Early Child Development and Care

ISSN: 0300-4430 (Print) 1476-8275 (Online) Journal homepage: https://www.tandfonline.com/loi/gecd20

The development of father–child attachment:

associations between adult attachment
representations, recollections of childhood
experiences and caregiving

Laura McFarland-Piazza , Nancy Hazen , Deborah Jacobvitz & Erin Boyd-

Soisson

To cite this article: Laura McFarland-Piazza , Nancy Hazen , Deborah Jacobvitz & Erin Boyd-
Soisson (2012) The development of father–child attachment: associations between adult
attachment representations, recollections of childhood experiences and caregiving, Early Child
Development and Care, 182:6, 701-721, DOI: 10.1080/03004430.2011.573071

To link to this article: https://doi.org/10.1080/03004430.2011.573071

Published online: 13 Jun 2011.

Submit your article to this journal

Article views: 1735

View related articles

Citing articles: 10 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=gecd20
Early Child Development and Care
Vol. 182, No. 6, June 2012, 701 –721

The development of father –child attachment: associations

between adult attachment representations, recollections of
childhood experiences and caregiving
Laura McFarland-Piazzaa∗ , Nancy Hazenb, Deborah Jacobvitzb and
Erin Boyd-Soissonc
a
Murray School of Teacher Education, Charles Sturt University, PO Box 789, Albury,
NSW 2640, Australia; bSchool of Human Ecology, The University of Texas at Austin,
1 University Station, Austin, Texas 78712, USA; cDepartment of Human Development and
Family Sciences, Messiah College, PO Box 3047, Grantham PA, 17027, USA
(Received 13 December 2010; final version received 15 March 2011)

The association between fathers’ adult attachment representations and their

recollections of childhood experiences with their caregiving quality with their
eight-month-old infants and with father – infant attachment classification was
examined in a longitudinal study of 117 fathers and their infants. Sensitive
caregiving was related to secure-autonomous classification in the Adult
Attachment Interview (AAI), hostile caregiving was related to fathers’ dismissing
and unresolved attachment, and emotional disengagement and role-reversed
caregiving were both related to fathers’ unresolved attachment. Childhood
experiences of parental pressure to achieve were related to fathers’ hostile and
role-reversed caregiving and low sensitivity, independent of AAI classification.
However, fathers’ childhood experiences of maternal neglect were related to
high-quality caregiving. It was also found that fathers’ secure-autonomous AAI
classification was related to secure father – child attachment in the Strange
Situation Paradigm, and this relation was mediated by sensitive caregiving.
Keywords: fathers; attachment; infants; caregiving; adult attachment

According to attachment theory, adults’ current mental representations of attachment

are stronger predictors of their parenting than their actual childhood experiences of
being parented (Main, Kaplan, & Cassidy, 1985). However, studies on this topic
have focused almost exclusively on mothers. Recent research has found that mothers
and fathers may have very different roles to play in caring for their children (Grossmann
et al., 2002; Paquette, 2004). Specifically, fathers are less likely than mothers to serve as
primary caregivers who comfort the child in times of distress and more likely to provide
challenging play opportunities. Thus, it may not be accurate to assume that father–child
attachment and caregiving functions the same as for mothers. Therefore, the goal of this
study was to examine fathers’ attachment representations and recollections of caregiv-
ing experiences in childhood as predictors of their caregiving behaviour and quality of

∗
Corresponding author. Email: lpiazza@csu.edu.au

ISSN 0300-4430 print/ISSN 1476-8275 online

# 2012 Taylor & Francis
http://dx.doi.org/10.1080/03004430.2011.573071
http://www.tandfonline.com
702 L. McFarland-Piazza et al.

infant–father attachment in an effort to learn more about the unique workings of the
father–child attachment system.
Attachment theory suggests that infants who are exposed to consistently sensitive
caregiving during the first year of life will come to expect that their caregiver will be
responsive to their needs and will develop a sense of security and trust with the care-
giver (Bowlby, 1969). In contrast, infants who are exposed to inconsistent or insensitive
care will develop a sense of mistrust. It is theorised that the development of the parent–
infant attachment is intergenerational in nature, influenced by the adults’ internal
working model of attachment and caregiving.
Individual differences in adults’ internal working models of attachment have been
assessed using the Adult Attachment Interview (AAI) (George, Kaplan, & Main, 1985/
1996; Main, Goldwyn, & Hesse, 1984–2002). Based on the coherence with which they
describe their early attachment-related experiences with their parents, adults are
assigned to one of three major attachment classifications: secure-autonomous, dismiss-
ing, or preoccupied. Secure-autonomous adults appear to value early attachment and
provide coherent and objective descriptions about their attachment figures and experi-
ences. Dismissing adults often claim that they cannot remember much about their child-
hood and seem to devalue attachment. They may idealise their parents as being perfect,
but are not able to provide specific examples of their parents as being loving or suppor-
tive. Preoccupied adults speak in an incoherent manner about their early experiences
and often seem confused and/or presently angry about their past experiences with care-
givers. Both the dismissing and preoccupied categories are considered to be insecure
forms of adult attachment, and insecure adults are believed to be at risk for repeating
negative patterns of parenting. Adults may also be classified as unresolved for
trauma, in conjunction with their best-fitting major classification, if they show signs
of mental disorganisation when discussing childhood experiences of abuse or the
death of a loved one (Main et al., 1984–2002).
Research has found that parents’ secure-autonomous working models of attachment
predict secure parent–child attachment, as assessed by the Strange Situation Paradigm
(SSP) (Hesse, 2008). However, one meta-analysis examining the intergenerational
transmission of attachment found that this link is much stronger in mothers than in
fathers (van IJzendoorn, 1995). van IJzendoorn (1995) found that the effect size for
the relationship between maternal AAI classification and mother–child attachment
for 656 mother–child dyads (r ¼ 0.50) was significantly different from the effect
size for the relationships between paternal AAI classification and father–child attach-
ment in 198 father–child dyads (r ¼ 0.35).
Attachment theory suggests that the link between adults’ internal working models of
attachment and parent–child attachment is mediated by caregiving quality. Parents’ rep-
resentations of attachment are theorised to influence their responsiveness to their
infants’ attachment signals (Main et al., 1985). A meta-analysis of 10 studies found
that secure mothers were more positive, sensitive, and responsive to their infants com-
pared with insecure mothers (van IJzendoorn, 1995). This meta-analysis also examined
three studies that included fathers, and the relation between adult AAI and caregiving
was even stronger when these studies were included. Although fewer studies have
investigated the association between fathers’ representations of attachment and their
caregiving, these studies also indicate that secure-autonomous fathers are more sensitive
than insecure fathers (Cohn, Cowan, Cowan, & Pearson, 1992; Grossman et al., 2002;
van IJzendoorn, 1995). In fact, the meta-analysis by van IJzendoorn (1995) indicated
that the relation between parents’ working models of caregiving and their sensitive
 Early Child Development and Care 703

