Attachment Representations and Early Adversity in Internationally Adopted Children from Russian Federation Using the Friends and Family Interview

Abstract

Extant literature suggests that adopted children can generate new attachment bonds with adoptive parents, although early adversity leaves an enduring mark. This research was focused on attachment representations when adoptees from Russia were reaching their adolescence. The participants were 29 adoptees and 38 children from a control group, aged between 8 and 13 years. The Friends and Family Interview, a semi-structured interview, assessed the narrative’s coherence, reflective function, internal working models, and attachment classification. Children’s narrative was coded into numerical data by two coders with high interrater reliability. Adopted children were classified 41% secure, 35% dismissing, 14% preoccupied, and 10% disorganized (82% secure in the control group). Children’s age, gender and verbal IQ, and family structure –but not mother’s academic level nor placement variables– were related to attachment variables in both groups. Although adoption offers a protective context that promotes secure attachment, insecurity and disorganization are still evident.

Keywords

attachment stress adaptation emotional development

For many years, Russia has been one of the main countries of origin worldwide for international adoptions (Selman, 2018). Almost 13,000 children have been adopted in Russia by Spanish families in the last two decades (1998–2018) (Observatorio de la Infancia, 2020). Children adopted from Russian institutions have been more frequently exposed to severe deprivation, in comparison with children adopted from other countries (Johnson, 2000). Nonetheless, scarce research has studied attachment representations in this population as they are reaching adolescence. Previous studies have usually combined subjects adopted from different countries and continents, and most of them have not included a community comparison group. Also, as indicated by Pace (2014), most of those studies have only reported global attachment classifications (secure, dismissing, preoccupied and disorganized), when other attachment dimensions (e.g., coherence, evidence of secure base) could also offer relevant information. The present research is focused on the study of attachment representations using the Friends and Family Interview in late childhood and early adolescent adoptees from Russia.

The Impact of Early Institutionalization on the Attachment System

A large proportion of children adopted internationally spent some time in residential care. Institutional care is characterized by the absence of a stable, primary caregiving figure, the frequent turnover of professional caregivers, and rigid routines unresponsive to individual needs (Berens & Nelson, 2015). Between 2002 and 2006 –time range of adoption of the current study–, there were over 700,000 children out of parental care in Russia (Biryukova & Sinyavskaya, 2017). In 2003 there were 249 Baby Homes in Russia, with almost 20,000 residents under 4 years of age (Ministry of Health of the Russian Federation, 2003). According to The St. Petersburg–USA Orphanage Research Team (2005) Baby Homes sizes range between 60 and 100, divided into wards of 9–14 children. Although there is well trained staff work, daily care is focused on children’s health and safety rather than their behavioral and socio-emotional development. A high proportion of children adopted from Russia often have experienced severe deprivation –e.g., inadequate nutrition, low emotional support, sensory and cognitive stimulation– (Johnson, 2000). Children adopted from Russian institutions generally presented higher difficulties in cognitive, emotional, and physical areas than children from other countries of origin (Johnson, 2000; van den Dries et al., 2009). The high numbers of children adopted from Russia worldwide, and their background, highlight the need of further studies focused on this population.

In terms of attachment development, the meta-analysis implemented by Lionetti et al. (2015) showed that being raised in an institution is a risk factor for insecure and disorganized attachment: only 18% of institutionalized children presented a secure attachment, while 28% were insecure, and 54% were disorganized.

How Attachment Representations Evolve When Institutionalized Children are Adopted?

For institutionalized children, adoption represents a significant change in caregiving circumstances (Palacios & Brodzinsky, 2010). New committed, individualized, permanent and loving caregivers create the relational context for a revision of the whole attachment system. The meta-analysis by van den Dries et al. (2009) included 39 adoption studies where adoptees’ age at assessment were classified in four categories (<4; 4–12; 12–18; >18 years of age) and age at adoption in two categories (under or above 12 months of age). Children adopted before 12 months of age were as securely attached as non-adopted controls and showed a lower prevalence of disorganized attachment in comparison with peers left behind in institutions, independently of their age at assessment. Nevertheless, findings point to an incomplete catch-up, because children adopted after 12 months were more likely to manifest insecure and disorganized attachment in comparison with their non-adopted peers.

How do Attachment Representations Develop in Late Childhood and Adolescence?

Although attachment representations are highly stable, they are active constructions, which, not without some resistance, can be restructured to continue being effective when circumstances (e.g., a significant modification of the caregiving environment) change (Steele et al., 2008). Although adolescents seek greater independence and differentiation from their primary caregivers, they continue to have a need for attachment figures (Allen, 2008). Bonds established with peers and other relevant adults may have a significant impact in the adoptees’ attachment representations. Therefore, this period of potential transformation of the attachment representations becomes relevant in the study of adoptees development (Palacios & Brodzinsky, 2010). The changes of attachment representations in adoptees during this developmental period is documented by Pace et al. (2015), but further studies are needed to broaden our understanding of whether the negative consequences of early adversity endure into adolescence.

From the methodological point of view, diverse instruments have been developed to evaluate attachment representations, especially interview tools (Bosmans & Kerns, 2015). Although other interview methods exist with a similar purpose, such as the Child Attachment Interview (CAI; Shmueli-Goetz et al., 2000), the Friends and Family Interview (FFI; Steele & Steele, 2005) is the most widely used technique. Theoretically guided by the Adult Attachment Interview (AAI; George et al., 1984), the Friends and Family Interview is scaled to the developmental abilities of older children and adolescents (8–16 years), featuring the relations with parents, siblings, peers, and teachers. The dimensional approach allows a more comprehensive and deep understanding of the attachment representations and avoids the limited approach of secure versus insecure classification.

