Content uploaded by Kazuko Y Behrens
Author content
All content in this area was uploaded by Kazuko Y Behrens on Feb 20, 2019
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=rahd20
Download by: [76.69.153.237] Date: 08 April 2016, At: 14:42
Attachment & Human Development
ISSN: 1461-6734 (Print) 1469-2988 (Online) Journal homepage: http://www.tandfonline.com/loi/rahd20
Infant attachment, adult attachment,
and maternal sensitivity: revisiting the
intergenerational transmission gap
Kazuko Y. Behrens, John D. Haltigan & Naomi I. Gribneau Bahm
To cite this article: Kazuko Y. Behrens, John D. Haltigan & Naomi I. Gribneau Bahm (2016):
Infant attachment, adult attachment, and maternal sensitivity: revisiting the intergenerational
transmission gap, Attachment & Human Development
To link to this article: http://dx.doi.org/10.1080/14616734.2016.1167095
Published online: 08 Apr 2016.
Submit your article to this journal
View related articles
View Crossmark data
Infant attachment, adult attachment, and maternal
sensitivity: revisiting the intergenerational transmission gap
Kazuko Y. Behrens
a
, John D. Haltigan
b
and Naomi I. Gribneau Bahm
c
a
Department of Social & Behavioral Sciences, State University of New York Polytechnic Institute, Utica, NY,
USA;
b
Faculty of Education & Faculty of Social Sciences, University of Ottawa, Ottawa, ON, Canada;
c
Psychology Departments, American River College and Cosumnes River College, Sacramento, CA, USA
ABSTRACT
This study investigated the intergenerational transmission of
attachment, utilizing the Adult Attachment Interview (AAI), the
Strange Situation Procedure (SSP), and the Maternal Behavioral
Q-Set (MBQS). We revisited fundamental questions in attachment
theory and research by examining: (1) the level of intergenera-
tional agreement between maternal attachment representations
and infant attachment security, and (2) whether maternal sensitiv-
ity serves as an intergenerational mediator between adult and
infant attachment security. Significant categorical matches
between the AAI and the SSP as well as mean differences for
MBQS scores between adult attachment secure-insecure groups
were found. Consistent with earlier intergenerational research,
maternal sensitivity only partially mediated the AAI-SSP link, indi-
cating the transmission gap remains. Consistent with recent med-
iation studies, using more contemporary analytical techniques, it
was confirmed that maternal sensitivity did mediate the direct
pathway between AAI security and SSP security. Thus, the trans-
mission gap appears somewhat different depending on the statis-
tical method used to measure mediation. Post hoc analyses
considered mothers’childhood experiences of separation/divorce
and this helped make sense of intergenerational mismatches.
ARTICLE HISTORY
Received 27 February 2015
Accepted 14 March 2016
KEYWORDS
Attachment; Adult
Attachment Interview;
Strange Situation; maternal
sensitivity; intergenerational
transmission
Intergenerational transmission of attachment security has been well-documented (van
IJzendoorn, 1995), primarily based on security status assessed by the Adult Attachment
Interview (AAI: protocol, George, Kaplan, & Main, 1996; scoring and classification system,
Main, Goldwyn, & Hesse, 2003), predicting infant attachment security assessed by the
Strange Situation Procedure (SSP: Ainsworth, Blehar, Waters, & Wall, 1978). Maternal
sensitive responsiveness to the child’s attachment needs has been assumed to play a
central role in the process of transmitting attachment intergenerationally. Mothers with
secure states of mind are expected to respond to their children’s needs swiftly and
appropriately, because these mothers (as opposed to those with insecure states of mind)
are believed more likely to value attachment relationships and to be reflective and
thoughtful (Main et al., 2003). However, despite the strong association found between
the SSP and AAI with a large effect size (d) of 1.06 (N= 854), the association between
CONTACT Kazuko Y. Behrens kazuko.behrens@sunyit.edu
ATTACHMENT & HUMAN DEVELOPMENT, 2016
http://dx.doi.org/10.1080/14616734.2016.1167095
© 2016 Informa UK Limited, trading as Taylor & Francis Group
Downloaded by [76.69.153.237] at 14:42 08 April 2016
maternal attachment security status and maternal sensitive responsiveness (n= 389) was
found less robust (d= .72: van IJzendoorn, 1995). van IJzendoorn thus coined the term
the “transmission gap,”referring to that aspect of the attachment transmission process
from parent to offspring which cannot be fully explained by maternal sensitivity.
Pederson, Gleason, Moran, and Bento (1998) pointed out that the inconsistent sensi-
tivity measures chosen for the studies included in van IJzendoorn’s(1995) meta-analysis
might partly explain the relatively modest transmission effects for maternal sensitivity.
Indeed, van IJzendoorn (1995) also posited that existing sensitivity measures might not
reliably capture certain aspects of mother–child interactions which were responsible for
the transmission of attachment security. Pederson et al. (1998) employed the SSP, AAI,
and a Q-sort based sensitivity measure, namely, the Maternal Behavioral Q-Set (MBQS:
Pederson & Moran, 1995), conducting the first study to test the full mediation model
among these three constructs –AAI secure representation, MBQS sensitivity, and SSP
security –concurrently. Pederson and Moran (1995) had previously reported a much
stronger association (r= .60) between attachment security and maternal sensitivity with
the MBQS than what had been reported (r= .24) in De Wolffand van IJzendoorn’s(1997)
meta-analysis. By using the MBQS, Pederson et al. (1998) sought to document a stronger
association between maternal attachment representation and sensitivity. However, the
link between the MBQS and the AAI was again found to be weaker than both the AAI-
SSP link and the MBQS-SSP link (Pederson et al., 1998). Somewhat surprisingly, however,
no study with a normative sample has replicated Pederson et al.’s study to date despite
the clear need to directly examine the mediation effect to help better understand the
transmission gap.
It is rather refreshing that Pederson, Bailey, Tarabulsy, Bento, and Moran (2014)
recently conducted a study in an attempt to re-conceptualize sensitivity from a more
relationship-focused perspective, following Ainsworth’s tradition of extremely detailed
note-taking. Pederson et al. (2014) have indeed shown a more robust association
(r= .65) to infant attachment security, affirming their appreciation of Ainsworth’s
insights as a tribute to her legacy. While a study such as this reminds us of the
importance of recognizing theoretical implications of associations among attachment
constructs, other recent studies have begun to investigate the largely unexplored or less
understood potential mechanisms that might affect the intergenerational transmission
of attachment security. For example, neural mechanisms underlying parental behaviors
have been examined and individual differences in processing certain cues such as
emotions, threat, or reward were found (e.g., Lenzi et al., 2013; Riem, Bakermans-
Kranenburg, van IJzendoorn, Out, & Rombouts, 2012; Swain, 2011). Similarly, in their
pilot EEG study, Behrens, Li, Gribneau Bahm, and O’Boyle (2011) found that a secure
mother responded to her own child’s images similarly to the way she responded to
positive images (e.g., puppies, smiling baby, happy couple), whereas mothers with
insecure representations responded to their child’s images similarly to the way they
responded to negative images (e.g., worms, emaciated child, corpse). These studies
appear to indicate that there are neurological correlates for parents’sensitive/insensitive
behaviors or approach/withdrawal tendencies when interacting with their children.