responsiveness to their infants actually tended to be stronger for samples of fathers than
for samples of mothers.
Research has found that sensitive caregiving mediates the relationship between
parental secure-autonomous internal working model of attachment and secure
parent–infant attachment for both mothers and fathers (van IJzendoorn & De Wolff,
1997). However, the relation between sensitive caregiving and secure parent–child
attachment appears to be much stronger for mothers than for fathers (De Wolff &
van IJzendoorn, 1997; Pederson, Gleason, Moran, & Bento, 1998; Ward & Carlson,
1995). Two studies have found a relation between father sensitivity and secure
father–child attachment (Cox, Owen, Henderson, & Margand, 1992; Goossens &
van IJzendoorn, 1990). However, other studies have not replicated this finding
(Caldera, Huston, & O’Brien, 1995; Easterbrooks & Goldberg, 1984; Grossmann &
Grossmann, 1992; Schneider Rosen & Rothbaum, 1993; Volling & Belsky, 1992).
Thus, there appears to be a ‘transmission gap’ for fathers. The present study attempts
to understand more fully this transmission gap by examining the links between
fathers’ internal working models of attachment, their childhood experiences, specific
types of paternal caregiving, and father–child attachment.
Although they are assumed to be less sensitive than secure-autonomous parents,
dismissing, preoccupied, and unresolved parents are thought to differ in their particular
styles of insensitive caregiving (van IJzendoorn, 1995). However, although studies
have found that dismissing parents are more likely to have avoidant infants, preoccu-
pied parents are more likely to have resistant infants, and unresolved parents are
more likely to have disorganised infants (Main et al., 1985; Main, 1995), there is a
dearth of studies that have examined the relation of particular types of insecure
working models of attachment to particular styles of insensitive caregiving. Dismissing
parents are theorised to dismiss or devalue their infants’ attachment needs, resulting in
rejecting or emotionally distancing behaviour, and thus leading to a greater likelihood
of forming insecure-avoidant attachment relationships with their infants. Although the
caregiving styles of dismissing parents per se have not been examined, mothers of
infants with avoidant attachments have been found to be rejecting and psychologically
unavailable in their caregiving (Belsky, Rovine, & Taylor, 1984; Egeland & Farber,
1984) and to avoid physical contact with their infants (Ainsworth, Blehar, Waters, &
Wall, 1978). In addition, studies examining mothers’ representations of cargiving indi-
cate that mothers of avoidant children are more rejecting and distancing (George &
Solomon, 1996). Thus, we expected that dismissing fathers may be more likely than
other fathers to show hostile rejection or emotionally disengaged caregiving patterns.
It is also possible that the link between fathers’ dismissing attachment and father–
infant avoidant attachment is mediated by hostile or emotionally disengaged
caregiving.
Preoccupied parents are theorised to be currently enmeshed with their own past
attachment experiences, and thus not consistently responsive to their infants’ attach-
ment needs, leading to resistant attachment. Again, the specific caregiving styles of pre-
occupied parents per se have not been examined, but mothers of children with resistant
attachments have been found to be inconsistent in their responsiveness or under-
involved and emotionally disengaged (Belsky et al., 1984; Egeland & Farber, 1984;
Isabella, Belsky, & von Eye, 1989; Smith & Pederson, 1988). In addition, in studies
assessing their representations of themselves as caregivers, mothers of resistant children
describe role-reversed relationship patterns, often indicating that they look to their child
to meet their own attachment needs (George & Solomon, 1996). Thus, we expected
704 L. McFarland-Piazza et al.

preoccupied fathers to show greater role-reversed or disengaged patterns with their

infants. It may also be the case that the link between fathers’ preoccupied attachment
and father–infant resistant attachment is mediated by role-reversed or emotionally
disengaged caregiving.
Finally, unresolved parents are theorised to experience fear as a result of past loss or
trauma which manifests itself in odd, unpredictable, frightening caregiving behaviours,
dissociated behaviours (e.g. trance-like staring) and/or behaviours indicating that the
parent may be frightened (Jacobvitz, Leon, & Hazen, 2006; Main & Hesse, 1990).
Infant disorganisation is thought to result when the attachment figure is at the same
time the source of their comfort and fear. Several studies have found empirical
support for the link between mothers’ unresolved state of mind and maternal frighten-
ing behaviour (Abrams, Rifkin, & Hesse, 2006; Jacobvitz et al., 2006; Madigan,
Moran, & Pederson, 2006; Schuengel, Bakermans-Kranenburg, van Ijzendoorn, &
Blom, 1999). However, this link has not been replicated for fathers. In a recent study
using the same longitudinal data set as that used in the present study, we focused
specifically on examining the antecedents and consequences of fathers’ frightening be-
haviour (Hazen, McFarland, Jacobvitz, & Boyd-Soisson, 2010). Our results indicated
that fathers’ frightening behaviour was not related to their representations of attachment
or to having disorganised attachment relationships with their infants. Rates of frighten-
ing behaviour were significantly higher in fathers than in mothers, but frightening beha-
viours were just as likely to be displayed by secure-autonous fathers as by insecure and
unresolved fathers. Given that Abrams et al. (2006) found that only the dissociative
subcategory of frightening behaviours was related to parents’ unresolved AAI status
and to infants’ disorganised attachment, we also examined the relation of fathers’
AAI categories and infants’ disorganised attachment to each of the subcategories of
frightening behaviour, and still we found no differences. Interestingly, however,
secure fathers were more likely than insecure and unresolved fathers to combine
their frightening behaviours with sensitivity.
In what ways, then, might the caregiving of unresolved fathers differ from that of
other fathers? Research by Lyons-Ruth, Bronfman, and Parsons (1999) has found
that mothers of disorganised infants show a broader set of maternal behaviours
characterised not only by frightening behaviour, but also by disrupted affective
communication with the infant. Disrupted and inappropriate affective responses
include responding to infant distress with marked emotional withdrawal and unavail-
ability or with hostile reactions such as laughing at the distressed baby. In addition,
mothers of disorganised children describe themselves as helpless caregivers who
cannot control their children, and describe their children as being either wild and out
of control, or exceptionally precocious and adult-like, able to care for themselves
and even for the parent (George & Solomon, 1996). Given that disorganised children
are more likely than other children to have unresolved parents, it may be that unre-
solved fathers are more emotionally disengaged, role-reversing, and hostile than
other fathers. If this is true, it may be that the link between fathers’ unresolved attach-
ment and father–infant disorganised attachment is mediated by emotional disengage-
ment, role-reversal, or hostility, rather than by fathers’ frightening behaviour.
Adults can be classified as secure-autonomous regardless of the content of the
childhood memories, as long as these memories are presented in a coherent, open,
and objective fashion. Thus, it is possible that adults’ recollections of childhood experi-
ences might impact their caregiving independent of their own internal working model
of attachment. Attachment researchers have not addressed how memories of particular
 Early Child Development and Care 705

childhood experiences might be related to particular caregiving styles, although

Bakermans-Kranenburg and van IJzendoorn have recently suggested that researchers
make more use of the continuous scales that assess childhood experience in the AAI
(Bakermans-Kranenburg & van IJzendoorn, 2009). Recent research does indicate
that mothers who reported negative childhood experiences on the AAI yet have a cur-
rently secure working model of attachment according to their AAI classification, were
just as likely to have to provide sensitive caregiving to their infants as were mothers
who recalled positive childhood experiences, and were classified as secure on the
AAI (Pearson, Cohn, Cowan, & Cowan, 1994; Phelps, Belsky, & Crnic, 1997).
Thus, we might expect that fathers’ AAI classifications should be stronger predictors
of their caregiving and their attachment quality with their infants than their memories
of childhood experiences. However, it may be that fathers’ recollections of childhood
experience may predict aspects of their caregiving quality independent of their AAI
classification.
In sum, this study had four primary goals: (1) to examine the association between
fathers’ working models of attachment (secure, dismissing, preoccupied, and unre-
solved) and different aspects of their caregiving quality (sensitivity, hostility, emotional
disengagement, and early forms of role-reversed care); (2) to examine the relation of
fathers’ working models of attachment and their attachment security with their
infants; (3) to examine whether the relation between fathers’ working models of attach-
ment and their attachment security with their infants is mediated by caregiving quality;
and (4) to examine the extent to which fathers’ recollections of childhood experiences
and their AAI classifications make unique contributions to predicting their caregiving
quality with their infants.

Method
Participants
Participants were part of a longitudinal study investigating parenting and family
relationships. Data from 117 fathers, from the larger sample of 125 couples, were
used. Couples were recruited when the women were in their third trimester of preg-
nancy through birthing classes, public service radio announcements, newspaper press
releases, and flyers distributed at maternity stores in a metropolitan area and surround-
ing communities. Criteria for participation were that the couples were living together at
the time of recruitment, were having their first child, and spoke English as their primary
language. Each family received a $50 savings bond for their child at the completion of
each of the three phases of data collection, bimonthly newsletters with updates on the
project, a T-shirt for their child, and an audiotape of lullabies.
The median age of fathers was 30, ranging from 19 to 51. Most participants were
Caucasian (82%); the remainder were Hispanic (9%), African-American (3%), or
Native American, Middle Eastern, or Indian (6%). The median family income was
$30,000–45,000. The combined income of 80% of the couples was $30,000–45,000
per year and 13% of the couples earned more than $45,000 per year. Families
earning less than $15,000 per year comprised 7% of the sample. Participants were
well-educated; 69% had some college and 23% held an advanced degree. By the
time infants were eight months, 118 of the original 125 families remained in the
study, and by 12 –15-month SSP assessment, 112 families remained in the study.
Eight families had moved away, two could not be located, and three declined to partici-
pate for personal reasons. Families who remained did not differ significantly in age or
706 L. McFarland-Piazza et al.

income from the original group. Seven fathers did not have AAI data and seven fathers
did not have SSP data. In each case, five fathers did not come to the lab to participate,
and two more did not have usable data due to audio malfunction. Fathers who had
missing data did not differ significantly in their caregiving scores from the other fathers.