Previous Studies of Adoptees’ Attachment Representations using Interview Methods

The review of previous research has shown 12 studies using interview methods to assess attachment representations in adoptees, 10 of which used the Friends and Family Interview (Table 1). The results cannot always be compared due to methodological differences. The studied samples were heterogeneous, with five studies on domestic adoption, and the rest involving children adopted from different continents. Participants’ ages varied widely (7–24 years old) and only three studies included a non-adopted comparison group. The average rate of secure attachment was 47.0%, dismissing 38.2%, preoccupied 12.6%, and disorganized 2.8%. Studies of children adopted in Eastern European countries using the Friends and Family Interview have reported relatively similar results: 41.7% secure, 29.2% dismissing, 25% preoccupied and 4.2% disorganized (Abrines et al., 2012). Previous literature highlights the paucity of studies which have used interview methods, despite their advantages to assess attachment representations in adolescence. Furthermore, despite being one of the main countries of origin in international adoption, only one project has reported attachment representations in early adolescent adoptees from Eastern Europe (see Abrines et al., 2012). Heterogeneous samples that mix children from different countries, cultures, and early life paths, can add confounding variables to the study of attachment representations. The study of a more homogeneous sample may bring a deeper understanding of how attachment representations are evolving in children adopted from Russia in their late childhood and early adolescence.

Table 1.

Attachment Representations in Adopted Children in Late Childhood and Adolescence Using Interview Methods.

Source	Na – Nc	Gender (% boys)	Age at study, y	Age at arrival, m	Country of origin	Measure	% Secure	% Dismissing	% Preoccupied	% Disorganized
Abrines et al. (2012) ^a	58– No	34.5	7–8	25.4	EE, China, Ethiopia	FFI	60.3 (EE 41.7)	25.9 (EE 29.2)	12.1 (EE 25.0)	1.7 (EE 4.2)
Barcons et al. (2012) ^a	116–No	46.6	8–11	30.6	Asia, EE, Latin-America, Africa	FFI	60.4 (EE 44.4)	25.0 (EE 31.5)	12.9 (EE 20.4)	1.7 (EE 3.7)
Barcons et al. (2014) ^a	168– No	53.6	7–11	29.0	Asia, EE, Latin-America, Africa	FFI	58.9 (EE 44.0)	25.0 (EE 32.0)	13.1 (EE 20.0)	3.0 (EE 4.0)
Groza et al. (2012) ^b	39– No	38.5	11–16	22.6	Domestic-Romania	FFI	51.3	48.7		0.0
Muntean and Ungureanub (2012) ^b	40– No	37.5	11–16	22.6	Domestic-Romania	FFI	50.0	50.0
Pace et al. (2015) ^c	46– No	50	11–16	74.9	South America, EE, Asia, Africa, domestic-Italy	FFI	67.4	26.1	6.5	0.0
Pace et al. (2018) ^c	46–34	50	11–16	74.9	South America, EE, Asia, Africa, domestic-Italy	FFI	65.2	28.3	6.5	0.0
Balenzano et al. (2018)	59–No	50.8	11–24	>60	Domestic-Italy	AAI	40.7	59.3
Beijersbergen et al. (2012)	125– No	44.8	14	2.5	Sri Lanka, South Korea, Colombia	AAI	39.2	41.6	19.2	n.r.
Escobar and Santelices (2013)	25–25	56	11–18	26.0	Domestic-Chile	FFI	32.0	52.0	16.0	0.0
Molina et al. (2015)	27– No	51.9	10–16	44.0	Asia, South America, EE, Africa	FFI	40.0	44.0	8.0	8.0
Vorria et al. (2015)	52–36	51.9	12–13	n.r.	Domestic-Greek	CAI	50.0	44.3		5.8

Note. Na = number of adoptees; Nc = number of non-adopted controls; Age at study, y = Age at study in years; Age at arrival, m = Age at arrival in months; n.r. = data were not reported; EE = Eastern Europe; FFI = Friends and Family Interview; AAI = Adult Attachment Interview; CAI = Child Attachment Interview.

^aStudies with shared samples.

^bStudies with shared samples.

^cStudies with shared samples.

What Factors Affect the Quality of Attachment in Adolescent Adoptees?

The quality of the preadoption experiences of intercountry adoptees are largely unknown. To assess the impact of different preadoption circumstances on the attachment representations, researchers usually operationalize early adversity using proxy indicators such as age at entry in institutions or age at adoption placement. For the purposes of this research, early adversity is defined as early placement and length of time spent in institutional care, and age at adoption. A younger age at entry into the institution (before the first birthday) has been associated with higher disorganization (Lionetti et al., 2015 –several countries of origin–). The length of institutionalization and age at adoption, usually used as quantitative indices of exposition to early adversity, have been linked to secure representations in the first wave of the present longitudinal project, when adoptees were aged between 4 and 8 years (Román et al., 2012). However, other studies that sampled late childhood and early adolescent adoptees have not found any link between length of institutionalization, age at adoption and attachment representations (Ballús et al., 2019; Barcons et al., 2012 –several countries of origin–).

Most studies using interview methods and heterogeneous countries of origin have not reported significant relations between attachment representations and the children’s gender, age at assessment or continent of origin (Beijersbergen et al., 2012; Groza et al., 2012; Vorria et al., 2006), although other studies have found higher security in girls than in boys (Pierrehumbert et al., 2009). Moreover, children’s IQ appeared to be related to attachment representations, as attested by studies on both adopted (Pace et al., 2015) and non-adopted children (Stievenart et al., 2011). Adoptees’ attachment behavior seems to be independent of adoptive parent’s variables such as parental educational level, mother’s employment, and family structure (Juffer & Rosenboom, 1997), but this associations have not been so far explored for attachment representations.

The Current Study

Based on the existing gaps in the previous literature, two research questions arose and guided this study: what sociodemographic and placement variables influence attachment representations' development? How attachment representations develop in adoptees from Russian institutions as they are reaching adolescence, in comparison with a community group? Hence, the aims of this research were (1) to study the relation between attachment representations and (a) children’s characteristics (age, gender, IQ, family variables), and (b) pre- and postadoption placement variables; (2) to explore attachment representations in children adopted from the Russian Federation in their late childhood and early adolescence.