These new findings have provided valuable insights regarding factors that may affect
the transmission process but do not directly address the transmission gap, nor offer
conclusive evidence for its resolution. Therefore, in our view, continued research into the
2K. Y. BEHRENS ET AL.
Downloaded by [76.69.153.237] at 14:42 08 April 2016
mediation processes involved in the intergenerational transmission of attachment secur-
ity remains necessary and justified.
New questions have been posed as to whether the effect size reported in 1995 is still
representative, such as whether the level of attachment transmission may be influenced
by the type of sample, the methodology used, or the inclusion of unpublished data
(Verhage et al., 2016). To address these questions, a meta-analysis pooling all extant data
over three decades, both published and unpublished, was recently completed (Verhage
et al., 2016). Verhage et al.’s meta-analysis included a substantially larger total sample
size (95 samples, N= 4819) than van IJzendoorn’s meta-analysis (18 samples, N= 854)
two decades earlier. In this meta-analysis, Verhage et al. confirmed the intergenerational
transmission of attachment found by van IJzendoorn but with overall smaller effect
sizes. Both publication bias and decline effect (Schooler, 2011) were identified, indicating
smaller effect sizes for unpublished studies as well as more recent studies. Despite the
extensiveness and thoroughness of this project, the transmission gap remained
although more recent studies did narrow the gap. Verhage et al. noted that one of
the potential reasons for the decline effect could be that replication studies tend to
include more diverse samples than the original studies. This led us to ask: if the sample
characteristics were to remain similar, would we find an effect size comparable to that of
Pederson et al. (1998) original study? We believe we can best contribute to the field by
returning to the primary fundamental questions inherent in the intergenerational trans-
mission process: (1) What is the actual level of intergenerational agreement between
maternal AAIs and infant–mother SSPs? and (2) Does maternal sensitivity (as measured
by the MBQS) mediate this link? The current study re-visits these questions as concurrent
examinations of these three attachment constructs are still needed to better understand
the mediating mechanisms between adult representational attachment security and
infant behavioral attachment security.
The present study
The current study examined the intergenerational transmission of attachment security
utilizing the AAI and SSP. To investigate whether maternal sensitivity mediates the AAI-
SSP link, we employed the MBQS as had Pederson et al. (1998). We believe the present
report represents an important addition to the literature in the form of a much needed
replication of Pederson et al. (1998) study examining these three core attachment
constructs concurrently. We also investigated the same constructs by conducting an
additional mediation analysis to explore potential differences in our results to inform
how we interpret these findings. Further, we consider possible factors not previously
explored in the attachment literature and offer potential theoretical explanations that
may inform the gap, specifically when a discrepancy exists between mothers’attach-
ment representations and their interactive behaviors with the child. Finally, we pose
questions: What does closing the transmission gap truly mean? Is it in fact necessary to
close the gap? Do we stop our investigation once the gap is closed empirically? We
revisit these questions and discuss the implications from the current findings.
We hypothesize that: (1) A strong AAI-SSP categorical match will be found, replicating
numerous studies; (2) MBQS scores will be significantly higher in the AAI secure group
than in the AAI insecure group, as has been found with SSP categories (e.g., Behrens,
ATTACHMENT & HUMAN DEVELOPMENT 3
Downloaded by [76.69.153.237] at 14:42 08 April 2016
Parker, & Haltigan, 2011); and (3) A mediation will be confirmed, but the transmission
gap will nonetheless remain, as Pederson et al. (1998) reported.
Methods
Participants
Seventy-four mothers and their 12-month-olds were recruited from the local community
through public announcements to participate in the SSP as part of a longitudinal social-
emotional development project (Hart & Behrens, 2013). Of these, eight mothers were
unable to participate in the AAI session, resulting in a total sample of 66 dyads. Infants’
mean age was 53.8 weeks (SD = 3.1). Thirty-four (52%) infants were boys, and 36 (55%)
were first-borns. Mother’s age ranged from 20 to 40 years old (M= 29.2, SD = 4.8). The
majority of mothers (83%) were Caucasian with 12% Hispanic and 5% Other. The sample
was considered middle class using Hollingshead’s Index of Social Position (Hollingshead,
1971) to create five class categories, with 1 being the highest and 5 being the lowest
(M= 2.58, SD = .84).
Procedures
Mothers and their infants arrived at a university laboratory to undergo the standard SSP.
Before proceeding, adequacy of the child’s general health was established and mothers
were briefed about the procedure. Mothers consented for the dyads to be video-taped
during the entire procedure. Mothers returned alone at a separate visit to participate in
the standard AAI, during which they consented to be both audio-taped and video-taped
for the entire duration of the interview. At each visit, participants received a gift card for
a major retail store.
Measures
Strange Situation Procedure
The SSP (Ainsworth et al., 1978) is a laboratory procedure consisting of eight epi-
sodes, each three minutes long (except for a brief introductory episode), which
include two separations from, and two reunions with, mother. Infants who actively
sought contact with the mother upon her return (thus receiving high scores on
Proximity Seeking (PS) and/or Contact Maintenance (CM) scales), but settled down
quickly and resumed exploration, were judged secure (B). Infants who largely avoided
the mother upon reunion and were more focused on toys (thus receiving high scores
on the Proximity Avoidance (PA) scale), were judged insecure-avoidant (A). Infants who
were focused on their mother’s whereabouts throughout the procedure and, upon
reunion, failed to settle down and showed anger (thus receiving high scores on the
Contact Resistance (CR) scale), were judged insecure-ambivalent (C). Infants who
exhibited brief bouts of disorganized, often incomprehensible, behaviors in the pre-
sence of the mother (thus receiving high scores on the Disorganized/disoriented scale),
were judged disorganized (D). Those who were judged primarily D were also given a
secondary organized classification(B,A,C).
4K. Y. BEHRENS ET AL.
Downloaded by [76.69.153.237] at 14:42 08 April 2016
Adult Attachment Interview
The AAI is a semi-structured, one-hour interview, asking interviewees to describe and
reflect on childhood experiences with their parents. Verbatim transcripts were coded
according to the Main, Goldwyn, and Hesse method of classification (Main et al., 2003).