Procedure
The AAI was administered separately to fathers and mothers when the mothers were in
their third trimester of pregnancy. When the infants were eight months old, mother–
infant and father–infant interactions were videotaped in home visits during 30-minute
interactions in which the parent was instructed to play with the infant as they ordinarily
would, to change the infant’s clothes, and to feed the infant. No specific instructions were
given on how to play or what should be played with. Although restricted to one particular
room chosen by the families at a pre-scheduled time convenient to the families, it was left
up to the mothers and fathers as to how they chose to play with their infant. A researcher
who operated the video camera was present in the room during this 30-minute session.
When infants were 12 and 15 months old, they were assessed for security of attachment
with both mother and father using the Strange Situation Procedure (SSP). The order in
which mothers and fathers were observed in the SSP with their infant was counterba-
lanced such that half of the infants were observed in this procedure with their fathers
at 12 months of age and half at 15 months.

Measurement
Fathers’ attachment representations
Fathers’ representations of attachment were assessed using the AAI (George et al.,
1985/1996), a semi-structured interview designed to assess adults’ current states of
mind regarding attachment based on their recollections of their childhood relationships
with their parents. Participants were asked to describe and evaluate attachment-related
childhood experiences, including loss of attachment figures through death and threaten-
ing experiences such as abuse, and to discuss their current perceptions of the effects of
these experiences on their development and why their parents may have behaved as
they did. Several trained faculty members and graduate students conducted the AAI,
memorising the questions so they could use a relaxed conversational style. All inter-
views were completed by the third author, who is a certified AAI trainer, or by inter-
viewers trained by the third author to administer the AAI. Training involved first
being interviewed by the trainer, then interviewing the trainer and receiving feedback,
and finally interviewing two pilot subjects and receiving feedback from the trainer on
the audiotaped interviews. Interviews lasted from 60 to 90 minutes and included probes
to increase the specificity of responses and facilitate recall.
Interviews were audiotaped and transcribed verbatim prior to coding. First,
interviews were rated on two types of scales: childhood experience scales and state
of mind scales. The childhood experiences scales consisted of loving, rejecting, invol-
ving/role reversing, pressure to achieve, and neglecting when present. Scales range
from 1 (low or absent) to 9 (high). Each of these scales is rated separately for the child-
hood experiences with mother and childhood experiences with father. The loving scale
rates the extent to which individuals experienced each parent as loving and supportive
in childhood. The rejecting scale represents the extent to which each parent appears to
 Early Child Development and Care 707

have rejected or avoided the individual’s bids for comfort and attention in childhood.
The involving/role reversing scale assesses the extent to which each parent makes
himself or herself the object of the child’s attention or depends on the child’s presence
for the maintenance of his or her own sense of security. The neglecting when present
scale assesses the extent to which parents were inattentive or psychologically inaccess-
ible though physically nearby. The pressure to achieve scale reflects the extent to which
a child felt pressured to achieve in terms of status or position, adult labour responsibil-
ities, school grades, or performance in other areas such as music and sports.
Next, the interview was assigned an overall attachment classification of secure-
autonomous, dismissing, or preoccupied, based largely on the ratings of the state of
mind scales. The transcripts of secure-autonomous individuals are characteristically
coherent and fluid, indicating autonomy, balance, and openness to the topics being dis-
cussed; thus, these individuals score high on the coherence of mind scale. In contrast,
individuals who are assigned to one of the two insecure classifications (dismissing and
preoccupied) share an inability to fully integrate their early memories in a coherent and
believable manner. The narratives of dismissing adults typically lack coherence due to
repeated insistence on a lack of memory, and either idealisation or derogation of attach-
ment figures; thus, they score high on the state of mind scales for lack of memory, ideal-
isation of attachment (scored separately for experiences with mother and with father),
and/or derogation of attachment. The transcripts of preoccupied individuals are
characterised by excessive and angry current involvement in relationships with attach-
ment figures, and/or by confused, vague, or passive speech towards a parent; thus, they
receive high scores on the involving anger (scored separately for experiences with
mother and with father) and/or passive speech state of mind scales. The unresolved cat-
egory is assigned in conjunction with the best-fitting overall category to individuals
who show lapses in the monitoring of reasoning or discourse during the discussion
of loss and/or trauma, e.g. speaking of a dead attachment figure as if he or she were
still alive. Adults receiving a score of 5 or higher on the unresolved loss or unresolved
trauma state of mind scales were placed in the unresolved category. Finally, adults were
placed into a ‘cannot classify’ (CC) group when they showed a mix of secure-auton-
omous, dismissing, and/or preoccupied strategies (Hesse, 2008). Thus, the state of
mind scales largely overlap with final AAI classifications, whereas the childhood
experience scales are more likely to be independent of AAI classifications; e.g. an indi-
vidual classified as secure-autonomous must show a coherent state of mind but may
recall any type of childhood experiences.
All of the interviews were coded by two graduate students who had successfully
completed training and were certified to code the AAI. One of the students coded all
of the AAIs and the second coded 28% of the interviews (N ¼ 32) for reliability.
The student coders were trained by the third author who was certified to conduct
AAI training workshops by Mary Main. After completing a two-week training work-
shop, both students were required to pass three tests taken at six-month intervals,
each of which consisted of about 10 transcripts. Exact agreement between the two
coders on the five-way AAI classification – secure-autonomous, dismissing, preoccu-
pied, unresolved with respect to trauma, and CC but not unresolved – was 88% (k ¼
0.85). Disagreements between the two coders were resolved by the third author. Intra-
class correlations for all of the childhood experience and state of mind scales ranged
from 0.88 to 0.65, except for Mother Rejecting (r ¼ 0.42) and Father Involving/
Role Reversing (r ¼ 0.31), so these two scales were excluded from further analysis
because of unacceptably low reliabilities. The AAI categories were coded by the
708 L. McFarland-Piazza et al.

same graduate students who rated childhood experiences, and 28% of the AAI tran-
scripts (N ¼ 32) were double coded. Exact agreement between the two coders on the
five-way classification – secure-autonomous, dismissing, preoccupied, unresolved
with respect to trauma, and CC but not unresolved – was 88% (k ¼ 0.85). Disagree-
ments between the two coders were resolved by the third author.

Father caregiving behaviours

The 30-minute father–infant interactions comprising play, feeding, and a clothes
change were rated on the Infant Caregiving Scales (ICS) developed by two of the
lead researchers. Items were derived from examples of sensitive and insensitive care-
giving described in Ainsworth’s global scales for assessing sensitivity, acceptance,
and cooperation (Ainsworth et al., 1978). The ICS includes 90 items rated on a
seven-point scale. Four caregiving scales were developed: sensitivity, hostility, disen-
gagement, and role reversal, using a criterion sort method (Waters & Deane, 1984).
Seven expert judges rated each of the 90 items according to the extent to which they
were diagnostic of each of the four caregiving constructs. Items which judges agreed
were highly diagnostic of a particular caregiving construct were used to create a
scale to assess that construct. Ratings on the items that made up each scale were aver-
aged to create scores for each scale. Cronbach’s alphas for the four caregiving scales
were 0.94 for sensitivity, 0.95 for role reversal, 0.91 for hostility, and 0.71 for emotion-
al disengagement.
The sensitivity scale assessed the extent to which parents responded quickly and
appropriately to their infant. Examples included, ‘Parent responds to baby when he
or she cries’ and ‘Parent’s actions are finely tuned to the baby’s wishes’. The role rever-
sal scale assessed the extent to which parents turned toward their infants to meet their
own emotional needs. Examples included, ‘Parent becomes annoyed or anxious when
baby does not comply with her wishes’ and ‘Parent does not respect baby as an auton-
omous person’. The hostility scale assessed the extent to which parents expressed nega-
tive emotions toward their infants, either verbally or physically. Examples included,
‘Parent calls the baby unflattering names’ and ‘Parent handles baby in a physically
hostile manner’. The disengagement scale assessed the extent to which parents with-
drew emotionally or physical from their infant. Examples included, ‘Parent and
baby’s interaction seem flat and disengaged’ and ‘Parent seems to have his/her mind
elsewhere’. Five coders rated fathers on all 90 items. Coding was done based on the
entire 30-minute videotaped father–infant interaction. Two coders rated 102 of
the 118 tapes (86%) for reliability, and seven tapes with low inter-rater reliability
were coded by a third coder. The ratings were averaged across all coders. Inter-rater
reliability, assessed using intra-class correlations, was 0.64 for sensitivity, 0.65 for
role reversal, 0.64 for disengagement, and 0.74 for hostility.

