ArticlePDF Available

Childhood Physical and Emotional Abuse by a Parent: Transference Effects in Adult Interpersonal Relations


Abstract and Figures

Extending research on transference and the relational self (Andersen & Chen, 2002), female undergraduates with or without a history of physical and emotional abuse by a loved parent participated in an experiment manipulating parental resemblance and threat-relevant interpersonal context in a new person. Transference elicited differences not evident in the control condition between abused and nonabused participants' responses, with greater rejection expectancy, mistrust, dislike, and emotional indifference reported by abused participants. Immediate implicit affect was more positive in transference than in the control condition regardless of abuse history. Yet, abused participants in transference also reported increased dysphoria that was markedly attenuated when interpersonal threat was primed, and no such pattern occurred among nonabused participants. Evidence that interpersonally guarded and affectively complex responses are triggered in transference among previously abused individuals suggests that this social-cognitive process may underlie long-term interpersonal difficulties associated with parental abuse.
Content may be subject to copyright.
Childhood Abuse and Transference 1
Childhood Physical and Emotional Abuse by a Parent:
Transference Effects in Adult Interpersonal Relations
Kathy R. Berenson and Susan M. Andersen
New York University
Word count: 9, 749
Key words: Child abuse, social cognition, transference
Published as:
Berenson K.R., & Andersen S.M. (2006). Childhood physical and emotional abuse by a parent:
Transference effects in adult interpersonal relations. Personality and Social Psychology Bulletin,
32, 1509-1522.
Childhood Abuse and Transference 2
Extending research on transference and the relational self (Andersen & Chen, 2002), female
undergraduates with or without a history of physical and emotional abuse by a loved parent
participated in an experiment manipulating parental resemblance and threat-relevant
interpersonal context in a new person. Transference elicited differences not evident in the control
condition between abused and nonabused participants’ responses, with greater rejection
expectancy, mistrust, dislike, and emotional indifference reported by abused participants.
Immediate implicit affect was more positive in transference than in the control condition
regardless of abuse history. Yet abused participants in transference also reported increased
dysphoria that was markedly attenuated when interpersonal threat was primed, and no such
pattern occurred among nonabused participants. Evidence that interpersonally guarded and
affectively complex responses are triggered in transference among previously abused individuals
suggests that this social-cognitive process may underlie long-term interpersonal difficulties
associated with parental abuse.
Key words: Child abuse, social cognition, transference
Childhood Abuse and Transference 3
Childhood Physical and Emotional Abuse by a Parent:
Transference Effects in Adult Interpersonal Relations
Physical abuse by parents is widely recognized to put children at risk for later
interpersonal problems, as well as a range of psychological symptoms and diagnoses (Knutson,
1995). Compared with their peers, abused children exhibit diminished social competence (Jaffe,
Wolfe, Wilson, & Zak, 1986). Mistrustful and vigilant, they may be overly constricted and/or
overly domineering in what appear to be attempts to anticipate and control others’ behavior
(Crittenden & DiLalla, 1988; Jacobvitz & Hazen, 1999; Lyons-Ruth, Bronfman, & Atwood,
1999). From infancy through adulthood, abused individuals are more withdrawn and emotionally
disengaged than others, exhibiting less social interaction, prosocial behavior, and affective
attunement (Mueller & Silverman, 1989; Solomon & George, 1999). They also report greater
interpersonal discomfort (Briere & Runtz, 1988). Ultimate costs for the individual and society
are high: In their relationships, people with a history of parental physical abuse are at increased
risk for aggressive behavior (Malinosky-Rummell & Hansen, 1993), rejection (Dodge, Pettit, &
Bates, 1994), and re-victimization (Wolfe, Wekerle, Reitzel-Jaffe & Lefebvre, 1998).
Difficulties associated with abuse history are likely to have both stable and contextually
dependent components. For example, vulnerabilities to overwhelming affect could promote
vigilant and avoidant responses to the possibility of threat even in benign situations. At the same
time, theory and research suggest that processes of construct activation underlie considerable
within-person variability in personality (Mischel & Shoda, 1995), and do so for interpersonal
patterns as well (Andersen & Chen, 2002).
Mental Representations and Correlates of Childhood Physical Abuse
Research demonstrates that previously abused people bring negative interpretations to
novel interpersonal situations more than do most and that this contributes to maladaptive social
behavior. For example, a hostile attribution bias, in which ambiguous behavior is interpreted as
maliciously intended, partially mediates aggression in abused individuals (Dodge, Petitt, Bates,
& Valente, 1995). In addition, rejection sensitivity, a processing disposition involving anxious
expectations for rejection, partially mediates abused individuals’ self-reports of insecure
attachment (Feldman & Downey, 1994). Abused children depict interpersonal relations as more
malevolent and painful than do nonabused children (Dean, Malik, Richards, & Stringer, 1986), a
phenomenon proposed to originate in more negative "working models" of caregivers (e.g.,
Childhood Abuse and Transference 4
Cicchetti, Toth & Lynch, 1995). Consistent with this, the degree to which children’s descriptions
of their parents lack coherence and positive affective tone mediates the association between
parental maltreatment and peer rejection (Shields, Ryan, & Cicchetti, 2001).
Yet, presumably due to basic motivations for attachment, most children love their
parents. Hence, the affect linked to representations of abusive parents may not typically be
negative but complicated by opposing feelings. In relating to their caregivers, up to 80% of
abused infants but less than 20% of the general population show disorganized attachment
characterized by confusing, asynchronous behavior (Carlson, Cicchetti, Barnett & Brunwald,
1989). Simultaneous or sequential contradictory behaviors are frequently observed in abused
children’s peer interactions as well, including approach and avoidance (George & Main, 1979),
caretaking and cruelty (Main & George, 1985; Jacobvitz & Hazen, 1999), and inhibition and
unmodulated impulsivity (van der Kolk, 1996). Attachment theorists propose that abused
individuals’ problematic interpersonal patterns reflect less coherently integrated representations
of the self and parent, the legacy of extensive efforts to regulate painful emotions in the parent-
child relationship through suppression and distortion (Bowlby, 1988; Main & Goldwyn, 1984;
Shields et al., 2001).
The notion that patterns of relating to important others later emerge with new people
through the operation of mental representations is often invoked in both social-cognitive and
attachment perspectives (e.g., Andersen, Reznik, & Chen, 1997; Baldwin, Keelan, Fehr, Enns, &
Koh-Rangarajoo, 1996; Bowlby, 1973). The present study directly examines this process among
individuals physically and emotionally abused by a parent.
Transference and the Relational Self
In the social-cognitive process of transference (see Andersen & Chen, 2002 for a review),
information stored in memory about people of significance to us (including parents among many
others over a lifetime) comes to influence how we experience new social encounters. To study
this experimentally, researchers manipulate activation of participants’ significant-other
representations as they learn about a new person. In an experimental condition, participants read
statements allegedly describing a new person that are actually derived from sentences they wrote
weeks earlier about someone significant to them, whereas in the control condition, the
descriptions have no specific relevance to anyone from their own lives. Research has repeatedly
demonstrated that this manipulation of significant-other resemblance leads the significant-other
Childhood Abuse and Transference 5
representation to be activated and used to “go beyond the information given” (Bruner, 1957)
about the new person, and alternative explanations involving the familiarity of the person’s
features have been ruled out in this effect. The research has also repeatedly shown that
interpersonal responses in transference reflect the affective tone associated with the original
significant other. For example, when the features attributed to a new person have been derived
from a positively regarded significant other rather than a negatively regarded one, an increase
occurs in participants’ self-reported liking, motivation for closeness, and expectancy for
acceptance in relation to him/her (Andersen, Reznik & Manzella, 1996), and even a self-
fulfilling prophecy effect in which this emotional quality emerges in the actual behavior of the
new person (Berk & Andersen, 2000).
The interdependence between people and their significant others involves strong
emotional bonds and an extensive social learning history of experiences with emotional
implications. Hence, emotional responses are readily evoked in transference, and, like other
affectively laden interpersonal reactions, they have been shown to typically reflect the overall
tone of the original significant-other representation. When learning about a new person who
resembles a positive significant other, participants’ immediate facial expressions are more
positive than when learning about a person who resembles a negative significant other, or in a
control condition. Even the negative features of a positive significant other evoke positive facial
affect when encountered in a new person (Andersen et. al., 1996), suggesting that the
transformation of loved ones’ faults into virtues, an emotional process characteristic of close
relationships (e.g., Murray & Holmes, 1999), can emerge in relation to a new person in
transference. Beyond this, self-reported mood states have also been shown to arise, and appear to
be influenced by aspects of the anticipated interpersonal encounter as well as by what is
anticipated based on the significant-other representation. For example, negative mood arises in a
positive transference when contextual cues indicate that the interpersonal roles typically assumed
in relation to the positive significant other are violated (Baum & Andersen, 1999).
No prior transference research has examined the repercussions of activating
representations of positive significant others associated with as much pain as abusive parents are
likely to be. Nonetheless, we assume that the process of transference itself should be invariant
(Andersen & Berk, 1998). Cues of parental resemblance in a new person should activate the
parental representation and universally bring affectively loaded material from the parent-child
Childhood Abuse and Transference 6
relationship to mind. It is the content of this relational material that should differ on the basis of
abuse history -- because among abused people relative to their nonabused peers, expectancies in
the relationship with a loved parent are likely to be negative -- and evoke divergent responses to
a new person in transference.
Current Research
We used an experimental paradigm to examine the social-cognitive process of
transference as a potential mechanism underlying interpersonal patterns associated with parental
abuse. Female college students with or without a history of physical and psychological abuse by
a parent they still love (as reported in advance) were or were not induced to experience
transference. That is, we arranged that features presented to participants, allegedly describing a
new person they expected to meet, were actually derived from descriptions of the participant’s
own parent (obtained during a pretest session) or from another participant’s parent (a control
condition). Crossed with this manipulation was one in which the new person was or was not
described as “getting increasingly tense and irritable”, under the assumption that this might
prime threatening aspects of the relationship with an abusive parent.