To address these research questions, we studied a sample of children adopted from Russia into Spanish families from the longitudinal study LAIS.US, in comparison with a community comparison group. Attachment representations were assessed using the Friends and Family Interview. Child and family variables were assessed to study their impact on adoptees’ attachment representations: verbal IQ was assessed through the Kaufman Brief Intelligence Test, and sociodemographic and placement variables were registered using the Interview about International Adoption Process. We expect that adoptees with a higher age at assessment, higher verbal IQ, shorter institutionalization and younger age at adoption would show more security and better attachment development. Furthermore, adoptees will present higher rates of insecure attachment in comparison with the control group.

Method

Participants

The sample was recruited from the longitudinal study LAIS.US carried on at the University of Seville since 2007. This project follows up the processes of recovery of two groups conformed by children adopted from Russia into Spanish families and children in the Spanish residential care system (not included in this article), and compared with a comparison community group. For the current study, all the families participating in the second wave were assessed. Attachment representations were examined in 8- to 13-year-old adopted children in comparison with the non-adopted group (the study of attachment representations in the residential care group is underway). The sample was composed by 29 children adopted by 27 Spanish families from institutional care in Russian orphanages, and a control group formed by 38 children who lived with their biological families and had no relation to child protection services. Children were adopted at an average age of 3.11 years (SD = 1.41) and at time of study had been with their adoptive families for an average of 7.24 years (SD = 1.17). All the children had lived in an institution prior to their adoption, entering at an average age of 10.41 months (SD = 16.81), and staying in the institution for an average of 26.97 months (SD = 14.45). Around half of the adoptive children (51.7%, n = 15) had at least a year of family experience prior to their institutionalization, mainly with their biological families. Other demographic characteristics are shown in Table 2.

Table 2.

Child and Family Characteristics of Adopted and Non-Adopted Children.

Study variables	Adopted children	Non-adopted children	Statistical tests
Study variables	M (SD)	M (SD)	Statistical tests
Age at assessment (years)	10.36 (1.27)	11.04 (1.36)	t(65) = 2.10*, d = −0.52
Gender (% boys)	75.9	55.3	χ²(1) = 3.04^ns, V = .21
Verbal IQ	97.3 (12.88)	104.1 (12.67)	t(65) = 2.14*, r = .26
Mother’s academic status (% university level)	82.8	47.4	χ²(1) = 8.81**, V = .36
Family structure (% two-parent families)	65.5	86.8	χ²(1) = 4.31*, V = .25

*p < .05; ** p < .01; ***p < .001; ^ns = not significant.

Instruments

Friends and family interview

Attachment representations were assessed using the Spanish version of the Friends and Family Interview (FFI; Steele & Steele, 2005). The initial phase helps stablish a rapport between the interviewer and the child, asking the child with whom he/she lives, other close persons, and favorite hobbies. Then it consists of most and least favorite qualities in the relationships with parents, siblings, teachers, and peers, and discussing how each person could describe him/her. Other relevant questions are related to what the child does when he/she is upset, earliest memory of separation from caregivers, and the relationship with his/her caregivers in the past and future. These are some of the most relevant dimensions (Kriss et al., 2012):

• Coherence is a central concept in narrative assessments of attachment, it reflects the individual’s overall organization across levels of representation. It is rated according to four dimensions: truth –convincing evidence to support the appraisal of self and others–; economy –right amount of information–; relation –degree to which given examples are relevant–; and manner –age-appropriate level of attention, politeness, and interest–. Following Stievenart et al., (2012), the overall coherence variable can be used as a unique marker of this construct.

• Reflective functioning refers to the psychological processes that allow children to appreciate the existence of other people’s mental states, as well as their own. It is operationalized across two subdomains: developmental perspective –ability to contrast current thoughts and feelings concerning relationships with past attitudes–, and diversity of feelings –understanding of diverse feelings being present in significant relationships–.

• Internal working models are coded according to evidence of secure base –what the child does when he/she is upset, and the most and least favorite aspects of each parent–; and adaptive response –what the child does when distressed–.

• Attachment classification. The overall transcript is rated for the presence of security, dismissiveness, preoccupation, and disorganization on Likert-type scales; the final attachment classification represents the dominant strategy observed in the transcript. This double coding system captures attachment representations both at categorical and dimensional levels, reflecting that a child may exhibit diverse strategies when the attachment system is activated.

Variables are scored on a 7-point scale from 1 to 4 (1 = No evidence; 2 = Mild evidence; 3 = Moderate evidence; and 4 = Marked evidence), including midpoints. The Friends and Family Interview has demonstrated robust inter-rater reliability and construct validity in both community (Steele & Steele, 2005) and at-risk samples (Kriss et al., 2012). In this work, collected interviews were recorded, transcribed, and then independently rated according to a standardized coding manual (Steele et al., 2009). The first two authors of this work (accredited by the New School for Social Research) blindly coded 18 of the 67 interviews (26%), while the rest of interviews were coded by the first author. The inter-rater agreement was high (k > 0.70) in most variables, and only four variables showed a lower inter-rater agreement (.69 > k > 0.57): diversity of feelings self and mother, secure base father; and adaptive response.

Kaufman brief intelligence test

Children’ verbal IQ was measured with Kaufman Brief Intelligence Test (K-BIT; Kaufman & Kaufman, 1990). K-BIT is an intelligence test for people between the ages of 4 and 90, it is motor-free and is individually applied in approximately 15 to 30 minutes. It consists of two subtests, Vocabulary and Matrices. Vocabulary consists of expressive vocabulary (45 items to name pictured objects) and definitions (37 items to identify words using a verbal description and a partial spelling of the word). Matrices requires the comprehension of relationships among visual stimuli (48 items). K-BIT provides verbal and non-verbal IQ, and an overall IQ composite in accordance with available standards for age. Items are grouped in blocks, with starting and finishing criteria being determined for each subtest. Right answers are scored with one point, while wrong answers get zero points. K-BIT has good psychometric properties in its Spanish validation (Cordero & Calonge, 2000): test-retest results showed correlations of 0.94 in the Vocabulary subtest, and 0.86 in the Matrices subtest.

