CoChat
Sign Up
Cited by 64Child Welfare and AdoptionEarly Childhood Education and DevelopmentChild Abuse and Trauma

IQ at age 12 following a history of institutional care: findings from the Bucharest Early Intervention Project

Open access
Almas, Alisa N., Degnan, Kathryn A., Nelson, Charles A., Zeanah, Charles H., Fox, Nathan A.·2016·Developmental Psychology
DOI 10.1037/dev0000167PMC PMC5083169
Open this paper in CoChat — free account

Abstract

Young children removed from institutions and placed into foster care or adoptive homes have been shown to experience significant gains in IQ relative to children who remain in institutions. Less is known about the long-term impact of severe early deprivation on development in late childhood. Data are presented from a follow-up of children at 12 years of age in the Bucharest Early Intervention Project, a randomized clinical trial of foster care for institutionally reared children. Of the original 136 children in the study, 107 were tested with the Wechsler Intelligence Scale for Children-IV. Results revealed continued benefit from the intervention through age 12, with 12-year-old children placed in foster care scoring significantly higher on full-scale IQ compared to their peers who received prolonged institutional care. Longitudinal IQ data revealed 2 IQ profiles from early to late childhood. Attachment security emerged as a significant predictor of a profile of stable, typical IQ scores over time. We demonstrate the continued importance of foster care intervention and the negative effects of severe, early psychosocial deprivation on IQ into late childhood. (PsycINFO Database Record

Integrity signals

Registry

Crossref · 10.1037/dev0000167

Cites

41 of 41 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

0 of 28 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    At age 12, children originally assigned to foster care had significantly higher full-scale IQ than children assigned to care as usual.

    “Results revealed continued benefit from the intervention through age 12, with 12-year-old children placed in foster care scoring significantly higher on full scale IQ compared to their peers who received prolonged institutional care.”
    • Also stated in: § Intent-to-Treat Analyses: Comparison of Intervention Groups · hedged
      “The result revealed a significant effect of group, F (1,105) = 4.55, p = .035, η2 = .04, indicating that children in the FC group had significantly higher full scale IQ scores than children in the CAU group.”
    Quote in the paperSubstantial claim
  • Findingown data

    Attachment security significantly predicted membership in a profile characterized by stable, typical IQ scores over time.

    “Attachment security emerged as a significant predictor of a profile of stable, typical IQ scores over time.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: IQ scores from different instruments across ages are comparable enough to model a single longitudinal IQ profile.
  • Findingcitation

    Across BEIP assessments, children who received institutional care early in life had lower IQ than a community comparison group.

    “At each assessment in the RCT, we consistently found negative effects of early psychosocial deprivation on IQ in the BEIP sample when comparing all children who received institutional care early in life to a comparison group of children living in the community (Fox et al., 2011; Nelson et al., 2007; Smyke et al., 2007).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Fox et al., 2011

      Not checked in the free run.

    • cites Nelson et al., 2007

      Not checked in the free run.

    • cites Smyke et al., 2007

      Not checked in the free run.

  • Findingown data

    Greater attachment security at 42 months was associated with a greater likelihood of following the Typical IQ profile.

    “The two IQ profiles were significantly differentiated by levels of birth weight and 42-month attachment security, such that greater birth weight, b = .002, t = 3.66, p < .001, d = .63, and attachment security, b = .948, t = 5.21, p < .001, d = .89, were associated with a greater likelihood of following the “Typical” IQ profile.”
    Quote in the paperSubstantial claim
  • Findingown data

    The foster care group outperformed the care-as-usual group on verbal comprehension.

    “Follow-up multivariate comparisons revealed a significant effect of group for the verbal subscale, F (1, 105) = 7.82, p = .006, η2 = .07, indicating that children in the FC group scored higher on verbal comprehension than those children in the CAU group.”
    Quote in the paperSubstantial claim
  • The quote only partly states this claim. 2
  • May combine two claims. 14
  • Classified as probably not a substantial claim. 11

  • The quote only partly states this claim.