The study was designed to examine two primary hypotheses. First, we tested the
prediction that regardless of abuse history, biased inferences about a new person should
demonstrate the basic social-cognitive process of transference as a function of the person’s
resemblance to the participant’s own parent. Second, because activation of a parental
representation should spread to the relationship between the parent and the self, interpersonal
responses in transference should differ with abuse history. Although prior studies have shown
that positive significant-other representations evoke positive responses to a new person in
transference, this has not been explored in participants preselected for particularly painful
experiences in relation to an otherwise positive significant other. We predicted that the positive
responses to a new person who resembles a loved parent would be diminished or even reversed
among those who were abused by this parent, perhaps especially in combination with
contextually cued interpersonal threat. Specifically, we hypothesized that the process of
transference would evoke more negative, wary responses to the new person (e.g., expectancy for
rejection, mistrust, emotional indifference, dislike, and avoidance motivation) among abused
participants relative to those who were not abused.
Additionally, we examined exploratory hypotheses about implicit and explicit affect. The
Childhood Abuse and Transference 7
relational self model and prior evidence clearly suggest that emotional responses associated with
the significant other arise in transference. Whether transference would evoke the same emotional
responses among previously abused individuals remains an open question, however, especially
given the clinical literature suggesting that abuse history may be associated with altered patterns
of affect regulation. We thus asked independent judges to rate the pleasantness of participants’
facial expressions at the moment they encoded each feature of the new person (the index of
implicit affect used in prior research, e.g., Andersen et al., 1996), and assessed self-reports of
dysphoric mood as well.
Female undergraduates (N = 144) completed the two-session study in partial fulfillment
of a course requirement. The ethnically diverse sample (50.0% Caucasian, 18.3% Asian, 16.9%
Latina, 11.3% African-American) had a median age of 19 (range 18-24).
Session 1 was completed by 225 preselected participants who reported physical and
emotional abuse by a parent while growing up, or no such abuse (details to follow). All but 13
(5.8%, 6 abused, 7 nonabused) agreed to participate in further studies (i.e., Session 2).
Participants who had not attended Session 2 before the semester’s end were lost to follow-up (n
= 45), and experimenter or equipment error invalidated another 4 cases. We also excluded 12
participants who did not view the relevant parent as positive and significant in their lives, and 7
who recognized the link between the sessions, as described below.
Female undergraduates preselected on self-reported abuse history were recruited for a
study in which they described both parents. Weeks later they participated in an allegedly
independent experiment in which they learned about a new person who resembled either their
own or another participant’s parent. Crossed with this manipulation, interpersonal threat was
primed for half the participants by describing the new person as “increasingly tense and
irritable,” for a 2 x 2 x 2 (Abuse History x Parental Resemblance x Threat Prime) design with n =
18 in each cell.
Preselection of Abused and Nonabused Participants
During a mass-testing session, participants reported how many times each parent directed
particular behaviors toward them during their upbringing (with response options from 0 to 40+).
Childhood Abuse and Transference 8
Physical abuse was assessed using the Severe Physical Violence subscale of the parent-to-child
version of the Conflict Tactics Scale (CTS, Gelles & Straus, 1988). This frequently used six-item
measure of typically injurious violent acts is well supported by reliability and validity data
(Straus & Gelles, 1990). To preselect participants whose treatment could clearly be characterized
as abusive or nonabusive, we also created a composite index of emotionally threatening parental
behavior by summing the Minor Physical Aggression and Verbal/Symbolic Aggression subscales
of the CTS (Gelles & Straus, 1988) with the Psychological Maltreatment Scale (Briere & Runtz,
1988). Internal consistency of this 16-item composite was .96 for mothers and .94 for fathers.
Eligibility for the abused group required at least one incident of Severe Physical Violence
by a parent who also engaged in frequent emotionally threatening behavior (above the median
before preselection).
The nonabused group was required to have reported no Severe Violence by
either parent, and below the median frequency of emotionally threatening behavior for both
parents. Table 1 presents descriptive statistics for parental behavior reported by the 144
participants in our final sample.
Screening questions ensured that no participants designated nonabused by our
preselection measure reported parental sexual abuse. After the experiment, participants rated the
occurrence of sexual abuse by parents (and others) on 5-point scales (from “not at all” to “a great
deal”) in response to the question, “To what extent, in your view, have you ever experienced
unwanted touching or coerced/forced sexual activity of any kind by the following people?” Six
participants preselected for parental physical abuse (8.3%) reported some degree of parental
sexual abuse, whereas none in the nonabused group did. Hence, no participants were excluded on
this basis.
Although all participants were drawn from a non-clinical population, we assumed and
observed clinically relevant between-group differences. Participants completed a 53-item
measure of psychological problems (Brief Symptom Inventory, BSI; Derogatis, 1993), rating
how distressed they were by symptoms in the past week on a 5-point scale. Using the Global
Severity Index (averaging all items), abused participants reported significantly more distress (M
= .91) than nonabused participants (M = .65), t (142) = 2.71, p < .01.
Session 1: Materials and Procedure
The first session consisted primarily of feature-listing tasks. After freely describing
themselves, participants completed sentences to describe the unique characteristics of their
Childhood Abuse and Transference 9
primary mother and father figures (with the prompts “My mother…” and “My father…”). They
were instructed to complete each sentence with up to 6 words and not to mention the self or other
specific people. Participants wrote 7 positive and 7 negative sentences about each parent, and
rank-ordered the sentences in each set by importance. Next, from a list of 42 relatively neutral
trait-adjectives (Anderson, 1968), they identified 10 descriptive traits, 10 counterdescriptive
traits, and 12 traits that were irrelevant to each parent.
Participants then indicated (“yes” or “no”) whether each parent was a) someone they
knew well, b) significant to them, and c) part of their ongoing lives. They also rated the overall
affective tone of each parental representation a positive tone was indicated by ratings above the
center of a 6-point scale from “Don’t like/love very much and don’t feel very good about” to
“Like/love very much and feel very good about.” Variation in these factors was important to
control because the relational self model highlights the special affective relevance of
representations of significant others (as opposed to more distant persons), that drives emotionally
and motivationally charged responses when the significant-other representation is activated
(Andersen & Chen, 2002). As we anticipated among college-age participants, few described their
parent figures as negative significant others. Because 90.4 % of abused participants evaluated at
least one abusive parent positively, and 99.1% of nonabused participants evaluated at least one
parent positively, focusing on positively evaluated parents seemed appropriate. Hence, 12 (5.3%)
participants for whom no relevant parent was a positive significant other were excluded.
Participants were told that the experiment was over and were partially debriefed. They
were invited to enroll in other studies in the department, described as entirely optional, that were
actually the second session of this research.
Session Two: Materials and Procedure
The second session was held at least 14 days later (median = 24, range 14-111) in a
different location with a different experimenter who was blind to abuse status and assigned
conditions. Each participant was met individually and led to believe that another participant was
already in session with another experimenter in an adjoining room. After assessing baseline
mood at the present moment using the Profile Of Mood States (POMS-R, McNair, Lorr, &
Droppleman, 1992), we explained our cover story, i.e., that we were helping the university
evaluate a prospective “buddy system” for incoming students by examining impressions of a new
person (“potential buddy”) in the course of learning about and interacting with him/her.
Childhood Abuse and Transference 10
Participants then read a series of handwritten statements in a fixed random order, one at a
time, about the person next door. We explained that a trained interviewer had prepared the
statements to provide a balanced characterization of the person’s positive and negative qualities,
and to note anything that stood out about his/her present frame of mind.
Manipulations of parental resemblance and interpersonal threat. In the parental
resemblance condition, the first 10 statements were based on information the participant had
provided about a parent during the previous session: 6 were derived from sentences about the
parent (ranked 4-6 on the positive and negative lists), and 4 were fillers based on traits
designated irrelevant to the parent. To reduce the risk of participants recognizing the source of
the features, idiographic statements were systematically paraphrased.
As in previous studies, each participant in the control condition was exposed to the same
set of descriptors as another participant in the resemblance condition paired 1-to-1 at random
and without replacement. Yoking thus held constant the content of the descriptors across
conditions, and was done within prime condition and the gender of the new person (matched to
the relevant parent’s gender). Because a generic category (such as “abusive parent”) may be
activated by features resembling this category independent of whether or not the significant-other
representation is activated (e.g., Karylowski, Konarzewski, & Motes, 1999; Baum & Andersen,
1999), we counterbalanced the stimulus content encountered in the control condition. That is,
half the participants in the control condition were paired with another in the resemblance
condition who had the same abuse history, and half were paired with another having a different
abuse history.
After participants read 10 statements about the new person, half read an additional
statement: “He/she is getting increasingly tense and irritable.” This was intended to prime
threatening aspects of an abusive parent-child relationship, in case these aspects were not
otherwise activated. This threat prime did not refer to abusive behavior, but rather a state of mind
with implications for what behavior could be expected from the new person in an imminent
Assessment of facial affect and mood. A hidden video camera captured participants’
facial expressions while reading each statement. Two independent judges (blind to condition)
rated the expressions for pleasantness, from 1 (Not at all) to 7 (Extremely). Inter-rater reliability
was .82, and the two judges’ ratings were then averaged.
Childhood Abuse and Transference 11
After learning about the new person, participants were instructed to visualize the
upcoming interaction with him/her, and again reported their present mood. The POMS-R
(McNair, et al., 1992) consists of emotion adjectives rated on five-point scales. We constructed
an index of dysphoric mood (alpha = .94) from the 40 items measuring Tension, Depression,
Anger, and Vigor (reversed).
Self-reported interpersonal responses. Participants rated 24 items about their responses to
the new person on 7-point scales (from “not at all” to “extremely”). A factor analysis of these
items with a Varimax rotation extracted 5 factors with eigenvalues greater than 1, which
corresponded to 5 a priori sets. Three factors were comprised of items from previous research
(see Andersen et al., 1996). We measured negative evaluation or dislike using 4 items (e.g.,
“How well do you think you will like this person?”) with an internal consistency of .87. They
loaded on a factor accounting for 9.8% of the total variance, with factor loadings of at least .70.
Our 8 items assessing motivation to avoid emotional closeness (e.g., “How much would you
want to share your feelings with this person?”) also had high internal consistency (alpha = .93).
With factor scores of .63 or higher, they loaded on a factor explaining 37.9% of the total
variance. We measured expectancy for rejection with 4 items (e.g., “How accepting do you think
this person would be of you?”) having an alpha of .88. They had loadings of .66 or higher on a
factor explaining 8.6% of the variance in our analysis.