Quality of attachment in infancy

The SSP (Ainsworth et al., 1978) was used to assess infant’s attachment security with both
mothers and fathers. Videotapes were coded for the following attachment patterns: secure
(B), insecure-avoidant (A), and insecure-resistant (C). Infants were classified as secure if
they used their parent as a secure-base, seeking comfort when distressed by a brief separ-
ation, and the comfort received successfully reducing their distress, such that they explored
and played again. Infants were classified as avoidant if they avoided proximity or eye
 Early Child Development and Care 709

contact with the mother upon reunion, and as resistant if they mixed seeking proximity-
seeking with angry resistance. Infants whose behaviour in the SSP showed a mix of
behaviours were designated as ‘CC’. In addition, the strange situations were coded on a
nine-point Likert-type disorganised/disoriented (D) scale (Main & Solomon, 1990).
This scale assessed apprehensive, trance-like, or sequential or simultaneous contradictory
behaviour occurring anytime in the presence of the parent. Infants were placed in a disor-
ganised group if they scored above 5 on the nine-point scales. If infants were assigned a
score of 5, the coder decided whether the behaviour warranted placement into the
disorganised category. Infants were also assigned a secondary classification.
Classification of organised (secure, avoidant, and resistant) attachment was based on
Ainsworth et al.’s (1978) coding system, and classification of disorganised attachment
was based on Main and Solomon’s (1990) system. Inter-rater reliability was obtained
using two trained coders who had achieved high reliability (r . 0.90) on 30 test
cases. One key coder first coded all of the videotapes for A, B, C, and D. Next, three
expert coders classified a subset of the videos again for reliability, such that ultimately
every strange situation was coded by two people. The primary coder’s inter-rater
reliability with the three other coders on the three-way classification (A, B, C categories)
was k ¼ 0.88. The reliability between the two coders for the four-way classification
(A, B, C, or D) was k ¼ 0.60, p , 0.001. The coder trainer resolved disagreement on
15 particularly difficult cases, many of which could not be classified (called ‘CC’).

Results
Descriptive statistics
The means and standard deviations for the childhood experience and caregiving scales
are reported in Table 1. On the AAI, 54 fathers were classified as secure-autonomous

Table 1. Childhood experience scales and caregiving scales: descriptive statistics.

Mean Standard deviation
a
Childhood experiences scales
Mother loving 4.1 1.56
Father loving 3.41 1.6
Father rejecting 5.11 1.48
Mother involving 2.51 1.71
Mother pressure to achieve 1.55 1.15
Father pressure to achieve 1.86 1.37
Mother neglect 1.7 1.46
Father neglect 2.4 2.11
Caregiving scalesb
Sensitive 4.44 0.83
Hostile 2.89 0.94
Role reversed 2.92 1.05
Disengaged 3.6 0.61
Frightening 4.05 2.3

N ¼ 118.
a

N ¼ 119.
b
710 L. McFarland-Piazza et al.

(46%), 36 as dismissing (31%), 4 as preoccupied (3%), 17 as unresolved (14%), and 7

as CC (6%). Of the unresolved fathers, six had a secondary classification as secure-
autonomous, six as dismissing, and five as preoccupied. The distribution of secure-
autonomous and unresolved AAI classifications in this sample is similar to that
found in a meta-analysis of the AAI classifications of a multi-national sample of 13
studies of fathers (N ¼ 439), which found 50% secure-autonomous, 24% dismissing,
11% preoccupied, and 15% unresolved or CC fathers (Bakermans-Kranenburg &
van IJzendoorn, 2009). Although the percentage of dismissing fathers is somewhat
higher in our sample, it was not significantly different from the percentage of fathers
found in the multi-national sample, according to a z-test for two proportions.
However, the proportion of preoccupied fathers in our sample was significantly
lower than in the multi-national sample, z ¼ 2.47, p , 0.05.
In the strange situation, 58 infants were classified as having secure attachment with
father (55.2 %), 12 as avoidant (11.4 %), 14 as resistant (13.3%), 14 as disorganised
(13.3%), and 7 as cannot-classify (CC) (6.7%). When disorganised and CC infants
were placed into their best-fitting alternative category, 68 were secure (64.8%), 16 avoi-
dant (15.0%), and 22 resistant (21%). Compared with the distribution of classifications
from a meta-analysis of nearly 2000 strange situations from a multi-national sample of
18 studies (70% secure, 23% avoidant, and 14% resistant) (van IJzendoorn & Kroonen-
berg, 1988), our sample has similar proportions of secure versus insecure infants, but a
significantly lower proportion of infants classified as avoidant, z ¼ 1.80, p , 0.05), and
a higher proportion of those classified as resistant (z ¼ 1.86, p , 0.05). The proportion
of infants classified as disorganised in our sample does not differ from the average of
about 15% found in studies done in North American and Europe (Jacobvitz et al., 2006).
Preliminary analysis was done to determine if family income level or fathers’ age
was significantly related to any of the study variables: AAI classification, fathers’ car-
egiving quality, SSP classification, and fathers’ childhood experience rating scales from
the AAI. Father income and age were not significantly correlated with any of the car-
egiving quality measures, nor did they differ according to AAI classifications.
However, an independent samples t-test indicated that infants who were securely
attached to their fathers in the SSP were significantly more likely to have older
fathers (mean age ¼ 32.2, SD ¼ 6.96) than those who were insecurely attached
(mean age ¼ 29.0, SD ¼ 5.25). Fathers’ age was also correlated with AAI ratings of
pressure to achieve from father, r(115) ¼ 20.22, p , 0.05 and neglect from mother,
r(115) ¼ 0.20, p , 0.05. Income was related only to AAI ratings of pressure to
achieve from mother, r(117) ¼ 20.29, p , 0.01, and pressure to achieve from
father, r(117) ¼ 20.33, p , 0.001. Therefore, we controlled for fathers’ age in ana-
lyses involving secure versus insecure SSP classification, pressure to achieve from
father, and neglect from mother; we controlled for family income in analyses involving
pressure to achieve from mother and from father.

Fathers’ AAI classification and caregiving quality

Two multivariate analyses of variance were used to examine the relation of AAI to car-
egiving. In both, the four dependant variables were the four caregiving scales (sensi-
tivity, hostility, disengagement, and role-reversal), the independent variable was
fathers’ AAI classifications, and the F-statistics and df error terms were determined
using Wilks’ lambda. In the first MANOVA, AAI classifications were broken down
into all five groups; however, preoccupied (N ¼ 4) and CC fathers (N ¼ 7) were
Table 2. MANOVAs for caregiving scales by adult attachment classification.
Mean (SD)
Secure-autonomous (F) Dismissing (D) Unresolved (U) Preoccupied∗∗ (E)
(N ¼ 51) (N ¼ 32) (N ¼ 17) (N ¼ 4) CC∗∗ (N ¼ 7) Univariate F
Sensitive 4.70 (0.69) 4.40 (0.91) 3.99 (0.63) 4.30 (0.80) 4.6 (0.98) 5.40∗
Hostile 2.63 (0.85) 3.03 (1.10) 3.22 (0.66) 3.16 (0.80) 2.64 (1.18) 3.67∗
Disengaged 3.51 (0.53) 3.54 (0.72) 3.94 (0.53) 3.61 (0.42) 3.59 (0.79) 3.40∗

Early Child Development and Care

Role-reversed 2.64 (0.89) 2.91 (1.17) 3.41 (1.02) 3.26 (0.72) 2.62 (0.93) 3.73∗
Frightening 4.06 (2.12) 3.53 (2.17) 4.18 (2.10) 6.5 (2.08) 3.86 (2.96) 0.76

Notes: Post hoc tests indicated that the secure-autonomous and unresolved groups differed significantly on sensitivity, hostility, disengagement, and role-reversal. Also,
the secure-autonomous and dismissing groups differed significantly in hostility.
∗
p , 0.05.
∗∗
Not included in this MANOVA due to small Ns.