To achieve a final classification, coders must first rate a number of scales for speaker’s
experiences and states of mind. Mothers who coherently and openly discussed their
childhood experiences with their parents (whether positive or negative) while demon-
strating that they clearly valued attachment relationships, were judged secure (F).
Mothers who positively described childhood relationships with their parents but did
not present supportive evidence were judged insecure-dismissing (Ds). Mothers who
often rambled and angrily discussed their childhood experiences and/or current rela-
tionships with (a) parent(s), were judged insecure-preoccupied (E). Mothers who showed
significant lapses in monitoring speech/reasoning during discussions regarding
deceased loved ones or traumatic experiences (thus receiving high scores on the
Unresolved scale) were judged unresolved (U). Those who were judged primarily U
were also given a secondary organized classification (F, Ds, E).
Maternal Behavioral Q-Set (Pederson & Moran 1995)
The MBQS consists of 90 cards, each with a descriptor of an interactive maternal
behavior with the child. Trained observers sort the cards, yielding a composite score,
ranging from −1 to +1, representing the similarity between a prototypically sensitive
mother and the mother being observed in the home. Shorter versions of the MBQS for
laboratory use are now available (e.g., Behrens, Parker, et al., 2011; Tarabulsy et al., 2009).
We used the 72-item version of the MBQS to code maternal behaviors during the SSP,
observing all episodes when mother was present (i.e., episodes 1–3, 5, and 8).
Validity
The SSP and the AAI are both widely-used and well-validated measures of attachment
with strong predictability of intergenerational transmission of attachment security (van
IJzendoorn, 1995). The AAI has been tested for a number of psychometric properties,
and found to be unrelated to memory, intelligence, social desirability, or discourse style
(e.g., Bakermans-Kranenburg & van IJzendoorn, 1993; Crowell et al., 1996).
The MBQS has been found to be associated with attachment security at levels (r= .60,
Pederson & Moran, 1996) which markedly exceed those reported in a meta-analysis of
other sensitivity measures (r= .24, De Wolff& van IJzendoorn, 1997). Findings from
studies utilizing the MBQS have been consistent across samples (Atkinson et al., 2005).
When the 72-item version of the MBQS was employed, strong associations between
MBQS scores and infants’attachment classifications, as well as interactive behavioral
scale scores, were reported (Behrens, Parker, et al., 2011). The 72-item version of the
MBQS was also tested for its comparability with a contingency-based measure used in
the same sample during another laboratory procedure two months prior to the MBQS
coding (Behrens, Hart, & Parker, 2012). A significant link was found, showing evidence of
convergent validity for this version of the MBQS in this particular context (SSP). The 72-
item version of the MBQS was further validated when a sub-sample was re-examined at
home with the MBQS-preschooler version (Pederson, Moran, & Bento, 2004) 30 months
later (Behrens, Parker, & Kulkofsky, 2014) and found significantly correlated.
ATTACHMENT & HUMAN DEVELOPMENT 5
Downloaded by [76.69.153.237] at 14:42 08 April 2016
Reliability
All SSPs were classified by a certified coder (KB).
1
Out of the full sample (N= 74), 38
tapes (53%) were double coded by other certified coders (JH, EC). Inter-rater agreement
for a two-way analysis (secure vs. insecure) was 84%, kappa = .68. Disagreements were
conferenced and resolved. For agreement between the primary coder and a reliability
coder who coded 25 cases (34%) on the eight reunion scales (4 scales x 2 reunions), the
intraclass correlation coefficients (ICC) ranged from .76 to .95 (M= .88).
All AAIs were coded by a certified coder (NB) who had attended multiple AAI
institutes and achieved over 80% agreement with Mary Main and Erik Hesse. Fifteen
(23%) AAI transcripts were double coded by other certified AAI coders (KM, KB) who
were also trained and certified reliable by Mary Main and Erik Hesse. Inter-rater agree-
ment for a two-way analysis (secure vs. insecure) was 93%, kappa = .86; three-way
analysis (F, Ds, E) was 93%, kappa = .89; four-way analysis (F, Ds, E, U) was 93%,
kappa = .90. All coders were blind to all other measures and participant information.
MBQS coders spent several months familiarizing themselves with the card descriptors
and the Q-sort measure, using several practice cases. Both coders, although familiar with
attachment theory, were untrained in the SSP or any other attachment measures and
blind to all information regarding the dyads (ICC = .93, based on 35% of the full sample
data).
Data analytic strategy
Following the preliminary analyses, attachment distributions for both the SSP and the
AAI are first reported. We then investigate a categorical match between the AAI and SSP.
We utilize a traditional categorical approach by examining (a) two-way (Secure-Insecure)
when Unresolved/Secure AAI and Disorganized/Secure SSP are considered Insecure, (b)
forced two-way (F-nonF vs. B-nonB) when Unresolved/Secure AAI is F and Disorganized/
Secure SSP is B, (c) three-way (F, Ds, E vs. B, A, C), and (d) four-way (F, Ds, E, U vs. B, A, C,
D) analyses.
Next, we examine mean differences of MBQS sensitivity scores between the AAI
secure and insecure groups. Mean differences of MBQS sensitivity scores between SSP
secure/insecure groups were previously reported elsewhere (Behrens, Parker, et al.,
2011). Finally, we examine whether MBQS sensitivity mediates the transmission of AAI
security to SSP security.
Results
Descriptive analyses
Preliminary analyses
A series of independent t-tests and correlational analyses showed that mother’s age, SES,
child’s gender, and child’s birth order had no effect on attachment group differences for
any of the measures: SSP, AAI, or MBQS. However, when we examined these variables
against infants’interactive behavior scores, child’s birth-order yielded significantly dif-
ferent CM scores, with later-born infants showing elevations in CM relative to first-borns
during the second reunion (t= 2.55, df = 64, p= .017).
6K. Y. BEHRENS ET AL.
Downloaded by [76.69.153.237] at 14:42 08 April 2016
Attachment distributions
The two-way distribution of the SSP classification for the current sample of 66 infants
was 42 (64%) secure and 24 (36%) insecure (avoidant, ambivalent, disorganized). The
forced two-way distribution of the SSP classification was 48 (72%) B and 18 (28%) non-B.
The three-way distribution of the SSP classification was 48 (72%) B, 9 (14%) A, and 9
(14%) C. The four-way distribution of the SSP classification was 42 (64%) B, 8 (12%) A, 8
(12%) C, and 8 (12%) D. The two-way distribution of the AAI classifications for the current
sample of 66 was 46 (70%) secure and 20 (30%) insecure (Dismissing, Preoccupied,
Unresolved). The forced two-way distribution of the AAI classification was 47 (71%) F
and 19 (29%) non-F. The three-way distribution of the AAI classification was 47 (71%) F,
15 (23%) Ds, and 4 (6%) E. The four-way distribution of the AAI classification was 46
(70%) F, 12 (18%) Ds, 3 (4%), E, and 5 (8%) U.