Demographic records

An abbreviated version of the Interview about International Adoption Process (EPAI; Sánchez-Sandoval et al., 2002) was used with the adoptive parents to register the characteristics of the family (e.g., single or two-parent) and of the child’s early experiences (e.g., time in the birth family, age at entry into institutional care and at adoption placement).

Recruitment and Screening

Data presented in this study were collected in the second wave of the Longitudinal Adoption & Institutionalization Study (LAIS.US) carried on at the University of Seville. Contact with families in this second wave was made by mail and telephone. The assessment took place over two visits carried out by two trained psychologists, one focused on the caregiver and the other on the child, working in separate rooms in the child’s home. The parent who spent the most time with the child was interviewed, in most cases the mother, with the only exception being one adoptive father who was more involved in the child’s day-to-day activities. Participating families did not receive any economic compensation. The recruitment in the first data collection was made through two Collaborating Agencies for International Adoption in Spain, and through the random selection of 10 schools from neighborhoods with different socio-economic levels. Inclusion criteria in the first wave were: children aged between 4 and 8 years, and, for the group of adopted children, having been adopted in Russia and having spent at least 9 months in their adoptive family at the time of the data collection. The Institutional Ethics Review Boards of the University of Seville and the Andalusian Government, according the Spanish and European regulations and the principles expressed in the Declaration of Helsinki, approved this study. The children’s parents gave written informed consent for participation in the study.

Analytic Strategy

Statistical analyses were conducted using the statistic software SPSS (version 25). Data distribution analysis showed that attachment variables and almost all demographic variables were non-normally distributed (Shapiro-Wilk tests, p_s < .05), and non-parametric tests were conducted in those cases.

For the first research question –how child, family, and placement variables impact on adoptees’ attachment variables–, between-groups comparisons of child and family variables were implemented through Student’s t and chi square analysis. ANCOVA analysis were implemented to check the influence of child and family variables on attachment variables. Lastly, placement variables were related to attachment variables using correlations, and to attachment classification using ANOVA and Student’s t.

The second research question was aimed to compare attachment representations between adoptees and a community group. To control the influence of confounding variables, ANCOVA analysis were implemented when (a) there were between-group differences in child and family variables, and (b) child and family variables had a significant impact on attachment variables. Between-groups comparisons in attachment variables were implemented through Mann Whitney U for those variables not related to child or family variables. Comparisons in attachment classifications were analyzed using chi square analysis.

Finally, three effect size analyses were conducted: Pearson’s r and Cramer’s V –where values of 0.1, 0.3 and 0.5 correspond to small, medium, and large effects respectively (Cohen, 1988)–, and partial eta squared –values of 0.01, 0.09 and 0.25 correspond to small, medium, and large effects respectively (Tabachnick & Fidell, 2007).

Results

Between-Group Differences in Child and Family Variables

Child and family characteristics and group comparisons are shown in Table 2. On average, adopted children were a few months younger than the control group, with a low effect size. No statistical differences were found in gender distribution between the two groups. The adopted children obtained significantly lower verbal IQ scores than the control group. Regarding the academic status of the mothers, adoptive mothers had 5.33 more odds of having reached a University level than the control group. In family structure (single-parent or two-parents), there was a higher prevalence of single-parent families in the adoptive group (OR = 3.47).

Impact of Child and Family Variables on Attachment Variables Across Groups

ANCOVA analysis were implemented to check the impact of child and family variables on attachment variables across groups. Only statistically significant relations are reported here. The children’s age at assessment had significant impact on coherence, F(1, 64) = 5.68, p = .020, and preoccupied attachment, F(1, 64) = 5.67, p = .020. Regarding children’s gender, significant relations were found with coherence, F(1, 64) = 4.95, p = .030, and diversity of feelings (self), F(1, 64) = 10.47, p = .002. Children’s verbal IQ was only related to secure base (mother), F(1, 63) = 4.74, p = .033. No significant relations were found between mother’s academic level and attachment variables. Lastly, the family structure was related to the diversity of feelings (mother), F(1, 59) = 7.41, p = .009, secure attachment, F(1, 64) = 7.67, p = .007, and dismissing attachment, F(1, 64) = 5.06, p = .028.

Relations of Attachment Representations With Placement Variables

Placement variables, such as length of preadoptive family experience, age at entry and length of institutional care, age at placement and length of adoptive experience, were not significantly related to attachment representation variables, nor the 4- and 2-way attachment classification. There were no differences in the 4- and 2-way attachment classification between those adoptees institutionalized at birth and adoptees who lived with their biological families before being institutionalized, χ²(3) = 1.70, p = .74; χ²(1) = 0.36, p = .710.

Differences in Attachment Representation Variables Between Groups

The average scores in attachment variables for each group are shown in Table 3. Only attachment variables not related to child and family variables in the first two sections were analyzed using Mann-Whitney U, and presented in Table 3. In reflective function variables, adopted children obtained lower scores in diversity of feelings towards themselves and their fathers, with medium effect sizes. No significant differences were found in developmental perspective. In the dimensions of internal working models, no differences were found on their perception of secure base with their fathers, but adoptees showed poorer adaptive responses in stressful situations, with low effect sizes. Lastly, in attachment classification, adoptees obtained higher disorganization scores than the control group, with low effect size.

Table 3.

Scores in Attachment Variables Obtained by Adopted and Non-Adopted Children.

Study variables	Adopted children		Control group		Mann-Whitney U	Effect size r
Study variables	M	SD	M	SD	Mann-Whitney U	Effect size r
Reflective functioning
Develop. persp.	1.75	0.70	2.13	0.93	411.50 ^ns	−.20
DoF self	1.90	0.90	2.55	1.03	352.50**	−.32
DoF father	1.55	0.61	2.19	0.94	227.00*	−.33
Internal working models
Secure base father	1.90	0.72	2.13	0.81	325.00 ^ns	−.13
Adaptive response	1.86	0.79	2.29	0.84	388.00*	−.28
Attachment classification
Disorganization	1.41	0.78	1.13	0.58	433.50*	−.29

Note. DoF = Diversity of Feelings; Develop. persp. = Developmental Perspective.