Two other factors reflected sets of items developed for this research. We measured
mistrust using 4 items about the new person’s behavioral predictability and dependability
(Rempel, Holmes & Zanna, 1985), e.g., “How much could one feel certain of what this person
would do from one moment to the next?” and “How much does this person seem like someone
who could be counted on for comfort and understanding?” Internal consistency for these items
was .73, and they loaded with scores of .57 or higher on a factor explaining 5.4% of the total
variance. Finally, we measured emotional indifference to being liked/accepted to capture
interpersonal disengagement (see Dozier & Kobak, 1992). These 4 items (e.g., “How much
would it matter to you whether or not this person likes you?”) had an alpha of .75. With factor
loadings of .60 or higher, they comprised a factor accounting for 7.0% of the total variance.
Representation-derived inference / memory and final assessments. Participants rated their
confidence that each of 20 sentences had been presented earlier about the new person, using a 1-
to-4 scale. Sentence content was controlled across resemblance conditions: Participants in the
Childhood Abuse and Transference 12
parental resemblance condition rated paraphrases of their own parental descriptors, and those in
the control condition rated paraphrases of someone else’s parental descriptors. In both
conditions, these sentences included 8 that had actually been presented in describing the new
person (2 positive and 2 negative, ranked #5 and #6 in each list), plus 4 irrelevant fillers. The
remaining 12 sentences had not previously been presented and these included 6 fillers and 6
drawn from paraphrases of parental descriptors (ranked #1 #3 in each list). Recognition-
confidence for non-presented statements relevant to the parent is our basic index of inferences
biased by activation of the parental representation in transference.
Afterwards, participants listed their thoughts about the new person and the study’s
purpose. Because awareness of the link between the sessions could result in response bias, we
excluded participants who reported suspicions about it (5%: 4 abused, 3 nonabused). The
symptom checklist and sexual abuse screening (described earlier) were then completed, as were
measures beyond the scope of this paper. Finally, we debriefed participants and obtained their
consent to use the videotaped data.
Content of Parent Descriptions
The idiographic features presented about the new person were examined for possible
between-group differences in content, specifically, in terms of a) overall positive (versus
negative) valence; and b) interpersonal acceptance (versus rejection). Two judges blind to abuse
history independently rated each feature on two 9-point scales, with inter-rater reliabilities of .83
and .73 respectively. When the ratings were averaged across judges and analyzed in separate t-
tests with parent abusiveness as the independent variable, neither analysis reached significance
(t’s < 1.25). Therefore, statements derived from abusive versus nonabusive parents did not differ,
on average, in terms of their overall valence nor the interpersonal acceptance they conveyed. In
addition, no feature presented was explicitly abusive or particularly unique to abusive people. To
illustrate, the abusive parent features that coders rated as the least accepting included “is very
accusatory” and “is a controlling person.” A set of abusive parent features at the median
acceptance rating included: “does not have very good manners,” ”loves to watch a good
comedy,” “uses a lot of foul language,” “has a talent for cooking,” “enjoys going fishing on the
weekends,” and “has some rather childish tendencies.”
Activation of the Parental Representation
Childhood Abuse and Transference 13
The social-cognitive process of transference was predicted to occur based on a new
person’s manipulated resemblance to the participant’s own parent, regardless of abuse history.
As in prior transference research, we examined activation and use of the parental representation
to fill in the blanks about the new person by examining recognition-memory confidence for
representation-derived descriptors that had not actually been presented about the new person in a
2 x 2 x 2 x 2 (Abuse x Resemblance x Threat x Descriptor Valence) analysis of covariance
(ANCOVA). We controlled participant age in all our analyses to reduce noise variance due to
developmental changes in parent-child relations across the transition to young adulthood.
Recognition confidence for irrelevant non-presented items was another covariate to control
potential response bias, and positive/negative descriptor valence was a repeated measure.
Adjusted cell means from this analysis are shown in Table 2. The predicted main effect of
resemblance was the only significant effect to emerge, indicating greater recognition confidence
for descriptors derived from the participant’s own parent, M = 1.32, rather than someone else’s,
M = 1.22, F (1, 134) = 4.24, p < .05, d = .30.
Hence, regardless of abuse, parental resemblance
led the parental representation to be activated and used in making inferences about the new
person, as predicted.
Activation of the Relationship with the Parent: Interpersonal Responses
When the parental representation is activated in transference, aspects of the relationship
with the parent should be activated as well. Although positive significant-other representations
typically evoke positive interpersonal responses in transference, our primary hypothesis was that
such positive transference responses would be diminished or even reversed when the parent had
been abusive. To examine this, we created composites for each of the 5 factors that arose in our
factor analysis of participants’ interpersonal responses, by averaging the relevant items. As
shown in the top portion of Table 3, the moderate inter-correlations among these variables
suggest that they reflect conceptually related but distinct (non-redundant) constructs.
We analyzed each composite in a 2 x 2 x 2 (Abuse x Resemblance x Threat) analysis of
covariance (ANCOVA), with age as the covariate. Adjusted cell means from each analysis are
presented in Table 4. Significant and nearly significant interaction effects were further examined
with planned contrasts of adjusted marginal means, using the mean square error from the
omnibus analysis and two-tailed tests of statistical significance.
Rejection expectancy. As predicted, the ANCOVA with rejection expectancy as the
Childhood Abuse and Transference 14
dependent variable showed a significant two-way interaction between abuse and resemblance, F
(1, 135) = 4.09, p < .05, shown in Figure 1. Planned comparisons (MSE = .81) indicated that
when the parental representation was activated in transference, rejection expectancy was greater
among abused than nonabused participants, M = 3.92 vs. M = 3.34, t (135) = 2.65, p < .01, d =
.62. There was no significant difference in the control condition, M = 3.65 vs. M = 3.76, t <1,
ns., d = -.11. Parental resemblance in the new person also marginally reduced rejection
expectancy relative to the control condition among nonabused participants, t = -1.91, p < .06, d =
-.44, but not among abused participants, t = 1.23, ns, d = .28. Beyond this interaction, the only
other effect to arise revealed that the threat prime evoked more rejection expectancy than did the
no-prime condition, M = 3.86 vs. M = 3.47, F (1, 135) = 6.87, p < .05, d = .43.
Mistrust. When analyzing our mistrust composite, the predicted interaction between
abuse and resemblance was significant, F (1, 135) = 5.60, p < .05, as portrayed in Figure 2.
Comparisons (MSE = .88) showed that abused participants were significantly more mistrustful
than nonabused participants in transference, M = 4.33 vs. M = 3.72, t (135) = 2.74, p < .01, d =
.63, and no such difference was evident in the control condition, M = 3.88 vs. M = 4.01, t < 1,
ns, d = -.14. Further comparisons showed that transference led abused participants to become
more mistrustful, t (135) = 2.06, p < .05, d = .47, but did not do so for nonabused participants, t
(135) = -1.31, ns, d = -.30. Finally, the omnibus analysis revealed a main effect for threat prime
showing increased mistrust in this context relative to the no-prime condition, M = 4.21 vs. M =
3.77, F (1, 135) = 8.01, p < .01, d = .46. No other effects emerged.
Emotional indifference. As portrayed in Figure 3, the predicted interaction between abuse
and resemblance was significant, F (1, 135) = 4.40, p < .05. Planned comparisons (MSE = 1.25)
indicated that abused participants expressed more indifference than nonabused participants to
being liked by the new person in transference, M = 5.13 vs. M = 3.98, t (135) = 4.28, p < .001, d
= .97, whereas this difference was not significant in the control condition, M = 4.88 vs. M =
4.53, t (135) = 1.33, ns, d = .30. Nonabused participants also reported significantly less
emotional indifference as a function of parental resemblance in the new person, t (135) = -2.05, p
< .05, d = -.46, but abused participants did not, t < 1, ns, d = .21. Finally, the overall analysis also
yielded a main effect for abuse history, indicating that abused participants generally expressed
more emotional indifference than did nonabused participants, M = 5.00 vs. M = 4.25, F (1, 135)
= 16.02, p < .001, d = .63, although this was qualified by the interaction. Consistent with prior
Childhood Abuse and Transference 15
evidence that people abused in childhood are more interpersonally distant than others, these
results also support our hypothesis that transference plays a role in heightening this effect.
Dislike and motivation to avoid closeness. Evaluative and motivational responses to a
new person in transference typically parallel the participant’s self-reported regard for the
significant other in general samples (Andersen et al., 1996), and it was of interest to determine
whether these would be colored by abuse history. Both analyses yielded main effects for abuse
history consistent with other evidence of abused individuals’ diminished sociability. Abused
participants disliked the new person more than did nonabused participants, M = 3.40 vs. M =
3.06, F (1, 135) = 4.72, p < .05, d = .35, and also expressed more motivation to avoid closeness
to the person than did nonabused participants M = 4.59 vs. M = 4.19, F (1, 135) = 4.12, p < .05,
d = .34.
For the analysis of avoidance motivation no other effects emerged, but for dislike the 3-
way interaction, shown in Figure 4, was also significant, F (1, 135) = 6.01, p < .05. Planned
comparisons (MSE = .92) showed that in the no-prime resemblance condition, abused
participants disliked the new person significantly more than did nonabused participants, t (135) =
2.67, p < .01, d = .88, whereas no such pattern emerged in the no-prime control condition, t
(135) = -1.19, ns, d = -.38. In the context of the threat prime, there was no longer a significant
between-group difference in transference, t < 1, ns, d = .29, but a nearly significant difference in
the control condition showed more dislike among abused participants, t (135) = 1.90, p < .06, d
= .62.