Main analyses
Crosstabulation
To test hypothesis 1, crosstabulation analyses were run to investigate the categorical
match between the SSP and AAI. Table 1 shows that a significant association was found
for the secure-insecure two-way match, kappa = .46, p< .001, and for the forced two-
way match, kappa = .44, p< .001. Table 2 shows signification associations for the three-
way match, kappa = .25, p= .001, and for the four-way match, kappa = .24, p= .002.
Independent-samples t-tests
To test hypothesis 2, independent-samples t-tests revealed that MBQS scores for the
secure group (secure: M= .74, SD = .13) were significantly higher than the insecure
group (insecure: M= .61, SD = .29) for the two-way analysis of the AAI, t(64) = 2.56,
p= .013. However, MBQS scores did not significantly differ between the F group
Table 1. The two-way categorical match between the AAI classifications and the SSP
classifications.
AAI
SSP Secure Insecure Total
Secure Count 36 6 42
(Expected) (29) (13)
% of total 55% 9% 64%
Insecure Count 10 14 24
(Expected) (17) (7)
% of total 15% 21% 36%
Total Count 46 20 66
% of total 70% 30% 100%
AAI
SSP F Non-F Total
B Count 40 8 48
(Expected) (34) (14)
% of total 61% 12% 73%
Non-B Count 7 11 18
Expected 13 5 18
% of total 11% 17% 27%
Total Count 47 19 66
% of total 71% 29% 100%
ATTACHMENT & HUMAN DEVELOPMENT 7
Downloaded by [76.69.153.237] at 14:42 08 April 2016
(including Unresolved/Secure) and the non-F group (all organized insecure groups) for
the forced two-way analysis.
Mediational analyses
Hypothesis 3 was tested with mediational analyses. We first performed the classic Baron
and Kenny (1986) procedure to show the causal steps approach, followed up with the
Structural Equation Modeling (SEM) to test whether maternal sensitivity mediated the
direct AAI-SSP link.
Baron and Kenny test
To examine the question of whether maternal sensitivity (MBQS) mediates the transmis-
sion of AAI security
2
to SSP security, we followed the four-step approach outlined by
Baron and Kenny (1986) in which a sequence of regression analyses is conducted.
Because the outcome variable, or dependent variable, is assumed to be continuous for
regression analyses, it was necessary to convert SSP security into a continuous variable.
Eight SSP 7-point interactive behavior scales (four scales from two reunions) were used
to create a Richters’security score (Richters, Waters, & Vaughn, 1988). Richters et al.
developed a series of discriminant functions to objectively quantify infants’attachment
security based on interactive scores and crying behaviors, which was cross-validated
based on data from 255 SSPs. van IJzendoorn and Kroonenberg (1990) further simplified
Table 2. The three-way and four-way categorical match between the AAI clasifications and the SSP
classifications.
AAI
SSP F Ds E Total
B Count 40 6 2 48
(Expected) (34) (11) (3)
% of total 61% 9% 3% 73%
A Count 5 3 1 9
(Expected) (6) (2) (1)
% of total 8% 5% 2% 14%
C Count 2 6 1 9
(Expected) (6) (2) (1)
% of total 3% 9% 2% 14%
Total Count 47 15 4 66
% of total 71% 23% 6% 100%
AAI
SSP F Ds E U Total
B Count 36 4 1 1 42
(Expected) (29) (8) (2) (3)
% of total 55% 6% 2% 2% 64%
A Count 4 2 1 1 8
(Expected) (6) (2) (0) (1)
% of total 6% 3% 2% 2% 12%
C Count 2 3 1 2 8
(Expected) (6) (2) (0) (1)
% of total 3% 5% 2% 3% 12%
D Count 4 3 0 1 8
(Expected) (6) (2) (0) (1)
% of total 6% 5% 0% 2% 12%
Total Count 46 12 3 5 66
% of total 70% 18% 5% 8% 100%
8K. Y. BEHRENS ET AL.
Downloaded by [76.69.153.237] at 14:42 08 April 2016
the coefficients by removing the crying scale, re-computing the discriminant functions,
and rounding the coefficients to two decimals for easier use.
In the first step, to test for the AAI-SSP path alone, a simple regression analysis was
conducted with AAI security predicting SSP security. In the second step, to test for the
AAI-MBQS path alone, a simple regression analysis was run with AAI security predicting
maternal sensitivity. In the third step, to test for the MBQS-SSP path alone, a simple
regression analysis was run with maternal sensitivity predicting SSP security. In the final
step, a multiple regression analysis was conducted with AAI security and maternal
sensitivity predicting SSP security. If full mediation is present, once the mediator –
maternal sensitivity –is entered, the direct path between AAI security and SSP security
will no longer be significant. Table 3 shows the significance of the coefficients for each
step. In the final step when both AAI security and maternal sensitivity (mediator) were
entered, the effect of AAI security remained significant, indicating that full mediation
was not supported. In this model, the effect of maternal sensitivity also remained
significant after controlling for AAI security (thus both AAI security and maternal
sensitivity significantly predicted SSP security), supporting partial mediation. The direct
effect between AAI security and SSP security β¼:41Þwas stronger than the mediated
path β¼:31Þ. Specifically, the effect of AAI security on SSP security mediated by
maternal sensitivity was .10 (i.e., .31 × .33).
Structural equation modeling
Using Mplus statistical software version 7.11 (Muthen & Muthen, 1998–2016), we con-
ducted a mediational model in which AAI security was the categorical predictor, mater-
nal sensitivity was the continuous mediator, and SSP security was the categorical
outcome, which resulted in a probit scale. As shown in Figure 1, results demonstrated
significant associations between AAI security and maternal sensitivity, between maternal
sensitivity and SSP security, and between AAI security and SSP security.
To examine the indirect effect from AAI security through maternal sensitivity to SSP
security, we used bootstrap estimations. Bootstrapping is a non-parametric method
based on repeated, random resampling with replacement (Preacher & Hayes, 2004).
Because the bootstrap method does not violate normality assumptions, it is
Table 3. Summary of linear multiple regression: four-step mediation analyses.