*p < .05; ** p < .01; ***p < .001; ^ns = non-significant.

Due to the influence of child and family variables on attachment representations showed in the previous two sections, ANCOVA analysis were implemented. Only those child and family variables significantly related to attachment variables in both previous sections were considered. As it is shown in Table 4, adoptees still showed lower scores in coherence –after controlling for their age–, diversity of feelings mother and security –after controlling for family structure–. When verbal IQ was controlled, no differences arose in secure base mother. Furthermore, no differences were found in preoccupied attachment when children’s age was controlled, but adoptees showed higher dismissiveness scores after controlling for family structure.

Table 4.

Scores in Attachment Variables After Controlling for Child and Family Variables.

Study variables	Adopted children		Control group		ANCOVA analysis	Effect size partial η²
Study variables	M	SE	M	SE	ANCOVA analysis	Effect size partial η²
Overall coherence^a	2.41	0.13	2.88	0.11	F(1,63)=7.12*	.10
Reflective functioning
DoF mother^b	1.55	0.17	2.20	0.15	F(1,59)=8.13**	.12
Internal working models
Secure base mother^c	2.25	0.95	2.68	1.00	F(1,63)=3.06 ^ns	.05
Attachment classification
Security^b	2.24	0.13	3.03	0.11	F(1,64)=21.60***	.25
Dismissiveness^b	2.15	0.18	1.57	0.15	F(1,64)=5.91*	.09
Preoccupied^a	1.58	0.13	1.40	0.11	F(1,64)=1.00 ^ns	.02

Note. DoF = Diversity of Feelings.

*p < .05; **p < .01; ***p < .001; ^ns = non-significant.

^aControlling for age.

^bControlling for family structure.

^cControlling for verbal IQ.

The attachment classification for each group is shown in Table 5. In the 4-way attachment classification there were significant differences between adopted and control children, with a medium effect size, χ²(3) = 11.96, p = .004, V = .42. Specifically, the odds of adopted children of presenting dismissing versus secure attachment were 8.57 times higher than in the control group. In the 2-way attachment classification, statistically significant differences arose, with a medium effect size, χ²(1) = 11.56, p = .001, V = .42; children in the control group had 6.28 more odds of being classified with secure attachment in comparison with adopted children.

Table 5.

Attachment Classification Distribution in the Adopted and Control Groups.

Attachment pattern	Adopted children		Control group
Attachment pattern	n	%	n	%
Secure	12	41.4	31	81.6
Insecure	17	58.6	7	18.4
Dismissing	10	34.5	3	7.9
Preoccupied	4	13.8	2	5.3
Disorganized	3	10.3	2	5.3

Discussion

The purpose of this study was to investigate the attachment representations among 8- to 13-years-old children adopted from institutional settings in Russia into Spanish families. Regarding the relation between attachment representations and children and family characteristics, our first goal in this study, several attachment representation variables were related to children’s age, gender, verbal IQ, and family structure, although no significant relations were found for mother’s academic level and placement variables. Previous research has found that older children obtained higher coherence and less disorganization in the Manchester Child Attachment Story Task (Green et al., 2000), which could be explained by the development of narrative elaboration with age (Nelson & Gruendel, 1986). Although attachment literature has shown that attachment representations are not gender specific, Pierrehumbert et al. (2009) found a higher expression of emotions related to attachment experiences in girls. These authors hypothesized that fully continuous coding procedures were more sensitive to capture subtle gender differences, in comparison with categorical coding procedures. McLaughlin et al. (2012) hypothesized that girls may be better able to elicit sensitive caregiving than boys in foster care after institutional rearing or, alternatively, boys may have higher difficulties to develop a secure attachment post-institutionalization due to a more profound deleterious effect of institutional deprivation.

Oppositely to previous research that documented a strong relation of Friends and Family Interview scales with verbal IQ (Pace et al., 2015; Stievenart et al., 2011), only secure base regarding the mother was related to verbal IQ in this study. This discrepancy could be due to differences in the characteristics of the samples –countries of origin, range of IQ scores– and different instruments to assess IQ –such as the Wechsler Intelligence Scale for Children–. Finally, regarding family structure, the higher presence of attachment security in the attachment representations of children growing up in single-parent families has not been reported previously. Perhaps this confirms the hypothesis put forward long ago by Shireman and Johnson (1985) in their analysis of single-parent adoptions, suggesting that this family configuration may favor more intense parent-child relationship that could be particularly valuable for some adoptees. Other studies are needed to reach a more definite conclusion.

Previous literature has reported the link between attachment representations and placement variables. The length of institutionalization and age at adoption, usually used as quantitative indices of exposition to early adversity, have been linked to attachment security in studies that sampled school-age adoptees (Román et al., 2012; van den Dries et al., 2009), but not in studies with late-childhood and early adolescent adoptees (Barcons et al., 2012; Groza et al., 2012; Pace et al., 2015). The data reported in this article align with the latter studies. It could be that the complexity of preadoption experiences is not well captured by proxy variables (time in institutions, age at adoption). Another explanation is that during the adolescence, distal factors are less relevant than more proximal ones, like the parents’ attachment styles and the quality of postadoption family experiences, as in Barone et al. (2017).

The second research question considered whether internationally adopted children in late childhood performed differently on attachment representations in comparison with non-adopted children. Findings reported in this article reveal that, several years after their adoptive placement, adopted children obtained lower scores in coherence, reflective function, internal working models, and security, and higher scores in dismissiveness and disorganization, showing the long-lasting consequences of early adversity on attachment representations. These results should nevertheless be considered cautiously due to possible influence of other factors, such as the relationships with adoptive parents or post-adoptive stressful events.