    Findingown data

    The longitudinal IQ data from 30 months to age 12 revealed two distinct IQ profiles.

    “Longitudinal IQ data revealed two IQ profiles from early to late childhood.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: IQ scores from different instruments across ages are comparable enough to model a single longitudinal IQ profile.

  • May combine two claims.

    Findingown data

    Group significantly predicted the probability of membership in the Typical IQ profile, with NI children having the highest probability, followed by FC children, and then CAU children.

    “The result revealed a significant effect of group (F (2,143) = 116.83, p < .001, η2 = .63). Post-hoc pairwise comparisons with a Bonferroni correction revealed that children in the NI group had a significantly higher probability of membership in the Typical profile than did children in the CAU or FC groups, while children in the FC group had a significantly higher probability of membership in the”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    Prior BEIP work identified factors predicting distinct IQ profiles up to age 8, including early language abilities and, for foster care children, caregiver-child relationship characteristics.

    “In some of the most recent work, we have identified factors that predicted distinct patterns or profiles of IQ across development up to 8 years of age, including language abilities in early childhood and, for children in the FC group, characteristics of the caregiver-child relationship (Fox et al., 2011).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Fox et al., 2011

      Not checked in the free run.

  • The quote only partly states this claim.

    Findingcitation

    At age 8, children placed in foster care scored significantly higher than children randomized to care as usual.

    “Children placed in foster case scored significantly higher than those randomized to care as usual.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The distribution of children across the two IQ profiles differed by group: the Low profile contained 45 CAU, 37 FC, and 5 NI children, while the Typical profile contained 0 CAU, 12 FC, and 44 NI children.

    “The Low profile included 45 children from the CAU group, 37 children from the FC group, and 5 children from the NI group.”
    • Also stated in: § Longitudinal Profiles of IQ from 30 Months of Age to 12 Years of Age · plain
      “The Typical profile included 0 children from the CAU group, 12 children from the FC group, and 44 children from the NI group.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    There were no significant effects of age at placement into foster care on IQ at age 12, whether age was analyzed as dichotomous cutoffs or as a continuous variable.

    “Results revealed no significant effects of age at placement into foster care. Partial correlations were calculated for age at placement into foster care as a continuous variable and full scale IQ as well as each of the subscale scores, controlling for gender. There were no significant results.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Preliminary analyses found no gender differences on age-12 IQ except on Processing Speed, where females scored higher than males.

    “Preliminary analyses across all groups revealed no gender differences on IQ scores at age 12, except for the subscale of Processing Speed, t (157) = 2.13, p = .04, with females scoring higher than males.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    In the ERA project, children adopted before six months of age fell within the normal range of cognitive functioning at age 11, while those adopted after six months had cognitive deficits.

    “they found that when children were 11 years of age, those who were adopted prior to six months of age fell within the normal range of cognitive functioning while children adopted after six months of age had cognitive deficits (Beckett et al., 2006).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Beckett et al., 2006

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Dennis (1973) found that adolescents who entered institutions later had significantly higher IQ than those institutionalized since birth, and both groups scored lower than school-attending children from more typical families.

    “He found that adolescents who had come to live in the institutions at later ages had significantly higher IQ scores than those who had lived in institutional care since birth, and that both groups were lower than children who attended school with the sample children and came from more “typical” families.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Dennis (1973)

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    Foster care intervention has continued importance and severe early psychosocial deprivation has negative effects on IQ into late childhood.

    “We demonstrate the continued importance of foster care intervention and the negative effects of severe, early psychosocial deprivation on IQ into late childhood.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Few studies have followed post-institutionalized children into late childhood and adolescence, and even fewer have examined children who remain in institutional care during that period.

    “Only a handful of studies have followed post-institutionalized children into late childhood and adolescence (e.g., Beckett, Castle, Rutter, & Sonuga-Barke, 2010; Vorria, Mtouna, & Rutter, 2015), and even fewer have examined the development of children who remain in institutional care for part or all of this time (e.g., Dennis, 1973; Humphrey et al., 2015; McLaughlin et al., 2015).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Beckett, Castle, Rutter, & Sonuga-Barke, 2010

      Not checked in the free run.