BSEβtP
Step 1 (SSP as DV)
(Constant) −1.625 .620 −2.619 .011
AAI security 2.675 .743 .410** 3.601 .001
Step 2 (MBQS as DV)
(Constant) .608 .043 14.300 .000
AAI security .130 .051 .305* 2.558 .013
Step 3 (SSP as DV)
(Constant) 4.256 1.253 −3.396 .001
Maternal sensitivity 6.431 1.726 .422** 3.726 .000
Step 4 (SSP as DV)
(Constant) −4.659 1.203 −3.873 .000
AAI security 2.025 .739 .311** 2.741 .008
Maternal sensitivity 4.989 1.726 .328** 2.891 .005
** Model is significant at the 0.01 level. * Model is significant at the 0.05 level. AAI security = dichotomous, SSP
security = continuous, Maternal sensitivity = continuous.
ATTACHMENT & HUMAN DEVELOPMENT 9
Downloaded by [76.69.153.237] at 14:42 08 April 2016
recommended for small sample sizes such as ours. Following Hayes (2009) recommen-
dation of at least 5000 repetitions, we conducted 20,000 bootstraps repetitions, in which
the 95% confidence interval corresponds to the p< .05 alpha level. The unstandardized
indirect effect was .256, and the 95% confidence interval of the indirect effect did not
contain zero (range from .002–.720), thus providing evidence for the mediation role of
sensitivity. These findings from two mediation models seem to indicate that the trans-
mission gap appears somewhat different depending on the statistical method used to
measure mediation.
Additional post hoc investigation
To better understand intergenerational mismatched cases between mothers’AAIs and
their children’s SSPs (see Table 2), post hoc considerations of mothers’experiences of
divorce or separation in their home of origin led us to examine the relevant AAI
transcripts (i.e., Dismissing AAI cases). We specifically focused on the first AAI questions
in which the speaker’s early family constellation is asked about and separation experi-
ences or the absence of a parental figure become evident.
For Ds speakers whose children were judged B, five out of six mothers experienced
their parents’divorce in early years. For Ds mothers whose children were judged C, one
experienced parental divorce and two lost their fathers in early years. Taken together,
eight out of 12 (67%) Ds mothers with mismatched infants experienced permanent
separations from their fathers due to divorce or death. All of these mothers stayed with
their own mothers following the separation in childhood, and had been judged Ds due
to high idealization scores regarding descriptions of their relationship with their
mothers.
Discussion
The present study revisited the question of the intergenerational transmission of attach-
ment by examining the mediating role of maternal sensitivity as measured by the MBQS
in associations between adult attachment security as measured by the AAI and infant
b = .131(SE = .052)* b = 1.952 (SE = .819)*
r = .31 r = .42
r = .41
b = 1.049 (SE = .373)**
Maternal
Sensitivity
AAI Security SSP Security
Figure 1. Path diagram for the association between AAI security and SSP security mediated by
maternal sensitivity. The unstandardized indirect effect from AAI security through maternal sensi-
tivity to SSP security was .256, and the 95% confidence interval of the indirect effect estimated by
the bootstrap method was from .002 to .720. b(SE) = unstandardized beta (standard error).
r= correlation coefficient. *** p< .001; ** p< .01; * p< .05.
10 K. Y. BEHRENS ET AL.
Downloaded by [76.69.153.237] at 14:42 08 April 2016
attachment security as measured with the SSP. The concept of a transmission gap
introduced by van IJzendoorn (1995) continues to remain in need of further empirical
scrutiny and testing, in part due to questions concerning the robustness of AAI-SSP
associations. Pederson et al. (1998) study was one of the first studies to specifically
investigate this question by concurrently examining AAI security, SSP security, and
maternal sensitivity. Following their prior success in showing stronger associations
between maternal sensitivity and SSP security using the Q-score method of sensitivity
measurement, Pederson et al. (1998) utilized the Q-sort based sensitivity measure
(MBQS) rather than a more traditional contingency-based sensitivity measure, reporting
that maternal sensitivity partially mediated the AAI-SSP link. Pederson et al. (1998) noted
that van IJzendoorn (1995) found, rather surprisingly, only three studies (with incon-
sistent results), examining these three attachment constructs to include in his meta-
analysis. Moreover, among the 95 samples Verhage et al. (2016) included in their recent
meta-analysis, again only three recent studies employed the same three measures (AAI,
SSP, MBQS) included in both Pederson et al. (1998) study and the current study.
Additionally, these three studies (Bailey, Moran, Pederson, & Bento, 2007; Chin, 2013;
Tarabulsy et al., 2005) all sampled populations that are considered at-risk (e.g., teenage
mothers, premature babies, adolescent mothers, respectively). As such, the current study
represents a much needed replication of Pederson et al. (1998) study.
Our results show that, as expected, significant categorical matches between the AAI
and the SSP were found, thus supporting hypothesis 1. Consistent with Pederson et al.
(1998)findings, mean differences of MBQS scores between AAI categorical groups were
also statistically significant, as expected: mothers with secure states of mind were
significantly more sensitive than were mothers with insecure states of mind, supporting
hypothesis 2. Finally, mediation analyses revealed that maternal sensitivity did not fully
mediate the transmission of attachment security between the AAI and the SSP with an
indirect effect of .10 (i.e., .31 x .33) using the Baron and Kenny test, highly similar to what
Pederson et al. (1998) reported. The SEM analysis with bootstrapping further confirmed
that there is an undeniable indirect effect of maternal sensitivity mediating the transmis-
sion of maternal attachment to child attachment; hypothesis 3 was thus supported.
Hence, the present study largely replicated Pederson et al. (1998)findings despite some
methodological differences (e.g., DV as a continuous variable using the Attachment
Q-set (AQS: Waters, 1995) score vs. DV as a binary variable, using the SSP category,
original MBQS vs. 72-item MBQS for maternal sensitivity). The three core attachment
constructs studied were significantly related to each other, but maternal sensitivity only
partially mediated the relationship between maternal attachment representations and
infant attachment security. This means AAI security still independently predicts SSP
security, and other factors besides sensitivity play a part in transmitting attachment
security.
In making sense of this partial mediation, it is useful to consider whether maternal
sensitivity is best operationalized as discrete maternal behaviors or is better empirically
construed as a relational construct. This question was recently posed by Pederson and
his colleagues (2014), in part to explain the lack of a stronger mediation effect of
maternal sensitivity. In our study, AAI Coherence (dimensional), rather than AAI security
(categorical), was not significantly correlated with maternal sensitivity, (p= .08). AAI
Coherence, however, may not represent exactly the same thing as AAI security does in
ATTACHMENT & HUMAN DEVELOPMENT 11
Downloaded by [76.69.153.237] at 14:42 08 April 2016
the transmission process. In particular, AAI Coherence does not uniquely discriminate U
status which may separately affect sensitivity, sometimes in otherwise secure parent. The
benefit of reporting attachment data with a categorical approach versus a dimensional
approach has been rigorously debated (e.g., Fraley & Spieker, 2003; Roisman, 2014).