Our attachment distribution is similar to other studies reporting data on children adopted in Eastern European countries using interview methods (see Table 1). Previous research has pointed out that adoptees from Eastern Europe had more than twice the risk of insecure attachment than children from other zones of origin, both in attachment representations (Barcons et al., 2014) and in attachment behavior (Lionetti et al., 2015; van den Dries et al., 2009). Regarding Friends and Family Interview specific variables, compared with the results found here, a previous study with children from Eastern Europe reported slightly higher scores in coherence (Abrines et al., 2012). The lower scores in the present study could be partially explained by differences in age at adoption, a year higher in our sample. Another partial explanation could be the focus on Russian adoptees in our current study, while Abrines et al. (2012) included adoptees from Russia, Ukraine, and Bulgaria.

The high rates of attachment insecurity among children adopted in Eastern European countries could result from the accumulative detrimental combination of different factors, such as the unfavorable caregiver-to-child ratio and deficient relationships, long stays in institutions and higher age at adoption (Julian et al., 2019). It is during the first year of life that attachment patterns are consolidated and become observable; being placed in an institution early in life may threaten or prevent the development of an organized pattern, leading to a higher rate of disorganized/cannot classify category in younger children (van den Dries et al., 2009). The fact that in our study children with and without family experience in their first year of life do not differ in their attachment representations could indicate that the experiences in the family before institutionalization were as negative as the institutional experiences.

In line with previous literature using interview methods, the percentage of disorganized attachment found in this study was surprisingly low −10.3%– in comparison with the high rates of disorganized behavior (van den Dries et al., 2009) and disorganized attachment representations found using story completion tasks (Steele et al., 2003). Following Pace (2014) it is unlikely that disorganized behavior and representations evaluated among adopted children in early stages ‘disappear’ at later stages. This author suggested that the Friends and Family Interview coding system could be improved to capture subtle cues of disorganization at a narrative level by the inclusion in future lines of research of two non-verbal codes: distress and fear, and frustration and anger.

Further, it is important to place the attachment classification of adopted children in context. If compared with their non-adopted peers there is a clear over-representation of insecurity and disorganization, the adoptees must also be compared with children left behind in institutional care. The 41% of secure and the 10% of disorganized attachment representations reported here are in contrast with the 18% and 54%, respectively, reported by Lionetti et al. (2015) in their meta-analysis of institutionalized children’s attachment behavior. While the negative consequences of early adversity are still detectable several years after adoption, the improvements favored by the new family relationships are also remarkable and the advantages of family versus institutional raising are highlighted.

This study has some limitations, and its results must be interpreted with caution. The sample size was reduced and came exclusively from Russia, so the results may not be representative of children adopted from other countries. This study presents a cross-sectional research design, with data collected 7 years after adoption. Different interpretations need to be considered, such as the quality of relationships with adoptive parents or stressful life events after adoption. Furthermore, this study may have been underpowered to detect some group differences, and we did not include a non-adopted institutionalized comparison group. Currently, we are undertaking two sets of analysis: the adoptive parents’ attachment styles to explore their relationship with their children’s attachment representations; and the attachment representations of the group of institutionalized children in the LAIS.US project.

Future lines of research would include a longitudinal evaluation of attachment to trace the complete sequence of development, like other longitudinal studies have done with other psychological domains (e.g., Sonuga-Barke et al., 2017). Our research team has implemented recently a third longitudinal assessment, when the sample was between 13- and 18 years old, and the corresponding attachment results will be available soon. Furthermore, future samples could be expanded to include groups of children adopted from other countries and with distinct developmental characteristics.

Practical implications suggest the need to train clinicians in the use of interview methods to allow detailed exploration of a child’s internal reality, but also the need to equip adoptive parents with strategies that allow them to satisfy their child’s long-term needs most adequately (Palacios, 2007). Additionally, adoptees should be evaluated early for attachment difficulties. Support for attachment recovery should be provided for those adoptees in need to help prevent future difficulties. Otherwise, the results of this study could point out the need to improve the effectiveness of social services and other care-related professionals in the sending countries, to enhance early detection of children exposed to depriving family contexts. The results of this study, and other similar ones, point out the need to promote protection policies alternative to institutional care, such as the placement in foster families, and, if institutionalization is unavoidable, improve the quality of institutions and decrease the length of institutional care.

Conclusions

To our knowledge, this is the first paper to study the attachment representations using interview methods with children adopted exclusively in Russia into Spanish families. Findings showed that adoptees were more likely to present insecure attachment than the control group, together with lower scores in variables related to coherence, reflective function, and internal working models. The attachment representation variables were related to adoptees’ age at assessment, gender, verbal IQ, and family structure, although no significant relations were found with mother’s academic level and placement variables. This research supports and extends previous findings of insecure attachment representations among post-institutionalized children and highlights the long-lasting impact of early adversity on attachment representations, as well as the heterogeneous impact on attachment representations of children variables such as age, gender, IQ, and placement variables.

Previous literature has shown that children reared in institutions have a higher prevalence of insecure attachment, in addition to the negative long-term consequences that non-secure patterns, especially disorganized attachment, have on children’s development. The results of our study highlight the benefits of adoption and the need to ensure that, in the absence of a stable and caring birth family, children are moved as early as possible into sensitive and supporting families.

Footnotes

Declaration of Conflicting Interests

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

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Spanish Ministry of Economy and Competitiveness [grant numbers SEJ2006-12216, PSI2010-19287], the Spanish Ministry of Economy, Industry and Competitiveness [grant number PSI2015-67757-R], and the European Regional Development Fund (ERDF).

ORCID iDs

María Peñarrubia

Maite Román

Jesús Palacios

Author Biographies

Jesús Palacios is honorary Professor at the Department of Developmental and Educational Psychology of the University of Seville (Spain). He does research in the areas of adoption and foster care from the perspective of Developmental Psychology and social policy.

Mari Peñarrubia is researcher at the Department of Developmental and Educational Psychology of the University of Seville (Spain). She does research in child protection –adoption and residential care– and she has specialized in attachment assessment.

Maite Román is tenured lecturer at the Department of Developmental and Educational Psychology of the University of Seville (Spain). She does research in child protection, and she is specialist in attachment, developmental contexts and children recovery after experiences of early adversity.