    • cites Vorria, Mtouna, & Rutter, 2015

      Not checked in the free run.

    • cites Dennis, 1973

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Dennis (1973) found that girls aged 12-16 in Lebanese institutions had a mean IQ of 54, while boys aged 10-14 had a mean IQ of 81.

    “He found a mean IQ score of 54 for girls aged 12 to 16, while boys aged 10 to 14 with the same early institutional experience had a mean IQ score of 81.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Dennis (1973)

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Hodges and Tizard found that at age 16, adolescents adopted before age 4.5 years had significantly higher IQs than adolescents in all other groups.

    “Hodges and Tizard (1989), in their follow up of children from the UK who were institutionalized and later placed in foster care, reunited with their biological families or adopted, found that, at age 16, adolescents who were adopted before age 4.5 years had significantly higher IQs than adolescents in all other groups.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Hodges and Tizard (1989)

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Prior research consistently indicates that severe psychosocial deprivation negatively affects children's cognitive, social, and emotional development in childhood.

    “The impact of severe psychosocial deprivation on children’s development has long been studied, with conclusions consistently pointing to negative effects across cognitive, social and emotional domains in childhood (e.g., Bowlby, 1951; Dennis & Najarian, 1957; Goldfarb, 1943).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Bowlby, 1951

      Not checked in the free run.

    • cites Dennis & Najarian, 1957

      Not checked in the free run.

    • cites Goldfarb, 1943

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    The group comparison results were obtained using a conservative intent-to-treat approach, analyzing data based on original group assignment.

    “It is important to note that the group comparison results were found using a conservative “intent-to-treat” approach, meaning the data were analyzed based on original groups to which children were assigned to at the outset of the study.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductnone given

    The BEIP maintained a non-interference policy throughout the study, so children changed placements based on child protection authorities' decisions regardless of group assignment.

    “the BEIP maintained a non-interference policy throughout the course of the study, meaning children changed placements based on decisions made by child protection authorities regardless of group assignment.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Prior to randomization, infants were screened for obvious physical handicaps or genetic anomalies.

    “Prior to randomization, while infants were living in the institution, we screened each child for obvious physical handicaps or genetic anomalies.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    The never-institutionalized comparison group was not matched to the CAU or FC children on socioeconomic status, so direct comparisons between groups should be interpreted with caution.

    “It is also important to note the comparison group of never institutionalized children were not matched to the CAU or FC children on socioeconomic status, and therefore direct comparisons between the groups should be interpreted with caution.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisereasoning

    LPA relies on a formal statistical model based on probabilities to classify cases, unlike traditional cluster algorithms.

    “Second, unlike traditional cluster analysis algorithms, which group cases near each other by some definition of distance (e.g., Euclidean distance for k-means cluster analysis), the LPA approach relies on a formal statistical model based on probabilities to classify cases.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    Trained and reliable Romanian psychologists administered all IQ tests, with U.S. clinicians supervising.

    “All of the IQ assessments were conducted by trained and reliable Romanian psychologists and supervised by U.S. clinicians.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    LPA allows estimation of qualitatively different groups from continuous data when group membership cannot be determined a priori.

    “Latent Profile Analysis, a form of Structural Equation Mixture Modeling, was used because it allows for the estimation of qualitatively different groups (i.e., latent profiles) from continuous data points when group membership cannot be determined a priori (Bauer & Curran, 2004).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Bauer & Curran, 2004

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    The study maintained a policy of not interfering in placement decisions beyond the initial randomization.

    “We maintained a policy of not interfering in any placement decisions beyond the initial randomization.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    In previous work, earlier placement into foster care predicted higher IQ scores in early childhood (Nelson et al., 2007).

    “Although in previous work we have found the timing of placement into foster care was important in early childhood, with earlier placement predicting higher IQ scores (Nelson et al., 2007), we did not find similar results at age 12.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Nelson et al., 2007

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Scale reliability for the Quality of Caregiving score was acceptable (Cronbach's alpha = .82).