Further contrasts examined effects of the experimental manipulations in the two groups
of participants. Nonabused participants in the no-prime condition showed the classic schema-
triggered evaluation effect in transference, in that they evaluated the new person somewhat more
positively when he/she resembled their own parent rather than someone else’s, t (135) = -1.96, p
< .06, d = -.63, but this effect disappeared in the threat prime condition, t (135) = 1.34, ns, d =
.44. Moreover, the threat prime led nonabused participants in transference to dislike the new
person significantly more, t (135) = 2.15 p < .05, d = .70, but did not do so for nonabused
participants in the control condition, t (135) = -1.13, ns, d = -.36. Abused participants showed a
quite different pattern. Transference increased their dislike for the new person in the no-prime
condition, t (135) = 1.96, p < .06, d = .63, but their evaluation of him/her was equivalently
negative in both threat prime cells, t < 1, ns, d = .11. The prime had no significant effect on
Childhood Abuse and Transference 16
abused participants’ evaluative responses in transference, t < 1, ns, d = .11, but did increase
dislike among these participants in the control condition, t (135) = 1.97, p < .06, d = .64. Thus,
nonabused participants showed the schema-triggered evaluation that typically occurs in
transference with positive significant others, although the context of threat interfered with it.
Abused participants, however, did not show this effect at all their dislike for the new person
increased whenever either parental resemblance or interpersonal threat was present.
Activation of the Relationship with the Parent: Implicit and Explicit Affect
Affect in immediate facial expressions. As shown in the lower portion of Table 3, the
pleasantness of participants’ facial expressions while encoding representation-derived features of
the new person showed significant negative correlations with the rejection expectancy, dislike,
avoidance motivation, and dysphoric mood they reported while anticipating an interaction with
him/her. Evidence that across the entire sample, immediate facial expressions were calibrated to
participants’ interpersonal reactions and emotional experiences suggests that they reflected an
implicit affective response that arises immediately and relatively automatically upon activation
of the significant-other representation.
Like their nonabused peers, abused participants might be expected to reveal positive
affect in their facial expressions while encoding the new person’s features in transference based
on the positive tone of the activated representation but they might also be expected to respond
with more wariness, therefore reducing or even reversing such an effect. To explore these
alternatives, we analyzed the pleasantness of participants’ facial responses to representation-
derived features and to the threat prime in separate 2 x 2 (Abuse x Resemblance) ANCOVAs.
(Threat is not a factor because when presented the prime always followed presentation of the
features, and no control cue was presented in the no-prime condition.) Facial responses to
irrelevant features and baseline mood (as well as age) were covariates in each analysis.
Adjusted cell means are shown in Table 5.
The analysis of responses to the new person’s features showed the predicted resemblance
effect, although it was only marginally significant, F (1, 127) = 3.46, p < .07, d = .21. No other
effects emerged. For both abused and nonabused participants, then, facial affect was somewhat
more positive to descriptors derived from the participant’s own parent, M = 3.65, rather than
someone else’s, M = 3.59. For participants in the threat prime condition, facial affect while
reading the statement about the new person’s alleged irritability was examined in the same 2 X 2
Childhood Abuse and Transference 17
ANCOVA. Only a main effect of resemblance emerged F (1, 58) = 4.59, p < .05, d = .49,
indicating significantly more positive reactions in the context of transference, M = 3.80 than in
the control condition, M = 3.60. Hence, immediate facial responses to the threat prime reflected
the positive tone of the parental representation in transference an effect not dampened among
abused participants, even though increasing irritability in their parent would perhaps have been a
signal for danger.
Self-reported dysphoric mood. To examine how participants’ mood states in transference
may vary with abuse history and interpersonal threat, we analyzed self-reported dysphoric mood
in a 2 x 2 x 2 (Abuse x Resemblance x Threat) ANCOVA with age and baseline mood as
covariates. Adjusted cell means are listed in Table 5. This analysis yielded a marginally
significant 2-way interaction between abuse history and threat prime, F (1, 129) = 3.53, p < .07,
but none of the corresponding contrasts reached statistical significance. The effect was also
qualified by the 3-way interaction shown in Figure 5, F (1, 129) = 3.92, p < .06, the only other
effect to emerge from the analysis.
Planned contrasts (MSE = 74.81) indicated that the dysphoria scores of abused and
nonabused participants differed from each other only in the context of transference. Specifically,
parental resemblance led abused participants to report more dysphoria than nonabused
participants in the absence of contextual threat, t (129) = 2.27, p < .05, d = .33, but less dysphoria
when threat was primed, although this reversal did not reach statistical significance, t (129) = -
1.60, ns, d = -.24. Dysphoria did not differ with abuse history in either control condition (t’s < 1).
Contrasts comparing resemblance effects for each group showed that transference
produced no change in dysphoria among nonabused participants (t’s <1). For abused participants,
however, transference led to a significant increase in dysphoria when threat was not primed, t
(129) = 2.02, p < .05, d = .30, and a nonsignificant decrease in dysphoria when threat was
primed, t (129) =-1.52, ns, d = -.23. Furthermore, among abused participants the threat prime
resulted in a significant decrease in dysphoria in the context of transference, t (129) = -2.71, p <
.01, d = -.40, but not in the control condition, (t < 1, ns, d = .13), and no such effect was evident
among nonabused participants (t’s < 1.25).
Transference Elicits Patterns Linked with Parental Abuse
As predicted, the process of transference occurred in both abused and nonabused
Childhood Abuse and Transference 18
participants when the parental representation was activated, as reflected in biased inferences and
immediate positive facial affect. Because activation of the parental representation should then
spread to aspects of the relationship between the parent and the self, abused and nonabused
participants’ responses to the new person were predicted to diverge in transference. Indeed, they
did -- in ways not evident in the control condition, and not accounted for by pre-existing mood.
With the exception of greater avoidance motivation among abused participants, interpersonal
patterns that differed on the basis of abuse history were not expressed as general traits but as
context-dependent. This is consistent, for example, with a cognitive-affective processing system
approach to personality (e.g., Mischel & Shoda, 1995) and also with our own conceptions of the
relational self (Andersen & Chen, 2002).
The process of transference heightened self-reported wariness among abused participants.
They came to mistrust and dislike the new person more in transference, and unlike their
nonabused peers, they showed no decrease in rejection expectancy and emotional dismissiveness
in this context. Wariness may well be adaptive for formerly abused individuals if it reduces their
likelihood of developing new attachments to abusive others, yet it may also be relevant to the
social and intimate relationship difficulties associated with a history of abuse. Readiness to
anticipate poor treatment, such as rejection or a breach of trust, has been shown to predict
ineffective and even aversive interpersonal behavior that in turn increases the likelihood that the
feared negative relationship outcomes will actually occur (e.g., Downey, Freitas, Michaelis, &
Khouri, 1998). Interpersonal disengagement also undermines social support and precludes
exposure to relationships that might challenge negative expectancies (Shields et al., 2001).
Yet despite the wariness they expressed, abused participants’ emotional responses in
transference were not consistently negative. Immediate facial expressions while learning about
the new person’s features, and then his/her irritable state of mind, became more positive under
conditions of parental resemblance than otherwise, both for abused and nonabused participants.
These results replicate previous evidence for automatic positive affect when a positive
significant-other representation is activated, and this prior research also suggests this would not
have occurred had the parent been considered a negative significant other (Andersen et al., 1996;
Andersen, Reznik, & Glassman, in press). The transference responses of abused participants
were therefore characterized by both negative and positive feelings, a classic definition of
ambivalence in the attitude literature. Observational studies of disorganized attachment (e.g.,
Childhood Abuse and Transference 19
Jacobvitz & Hazen, 1999) provide compelling examples of contradictory interpersonal responses
in the interactions of abused children with their caregivers and others, and the operation of
affectively complex mental representations has been posited to account for continuity in these
patterns across relationships. No prior study we know of, however, has demonstrated this in a
controlled experiment.
We do not argue that abuse played a causal role in the between-group differences
triggered in transference because abuse history is correlational, and cannot be teased apart from
other aspects of the lives in which it occurred. We instead argue that these data enable causal
conclusions about the effects of activating mental representations of attachment figures from
childhood in abused and nonabused participants, and thus contribute the first experimental
evidence for this as a mechanism eliciting and/or augmenting individual differences associated
with childhood abuse in adult relations. The effects observed were predominantly small to
moderate in effect size, but this is neither surprising nor detracts from their importance given the
subtle nature of the manipulations that produced them (see Prentice & Miller, 1992).
Responses to Interpersonal Threat in Transference
To the extent that the nature of a specific relationship varies with context, different
versions of the relationship may be elicited in transference depending on relevant cues in the
immediate situation. For this reason it seemed possible that an ordinary encounter with a new
person who shares features with the parent may not trigger the more threatening aspects of an
abusive parent-child relationship, and to facilitate activation of these aspects, we manipulated
whether or not the new person was said to be in an irritable, conflict-prone state of mind. For
several of our central dependent variables, this additional cue was apparently unnecessary for
evoking differences between abused and nonabused participants in transference. Rejection
expectancy, mistrust, and emotional indifference toward the new person were increased among
abused (vs. nonabused) participants purely as a function of activating the parental representation,
indicating that activation spread to negative aspects of the parent-child relationship even without
a more explicit contextual trigger. Dislike for the person also increased among abused
participants in transference and whenever they were told the person was in an irritable state of
mind (relative to the no-prime control condition). It is unclear whether this occurred because the
prime activated a general wariness about people in potentially volatile moods, or if it may even
have activated this aspect of the relationship with the parent in the absence of any other specific
Childhood Abuse and Transference 20
parental resemblance in the new person, an open question for future research.
Given the considerable negativity of abused participants’ self-reported interpersonal
responses in transference, their emotional reactions to the threat prime were intriguing. As noted,
for all participants, facial affect was more positive when reading this cue in transference than in
the control condition. Yet, only abused participants in the transference context reported
significantly less dysphoria after learning about an irritable person than a nonirritable one. A
possible interpretation of this finding, consistent with our evidence that transference heightened
the emotional indifference associated with abuse history, is that it facilitated well-practiced
affective responses to threat. For example, perhaps if participants had learned not to express
negative emotions so as not to provoke the parent in a volatile mood, or if they had coped by
disconnecting themselves emotionally, the combination of parental resemblance and threat-
relevant interpersonal context may have triggered similar responses. Another plausible
interpretation is that perhaps the mood of abused participants in transference was changed not by
the threatening quality of the primed interpersonal context but by its familiarity. Having a
familiar script for what to expect and how to behave when the parent is in an irritable mood may
have made this condition less aversive than the more ambiguous no-prime transference condition
for abused participants. Another possible explanation is that because the cue introduced the
immediate state of mind of the new person, it brought abused participants’ attention back to the
here-and-now, reducing the emotional hold of the past on the present in transference.