711
712 L. McFarland-Piazza et al.

excluded from this analysis because of the small Ns for these groups. The data for all
subjects were analysed in the second MANOVA. A significant multivariate effect for
AAI classification was found, F (8, 188) ¼ 2.20, p , 0.05. Table 2 shows the univari-
ate effects, means, and standard deviation scores. Duncan post hoc tests revealed that
secure-autonomous fathers had significantly higher sensitivity scores than did unre-
solved fathers, although dismissing fathers did not differ significantly from either of
the other two groups. Dismissing fathers and unresolved fathers had higher hostility
scores than secure-autonomous fathers. Finally, unresolved fathers had higher scores
for disengagement and role-reversal compared with secure-autonomous fathers.
In the second MANOVA, secure-autonomous fathers were compared with insecure
fathers (all other AAI groups, including all unresolved and CC fathers) on all measures of
paternal caregiving quality. According to Hesse (2008), since CC interviews show a mix
of attachment strategies, they have very low coherence and ‘they are necessarily defined
as insecure’ (Hesse, 2008, p. 572). The multivariate effect was not significant, F (10, 105)
¼ 1.65, ns. However, univariate analyses indicated that secure-autonomous fathers (N ¼
59) and insecure (N¼ 52) fathers differed significantly in sensitive caregiving, means ¼
4.65 and 4.31, SD ¼ 0.71 and 0.84, respectively, F (1,109) ¼ 4.89, p , 0.05), and in
hostility, means ¼ 2.65 and 3.04, SD ¼ 0.86 and 0.97, respectively, F (1, 109) ¼
4.91, p , 0.05. In addition, secure-autonomous fathers differed marginally from inse-
cure fathers on role-reversal, means ¼ 2.68 and 3.02, SD ¼ 0.92 and 1.11, respectively,
F (1, 109) ¼ 3.01, ns. Fathers’ disengaged behaviour did not differ according to whether
they had a secure-autonomous versus insecure relationship with their infant.

Fathers’ AAI classification and father–infant attachment

The relation of the five major AAI categories (secure-autonomous, dismissing, preoc-
cupied, unresolved, and CC, and the five major SSP categories (secure, avoidant, resist-
ant, disorganised, and CC) are shown in Table 3. A series of logistic regressions were
run to examine whether each AAI category predicted the correspondent SSP category,
after controlling for fathers’ age. In the first logistic regression, we first entered father’s
age as a control, then the secure-autonomous (versus all others) AAI classification, to
predict secure (versus all others) SSP classification. In the second logistic regression,
we entered fathers’ age, then dismissing AAI (versus all other), to predict avoidant

Table 3. Chi-square: relation of father AAI classifications and father – infant SSP
classifications collapsed into secure versus insecure.
Father – infant strange situation
classifications
Secure Insecure Total
Father AAI classifications Insecure
Actual frequencies 23 29 52
Expected frequencies 27.9 24.1 32
Autonomous
Actual frequencies 29 16 45
Expected frequencies 24.1 20.9 64
Total 52 45 97

Note: All unresolved, disorganised, and CC cases counted as insecure.

 Early Child Development and Care 713

SSP classification. In the third logistic regression, we entered fathers’ age, then preoc-
cupied AAI (versus all others), to predict resistant SSP classification. Finally, we
entered fathers’ age, then unresolved AAI (versus all others), to predict disorganised
SSP classification. Significant results were found only for the first logistic regression.
Father’s age was a significant predictor of secure SSP attachment, B ¼ 0.09, SE ¼
0.04, p , 0.05, but secure AAI classification also predicted secure SSP attachment
even after controlling for father’s age, B ¼ 20.76, SE ¼ 0.44, p , 0.05.
Because relations were found only between secure versus insecure AAI and SSP
attachment, we were only able to examine our first mediational hypothesis, i.e. that
fathers’ caregiving quality would mediate the relation between fathers’ secure-auton-
omous attachment and secure father–infant attachment. To examine this hypothesis,
we first needed to examine whether fathers’ caregiving quality predicted secure
versus insecure SSP attachment. A series of logistic regressions was conducted in
which differences between secure versus insecure SSP classifications were examined
for each of the four caregiving quality scales (sensitive, hostile, disengaged, and
role-reversed), after controlling for fathers’ age. Again, all father–infant dyads classi-
fied as disorganised and CC were designated as insecure. Results indicted that only
fathers’ sensitive caregiving significantly predicted secure SSP attachment (B ¼
0.58, SE ¼ 0.27, p , 0.05) after controlling for fathers’ age.
To examine whether fathers’ sensitive caregiving mediates the relation between
fathers’ secure-autonomous AAI classifications and father–infant SSP security, we fol-
lowed the steps recommended by Baron and Kenny (1986). Having already determined
that secure AAI classification predicts secure SSP classification even after controlling
for father’s age, we next used an ordinary least squares (OLS) regression to determine
that fathers’ secure AAI classification also predicts the mediator variable, sensitive car-
egiving, b ¼ 0.33, SE ¼ 0.15, p ,0.05. Finally, we conducted a logistic regression
with the predictor variable (secure-autonomous versus insecure AAI classification) and
the mediator (sensitive caregiving) as independent variables and secure (versus inse-
cure) SSP attachment as the dependent variable. If the relation between the predictor
variable and dependent variable significantly declines when the mediator is included
in the regression, then the mediational hypothesis is supported. The association
between fathers’ secure-autonomous AAI classification and father–infant secure
attachment did decline to non-significance when sensitive caregiving was entered as
a mediator, b ¼ 20.69, SE ¼ 0.43, p ¼ 0.11, supporting the hypothesis that the
relation between fathers’ secure-autonomous AAI status and father–infant secure
attachment is mediated by father’s sensitive caregiving.

Fathers’ childhood experiences and caregiving quality

We first examined the relation of fathers’ caregiving to their ratings for childhood
experiences (mother and father loving, mother and father rejecting, mother and father
involving, mother and father neglecting, and mother and father pressure to achieve)
using first-order bivariate correlations. Few significant relations were found. Pressure
to achieve from mothers was negatively related to fathers’ sensitivity, r (111) ¼
20.23, p , 0.05, and positively related to their hostility, r(111) ¼ 0.37, p , 0.001,
and role-reversed caregiving, r (111) ¼ 0.28, p , 0.01. Pressure to achieve from
fathers was also negatively related to fathers’ sensitivity, r (111) ¼ 20.23, p , 0.05,
and negatively related to their hostility, r(111) ¼ 0.20, p , 0.05, Surprisingly,
neglect from mothers was positively related with fathers’ sensitivity, r (111) ¼ 0.28,
714 L. McFarland-Piazza et al.