Bernier, Matte-Gagne, Belanger, and Whipple (2014) recently showed that MBQS sensi-
tivity, together with another mediator (maternal autonomy support), fully mediated the
transmission of AAI coherence to child’s attachment security (assessed by AQS) when all
variables were entered as continuous. Perhaps systematically incorporating both
approaches in mediation analyses may yield additionally informative results, as was
also recommended by Verhage et al. (2016).
It is to be noted that the current sample may be somewhat unique, although it is
non-clinical and considered normative based on the attachment categorical distribution:
a majority of mothers (70%) and children (64%) were judged secure (Table 1). However,
the distribution within SSP insecure categories deviates from the global distribution (van
IJzendoorn & Kroonenberg, 1988) in that this sample had equal numbers of children in A
and C categories for the three-way distribution as well as A, C, and D categories for the
four-way distribution, whereas the global norm predicts more A than C children in most
cultural contexts.
Furthermore, Table 2 shows that Dismissing (Ds) mothers in this sample predicted
more babies in the B and C categories than in the expected A category. Post hoc
considerations of mothers’divorce or separation experiences as children led us to
examine each Ds transcript for evidence of such events. As presented earlier (see
Results), a majority of Ds mothers (67%) who had mismatched SSP cases indeed
experienced early separation from their fathers. It is possible that those who experi-
enced parental divorce (or loss of father) in early childhood and stayed with their mother
throughout that difficult time may understandably be grateful to their mothers, perhaps
a bit excessively, thus scoring high on idealization, without necessarily subsequently
engaging in the rejecting or rebuffing parenting behaviors that are believed to predict
avoidance in offspring. In fact, the mean MBQS score of these cross-over Ds cases was
.72, which is nearly the same as the mean MBQS score of .74 among the secure mothers
(see Results). Such discrepancies or counter-intuitive findings could partly contribute to
explaining the transmission gap.
Murphy et al. (2014) investigated the association between AAI states of mind and
Adverse Childhood Experiences (ACEs), comparing clinical and community samples. As
expected, a majority of the clinical sample (84%) reported significantly more exposures
to ACEs and had significantly more U/CC classifications (76%) than the community
sample (27%, 9%, respectively). Importantly, among the ACE categories of adversity,
Parental separation or divorce was experienced by nearly 90% of the clinical sample.
Steele et al. (2016) further demonstrated how ACEs significantly affected parental
distress, indicating that adverse childhood experiences (including divorce and separa-
tion) are likely to influence parenting practices, even though our particular sample
showed more of an impact on states of mind. Thus, future AAI studies may consider
employing the ACE questionnaire or simply creating a checklist to systematically record
early separation experiences due to parental divorce as a separate variable or covariate
to be considered for analyses.
12 K. Y. BEHRENS ET AL.
Downloaded by [76.69.153.237] at 14:42 08 April 2016
Of note, even though MBQS scores differed significantly between the AAI secure and
insecure groups, the current sample no longer showed significant differences between
the groups when Unresolved/Secure (U/F) mothers were included in the secure group.
This is not surprising given that mothers with U status have been shown to be more
likely to display frightened/ frightening (FR: Hesse & Main, 2000,2006; Main & Hesse,
1992–2006) behaviors, possibly hindering them from engaging in sensitive interactions
with their children. Indeed, somewhat ironically in our view, the pathway from U to FR,
subsequently predicting D (see Hesse & Main, 2006) can perhaps be considered the best
demonstration of attachment transmission both empirically and theoretically. It is still
not clear, however, what exactly caregivers do, rather than what they do not do (FR
behaviors), that will likely promote secure attachment.
Tarabulsy et al. (2005) investigated ecological contributions such as maternal educa-
tion, depression, and family support, and found that maternal sensitivity was a signifi-
cant mediator of infant security when ecological variables were controlled for, whereas
maternal attachment representation no longer predicted infant security. Given that we
have also discussed a potential effect of early parental divorce on the link between
maternal attachment representation and maternal sensitivity, various ecological factors
need systematic investigation as discussed above. Furthermore, continuing neurophy-
siological investigations of parental behaviors such as those presented earlier may shed
light on understanding how mothers’individual differences in the way they perceive
their child or interpret their child’s cues at the neurological level may guide their
interactive behaviors. Once we control all those factors with a large dataset such as
the one Verhage et al. (2016) synthesized for their meta-analysis, it may be possible to
begin to shed further light on “transmission gap.”Verhage et al. (2016) argue that it is
important to understand both the conditions of when transmission occurs and when it
does not, studying varied transmission levels. The present study’sfindings appear to
reflect this sentiment –the similarities of our sample characteristic with those of
Pederson et al. (1998) may have yielded the similar transmission effect sizes, while
additional mediation tests confirmed that maternal sensitivity mediates the AAP-SSP
link but nonetheless a transmission gap still remained.
This study has a number of limitations. First, the sample is small (N= 66) and is
largely homogenous, which compromises generalizability to diverse groups. The
current study utilized the MBQS during the SSP, observing maternal behaviors and
children’s behaviors concurrently. Although coding of each observation was done
independently by different coders who were not trained in any attachment assess-
ment measures, lack of independence in the context itself –observing behavior of
both child and mother in the same context –may present potential problems. The
range of maternal interactive behaviors was also quite limited based on approxi-
mately 13 minutes of observation compared to a much longer home observations.
However, successful employment of the MBQS in brief laboratory observations of
maternal behaviors (12-minutes Atkinson et al., 2005, 10-minutes; Tarabulsy et al.,
2009) has been reported. In particular, Tarabulsy et al. validated their shorter version
(25-item) MBQS with home observation. Likewise, Behrens, Parker, et al. (2011)
reported consistency between sensitivity measured with the 72-item version MBQS
during the SSP and another contingency-based sensitivity measure from a separate
laboratory context for the same sample. Behrens et al. (2012) also reported stability
ATTACHMENT & HUMAN DEVELOPMENT 13
Downloaded by [76.69.153.237] at 14:42 08 April 2016
when the sample was followed up 30 months later during home observations. The
validity of the 72-item version of the MBQS would be stronger, however, with
additional concurrent home observations. The current study also focused exclusively
on infant–mother attachment. Further investigation of father–child attachment, or
the effect of paternal sensitivity on the child’s attachment, is urgently needed.
Despite these limitations, however, the current study contributes to the attachment
literature by reporting findings that: (1) are consistent with previous reports utilizing a
traditional categorical approach to attachment classifications, showing a strong match
between the AAI and the SSP; (2) replicate Pederson et al. (1998)findings with respect to
maternal sensitivity scores being significantly higher for the AAI secure group than the
AAI insecure group; and (3) replicate Pederson et al. (1998)finding regarding maternal
sensitivity partially mediating AAI security to SSP security.