Due to the limitations of our study’s design, conclusions about this finding must remain
speculative. Our data do not address precisely by what means the change in self-reported mood
occurred, and having no control contextual cue in the no-prime condition leaves ambiguity about
the prime’s “active ingredients.” Because increased understanding of emotion regulation
processes among formerly abused people could have direct implications for clinical intervention,
this finding warrants further study. Additional research is also needed on precisely how
contextual factors interact with the process of transference in altering mood states (see also
Baum & Andersen, 1999).
The sample of college women included in the study presents a clear limitation for its
generalizability. Although we have no reason to expect that the central findings would have been
different if men had also been included, this is an open empirical question. The developmental
Childhood Abuse and Transference 21
significance of attachment or detachment in relation to an abusive parent may also vary with
gender if men are less likely than women are to internalize parental rejection at a cost to the self
(Berenson, Crawford, Cohen & Brook, 2005). Because perceptions of parents and parent-child
relationships should change with age and exposure to alternative interpersonal interactions with
significant others, restricted variability in age also limits generalizability. An older sample may
have less idealized views of their parents (Hazan & Shaver, 1987), and perhaps have developed
greater ability to interrupt ineffective transference patterns a matter of interest for future
investigation. The high functioning of this sample (attending college and able to complete both
sessions of the study) likely restricted variability in debilitating psychological disturbance, and
rates of psychiatric diagnoses and treatment are unknown. The observed transference responses
therefore may not necessarily generalize to individuals manifesting more severe clinical
outcomes associated with early abuse, such as those who might experience extremely intense
affective instability or dissociative states.
We focused on positively evaluated parent figures because we expected that the vast
majority of participants would evaluate their parents positively based on attachment needs, even
if mixed or negative feelings were also present. Both our pretest ratings of parent figures and
evidence for positive implicit affect in transference suggest that outright devaluation of a parent
is rare in these college students atypical even in those whom a parent abused, though it has
been shown that college students can readily identify negative significant others who are not
parents. Transference responses of individuals who blatantly dislike and reject the parent would
be expected to differ from those in the present sample (as in prior research on negative
significant others, e.g., Andersen et al., 1996). In short, it is to the population of quite high-
functioning, college-age women who continue to love their parent (even if abusive) that our
results should generalize.
Concluding comments
The present study demonstrated the role of transference in eliciting differences between
abused and nonabused participants in their responses to a new person. Although biased
inferences and automatic positive affect were evoked among both abused and nonabused
participants in transference, abused participants became more guarded than nonabused
participants in this context as well. Abused participants also reported complex mood changes in
transference, specifically, an increase in dysphoria that was markedly attenuated when the new
Childhood Abuse and Transference 22
person was described as being in an irritable, potentially volatile frame of mind. Even if they are
adaptive for a child attempting to preserve and endure a relationship with an abusive parent,
patterns like these may well be ineffective for negotiating personal and interpersonal needs in
ordinary encounters with others later in life, and the potential relationship of these patterns to
clinically-relevant difficulties warrants examination in future research. The first to examine the
transference process with regard to profoundly problematic relationships, this study highlights
the potential value of this line of work for understanding the pains and pleasures that past
significant relationships can bring when experienced in the present.
Childhood Abuse and Transference 23
Andersen, S. M., & Berk, M. S. (1998). Transference in everyday experience: Implications of
experimental research for relevant clinical phenomena. Review of General Psychology, 2,
Andersen, S.M. & Chen, S. (2002). The relational self: An interpersonal social-cognitive theory.
Psychological Review, 109, 619-645.
Andersen, S. M., Reznik, I., & Glassman, N.S. (in press). The unconscious relational self. In R.
Hassin, J.A. Bargh, & J.S. Uleman (Eds.), The new unconscious. New York: Oxford
Andersen, S. M., Reznik, I., & Manzella, L. M. (1996). Eliciting facial affect, motivation, and
expectancies in transference: Significant-other representations in social relations. Journal
of Personality and Social Psychology, 71, 1108-1129.
Anderson, N. H. (1968). Likability ratings of 555 personality-trait words. Journal of Personality
and Social Psychology, 9, 272-279.
Baldwin, M. W., Keelan, J. P. R., Fehr, B., Enns, V., & Koh-Rangarajoo, E. (1996). Social-
cognitive conceptualization of attachment working models: Availability and accessibility
effects. Journal of Personality and Social Psychology, 71, 94-109.
Baum, A. & Andersen, S. M. (1999). Interpersonal roles in transference: Transient mood states
under the condition of significant-other activation. Social Cognition, 17, 161-185.
Berenson, K.R., Crawford, T.N., Cohen, P. & Brook, J. (2005). Implications of identification
with parents and parents’ acceptance for adolescent and young adult self-esteem. Self and
Identity, 4, 289-301.
Berk, M. S., & Andersen, S. M. (2000). The impact of past relationships on interpersonal
behavior: Behavioral confirmation in the social-cognitive process of transference.
Journal of Personality and Social Psychology, 79, 546-562.
Bowlby, J. (1973). Attachment and loss: Vol.2. Separation. New York: Basic Books.
Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human development. New
York: Basic Books.
Briere, J., & Runtz, M. (1988). Multivariate correlates of childhood psychological and physical
maltreatment among university women. Child Abuse and Neglect, 12, 331-341.
Childhood Abuse and Transference 24
Bruner, J. S. (1957). Going beyond the information given. In H. E. Gruber, K. R. Hammond &
R. Jessor (Eds.), Contemporary approaches to cognition (pp. 41-69). Cambridge, MA:
Harvard University Press.
Carlson, V., Cicchetti, D., Barnett, D., & Braunwald, K. (1989). Disorganized/disoriented
attachment relationships in maltreated infants. Developmental Psychology, 25, 525-531.
Cicchetti, D. Toth, S. L. & Lynch, M. (1995). Bowlby’s dream comes full circle: The application
of attachment theory to risk and psychopathology. In T. H. Ollendick & R. J. Prinz
(Eds.), Advances in clinical child psychology (Vol. 17, pp. 1-75). New York: Plenum
Crittenden, P. M., & DiLalla, D. L. (1988). Compulsive compliance: The development of an
inhibitory coping strategy in infancy. Journal of Abnormal Child Psychology, 16, 585-
Dean, A. L., Malik, M. M., Richards, W., & Stringer, S. A. (1986). Effects of child maltreatment
on children’s conceptions of interpersonal relationships. Developmental Psychology, 22,
Derogatis, L. R. (1993). Brief symptom inventory (BSI): Administration, scoring, and
procedures manual. Minneapolis, MN: National Computer Systems, Inc.
Dodge, K. A., Pettit, G. S., Bates, J. E. (1994). Effects of physical maltreatment on the
development of peer relations. Development and Psychopathology, 6, 43-55.
Dodge, K. A., Pettit, G. S., Bates, J. E., & Valente, E. (1995). Social information-processing
patterns partially mediate the effect of early physical abuse on later conduct problems.
Journal of Abnormal Psychology, 104, 632-643.
Downey, G., Freitas, A. L., Michaelis, B., & Khouri, H. (1998). The self-fulfilling prophecy in
close relationships: Rejection sensitivity and rejection by romantic partners. Journal of
Personality and Social Psychology, 75, 545-560.
Dozier, M., & Kobak, R. R. (1992). Psychophysiology in attachment interviews: Converging
evidence for deactivating strategies. Child Development, 63, 1473-1480.
Feldman, S., & Downey, G. (1994). Rejection sensitivity as a mediator of the impact of
childhood exposure to family violence on adult attachment behavior. Development and
Psychopathology, 6, 231-247.
Gelles, R., & Straus, M. A. (1988). Intimate Violence. New York: Touchstone.
Childhood Abuse and Transference 25
George, C., & Main, M. (1979). Social interactions of young abused children: Approach,
avoidance, and aggression. Child Development, 50, 306-318.
Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal
of Personality and Social Psychology, 52, 511-524.
Jacobvitz, D. & Hazen, N. (1999). Developmental pathways from infant disorganization to
childhood peer relationships. In Solomon, J., & George, C. (Eds.), Attachment
disorganization (pp. 127-159). New York: Guilford Press.
Jaffe, P., Wolfe, D., Wilson, S., & Zak, L. (1986). Similarities in behavioral and social
maladjustment among child victims and witnesses to family violence. American Journal
of Orthopsychiatry, 56, 142-146.
Karylowski, J. J., Konarzewski, K., & Motes, M. (1999). Recruitment of exemplars as reference
points in social judgments. Journal of Experimental Social Psychology, 36, 275-303.
Knutson, J. F. (1995). Psychological characteristics of maltreated children: Putative risk factors
and consequences. Annual Review of Psychology, 46, 401-431.
Lyons-Ruth, K., Bronfman, E., & Atwood, G. (1999). A relational diathesis model of hostile-
helpless states of mind: Expressions in mother-infant interaction. In J. Solomon & C.
George (Eds.), Attachment disorganization (pp. 33-70). New York: Guilford Press.
Main, M., & George, C. (1985). Response of abused and disadvantaged toddlers to distress in
agemates: A study in the day care setting. Developmental Psychology, 21, 407-412.
Main, M., & Goldwyn. (1984). Predicting mothers rejection of her infant from representations of
her own experience. Child Abuse and Neglect, 8, 203-217.
Malinosky-Rummell, R., & Hansen, D. J. (1993). Long-term consequences of childhood physical
abuse. Psychological Bulletin, 114, 68-79.
McNair, D. M., Lorr, M., Droppleman, L. F. (1992). EdITS manual for the Profile of Mood
States, Revised. San Diego: Edits Educational and Industrial Testing Service.
Mischel, W. & Shoda, Y. (1995). A cognitive-affective system theory of personality:
Reconceptualizing situations, dispositions, dynamics, and invariance in personality
structure. Psychological Review, 102, 246-268.
Mueller, N., & Silverman, N. (1989). Peer relations in maltreated children. In D. Cicchetti & V.
Carlson (Eds.), Child maltreatment: Theory and research on the causes and
consequences of child abuse and neglect (pp. 529-578). New York: Cambridge
Childhood Abuse and Transference 26
University Press.