Table 4. Multiple regressions examining unique variance accounted for by predictors of

fathers’ caregiving ratings.
B SE B b
Predictors of fathers’ sensitive caregiving
Secure versus insecure AAI classification 0.30 0.14 0.19∗
Mother pressure to achieve 20.09 0.08 20.13
Father pressure to achieve 20.08 0.06 20.13
Mother neglect 0.15 0.05 0.28∗∗
Predictors of fathers’ hostile caregiving
Dismissing versus not-dismissing AAI classification 0.19 0.17 0.10
Unresolved versus not unresolved AAI classification 0.44 0.22 0.17∗
Father pressure to achieve 20.03 0.07 20.05
Mother pressure to achieve 0.31 0.09 0.38∗∗
Mother neglect 20.14 0.06 20.22∗∗
Predictors of fathers’ role-reversed caregiving
Unresolved versus not unresolved AAI classification 0.66 0.24 0.24∗∗
Mother pressure to achieve 0.23 0.08 0.27∗∗
Mother neglect 20.19 0.06 20.27∗∗
∗
p , 0.05.
∗∗
p , 0.01.

p , 0.01, and negatively related to hostility, r(111) ¼ 20.23, p , 0.05, and role-
reversed caregiving, r (111) ¼ 20.26, p , 0.01. That is, fathers who recalled
their mothers as more neglecting were likely to be more sensitive and less hostile and
role-reversing with their infants. No other childhood experience scales were related to
caregiving, and none of the childhood experience scales were correlated with fathers’
emotional disengagement or FR caregiving behaviours.

Unique predictors of fathers’ caregiving quality

To examine whether fathers’ recollections of childhood experiences predicted their car-
egiving quality independent of their attachment representations, and to control for
shared variance between all of the significant predictors of each type of caregiving, mul-
tiple regressions were performed in which all significant predictors of each caregiving
quality measure were entered simultaneously to predict that type of caregiving quality.
Separate regressions were run for each of the three caregiving scales (sensitive, hostile,
and role-reversed caregiving) that were significantly correlated with one or more child
experience scales. In each of the three regressions, fathers’ age and family income were
entered first to control for these covariates. As shown in Table 4, after simultaneously
controlling for all other predictors of fathers’ sensitive caregiving, fathers’ secure
attachment classification, as well as fathers’ recollections of high neglect from
mother, were still significant predictors of sensitive caregiving. In addition, both
fathers’ hostile caregiving and fathers’ role-reversed caregiving were still significantly
predicted by unresolved attachment classification, mothers’ high pressure to achieve
and mothers’ low neglect after controlling for all predictors simultaneously.
 Early Child Development and Care 715

Discussion
The aim of this study was to examine how fathers’ attachment representations and their
recollections of childhood experiences relate to specific aspects of their caregiving and
to infant–father attachment. Fathers’ secure-autonomous adult attachment classifi-
cation was related to higher sensitive and lower hostile caregiving, dismissing and unre-
solved adult attachment classifications were related to higher hostility, and unresolved
adult attachment classification was related to higher disengaged and role-reversed car-
egiving. These results support other studies that found a relation between fathers’ adult
attachment representations and caregiving quality (Cohn et al., 1992; van IJzendoorn,
Kranenburg, Zwart-Woudstra, Van Busschback, & Lambermon, 1991) and also
provide some support for the idea that different insecure paternal states of mind are
related to different styles of insensitive caregiving.
In general, our findings show similar relations between fathers’ adult attachment
representations, caregiving, and SSP attachment to those that have been found for
mothers. In particular, secure-autonomous attachment was found to predict fathers’ sen-
sitive caregiving as well as secure father–infant attachment in the SSP, supporting
other research establishing a link between adult attachment security and infant attach-
ment security for mothers (De Wolff & van IJzendoorn, 1997) and for fathers (Cox
et al., 1992; Goossens & van IJzendoorn, 1990). In addition, we found that dismissing
fathers were more likely than secure-autonomous fathers to show hostile, rejecting
patterns of caregiving, similar to the caregiving patterns found for mothers who have
avoidant attachments with their infants (who have been found to be more likely to
be dismissing) (Belsky et al., 1984; Egeland & Farber, 1984).
We also found that the relation between fathers’ internal working models of
attachment and father–infant SSP classifications was mediated by fathers’ sensitive
caregiving, as has been found in research with mothers (van IJzendoorn, 1995).
However, we did not find that specific types of insecure attachment representations
and insensitive caregiving were related to specific types of insecure infant attachment.
Even though different types of insecure AAI classifications were related to specific
types of insensitive caregiving to some extent, we found no evidence that dismissing
fathers were more likely to have avoidantly attached infants or that unresolved
fathers were more likely to have disorganised infants, as has been found for mothers.
Rather, for fathers, insecure AAI classifications were related to insecure SSP classifi-
cations, collapsed over insecure types.
One reason for the lack of specificity in the relation of particular types of fathers’
insecure working models to particular types of insecure father–infant attachment
may be due to the fact that in this sample, there were only four preoccupied fathers,
so the relation of fathers’ preoccupied attachment representations to their caregiving
and SSP attachment could not be adequately assessed. In addition, we found that
both dismissing and unresolved attachment predicted fathers’ hostile caregiving.
Thus, it may be that for fathers, the caregiving patterns of dismissing and unresolved
fathers are somewhat similar, except that unresolved fathers seem to have even more
negative and problematic caregiving patterns than dismissing fathers, since they were
also more likely than other fathers to be role-reversing and emotionally disengaged
as well as more hostile. Also, when both unresolved and dismissing AAI classifications
were simultaneously entered into a regression to predict hostile caregiving, only
unresolved caregiving remained a significant predictor. Thus, the finding that fathers’
insecure representations of attachment predict insensitive caregiving in general,
716 L. McFarland-Piazza et al.

which in turn predicts insecure father–infant SSP classification, may be driven largely
by the unresolved fathers. Our findings also indicate that having an unresolved
representation of attachment may be particularly detrimental for fathers’ caregiving
quality, perhaps more detrimental than other types of insecure attachment.
One major difference in the relation of adult attachment, caregiving, and infant–
parent attachment for fathers versus mothers is that for mothers, unresolved AAI classi-
fication predicted infants’ disorganised attachment, and the link was mediated by
mothers’ frightening behaviours, in our sample (Jacobvitz et al., 2006) as well as in
other samples (Abrams et al., 2006; Schuengel et al., 1999) In contrast, in previous
research with the present sample, we found that for fathers, unresolved AAI classifi-
cation was not related to frightening behaviour with infants or to infants’ disorganised
attachment (Hazen et al., 2010). Interestingly, the unresolved mothers in our sample
have similar levels of frightening behaviour as the unresolved fathers (Hazen et al.,
2010). However, unlike mothers, fathers who were not unresolved were just as likely
to engage in frightening behaviour as unresolved fathers, so that fathers as a group
had significantly higher mean levels of frightening behaviour than did mothers
(Hazen et al., 2010). We found that many fathers, secure as well as insecure and unre-
solved, engaged in frightening behaviours with their infants such as grabbing them
from behind without warning and looming into their faces with bared teeth and
roaring. However, we found also secure-autonomous fathers were more likely than
insecure fathers to be sensitive and frightening at the same time (Hazen et al., 2010).
That is, if the infant became upset by the frightening behaviour, secure-autonomous
fathers were more likely to stop and comfort the infant. In contrast, insecure and
unresolved fathers were more likely to persist in the frightening behaviour. It may be
that fathers who engage in highly stimulating play (even to the point of seeming
frightening), but remain sensitive, may help their children learn to regulate their own
emotions. In support of this idea, we found that the interaction between frightening
behaviour and sensitive caregiving predicted children’s emotion regulation at
24 months, such that children whose fathers were both frightening and insensitive with
them during infancy showed the highest emotional under-regulation at 24 months,
whereas those whose fathers were frightening and sensitive had the lowest mean levels
of under-regulation (Hazen et al., 2010).
In the present study, we found that instead of showing more frightening behaviours,
unresolved fathers showed significantly more early signs of role-reversed caregiving
and emotional disengagement compared with fathers who were not unresolved, and
they were less sensitive as well. It may be that fathers with unresolved trauma withdraw
and abandon the parenting role when feeling overwhelmed by trauma, thus displaying
emotional disengagement or role-reversed caregiving. Since mothers are usually the
primary caregivers, they may be more likely to feel that they cannot disengage from
the caregiving role. When feeling overwhelmed by trauma, they may thus be more
likely to remain engaged and non-consciously lapse into a dissociated state, showing
frightened/frightening behaviour as a result instead of simply withdrawing from the
parenting role.
Although fathers’ adult attachment representations predicted their caregiving better
than did their recollections of childhood experiences, recollections of maternal and
paternal pressure to achieve and maternal neglect did relate to fathers’ caregiving
independent of their AAI classifications. Specifically, recollection of higher maternal
and paternal pressure to achieve predicted lower sensitivity and greater hostility, and
higher maternal pressure to achieve predicted greater role-reversed caregiving.
 Early Child Development and Care 717