    “Scale reliability was acceptable (Cronbach’s alpha = .82).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    Reliability for attachment security coding was excellent (r = .87).

    “As reported previously (Smyke et al., 2010), reliability for this coding was excellent (r = .87)”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Smyke et al., 2010

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Roy, Rutter and Pickles (2000) found that foster children who lived continuously in the same foster home had significantly lower rates of problem behaviors than institutional-reared children, but did not differ from institutional children on IQ.

    “Roy, Rutter and Pickles (2000), when examining a sample of children specifically selected for their continuity of placement, found similar results. Foster children in their sample, who had lived continuously in the same foster home, demonstrated significantly lower rates of problem behaviors compared to institutional-reared children, while the foster children, however, did not differ from the”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Roy, Rutter and Pickles (2000)

      Not checked in the free run.

26 of 26

  • The authors state this as an assumption.

    Missing data are missing at random.

    “use of SEM’s maximum-likelihood estimation (MLE) method assumes the data are missing at random”

  • The paper relies on this without support.

    The NI group is representative of typically developing Romanian children.

    “Unlike some other studies, the BEIP includes a representative comparison group, the NI group”

  • The paper relies on this without support.

    IQ scores from different instruments across ages are comparable enough to model a single longitudinal IQ profile.

    “IQ had been assessed at 30 and 42 months of age using the Bayley Scales of Infant Development, at 54 months of age using the WPPSI (see Nelson et al, 2007, for details), and at 8 years of age using the WISC-IV (see Fox, et al, 2011, for details).”
    “latent profiles of IQ were estimated for all children in the sample (CAU, FC, and NI groups) using latent profile analysis (LPA). This method allowed for the examination of longitudinal patterns of IQ from 30 months to 12 years of age.”

  • The paper relies on this without support.

    Selection bias may exist in which children are adopted early versus later from institutions.

    “For one, there may be a selection bias in which children are adopted early vs. later from institutions (also who gets adopted and who does not).”

  • The paper relies on this without support.

    The foster care intervention provided in the BEIP was of high quality.

    “These results highlight the positive impact of high quality, foster care experienced early in life on remediating some of the negative effects of severe psychosocial deprivation on cognitive development even 8 years after the RCT concluded.”
    “high quality care early in life has an impact on cognitive success through to late childhood.”

  • The paper relies on this without support.

    The never-institutionalized comparison group is representative of typical cognitive development.

    “children from the community (NI) who had never experienced institutional care.”
6 of 6

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Statistics recomputed

Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.

  • Not confirmedχ²(16) = 152.29reported p = .22 · recomputed p < .001

    “χ2(16)=152.29, p = .22”

  • Not confirmedχ²(16) = 60.39reported p = .06 · recomputed p < .001

    “χ2(16)=60.39, p = .06”

  • ConsistentF(1, 107) = 0.59reported n.s. · recomputed p = .444

    “F (1, 107) = .59, ns”

  • ConsistentF(1, 104) = 0.21reported n.s. · recomputed p = .648

    “F(1,104) = .21, ns.”

  • Consistentt(157) = 2.13reported p = .04 · recomputed p = .035

    “t (157) = 2.13, p = .04”

  • ConsistentF(1, 105) = 4.55reported p = .035 · recomputed p = .035

    “F (1,105) = 4.55, p = .035”

  • ConsistentF(1, 105) = 7.82reported p = .006 · recomputed p = .006

    “F (1, 105) = 7.82, p = .006”

  • ConsistentF(1, 105) = 0.88reported n.s. · recomputed p = .350

    “F (1,105) = .88, ns”

  • ConsistentF(1, 105) = 1.46reported n.s. · recomputed p = .230

    “F (1,105) = 1.46, ns”

  • ConsistentF(1, 105) = 2.1reported n.s. · recomputed p = .150

    “F (1,105) = 2.10, ns”

  • ConsistentF(1, 105) = 42.73reported p < .001 · recomputed p < .001

    “F (1,105) = 42.73, p < .001”

  • ConsistentF(4, 99) = 11.04reported p < .001 · recomputed p < .001

    “F (4, 99) = 11.04, p < .001”