Murray, S. L., & Holmes, J. G., (1999). The (mental) ties that bind: Cognitive structures that
predict relationship resilience. Journal of Personality and Social Psychology, 77, 1228-
Prentice, D.A. & Miller, D.T. (1992). When small effects are impressive. Psychological Bulletin,
112, 160-164.
Rempel, J. K., Holmes, J. G., & Zanna, M. P. (1985). Trust in close relationships. Journal of
Personality and Social Psychology, 49, 95-112.
Shields, A., Ryan, R. M., & Cicchetti, D. (2001). Narrative representations of caregivers and
emotion dysregulation as predictors of maltreated children’s rejection by peers.
Developmental Psychology, 37, 321-337.
Solomon, J., & George, C. (1999). The place of disorganization in attachment theory: Linking
classic observations with contemporary findings. In J. Solomon & C. George (Eds.),
Attachment disorganization (pp. 3-32). New York: Guilford Press.
Straus, M. A. & Gelles, R. J. (Eds.). (1990). Physical violence in American families: Risk factors
and adaptations to violence in 8145 families. New Brunswick, NJ: Transaction
Toth, S. L., & Cicchetti, D. (1996). Patterns of relatedness, depressive symptomatology, and
perceived competence in maltreated children. Journal of Consulting and Clinical
Psychology, 64, 43-41.
van der Kolk, B. A. (1996). The complexity of adaptation to trauma: Self-regulation, stimulus
discrimination, and characterological development. In B. A. van der Kolk, A. C.
McFarlane, & L. Weisaeth (Eds.), Traumatic stress: The effects of overwhelming
experience on mind, body, and society (pp. 182-213). New York: Guilford Press.
Wolfe, D. A., & Wekerle, C., Reitzel-Jaffe, D., Lefebvre, L., (1998). Factors associated with
abusive relationships among maltreated and nonmaltreated youth. Development and
Psychopathology, 10, 61-85.
Childhood Abuse and Transference 27
Author Note
Kathy R. Berenson and Susan M. Andersen, Department of Psychology, New York
This research was part of a dissertation submitted by the first author to the Department of
Psychology, New York University.
For their assistance, many thanks to Lauren Ben-Ezra, Christina Bogiatzi, Natasha Bonin,
Regina Bordieri, Radoika Disla, Beth Gobeille, Ivona Huszcza, Lauren Irwin, Vera Jones,
Tatiana Kamorina, Devin McKinney, Shannon Pujadas, and Katharine Velez.
Address correspondence to Kathy Berenson, who is now at the Department of
Psychology, Columbia University.
Childhood Abuse and Transference 28
We used a median split for emotionally threatening incidents with the parent (79 for mothers
and 46 for fathers) to avoid restricting the nonabused group to extreme respondents reporting no
conflict or discipline experiences with either parent.
Because participants completed this measure after the experiment, we conducted an ANOVA to
determine if their symptom reports varied with conditions, and they did not (F’s < 1).
Three participants preselected for the abused group were excluded because they did not know
their abusive parent(s) well (in two cases, due to parental death or estrangement during
childhood). The other 9 excluded participants (8 abused, 1 nonabused) had evaluated their
relevant significant parent figure(s) negatively. Whereas a balanced design including both
positive and negative parent representations would have been ideal, the negative cells would
have been exceedingly difficult to fill in the population: Only 2.8% of nonabused participants’
parents, and 12.0% of abusive parents, were rated as negative significant others.
Among abused participants, age showed nearly significant associations with two dependent
variables of analyses to follow: self-reported emotional indifference (r=.22) and positive implicit
affect (r=.23).
Interestingly, although no effects involving feature valence were significant, the adjusted cell
means suggest that the effect of resemblance is largely carried by the positive features. Indeed,
when the positive features were analyzed in a separate 2 x 2 x 2 ANCOVA, the predicted effect
of resemblance emerged, own, M = 1.37 vs. yoked, M = 1.23, F (1, 134) = 4.08, p < .05, d=.32,
and no other effects were significant. The comparable analysis for negative features showed no
significant effects at all. Similar to patterns observed for positive significant others in previous
research (Andersen et al., 1996), the basic transference effect in this study appears to
predominantly involve positive representation-derived features both for abused and for
nonabused participants.
Plots of baseline mood showed highly skewed distributions and widely fluctuating standard
Childhood Abuse and Transference 29
deviations across cells; thus, to increase the stability of this covariate for our analysis, we
excluded observations more than 3 standard deviations above the grand mean (i.e., mood scores
of 3 abused and 2 nonabused participants) for analyzing these data. Further analyses examined
whether or not inclusion of individuals with extreme baseline dysphoria scores influenced
interpersonal responses to the new person. First, we repeated the previously reported analyses
with these 5 participants excluded, and then again with baseline mood covaried out. Both times,
all previously reported main and interaction effects of resemblance in the omnibus analyses
showed similar significance levels and patterns of mean differences as before.
Childhood Abuse and Transference 30
Table 1
Mean Incidents of Parental Behavior toward Participants by Abuse History Groups
Severe Physical Violence
by relevant parent
by both parents combined
Emotionally Threatening Behavior
by relevant parent
by both parents combined
Childhood Abuse and Transference 31
Table 2
Adjusted Means and Standard Errors for Recognition Confidence for Representation-Derived Non-Presented Features
Participant’s Own
Yoked Participant’s
M (SE)
M (SE)
M (SE)
M (SE)
Positive features
Threat prime
1.46 (.10)
1.27 (.10)
1.21 (.10)
1.20 (.10)
No prime
1.37 (.10)
1.40 (.10)
1.32 (.10)
1.19 (.10)
Negative features
Threat prime
1.39 (.09)
1.21 (.09)
1.10 (.09)
1.19 (.09)
No prime
1.20 (.09)
1.25 (.09)
1.27 (.09)
1.26 (.09)
Note: Cell means are adjusted for age and recognition confidence for irrelevant non-presented features
Childhood Abuse and Transference 32
Table 3
Correlations among Interpersonal and Affective Responses
Interpersonal response variables
1. Rejection expectancies
2. Mistrust
.40 ***
3. Indifference
.27 **
4. Dislike
.51 ***
.20 *
.28 **
5. Avoidance
.53 ***
.33 ***
.29 ***
.62 ***
Affect variables
6. Positive facial response to
representation-derived features
-.21 *
-.22 **
-.22 **
7. Dysphoric mood
.37 ***
.19 *
.22 **
-.21 *
* significant at p<.05, ** p<.01, *** p<.001
Childhood Abuse and Transference 33
Table 4
Adjusted Means and Standard Errors for Interpersonal Responses
Participant’s Own
Yoked Participant’s
M (SE)
M (SE)
M (SE)
M (SE)
Rejection expectancy
Threat prime
4.12 (.21)
3.65 (.21)
3.86 (.21)
3.82 (.21)
No prime
3.71 (.21)
3.03 (.22)
3.45 (.21)
3.69 (.21)
Threat prime
4.55 (.22)
4.05 (.22)
3.91 (.22)
4.33 (.22)
No prime
4.12 (.22)
3.39 (.22)
3.85 (.22)
3.70 (.22)
Emotional indifference
Threat prime
5.25 (.26)
4.13 (.26)
4.92 (.26)
4.71 (.27)
No prime
5.00 (.27)
3.83 (.27)
4.84 (.26)
4.35 (.26)
Threat prime
3.64 (.23)
3.35 (.23)
3.53 (.23)
2.92 (.23)
No prime
3.53 (.23)
2.66 (.23)
2.90 (.23)
3.28 (.23)
Avoidance motivation
Threat prime
4.81 (.28)
4.53 (.28)
4.58 (.28)
4.29 (.28)
No prime
4.73 (.28)
3.78 (.28)
4.24 (.28)
4.15 (.28)
Note: Cell means are adjusted for age.
Childhood Abuse and Transference 34
Table 5
Adjusted Means and Standard Errors for Implicit and Explicit Affect
Participant’s Own
Yoked Participant’s
M (SE)
M (SE)
M (SE)
M (SE)
Positive facial affect
representation-derived features
3.66 (.03)
3.65 (.03)
3.59 (.03)
3.60 (.03)
threat cue (in prime condition)
3.81 (.09)
3.79 (.09)
3.61 (.09)
3.59 (.10)
Self-reported dysphoric Mood
Threat prime
40.69 (2.13)
45.45 (2.05)
45.23 (2.10)
46.43 (2.19)
No prime
48.64 (2.05)
41.98 (2.07)
42.70 (2.10)
44.12 (2.04)
Note: All cell means are adjusted for age and baseline mood; facial affect means are also adjusted for facial affect in response to
irrelevant features.
Figure Captions
Figure 1. Adjusted mean expectancy for rejection by the new person as a function of abuse
history and parental resemblance.
Figure 2. Adjusted mean mistrust in the new person as a function of abuse history and parental
Figure 3. Adjusted mean emotional indifference about being liked/accepted by the new person as
a function of abuse history and parental resemblance.
Figure 4. Adjusted mean dislike for the new person as a function of abuse history, parental
resemblance and threat prime
Figure 5. Adjusted mean self-reported dysphoric mood as a function of abuse history, parental
resemblance and threat prime
Figure 1. Adjusted mean expectancy for rejection by the new person as a function of abuse
history and parental resemblance.
Figure 2. Adjusted mean mistrust in the new person as a function of abuse history and parental
Figure 3. Adjusted mean emotional indifference about being liked/accepted by the new person as
a function of abuse history and parental resemblance.
Figure 4. Adjusted mean dislike for the new person as a function of abuse history, parental
resemblance and threat prime
Figure 5. Adjusted mean self-reported dysphoric mood as a function of abuse history, parental
resemblance and threat prime
... Specifically, childhood maltreatment may increase negative attachment representations (i.e., beliefs and expectations) of the self and close others, such as beliefs that oneself is worthless, "bad," and unlovable, and expectations that others will be rejecting, unsupportive, untrustworthy, or aggressive (Wright, Crawford, & Del Castillo, 2009). Indeed, experimental research finds that compared to nonabused women, women with histories of childhood maltreatment report greater rejection expectancy and mistrust in response to a new person who poses a mild interpersonal threat (Berenson & Andersen, 2006). ...