Perhaps, fathers who were under pressure to perform and achieve as children were
socialised to be more career-oriented in their adult life. Therefore, they may feel resent-
ment when they do spend time with their child and may not be able to sensitively
respond to their child’s needs because they are used to focusing on their own needs.
Another possibility is that fathers may transfer the pressure that they felt to achieve
to their expectations for their own child, expecting even their young infants to
perform up to a certain standard (i.e. becoming upset when the baby makes a mess
while eating or fails to put a shape into the correct place on a toy). Little empirical
research has examined pressure to achieve. One study, however, did find that adoles-
cents who feel unduly pressured to achieve and do well in school are more likely to
report deviant activity, feel incapable of reaching goals, and have low self-esteem com-
pared with those who do not feel such pressure (Eskilson, Wiley, Muehlbauer, &
Dodder, 1986). More research is needed to examine pressure to achieve and how it
relates to fathering and child outcomes.
Surprisingly, fathers’ recollections of neglect from their mothers in childhood pre-
dicted some aspects of positive caregiving, including higher sensitivity, lower hostility,
and lower role-reversal. This finding seems contradictory as most research on early
neglect suggests negative outcomes (van der Kolk, Perry, & Herman, 1991; Loos &
Alexander, 1997). Thus, the coding for mother neglect was checked carefully for
errors. To better understand this finding, the interview transcripts for fathers who
reported high mother neglect were examined. Most of these fathers indicated that
they had to be responsible for their own care and the care of younger siblings at a
young age. Thus, perhaps these fathers learn appropriate caregiving skills that may
carry over into their interactions with their own children. Because boys typically
have less caregiving experience compared with girls, these early caregiving experiences
may be particularly valuable for future fathers, even if obtained within a context of
perceived neglect. Still, sensitive caregiving would not be expected unless these men
were able to transform their early negative working models and become secure as
adults. However, we found no relationship between fathers’ ratings of high maternal
neglect and their AAI categorisation.
It is also unknown if the level of neglect reported by fathers in this sample was com-
parable to those in other studies finding strong links between early neglect and negative
outcomes. In our sample, only two fathers were rated above the midpoint of 5 on a nine-
point scales and another eight subjects scored a 5. The two fathers judged to have
experienced high neglect both experienced maternal abandonment at times during
childhood. One of them was below average in sensitivity and was classified as CC
and unresolved on the AAI, as might be expected. However, the other father who
experienced high neglect was more than one standard deviation above the mean on sen-
sitivity and was secure on the AAI. This subject was judged to have had a very loving
father who cared for him after his mother abandoned him, as well as a loving step-
mother. The eight fathers who scored 5 on maternal neglect generally reported that
their mothers were very busy out of necessity, i.e. one reported that his mother was
a single mother who held two jobs. These were the fathers who reported often caring
for younger siblings. Thus, in our sample, factors other than the actual neglect from
mothers, such as early caregiving experiences or loving care from other attachment
figures, may have been driving this finding. Clearly, future research on the role of
childhood experiences of neglect, levels of neglect, and experiences that may enable
to overcome the effects of early neglect is needed.
718 L. McFarland-Piazza et al.

The findings of this study, indicating that different insecure groups show different
patterns of negative caregiving, demonstrate the importance of examining attachment
subgroups separately. Perhaps, the most significant limitation of this study is the rela-
tively small sample size, which prevented us from examining the caregiving and attach-
ment patterns of preoccupied fathers. Future research should also examine the specific
caregiving styles of under-represented insecure subgroups, perhaps using meta-analytic
methods. These findings also need to be extended to more diverse samples of fathers,
including low-SES fathers, fathers of diverse cultural and ethnic backgrounds, and
fathers who are primary caregivers or single fathers.
To summarise, this study contributes to the literature on the relation between
fathers’ representations of attachment, recollections of childhood experiences, caregiv-
ing styles, and infant–father attachment. These relations, in general, were found to be
similar to those that have been found for mothers. However, unresolved mothers have
been found to show frightening patterns of caregiving, whereas we found that unre-
solved fathers were more likely to show role-reversed and emotionally disengaged pat-
terns of care. This study revealed that while a few types of childhood experiences were
related to their caregiving, fathers’ attachment representations were more strongly
related to their caregiving. Thus, it is possible for fathers to have had negative child-
hood experiences and still parent in sensitive and loving ways with their own children
if they have open, flexible and secure attachment representations.

Notes on contributors
Laura McFarland-Piazza, PhD, is a lecturer (equivalent to an assistant professor) of early child-
hood teacher education at Charles Sturt University. Her research focuses on parent – child
relationships, positive child guidance, and parent –teacher relationships. She teaches classes
in infant-toddler development, research methods, health and safety, and early childhood assess-
ment and observation.
Nancy Hazen, PhD, is an associate professor of human development and family sciences at the
University of Texas at Austin. Her research focuses on the role of parenting and family relation-
ships in the development of children’s social competence, with a particular focus on attachment,
peer relationships, and the role of fathers. She teaches classes in child development, peer
relationships, and family relationships.
Deborah Jacobvitz, PhD, is a professor of human development and family sciences at the
University of Texas at Austin. Her research focuses on the transmission of attachment relation-
ships across generations, parent – child relationships, relations among relationships within the
family, and developmental outcomes from infancy to middle childhood. She teaches classes in
child development, attachment across the lifespan, theories of child and family development,
and infancy.
Erin Boyd-Soisson, PhD, CFLE is an associate professor of human development and family
science and a certified family life educator at Messiah College. Her research interests include
using family systems theory and a family strengths perspective to understand family folklore
as well as Amish family relationships. She teaches classes in child, adolescent, and adult
development, as well as play and development and foundations of marriage and family.

References
Abrams, K.Y., Rifkin, A., & Hesse, E. (2006). Examining the role of parental frightened/
frightening subtypes in predicting disorganized attachment within a brief observational
procedure. Development and Psychopathology, 18, 345–361.
Ainsworth, M.D.S., Blehar, M.C., Waters, E., & Wall, S. (1978). Patterns of attachment.
Hillsdale, NJ: Erlbaum.
 Early Child Development and Care 719