Overall, then, the current study’sfindings were largely consistent with both earlier and
recent mediation studies that confirmed the role of maternal sensitivity in mediating the
pathway between adult attachment security and child attachment security. We now return
to our earlier questions. What does closing the gap truly mean? Is it in fact necessary to close
the gap? Do we stop our investigation once the gap is closed empirically? Bernier et al.
(2014) showed full mediation of maternal sensitivity when autonomy support was also
entered as a mediation variable and discussed the importance of a multidimensional
approach to better understand the transmission process, yet did not necessarily claim to
have “closed the gap.”We concur with their suggestion that it is more important to focus on
what constitutes the gap and simultaneously investigate multiple factors that may affect the
transmission process rather than focusing on how to close the gap. In fact, as our results
from two mediation analyses show, the transmission gap appears somewhat different
depending on the statistical method used to measure mediation. In conclusion, researchers
should continue to explore why parents engage in certain behaviors when interacting with
their children above and beyond how their states of mind affect them. Additionally, we
believe that small sample sizes should not automatically be dismissed as having less value.
Instead, studies such as the current one may enable contextual investigations that present
some benefits towards enriching our understanding of the transmission process of security
from mother to infant.
Note
1. KB = Kazuko Y Behrens (all SSP cases and all AAI cases were randomized and re-numbered
prior to coding onset to maintain blindness). JH = John D. Haltigan. EC = Elizabeth Carlson.
NB = Naomi I. G. Bahm, KM = Kirsten B. Mathews.
2. For our sample, the AAI Coherence scales (both Coherence of Mind and Coherence of
Transcript) were not significantly correlated with the MBQS scale. Thus, AAI security as a
continuous scale was not used in the analyses.
Acknowledgements
We thank the children and their mothers who participated in the study. We also thank Tomo
Umemura who provided statistical support in completing part of the statistical analyses. We are
grateful to Elizabeth Carlson who advised us on several difficult SSP cases and also coded part of
our SSP data.
14 K. Y. BEHRENS ET AL.
Downloaded by [76.69.153.237] at 14:42 08 April 2016
Disclosure statement
No potential conflict of interest was reported by the authors.
References
Ainsworth, M.D.S., Blehar, M., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological
study of the strange situation. Hillsdale, NJ: Erlbaum.
Atkinson, L., Goldberg, S., Raval, V., Pederson, D., Benoit, D., Moran, G., . . . Leung, E. (2005). On the
relation between maternal state of mind and sensitivity in the prediction of infant attachment
security. Developmental Psychology,41,42–53. doi:10.1037/0012-1649.41.1.42
Bailey, H.N., Moran, G., Pederson, D.R., & Bento, S. (2007). Understanding the transmission of
attachment using variable- and relationship-centered approaches. Development and
Psychopathology,19, 313–343. doi:10.1017/S0954579407070162
Bakermans-Kranenburg, M.J., & van IJzendoorn, M.H. (1993). A psychometric study of the Adult
Attachment Interview: Reliability and discriminant validity. Developmental Psychology,29, 870–
879. doi:10.1037/0012-1649.29.5.870
Baron, R.M., & Kenny, D.A. (1986). The moderator–mediator variable distinction in social psycho-
logical research: Conceptual, strategic, and statistical considerations. Journal of Personality and
Social Psychology,51, 1173–1182. doi:10.1037/0022-3514.51.6.1173
Behrens, K.Y., Hart, S.L., & Parker, A.C. (2012). Maternal sensitivity: Evidence of stability across time,
contexts, and measurement instruments. Infant and Child Development,21, 348–355.
doi:10.1002/icd.1747
Behrens, K.Y., Li, Y., Gribneau Bahm, N.I., & O’Boyle, M.W. (2011). EEG responses to photos: A case
study of three major AAI classifications. Psychological Reports,108, 993–1010.
doi:10.2466.02.07.09.09.21.PRO.108.3.993-1010
Behrens, K.Y., Parker, A.C., & Haltigan, J.D. (2011). Maternal sensitivity assessed during the Strange
Situation Procedure predicts child’s attachment quality and reunion behaviors. Infant Behavior
and Development,34, 378–381. doi:10.1016/j.infbeh.2011.02.007
Behrens, K.Y., Parker, A.C., & Kulkofsky, S. (2014). Stability of maternal sensitivity across time and
contexts with Q-sort measures. Infant and Child Development,23,532–541. doi:10.1002/icd.1835
Bernier, A., Matte-Gagne, C., Belanger, M.-E., & Whipple, N. (2014). Taking stock of two decades of
attachment transmission gap: Broadening the assessment of maternal behavior. Child
Development,85, 1852–1865. doi:10.111/cdev.12236
Chin, F. (2013). Cognitive and socio-emotional developmental competence in premature infants at 12
and 24 months: Predictors and developmental sequelae. State College, PA: The Pennsylvania State
University.
Crowell, J.A., Waters, E., Treboux, D., O’Connor, E., Colon-Downs, C., Feider, O., . . . Posada, G. (1996).
Discriminant validity of the Adult Attachment Interview. Child Development,67, 2584–2599.
doi:10.2307/1131642
De Wolff, M., & van IJzendoorn, M. (1997). Sensitivity and attachment: A meta-analysis on parental
antecedents of infant attachment. Child Development,68, 571–591. doi:10.1111/j.1467-
8624.1997.tb04218.x
Fraley, R.C., & Spieker, S.J. (2003). Are infant attachment patterns continuously or categorically
distributed? A taxometric analysis of strange situation behavior. Developmental Psychology,39,
423–429. doi:10.1037/0012-1649.39.3.387
George, C., Kaplan, N., & Main, M. (1996). Adult Attachment Interview (Unpublished protocols).
University of California, Berkeley.
Hart, S.L., & Behrens, K.Y. (2013). Affective and behavioral features of jealousy protest: Associations
with child temperament, maternal interaction style, and attachment. Infancy,18, 369–399.
doi:10.111/j.1532-7078.2012.00123.x
Hayes, A.F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium.
Communication Monographs,76, 408–420. doi:10.1080/03637750903310360
ATTACHMENT & HUMAN DEVELOPMENT 15
Downloaded by [76.69.153.237] at 14:42 08 April 2016
Hesse, E., & Main, M. (2000). Disorganization in infant and adult attachment: Description, corre-
lates, and implications for developmental psychopathology. Journal of the American
Psychoanalytic Association,48, 1097–1127. doi:10.1177/00030651000480041101
Hesse, E., & Main, M. (2006). Frightened, threatening, and dissociative parental behavior in low-risk
samples: Description, discussion, and interpretations. Development and Psychopathology,18,
309–343. doi:10.1017/S0954579406060172
Hollingshead, A.B. (1971). Commentary on the indiscriminate state of social class measurement.