... This study is among the first to demonstrate that the association between childhood maltreatment and prenatal psychopathology is mediated by romantic relationship quality. These results align with previous research findings that childhood maltreatment increases risk for negative interpersonal expectations and involvement in low-quality romantic relationships (Berenson & Andersen, 2006;Colman & Widom, 2004). It will be critical for future research to directly assess women's attachment-related beliefs and expecta- This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. ...
... In a seminal program of research, Andersen and Chen (2002) and Andersen and Przybylinski (2012) showed that transference is quite common and has important downstream consequences on interpersonal perception and interactions, in some cases leading to a self-fulfilling prophecy in which individuals elicit responses consistent with the important-other representation (e.g., Berk & Andersen, 2000). Of particular relevance for this review, it has also been shown that transference can occur for more mixed or ambivalent relationships (Berenson & Andersen, 2006;Berk & Andersen, 2008). In these studies, individuals who experienced transference stemming from an ambivalent relationship (e.g., loved parent but a history of abuse) showed signs of positivity (e.g., implicit affect, affection seeking) toward the relevant new relationship but also viewed that person more negatively (rejecting, less trust; Berenson & Andersen, 2006;Berk & Andersen, 2008). ...
... Of particular relevance for this review, it has also been shown that transference can occur for more mixed or ambivalent relationships (Berenson & Andersen, 2006;Berk & Andersen, 2008). In these studies, individuals who experienced transference stemming from an ambivalent relationship (e.g., loved parent but a history of abuse) showed signs of positivity (e.g., implicit affect, affection seeking) toward the relevant new relationship but also viewed that person more negatively (rejecting, less trust; Berenson & Andersen, 2006;Berk & Andersen, 2008). ...
The protective influence of social relationships on health is widely documented; however, not all relationships are positive, and negative aspects of relationships may be detrimental. Much less is known about the relationships characterized by both positivity and negativity (i.e., ambivalence). This article provides a theoretical framework for considering the influence of ambivalent relationships on physical health, including reasons why ambivalence should be considered separately from relationships characterized as primarily positive (supportive) or primarily negative (aversive). We introduce the social ambivalence and disease (SAD) model as a guide to understanding the social psychological antecedents, processes, and consequences of ambivalent relationships. We conclude by highlighting gaps in the literature and features of the SAD model that may serve as a guide to future research on potential health-relevant pathways of ambivalent relationships.
... In contrast with social support, individuals who experience significant negative social experiences (e.g. abuse, undermining), sometimes referred to as social pain or social stress, often report higher levels of somatic symptoms including physical pain (Berenson and Andersen, 2006;MacDonald and Leary, 2005;Tchalova and Eisenberger, 2016). Among negative social experiences, social undermining is distinctive, described as perceived aggressive or threatening behavior toward a person that results from maligned social interactions or chronic social conflict within interpersonal relationships (Abbey et al., 1985). ...
... For example, social undermining could be especially potent for those who have experienced significant social or physical pain early in life (e.g. child abuse, neglect, or trauma; Berenson and Andersen, 2006). As women offer greater support to their social networks compared to men (Taylor, 2011), women may consequently experience greater psychological and physical effects from negative social interactions like undermining. ...
Women may be disproportionately vulnerable to acute pain, potentially due to their social landscape. We examined whether positive and negative social processes (social support and social undermining) are associated with acute pain and if the processes are linked to pain via negative cognitive appraisal and emotion (pain catastrophizing, hyperarousal, anger). Psychosocial variables were assessed in inner-city women (N = 375) presenting to an Emergency Department with acute pain. The latent cognitive-emotion variable fully mediated social undermining and support effects on pain, with undermining showing greater impact. Pain may be alleviated by limiting negative social interactions, mitigating risks of alternative pharmacological interventions.
... Duygusal istismar bireyin olumsuz benlik algısı geliştirmesine yönelik etkisi genel bağlamda olumsuz algıların yerleşmesine de zemin hazırlayabilmekte (Goldsmith, & Freyd, 2005), duygusal istismara maruz kalan birey istismar içermeyen bir yaşantıyı bile istismarmış gibi değerlendirip, sosyal ilişkilerini bu doğrultuda düzenleme yoluna gidebilmektedir ve bu nedenle bireyin algıladığı duygusal istismar düzeyi, hayatında kuracağı ilişkilerin yapısını ve içeriğini etkileyen önemli etkenlerden birini oluşturmaktadır (Wright, Crawford, & Del Castillo, 2009). Duygusal istismara maruz kalan bireyler gündelik yaşamlarında alışılagelmiş durumlara yönelik daha olumsuz yorumlar yapabilme potansiyeline sahip olabilmektedirler (Berenson & Andersen, 2006). Çalışmalar kadınların gebelik öncesi dönemin yanı sıra gebelik dönemlerinde de duygusal istismara maruz kaldıklarını ve hatta duygusal istismarın gebelik döneminde fiziksel ve cinsel istismara kıyasla arttığını göstermektedir. ...
Conference Paper
Full-text available
Bu araştırmanın amacı, gebelerin doğum korkusu tutumlarını belirlenmesi, bu tutumlar üzerinde duygusal istismar ve kaygı faktörlerinin olası etkilerinin belirlenmesidir. Bu amaçla tanımlayıcı ve ilişki arayıcı araştırma özelliğinde şekillendirilen bu çalışmada sosyal medya aracılığı ile ulaşılan ve araştırmaya katılmaya gönüllü olan 280 gebeden, tanıtıcı bilgi formu, Wijma Doğum Beklentisi/Deneyimi Ölçeği (W-DEQ)-A, Beck Anksiyete Ölçeği ve Yetişkin Bireyler için Algılanan Duygusal İstismar Ölçeğinden oluşan ölçek formunu doldurmaları istenmiş, elde edilen veriler SPSS.20 programında çoklu regresyon analizi ile analiz edilmiştir. Tanıtıcı bilgi formunda gebelerin sosyo-demografik özelliklerinin yanı sıra daha önce var olan doğum, düşük, ölü doğum gibi deneyimlerine de yer verilen 24 soru bulunmaktadır. Yapılan analiz sonucunda kaygı ve duygusal istismar değişkenlerinin birlikte gebelerin doğum korkusunun %16 oranında yordadığı görülmüştür. Katılımcılardan daha önce doğum tecrübesi olmuş olan 155 gebenin daha önceki doğum tecrübeleri ve şu anki doğum kaygıları incelendiğinde 155 gebenin 79’u şu anki gebeliklerine dair doğum korkuları olduğunu dile getirirken 76’sı doğum korkuları olmadığını dile getirmişlerdir. Daha önce doğum tecrübesi olmuş grubun Wijma Doğum Beklentisi/Deneyimi Ölçeği (W-DEQ)-A, Beck Anksiyete Ölçeği ve Yetişkin Bireyler için Algılanan Duygusal İstismar Ölçeklerinden aldıkları toplam puanlar ile yapılan çoklu regresyon analizi sonucunda bu grup için duygusal istismar ve anksiyetenin doğum korkusunun %18 sini açıkladığı görülmüştür. Katılımcılar daha önceki doğum deneyimleri hakkında acı, mutluluk, heyecan, kaygı, stres, çaresizlik, korku ve ağrı duygularını tanımlarken çoğunlukla normal doğum planlarken sezaryen yöntemi ile doğumu gerçekleştirmiş olmanın verdiği hayal kırıklığından bahsetmişlerdir. Katılımcıların doğum süreci ve sonrasında yaşadıkları deneyime dair oluşturdukları yapıların doğum şekli, hastane personeli ve yakınlarının desteği ile şekillendiği görülmüştür.
... Previous work has typically operationalized ambivalence as a series of alternating positive and negative states (e.g., Mikulincer, Shaver, Bar-on, & Ein-dor, 2010), as incongruity between emotions and behaviors (e.g., Berenson & Andersen, 2006), or as incongruity between explicit and implicit evaluations (e.g., Petty, Tormala, Briñol, & Jarvis, 2006). Although these definitions reflect various ambivalent states, they do not reflect a situation in which the same stimulus simultaneously automatically triggers two different and opposing evaluations (Figure 1c). ...
Full-text available
Despite the rich literature on implicit partner evaluations, there has been scant attention to a defining feature of significant other mental representations—their affective complexity. Recent findings (Zayas & Shoda, 2015), however, provide an empirical demonstration that significant others automatically and simultaneously activate positive and negative evaluations—a phenomenon we refer to as implicit ambivalence. A primary aim of this paper is to extend extant theory by elaborating on the features of the dyadic context that may contribute to the formation of implicit ambivalence. Particularly, drawing from research from relationship science, social cognition, and social neuroscience, we focus on the ability of significant others to dynamically and simultaneously confer rewards and threats, the attunement of perceivers to potential social rewards and social threats, and aspects of sense-making of another person's mind that may give rise to implicit ambivalence. From this new perspective, implicit ambivalence is not a pathological or rare state. Quite the opposite, implicit ambivalence may be a normative, typical process, that is triggered even by people who are highly positive in one's network. We identify future directions for social cognition and relationship science.
... Theory (Cacioppo & Berntson, 1994;Norris, Gollan, Berntson, & Cacioppo, 2010) and empirical studies (Ersner-Hershfield et al., 2008;Griffin & Sayette, 2008;Harris & Alvarado, 2005;Larsen, McGraw, Mellers, & Cacioppo, 2004;Zayas & Shoda, 2015) indicate that stimuli can activate positive and negative affect in healthy individuals. However, greater activation of negative and positive emotions in response to a given stimulus has been observed among individuals with a history of abuse (Berenson & Andersen, 2006), borderline personality disorder (Ebner-Priemer et al., 2008), and schizophrenia (Cohen et al., 2010;Kring & Elis, 2013). Current findings, then, suggest that those with high levels of positive urgency respond to positively valenced stimuli in a manner that is parallel to responses observed in those with psychopathology. ...
Positive urgency, defined as a tendency to become impulsive during positive affective states, has gained support as a form of impulsivity that is particularly important for understanding psychopathology. Despite this, little is known about the emotional mechanisms and correlates of this form of impulsivity. We hypothesized that positive urgency would be related to greater emotional reactivity in response to a positive film clip. Seventy-five undergraduates watched a positive film clip, and a multimodal assessment of emotion was conducted, including subjective emotional experience, physiological activation (i.e., heart rate, respiratory sinus arrhythmia, skin conductance), and facial emotional behavior (i.e., objectively coded using the Facial Action Coding System). Positive urgency was not significantly related to greater positive emotional reactivity but rather a more complex array of emotions expressed in facial behavior, as indexed by similar levels of positive yet greater levels of negative behavior. These findings show that positive urgency may be linked to altered emotionality, but does not appear related to heightened positive emotional reactivity. Potential implications for functional outcomes are discussed.
