retained in the EFA was analyzed. The Mardia test suggested that the multivariate distribution was asymmetric (Mardia skewness = 945.85, p < .0001) and leptocurtic (Mardia kurtosis = 8.51, p < . 0001) and, therefore, neither the skewness nor the kurtosis were normal multivariate. The results of the Anderson-Darling test for the univariate tests of each of the 14 items showed that none of them was strictly normal at the univariate level. The CFA was specified using an estimator for ordinal variables, the Diagonal Weighted Least Squares (DWLS), and a two-factor structure.

The model results showed an appropriate fit, with a significant χ²-value (χ² = 140, df = 74, p < .001, χ²/df = 1.89), an RMSEA value of .06 (confidence interval [.05, .08]), an SRMR of .06, a CFI of .99, and a TLI of .99. Figure 1 shows the 2-factor model with its factor loads.

Convergent Validity

We also found that the Open Adoption Scale-Spanish (OAS-S) and its subscales were positively and strongly related to the four myths about the absence of benefits of open adoption for each member of the adoptive triad (see Table 3). In addition, we observed that the items referring to the absence of benefits for adoptive parents, adopted children, and the adoptive family had stronger relationships with the Adoptive Family subscale than with the Biological Parents subscale.

Discussion

Open adoption has positively impacted all members of the adoptive triad (Smith et al., 2020; Wolfgram, 2008). However, both in the international and national literature, positive and negative attitudes toward open adoption are reported by professionals in the world of adoption (Corral et al., in press; Rosser & Berástegui, 2017; Ryan et al., 2011). It is important to know the professionals’ attitudes because they can contribute significantly to a successful adoption process (Berástegui, 2016). The objective of this study was to adapt the Open Adoption Scale (Brown et al., 2007) to the Spanish language to have a validated instrument to explore attitudes toward open adoption. The EFA revealed that the 14-item Open Adoption Spanish (OAS-S) has a two-factor structure. The first factor contained the items related to the adoptive family (e.g., the adopted children and the adoptive parents). The second factor comprised the items related to the biological parents. The CFA supported the existence of this structure, with adequate fit indices. This structure contrasts with the theoretical structure of the questionnaire as proposed by the original authors, which includes three blocks associated with the three protagonists of the adoptive triad (Brown et al., 2007). However, the two-factor structure proposed for the Spanish version is consistent with that observed by Brown et al. (2007): the Adopted Child and Adoptive Parents scales would be included in the Adoptive Family scale, and the Biological Parents scale would remain similar to the original proposal of the instrument. This two-factor structure can be due to sampling limitations in both versions or a cultural artifact, resulting from the identification between child and adoptive family needs and views and the difficulty of differentiating these two perspectives in the general population. Both versions could benefit from further examination with actual members of the adoptive triad.

The internal consistency indices of the scale and its subscales were adequate. Regarding the evidence of convergent validity, the OAS-S and its subscales were strongly related to four myths about the benefits of open adoptions for each member of the adoptive triad. In addition, we observed that beliefs about the impact of open adoption for children were more strongly related to the perception of the impact on the adoptive family than to the impact on the biological family.

As limitations of the study, we note that convergent validity could not be examined optimally, as no other validated instruments were found to assess beliefs about open adoption, a limitation that this study shares with the validation study of the original scale (Brown et al., 2007). Other limitations of the study are related to the type of sampling and sample size used for the study. Most of our sample was formed by young women with university studies, so it would be relevant to evaluate these attitudes in more representative samples of the general population.

Some limitations are related to the adjustment of the scale to the adoption system in Spain. On the one hand, the scale identifies biological family with biological parents, while open adoption with siblings is proving very frequent in our country (Adroher et al., 2023). Therefore, it is important to include a subscale that collects beliefs about the impact of open adoption among siblings or other members of the biological family.

On the other hand, this questionnaire assesses attitudes toward contact or maintaining a relationship generically. However, part of the success of open adoption involves appropriate contact modulation for each case. Spanish respondents may be imagining visits between the biological family and the child, the preferred mode of contact in foster care, which represents the highest degree of contact. However, the possibility of evaluating differential attitudes toward different forms of lower-intensity contact (letters, telephone, videoconference) could provide a more nuanced view of the attitude toward open adoption in families and professionals in the face of intervention. It is, therefore, relevant in future research to incorporate different forms of contact to deepen the understanding of beliefs and attitudes toward open adoption.

Another future line of research related to attitudes toward open adoption is to apply this instrument to adoption professionals. This would add quantitative evidence to the qualitative evidence already existing in our cultural context on professionals’ attitudes toward open adoption (Corral et al., in press; Rosser & Berástegui, 2017).

Finally, we note that open adoption was developed in the Anglo-Saxon world, especially concerning private adoptions, whereas authorities have been more prudent in promoting adoptions from care (Faulkner & Madden, 2012; Frasch et al., 2000; Ward et al., 2022), which is more similar to the adoption system in Spain. Open adoption can be a measure to mobilize the development of permanency planning for older children or groups of siblings, but it must be accompanied by research and monitoring of the efficiency of the measures, an effort in which the tool validated herein and its future developments can play a crucial role.

In conclusion, the OAS-S is a 14-item scale that assesses attitudes toward open adoption in Spanish (see and Appendix A and B). The instrument has good psychometric properties, and its brevity and easy application make it useful for evaluating beliefs about open adoption in adoption technicians, adoptive families, and the general population. In addition, this instrument has the potential for research on changing social attitudes toward open adoption in other Spanish-speaking contexts, although further cross-cultural validation should be considered (Abell et al., 2006).