Social Forces,49, 563–567. doi:10.1093/sf/49.4.563
Lenzi, D., Trentini, C., Pantano, P., Macaluso, E., Lenzi, G.L., & Ammaniti, M. (2013). Attachment
models affect brain responses in areas related to emotions and empathy in nulliparous women.
Human Brain Mapping,34, 1399–1414. doi:10.1002/hbm.21520
Main, M., Goldwyn, R., & Hesse, E. (2003). Adult attachment scoring and classification system
(Unpublished manuscripts). University of California, Berkeley.
Main, M., & Hesse, E. (1992–2006). Frightened, threatening, dissociative, timid-deferential, sexualized,
and disorganized parental behavior: A coding system for frightened/frightening (FR) parent-infant
interactions (Unpublished manuscript). University of California, Berkeley.
Murphy, A., Steele, M., Dube, S.R., Bate, J., Bonuck, K., Meissner, P., . . . Steele, H. (2014). Adverse
Childhood Experiences (ACEs) questionnaire and Adult Attachment Interview (AAI): Implications
for parent child relationships. Child Abuse & Neglect,38, 224–233. doi:10.1016/j.
chiabu.2013.09.004
Muthen, L.K., & Muthen, B.O. (1998–2016). Mplus user’s guide (6th ed.). Los Angeles, CA: Muthen &
Muthen.
Pederson, D.R., Bailey, H.N., Tarabulsy, G.M., Bento, S., & Moran, G. (2014). Understanding sensitiv-
ity: Lessons learned from the legacy of Mary Ainsworth. Attachment & Human Development,16,
261–270. doi:10.1080/14616734.2014.900094
Pederson, D.R., Gleason, K.E., Moran, G., & Bento, S. (1998). Maternal attachment representations,
maternal sensitivity, and the infant-mother attachment relationship. Developmental Psychology,
34,925–933. doi:10.1037/0012-1649.34.5.925
Pederson, D.R., & Moran, G. (1995). A categorical description of attachment relationships in the
mode and its relation to Q-sort measures of infant-mother interaction. In E. Waters, B. Vaughn,
G. Posada, & K. Kondo-Ikemura (Eds.), Caregiving, cultural and cognitive perspectives on secure-
base behavior and working models: New growing points of attachment theory and research.
Monographs of the Society for Research in Child Development, 60 (2–3, Serial No. 244) (pp.
247–254).
Pederson, D.R., & Moran, G. (1996). Expressions of the attachment relationship outside of the
strange situation. Child Development,67, 915–927.
Pederson, D.R., Moran, G., & Bento, S. (2004). Preschool maternal sensitivity Q-set adapted from
Elizabeth Brown (Unpublished Manuscript). Department of Psychology, University of Western
Ontario, London, ON.
Preacher, K.J., & Hayes, A.F. (2004). SPSS and SAS procedures for estimating indirect effects in
simple mediation models. Behavior Research Methods, Instruments, & Computers,36, 717–731.
doi:10.3758/BF03206553
Richters, J.E., Waters, E., & Vaughn, B.E. (1988). Empirical classification of infant-mother relation-
ships from interactive behavior and crying during reunion. Child Development,59, 512–522.
doi:10.2307/1130329
Riem, M.M.E., Bakermans-Kranenburg, M.J., van IJzendoorn, M.H., Out, D., & Rombouts, S.A. (2012).
Attachment in the brain: Adult attachment representations predict amygdala and behavioral
responses to infant crying. Attachment & Human Development,14, 533–551. doi:10.1080/
14616734.2012.727252
Roisman, G.I. (2014). General discussion. In C. Booth-LaForce & G. Roisman (Eds.), The Adult
Attachment Interview: Psychometrics, stability and change from infancy, and developmental
origins. Monographs of the Society for Research in Child Development, 79(3, Serial No.314)
(pp. 126–137).
16 K. Y. BEHRENS ET AL.
Downloaded by [76.69.153.237] at 14:42 08 April 2016
Schooler, J. (2011). Unpublished results hide the decline effect. Nature,470, 437. doi:10.1038/
470437a
Steele, H., Bate, J., Steel, M., Dube, S.R., Danskin, J., Knafo, H., . . . Murphy, A. (2016). Adverse
childhood experiences, poverty, and parenting stress. Canadian Journal of Behavioural Science,
48,32–38. doi:10.1037/cbs0000034
Swain, J.E. (2011). The human parental brain: In vivo neuroimaging. Progress in Neuro-
Psychopharmacology & Biological Psychiatry,35, 1242–1254. doi:10.1016/j.pnpbp.2010.10.017
Tarabulsy, G.M., Bernier, A., Provost, M.A., Maranda, J., Larose, S., Moss, E., . . . Tessier, R. (2005).
Another look inside the gap: Ecological contributions to the transmission of attachment in a
sample of adolescent mother-infant dyads. Developmental Psychology,41, 212–224. doi:10.1037/
0012-1649.41.1.212
Tarabulsy, G.M., Provost, M.A., Bordeleau, S., Trudel-Fitzgerald, C., Moran, G., Pederson, D.R., . . .
Pierce, T. (2009). Validation of a short version of the maternal behavior Q-set applied to a brief
video record of mother-infant interaction. Infant Behavior and Development,32, 132–136.
doi:10.1016/j.infbeh.2008.09.006
van IJzendoorn, M.H. (1995). Adult attachment representations, parental responsiveness, and
infant attachment: A meta-analysis on the predictive validity of the Adult Attachment
Interview. Psychological Bulletin,117, 387–403. doi:10.1037/0033-2909.117.3.387
van IJzendoorn, M.H., & Kroonenberg, P.M. (1988). Cross-cultural patterns of attachment: A meta-
analysis of the strange situation. Child Development,59, 147–156. doi:10.2307/1130396
van IJzendoorn, M.H., & Kroonenberg, P.M. (1990). Cross-cultural consistency of coding the strange
situation. Infant Behavior and Development,13, 469–485. doi:10.1016/0163-6383(90)90017-3
Verhage, M.L., Schuengel, C., Madigan, S., Fearon, R.M.P., Oosterman, M., Cassibba, R., . . . van
IJzendoorn, M.H. (2016). Narrowing the transmission gap: A synthesis of three decades of
research on intergenerational transmission of attachment. Psychological Bulletin,142, 337–366.
doi:10.1037/bul0000038
Waters, E. (1995). The attachment Q-set, version 3.0. Monographs of the Society for Research in Child
Development,60, 234–246. doi:10.2307/1166181
ATTACHMENT & HUMAN DEVELOPMENT 17
Downloaded by [76.69.153.237] at 14:42 08 April 2016