Purpose While within-firm service failure and recovery have been studied extensively, the context in which a service failure at one firm “spills over” and provides an opportunity for another entity, an external firm, to recover a customer has received limited attention. This study aims to examine how the extent of a service failure plays a role in how external firms should shape their recovery efforts, and how customers’ evaluations of the recovering firm and their feelings of unhappiness are affected. Design/methodology/approach A pretest conducted on MTurk gauged participants’ perceptions of equitability of the external firm’s recovery effort. In the main study, a 3 × 3 between-subjects experiment examined the effects of failure extent and external recovery type on evaluations of the recovering firm and reduced feelings of unhappiness. Findings It is found that equity judgments remain consistent in the external recovery context; transferred negative affect is able to be mitigate only in low-failure scenarios, and customers’ evaluations of the external firm increase only in high-failure scenarios. Research limitations/implications The use of hypothetical scenarios, as opposed to the employment of a field study, is the primary limitation of the study. Originality/value This research finds that external firms can reap the benefits of another firm’s service failure by offering no-cost recoveries, rather than ones that carry some form of cost.
Objective People with Borderline Personality Disorder (BPD) may experience heightened rejection sensitivity (RS), a disposition developing from repeated childhood rejecting experiences. It is not known whether the full RS model accounts for the cognitive–affective experiences common in BPD. This systematic review extends upon previous reviews, firstly by assessing the link between childhood rejecting experiences and adult RS, and secondly by considering the link between BPD and RS in both non‐clinical and clinical samples. Method Two research questions were devised, and searches based on predetermined criteria were conducted using PsycNET, PubMed, SCOPUS, and Web of Science. Data were extracted by one researcher and 20% was inter‐rated, with high levels of agreement. Forty‐three papers were systematically reviewed, and 31 included in meta‐analysis and meta‐regression. Results Studies assessing the link between childhood rejection and RS are limited; however, emotional abuse and neglect appears linked with RS. Pooled effect sizes suggest RS is linked with BPD (r = .326), with strong effect sizes when comparing clinical and control samples (r = .655). Qualitative synthesis suggests this may be mediated by executive control, although further research is required. The small number of studies considering the full RS model with regard to BPD suggests the interaction between emotional abuse and neglect affects rejection sensitivity; however, outcomes are inconsistent. Conclusions Childhood rejection, particularly emotional abuse and neglect, appears to be linked to rejection sensitivity, and rejection sensitivity is linked to BPD. However, this may not be linear. Implications for clinical practice and research are discussed. Practitioner points • Rejection sensitivity is consistently linked with BPD, in clinical and non‐clinical samples. Supporting mentalization or improved theory of mind may offer a therapeutic target for this disposition. • Considering the causes and effects of rejection sensitivity may offer a non‐blaming explanation of interpersonal difficulties in BPD and could be utilized as part of formulation and the therapeutic relationship. • However, the possible interaction between emotional abuse and neglect and rejection sensitivity suggests rejection sensitivity is not always apparent for people with BPD. Idiosyncratic formulation should consider this. • The literature included in the review is limited to Western populations with a high proportion of females, which may limit generalizability. • Measures of rejection sensitivity included in the review were restricted to self‐report, which may be subject to bias. Furthermore, measures of childhood rejection were retrospective in nature due to the exclusion of child samples. Further research should consider longitudinal and observational study designs.
Close relationships afford us opportunities to create and maintain meaning systems as shared perceptions of ourselves and the world. Establishing a sense of mutual understanding allows for creating and maintaining lasting social bonds, and as such, is important in human relations. In a related vein, it has long been known that knowledge of significant others in one's life is stored in memory and evoked with new persons-in the social-cognitive process of 'transference'-imbuing new encounters with significance and leading to predictable cognitive, evaluative, motivational, and behavioral consequences, as well as shifts in the self and self-regulation, depending on the particular significant other evoked. In these pages, we briefly review the literature on meaning as interpersonally defined and then selectively review research on transference in interpersonal perception. Based on this, we then highlight a recent series of studies focused on shared meaning systems in transference. The highlighted studies show that values and beliefs that develop in close relationships (as shared reality) are linked in memory to significant-other knowledge, and thus, are indirectly activated (made accessible) when cues in a new person implicitly activate that significant-other knowledge (in transference), with these shared beliefs then actively pursued with the new person and even protected against threat. This also confers a sense of mutual understanding, and all told, serves both relational and epistemic functions. In concluding, we consider as well the relevance of co-construction of shared reality n such processes.
We propose that the self, as experienced in the moment, is embedded in interpersonal contexts and thus varies across contexts, in terms of relational selves stored in memory. When representations of significant others and the relationship one has with them are cued and thus automatically activated, the self shifts in how it is experienced and expressed. Research on the relational self, involving both the social-cognitive process of transference and simple priming, demonstrates the ways in which variability in the self arises based on pre-existing memory and specific interpersonal cues, and speaks as well to how the self nonetheless remains phenomenally coherent. We propose that this phenomenal coherence in the self is tied to longstanding, personally meaningful knowledge of self in relation to various significant others, just as variability is defined in part by these prior relationships as activated in context. We propose that the self, as experienced in the moment, is embedded in interpersonal contexts and thus varies across contexts, in terms of relational selves stored in memory. When representations of significant others and the relationship one has with them are cued and thus automatically activated, the self shifts in how it is experienced and expressed. Research on the relational self, involving both the social-cognitive process of transference and simple priming, demonstrates the ways in which variability in the self arises based on pre-existing memory and specific interpersonal cues, and speaks as well to how the self nonetheless remains phenomenally coherent. We propose that this phenomenal coherence in the self is tied to longstanding, personally meaningful knowledge of self in relation to various significant others, just as variability is defined in part by these prior relationships as activated in context.
This study examined whether maltreated children were more likely than nonmaltreated children to develop poor-quality representations of caregivers and whether these representations predicted children's rejection by peers. A narrative task assessing representations of mothers and fathers was administered to 76 maltreated and 45 nonmaltreated boys and girls (8-12 years old). Maltreated children's representations were more negative/constricted and less positive/coherent than those of nonmaltreated children. Maladaptive representations were associated with emotion dysregulation, aggression, and peer rejection, whereas positive/coherent representations were related to prosocial behavior and peer preference. Representations mediated maltreatment's effects on peer rejection in part by undermining emotion regulation. Findings suggest that representations of caregivers serve an important regulatory function in the peer relationships of at-risk children.
An attachment theory framework is applied toward understanding the emergence of depressive symptomatology and lower perceived competence in maltreated and nonmaltreated children. Hypotheses that maltreated children with nonoptimal patterns of relatedness evidence elevated depressive symptomatology and lower competence, whereas nonmaltreated children with optimal or adequate patterns of relatedness exhibit the least depressive symptomatology and higher competence, were confirmed. Additionally, differentiations between maltreated children with and without optimal or adequate patterns of relatedness emerged, suggesting that relatedness may mitigate against the adverse effects of maltreatment. Moreover, sexually abused children with confused patterns of relatedness evidenced clinically significant depressive symptomatology. Results are discussed with regard to mechanisms that contribute to adaptation or maladaptation in children with negative caregiving histories.
In formulating his theoretical perspective on the development of human attachment relationships, Bowlby incorporated knowledge from a variety of disciplines, viewpoints, and research paradigms (Ainsworth & Bowlby, 1991; Bretherton, 1992). Psychoanalysis, object relations theory, Sullivanian interpersonal psychiatry, social, experimental and developmental psychology, evolutionary theory, and ethology all exerted major impacts on Bowlby’s hypotheses regarding the origins, course, and sequelae of secure and insecure attachment relationships (Ainsworth, 1967, 1969; Ainsworth, Blehar, Waters, & Wall, 1978; Bowlby, 1969/1982, 1973, 1980; Bretherton, 1987, 1991).
This article explores the possibility that romantic love is an attachment process--a biosocial process by which affectional bonds are formed between adult lovers, just as affectional bonds are formed earlier in life between human infants and their parents. Key components of attachment theory, developed by Bowlby, Ainsworth, and others to explain the development of affectional bonds in infancy, were translated into terms appropriate to adult romantic love. The translation centered on the three major styles of attachment in infancy--secure, avoidant, and anxious/ambivalent--and on the notion that continuity of relationship style is due in part to mental models (Bowlby's "inner working models") of self and social life. These models, and hence a person's attachment style, are seen as determined in part by childhood relationships with parents. Two questionnaire studies indicated that relative prevalence of the three attachment styles is roughly the same in adulthood as in infancy, the three kinds of adults differ predictably in the way they experience romantic love, and attachment style is related in theoretically meaningful ways to mental models of self and social relationships and to relationship experiences with parents. Implications for theories of romantic love are discussed, as are measurement problems and other issues related to future tests of the attachment perspective.
Research has demonstrated that transference occurs in social perception - defined in terms of the activation and application of a significant-other representation to a new person - using memory confidence and evaluation as indices (e.g., Andersen & Baum, 1994; Andersen, Reznik, & Manzella, 1996). The present research examined interpersonal roles in transference, and the notion that transient mood in transference may be predicted by one's interpersonal role with the significant other and its congruence or incongruence with the new person's role. In a combined idiographic-nomothetic design, participants learned about a new person characterized by features descriptive of their own positively toned significant other or that of a yoked participant's. Importantly, this new person was cast in an interpersonal role congruent or incongruent with the significant other's. Given the positive significant-other relationship, we predicted that role congruence would be associated with positive affect and role incongruence with negative affect, and hence that this pattern should emerge in transference. Results confirmed this prediction. Participants' transient mood was relatively more positive (nondepressive) when the target had resembled their own significant other and occupied a congruent versus incongruent role, and was clearly negative (depressive) based on role incongruence. Memory confidence and evaluation effects verified that transference was triggered in the significant-other resemblance condition, and thus that interpersonal roles predicted self-reported mood in transference. Implications for self-other relations and relational schemas are discussed.