References

Abrines

Barcons

Marre

Brun

Fornieles

Fumadó

(2012). ADHD-like symptoms and attachment in internationally adopted children. Attachment & Human Development, 14(4), 405–423. https://doi.org/10.1080/14616734.2012.691656

Allen

(2008). The attachment system in adolescence. In Cassidy

Shaver

P. R.

(Eds.), Handbook of attachment (pp. 419–435). Guilford Press.

Balenzano

Coppola

Cassibba

Moro

(2018). Pre-adoption adversities and adoptees’ outcomes: The protective role of post-adoption variables in an Italian experience of domestic open adoption. Children and Youth Services Review, 85, 307–318. https://doi.org/10.1016/j.childyouth.2018.01.012

Ballús

Casas

Urrutia

Pérez-Testor

(2019). Attachment representations in international adolescent adoptees in Spain, over 8 to 17 years of placement. International Social Work, 62(6), 1507–1521. https://doi.org/10.1177/0020872819878484

Barcons

Abrines

Brun

Sartini

Fumadó

Marre

(2012). Social relationships in children from intercountry adoption. Children and Youth Services Review, 34(5), 955–961. https://doi.org/10.1016/j.childyouth.2012.01.028

Barcons

Abrines

Brun

Sartini

Fumadó

Marre

(2014). Attachment and adaptive skills in children of international adoption. Child & Family Social Work, 19(1), 89–98. https://doi.org/10.1111/j.1365-2206.2012.00883.x

Barone

Lionetti

Green

(2017). A matter of attachment? How adoptive parents foster post-institutionalized children’s social and emotional adjustment. Attachment & Human Development, 19(4), 323–339. https://doi.org/10.1080/14616734.2017.1306714

Beijersbergen

Juffer

Bakermans-Kranenburg

van IJzendoorn

(2012). Remaining or becoming secure: Parental sensitive support predicts attachment continuity from infancy to adolescence in a longitudinal adoption study. Developmental Psychology, 48(5), 1277–1282. https://doi.org/10.1037/a0027442

Berens

Nelson

(2015). The science of early adversity: Is there a role for large institutions in the care of vulnerable children? Lancet, 386(9991), 388–398. https://doi.org/10.1016/S0140-6736(14)61131-4

10.

Biryukova

Sinyavskaya

(2017). Children out of parental care in Russia: What we can learn from the statistics. The Journal of Social Policy Studies, 15(3), 367–382. https://doi.org/10.17323/727-0634-2017-15-3-367-382

11.

Bosmans

Kerns

(2015). Attachment in middle childhood: Progress and prospects. New Directions for Child and Adolescent Development, 148, 1–14. https://doi.org/10.1002/cad.20100

12.

Cohen

(1988). Statistical power analysis for the behavioral sciences. Lawrence Erlbaum Associates, Publishers.

13.

Cordero

Calonge

(2000). Test breve de inteligencia de Kaufman K-BIT. Versión española. TEA Ediciones.

14.

Escobar

Santelices

(2013). Attachment in adopted adolescents. National adoption in Chile. Children and Youth Services Review, 35(3), 488–492. https://doi.org/10.1016/j.childyouth.2012.12.011

15.

George

Kaplan

Main

(1984). Adult attachment interview [Unpublished manuscript]. University of California.

16.

Green

Stanley

Smith

Goldwyn

(2000). A new method of evaluating attachment representations in young school-age children: The Manchester Child Attachment Story Task. Attachment & Human Development, 2(1), 48–70. https://doi.org/10.1080/146167300361318

17.

Groza

Muntean

Ungureanu

(2012). The adoptive family within the Romanian cultural context: An exploratory study. Adoption Quarterly, 15(1), 1–17. https://doi.org/10.1080/10926755.2012.661327

18.

Johnson

D. E.

(2000). Medical and developmental sequelae of early childhood institutionalization in Eastern Europe Adoptees. In Nelson

(Ed.), The effects of early adversity on neurobehavioral development (pp. 113–162). Psychology Press.

19.

Juffer

Rosenboom

(1997). Infant-Mother attachment of internationally adopted children in the Netherlands. International Journal of Behavioral Development, 20(1), 93–107. https://doi.org/10.1080/016502597385469

20.

Julian

Wright

Jimenez-Etcheverria

(2019). Young children in institutional care: Characteristics of institutions, children’s development, and interventions in institutions. In Tulviste

Best

Gibbons

(Eds.), Children’s social worlds in cultural context (pp. 217–232). Springer.

21.

Kaufman

A. y.

Kaufman

(1990). Kaufman Brief intelligence test. American Guidance Service, Inc.

22.

Kriss

Steele

(2012). Measuring attachment and reflective functioning in early adolescence: An introduction to the friends and family interview. Research in Psychotherapy: Psychopathology, Process and Outcome, 15(2), 87–95. https://doi.org/10.7411/RP.2012.009

23.

Lionetti

Pastore

Barone

(2015). Attachment in institutionalized children: A review and meta-analysis. Child Abuse & Neglect, 42, 135–145. https://doi.org/10.1016/j.chiabu.2015.02.013

24.

McLaughlin

Zeanah

Fox

Nelson

(2012). Attachment security as a mechanism linking foster care placement to improved mental health outcomes in previously institutionalized children. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 53(1), 46–55. https://doi.org/10.1111/j.1469-7610.2011.02437.x

25.

Ministry of Health of the Russian Federation (2003). Services for mother and child protection in 2002. Russian Medical Science Academy.

26.

Molina

Casonato

Ongari

Decarli

(2015). Les représentations d’attachement chez les adolescents adoptés et leurs parents. Neuropsychiatrie de l’enfance et de l’adolescence, 63(6), 376–384. https://doi.org/10.1016/j.neurenf.2015.04.004

27.

Muntean

Ungureanub

(2012). Adoptions of children in Romania: Applying attachment theory. Procedia - Social and Behavioral Sciences, 33, 6–10. https://doi.org/10.1016/j.sbspro.2012.01.072

28.