Bakermans-Kranenburg, M., & van IJzendoorn, M. (2009). The first 10,000 adult attachment
interviews: Distributions of adult attachment representations in clinical and non-clinical
groups. Attachment and Human Development, 11(3), 223–263.
Baron, R.M., & Kenny, D.A. (1986). The moderator –mediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51, 1173–1182.
Belsky, J., Rovine, M., & Taylor, D. (1984). The Pennsylvania Infant and Family Development
Project: III. Origins of individual differences in infant – mother attachment: Maternal and
infant contributions. Child Development, 55(3), 718–728.
Bowlby, J. (1969). Attachment. New York, NY: Basic Books.
Caldera, Y.M., Huston, A., & O’Brien, M. (1995). Antecedents of father – infant attachment:
A longitudinal study. Paper presented at the meetings of the Society for Research in
Child Development, Indianapolis, IN.
Cohn, D.A., Cowen, P.A., Cowen, C.P., & Pearson, J. (1992). Mothers’ and fathers’ working
models of childhood attachment relationships, parenting styles, and child behavior.
Development and Psychopathology, 4, 417–431.
Cox, M., Owen, M., Henderson, V., & Margand, N. (1992). Prediction of infant – father and
infant –mother attachment. Developmental Psychology, 28(3), 474–483.
De Wolff, M.S., & van IJzendoorn, M.H. (1997). Sensitivity and attachment: A meta-analysis on
parental antecedents of infant attachment. Child Development, 68, 571–591.
Easterbrooks, M.A., & Goldberg, W.A. (1984). Toddler development in the family: Impact of
father involvement and parenting characteristics. Child Development, 55, 740–752.
Egeland, B., & Farber, E.A. (1984). Infant – mother attachment: Factors related to its develop-
ment and change over time. Child Development, 55(3), 753–771.
Eskilson, A., Wiley, M.G., Muehlbauer, G., & Dodder, L. (1986). Parental pressure, self-esteem,
and adolescent reported deviance: Bending the twig too far. Adolescence, 21(83), 501–515.
George, C., Kaplan, N., & Main, M. (1985/1996). Adult Attachment Interview (Unpublished
protocol, 3rd ed.). University of California, Berkeley.
George, C., & Solomon, J. (1996). Representational models of relationships: Links between car-
egiving and attachment. Infant Mental Health Journal, 17, 198–216.
Goossens, F.A., & van IJzendoorn, M.H. (1990). Quality of infants’ attachments to professional
caregivers: Relation to infant – parent attachment and day-care characteristics. Child
Development, 61, 832–837.
Grossmann, K., & Grossman, K.E. (1992). Newborn behavior, the quality of early parenting and
later toddler – parent relationships in a group of German infants. In J.K. Nugent, B.M. Lester,
& T.B. Brazelton (Eds.), The cultural context of infancy (Vol. 2, pp. 3–38). Norwood, NJ:
Ablex.
Grossmann, K., Grossmann, K.E., Fremmer-Bombik, E., Kindler, H., Scheuerer-Englisch, H., &
Zimmermann, P. (2002). The uniqueness of the child –father attachment relationship:
Fathers’ sensitive and challenging play as the pivotal variable in a 16-year longitudinal
study. Social Development, 11, 307–331.
Hazen, N., McFarland, L., Jacobvitz, D., & Boyd-Soisson, E. (2010). Father – infant attachment
and fathers’ frightening and sensitive infant caregiving: Implications for children’s later
development of emotion regulation and social-emotional problems. Early Child
Development and Care, 18(1), 51–69.
Hesse, E. (2008). The Adult Attachment Interview: Protocol, method of analysis, and empirical
studies. In J. Cassidy & P.R. Shaver (Eds.), Handbook of attachment theory and research
(2nd ed., pp. 552–558). New York, NY: Guilford Press.
Isabella, R., Belsky, J., & von Eye, A. (1989). Origins of infant – mother attachment: An exam-
ination of interaction synchrony during the infant’s first year. Developmental Psychology,
25, 12–21.
Jacobvitz, D., Leon, K., & Hazen, N. (2006). Does expectant mothers’ unresolved/disorganized
trauma predict frightened/frightening (FR) maternal behavior? Risk and protective factors.
Development and Psychopathology, 18, 363–379.
Loos, M.E., & Alexander, P.C. (1997). Differential effects associated with self-reported
histories of abuse and neglect in a college sample. Journal of Interpersonal Violence, 12,
340–360.
720 L. McFarland-Piazza et al.

Lyons-Ruth, K., Bronfman, E., & Parsons, E. (1999). Maternal frightened, frightening, or aty-
pical behavior and disorganized infant attachment patterns. In J. Vondra & D. Barnett (Eds.),
Atypical patterns of infant attachment: Theory, research, and current directions.
Monographs of the Society for Research in Child Development (pp. 67 – 96). Monographs
of the Society for Research in Child Development, 64(3 serial no. 258).
Madigan, S., Moran, G., & Pederson, D.R. (2006). Unresolved states of mind, disorganized
attachment relationships, and disrupted mother – infant interactions of adolescent mothers
and their infants. Developmental Psychology, 4(2), 293–304.
Main, M. (1995). Recent studies in attachment. In S. Goldberg, R. Muir, & J. Kerr (Eds.),
Attachment theory: Social, developmental and clinical perspectives (pp. 407–474).
Hillsdale, NJ: Analytic Press.
Main, M., Goldwyn, R., & Hesse, E. (1984 –2002). Adult attachment scoring and classification
systems. Unpublished manuscript, Department of Psychology, University College, Berkley.
Main, M., & Hesse, E. (1990). Parents’ unresolved traumatic experiences are related to infant
disorganized attachment status: Is frightened and/or frightening parental behavior the
linking mechanism? In M.T. Greenberg, D. Cicchetti, & E.M. Cummings (Eds.),
Attachment in the preschool years: Theory, research & intervention (pp. 161–182).
Chicago: University of Chicago Press.
Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, childhood, and adulthood:
A move to the level of representation. In I. Bretherton & E. Waters (Eds.), Growing
points of attachment: Theory and research (pp. 66 – 104). Monographs of the Society for
Research in Child Development, 50, 1 – 2, Serial no. 209. Chicago: University of Chicago
Press.
Main, M., & Solomon, J. (1990). Procedures for identifying infants as disorganized/disoriented
during the Ainsworth Strange Situation. In M. Greenberg, D. Cicchetti, & E. M. Cummings
(Eds.), Attachment in the preschool years: Theory, research and intervention (pp. 121–
160). Chicago: University of Chicago Press.
Paquette, D. (2004). Theorizing the father– child relationship: Mechanisms and developmental
outcomes. Human Development, 47, 193–219.
Pearson, J., Cohn, D., Cowan, P., & Cowan, C. (1994). Earned- and continuous-security in adult
attachment: Relation to depressive symptomatology and parenting style. Development and
Psychopathology, 6, 359–373.
Pederson, D.R., Gleason, K.E., Moran, G., & Bento, S. (1998). Maternal attachment represen-
tations, maternal sensitivity, and the infant – mother attachment relationship. Developmental
Psychology, 34(5), 925–933.
Phelps, J., Belsky, J., & Crnic, K. (1997). Earned security, daily stress, and parenting: A com-
parison of five alternative models. Development and Psychopathology, 10, 21–38.
Schneider Rosen, K., & Rothbaum, F. (1993). Quality of parental caregiving and security of
attachment. Developmental Psychology, 29, 358–367.
Schuengel, C., Bakermans-Kranenburg, M.J., van Ijzendoorn, M.H., & Blom, M. (1999).
Unresolved loss and infant disorganization: Links to frightening maternal behavior. In
J. Solomon & C. George (Eds.), Attachment and disorganization (pp. 71–94).
New York, NY: Guilford Press.
Smith, P.B., & Pederson, D.R. (1988). Maternal sensitivity and patterns of infant – mother
attachment. Child Development, 59(4), 1097–1101.
van IJzendoorn, M.H. (1995). Adult attachment representations, parental responsiveness, and
infant attachment: A meta-analysis on the predictive validity of the adult attachment inter-
view. Psychological Bulletin, 117(3), 387–403.
van IJzendoorn, M.H., & De Wolff, M.S. (1997). In search of the absent father: Meta-analysis of
infant –father attachment: A rejoinder to our discussants. Child Development, 68(4),
604–609.
van IJzendoorn, M.H., Kranenburg, M.J., Zwart-Woudstra, H.A., Van Busschback, A.M., &
Lambermon, M.W.E. (1991). Parental attachment and children’s socio-emotional develop-
ment: Some findings on the validity of the Adult Attachment Interview in the Netherlands.
International Journal of Behavioral Development, 15, 375–394.
van IJzendoorn, M., & Kroonenberg, P. (1988). Cross-cultural patterns of attachment: A meta-
analysis of the Strange Situation. Child Development, 59, 147–156.
 Early Child Development and Care 721

van der Kolk, B.A., Perry, J.C., & Herman, J.L. (1991). Childhood origins of self-destructive
behavior. American Journal of Psychiatry, 148, 1665–1671.
Volling, B.L., & Belsky, J. (1992). Infant, father, and marital antecedents of infant – father
attachment security in dual-earner and single-earner families. International Journal of
Behavioral Development, 15, 83–100.
Ward, M.J., & Carlson, E.A. (1995). Associations between adult attachment representations,
maternal sensitivity, and infant –mother attachment in a sample of adolescent mothers.
Child Development, 66(1), 69–79.
Waters, E., & Deane, K.E. (1985). Defining and assessing individual differences in attachment
relationships: Q-methodology and the organization of behavior in infancy and early child-
hood. In I. Bretherton & E. Waters (Eds.), Growing points of attachment theory and
research (pp. 41 – 65). Monographs of the Society for Research in Child Development,
50(1– 2, Serial No. 209).