  • ConsistentF(1, 105) = 36.38reported p < .001 · recomputed p < .001

    “F (1,105) = 36.38, p < .001”

  • ConsistentF(1, 105) = 36.54reported p < .001 · recomputed p < .001

    “F (1,105) = 36.54, p < .001”

  • ConsistentF(1, 105) = 23.79reported p < .001 · recomputed p < .001

    “F (1,105) = 23.79, p < .001”

  • ConsistentF(1, 105) = 26.86reported p < .001 · recomputed p < .001

    “F (1,105) = 26.86, p < .001”

  • Consistentr(102) = -0.26reported p = .008 · recomputed p = .008

    “r (102) = −.26, p = .008”

  • Consistentr(102) = -0.29reported p = .003 · recomputed p = .003

    “r (102) = −.29, p = .003”

  • Consistentr(102) = -0.23reported p = .02 · recomputed p = .019

    “r (102) = −.23, p = .02”

  • ConsistentF(1, 49) = 2.78reported n.s. · recomputed p = .102

    “F (1, 49) = 2.78, ns.”

  • ConsistentF(4, 46) = 1.97reported n.s. · recomputed p = .115

    “F (4, 46) = 1.97, ns.”

  • Consistentχ²(16) = 394.78reported p = .01 · recomputed p < .001

    “χ2(16)=394.78, p = .01”

    The paper prints p at a floor (such as p < .001)

  • ConsistentF(2, 143) = 116.83reported p < .001 · recomputed p < .001

    “F (2,143) = 116.83, p < .001”

23 of 23

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

  • Low-profile group counts fall short of its reported size

    Minor

    The listed CAU, FC, and NI members of the Low profile add to 87, not the reported 89.

    Reported 89 · recomputed 87

1 of 1

Cites

Flagged first, then by citations

Research works printed but not linked: 2

  1. Bzoch KR League R 1971 Assessing language skills in infancy Baltimore University Park Press
  2. van IJzendoorn MH Luijk MPCM Juffer F 2008 IQ of children growing up in children’s homes Merrill-Palmer Quarterly 54 341 366

Cited by

Sorted by citations

Related

Authors
Almas, A. N., Degnan, K. A., Nelson, C. A., Zeanah, C. H., & Fox, N. A.
Affiliations
  • University of British Columbia (CA)
  • University of Maryland, College Park (US)
  • Boston Children's Hospital (US)
  • Tulane University (US)
ORCID
Venue
Developmental Psychology
Publisher
American Psychological Association (APA)
Published
2016-11
Volume
52
Issue
11
Pages
1858-1866
ISSN
  • 0012-1649 (ISSN-L)
  • 1939-0599
DOI
10.1037/dev0000167
PMC
PMC5083169
OpenAlex
W2529287335
Cited by
63
References
41
Citations by year
2026: 4 · 2025: 6 · 2024: 3 · 2023: 10 · 2022: 8 · 2021: 5 · 2020: 10 · 2019: 5 · 2018: 7 · 2017: 5
FWCI
13.82
Type
journal article
Language
en
Open access
Licence
Locations
  • Developmental Psychology (publishedVersion, other-oa)
  • PubMed (publishedVersion)
  • PubMed Central (acceptedVersion)
Funders
  • Binder Family Foundation
  • National Institutes of Health (R01MH091363)
Topics
  • Social Sciences › Safety Research › Child Welfare and Adoption
  • Social Sciences › Education › Early Childhood Education and Development
  • Social Sciences › Psychology › Clinical Psychology › Child Abuse and Trauma
Keywords
institutional care, foster care, IQ, early deprivation, psychosocial deprivation, attachment security, Bucharest Early Intervention Project, child development, late childhood, adoptive placement
MeSH terms
Age Factors, Child, Child Development, Child, Institutionalized, Female, Foster Home Care, Humans, Intelligence, Intelligence Tests, Longitudinal Studies, Male, Intellectual Disability, Caregivers, Early Intervention, Educational

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-04.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-04.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review