Nelson

Gruendel

(1986). Children’s scripts. In Nelson

(Ed.), Event knowledge: Structure and function in development (pp. 21–46). Erlbaum.

29.

Observatorio de la Infancia (2020). Boletín de datos estadísticos de medidas de protección a la infancia. Boletín número 21. Datos 2018. Spanish Ministry of Social Rights & 2030 Agenda. https://observatoriodelainfancia.vpsocial.gob.es/productos/pdf/Boletin_Proteccion_21_Accesible.pdf

30.

Pace

(2014). Assessing attachment representations among adoptees during middle childhood and adolescence with the Friend and Family Interview (FFI): Clinical and research perspectives. Frontiers in Psychology, 5(1114), 1–4. https://doi.org/10.3389/fpsyg.2014.01114

31.

Pace

Di Folco

Guerriero

(2018). Late-adoptions in adolescence: Can attachment and emotion regulation influence behaviour problems? A controlled study using a moderation approach. Clinical Psychology & Psychotherapy, 25(2), 1–13. https://doi.org/10.1002/cpp.2158

32.

Pace

Di Folco

Guerriero

Santona

Terrone

(2015). Adoptive parenting and attachment: Association of the internal working models between adoptive mothers and their late-adopted children during adolescence. Frontiers in Psychology, 6, 1433. https://doi.org/10.3389/fpsyg.2015.01433

33.

Palacios

(2007). Después de la adopción: Necesidades y niveles de apoyo. Anuario de Psicología, 38(2), 181–198.

34.

Palacios

Brodzinsky

(2010). Adoption research: Trends, topics, outcomes. International Journal of Behavioral Development, 34(3), 270–284. https://doi.org/10.1177/0165025410362837

35.

Pierrehumbert

Santelices

Ibáñez

Alberdi

Ongari

Roskam

Stievenart

Spencer

Fresno-Rodríguez

Borghini

(2009). Gender and attachment representations in the preschool years: Comparisons between five countries. Journal of Cross-Cultural Psychology, 40(4), 543–566. https://doi.org/10.1177/0022022109335181

36.

Román

Palacios

Moreno

López

(2012). Attachment representations in internationally adopted children. Attachment & Human Development, 14(6), 585–600. https://doi.org/10.1080/14616734.2012.727257

37.

Sánchez-Sandoval

Palacios

León

(2002). Entrevista sobre el proceso de adopción internacional (EPAI). [Unpublished manuscript]. Faculty of Psychology, University of Seville.

38.

Selman

(2018). Global statistics for intercountry adoption: Receiving states and states of origin 2005-2018. Newcastle University.

39.

Shireman

Johnson

(1985). Single-parent adoptions: A longitudinal study. Children and Youth Services Review, 7(4), 321–334. https://doi.org/10.1016/s0190-7409(85)80005-0

40.

Shmueli-Goetz

Target

Datta

Fonagy

(2000). Child Attachment Interview (CAI) coding and classification manual. Version IV. University College London.

41.

Sonuga-Barke

Kennedy

Kumsta

Knights

Golm

Rutter

Maughan

Schlotz

Kreppner

(2017). Child-to-adult neurodevelopmental and mental health trajectories after early life deprivation: The young adult follow-up of the longitudinal English and Romanian adoptees study. The Lancet, 389(10078), 1539–1548. https://doi.org/10.1016/S0140-6736(17)30045-4

42.

Steele

(2005). The construction of coherence as an indicator of attachment security in middle childhood: The Friends and Family Interview. In Kerns

K. A.

Richardson

R. A.

(Eds.), Attachment in middle childhood (pp. 137–160). Guilford Press.

43.

Steele

Kriss

(2009). The Friends and Family Interview (FFI) coding guidelines. Unpublished manuscript. The New School for Social Research.

44.

Steele

Hodges

Kaniuk

Hillman

Henderson

(2003). Attachment representations and adoption: Associations between maternal states of mind and emotion narratives in previously maltreated children. Journal of Child Psychotherapy, 29(2), 187–205. https://doi.org/10.1080/0075417031000138442

45.

Steele

Hodges

Kaniuk

Steele

Hillman

Asquith

(2008). Forecasting outcomes in previously maltreated children: The use of the AAI in a longitudinal adoption study. In Steele

Steele

(Eds.), Clinical applications of the adult attachment interview (pp. 427–451). Guilford Press.

46.

Stievenart

Casonato

Muntean

van de Schoot

(2012). The friends and family interview: Measurement invariance across Belgium and Romania. European Journal of Developmental Psychology, 9(6), 737–743. https://doi.org/10.1080/17405629.2012.689822

47.

Stievenart

Roskam

Meunier

van de Moortele

(2011). The reciprocal relation between children’s attachment representations and their cognitive ability. International Journal of Behavioral Development, 35(1), 58–66. https://doi.org/10.1177/0165025410370790

48.

Tabachnick

Fidell

(2007). Using multivariate statistics. Pearson.

49.

The St. Petersburg-USA Orphanage Research Team (2005). Characteristics of children, caregivers, and orphanages for young children in St. Petersburg, Russian Federation. Journal of Applied Developmental Psychology, 26(5), 477–506. https://doi.org/10.1016/j.appdev.2005.06.002

50.

van den Dries

Juffer

van IJzendoorn

Bakermans-Kranenburg

(2009). Fostering security? A meta-analysis of attachment in adopted children. Children and Youth Services Review, 31(3), 410–421. https://doi.org/10.1016/j.childyouth.2008.09.008

51.

Vorria

Ntouma

Vairami

Rutter

(2015). Attachment relationships of adolescents who spent their infancy in residential group care: The Greek Metera study. Attachment & Human Development, 17(3), 257–271. https://doi.org/10.1080/14616734.2015.1028947

52.

Vorria

Papaligoura

Sarafidou

Kopakaki

Dunn

van IJzendoorn

Kontopoulou

(2006). The development of adopted children after institutional care: A follow-up study. The Journal of Child Psychology and Psychiatry, 47(12), 1146–1253. https://doi.org/10.1111/j.1469-7610.2006.01666.x