<< Back  

 Can't open the PDF? Click here for help.


Empathy and childhood maltreatment: A mixed-methods investigation

Simon C. Locher, B.SocSci. (Hons)

Department of Psychology, University of Cape Town, Cape Town, South Africa

Lisa Barenblatt, B.SocSci. (Hons)

Department of Psychology, University of Cape Town, Cape Town, South Africa

Melike M. Fourie, PhD

Department of Psychology, University of the Free State, Bloemfontein, South Africa

Dan J. Stein, MD, PhD

Department of Psychiatry, University of Cape Town, Cape Town, South Africa

Pumla Gobodo-Madikizela, PhD

Department of Psychology, University of the Free State, Bloemfontein, South Africa

BACKGROUND: Impaired empathy is regarded as a psychological consequence of childhood maltreatment, yet few studies have explored this relationship empirically. We investigated whether empathy differed in healthy and maltreated individuals by examining their emotional responses to people in distress.

METHODS: Forty-nine individuals (age 20 to 60) viewed short film clips from the South African Truth and Reconciliation Commission testimonies depicting dialogues between victims and perpetrators of gross human rights violations. Participants were divided into 3 groups based on their scores on the Childhood Trauma Questionnaire: control (n = 18), moderate maltreatment (n = 21), and severe maltreatment (n = 10). We employed a mixed-methods design to explore empathic responses to film clips both quantitatively and qualitatively.

RESULTS: Quantitative results indicated that self-reported empathy was lower in the moderate maltreatment group compared to the control group, but of similar strengths in the severe maltreatment and control groups. However, qualitative thematic analysis indicated that both maltreatment groups displayed themes of impaired empathy.

CONCLUSIONS: Our results support the notion that childhood maltreatment is associated with impaired empathy, and suggest that such impairment may differ depending on the level of maltreatment: moderate maltreatment was associated with emotional blunting and impaired cognitive empathy, whereas severe maltreatment was associated with emotional over-arousal and diminished cognitive insight.

KEYWORDS: empathy, child maltreatment, emotion regulation, attachment



Early family relations are arguably the most important, and most enduring social relationships that affect a child’s development,1 yet millions of children across the world are exposed to negative early interactions, including maltreatment.2,3 Although the prevalence of childhood maltreatment is high across the world, it is particularly prevalent in South Africa.4

Childhood maltreatment is associated with a variety of adverse psychological effects among adults,5 one reason being that maltreated children often grow up in an environment that fails to provide appropriate opportunities that guide development.6,7 A psychological ability that has been implicated consistently in descriptions of impairments due to maltreatment, is empathy.1,8,9 Notably, maltreated children tend to have difficulty sharing with, caring for, or taking the perspectives of others.10 Moreover, maltreating parents often have impaired empathy themselves.11

Impaired empathy is a trait impacting most important relationships, and one that may be modified by therapy. Understanding the role that maltreatment contributes to the development of impaired empathy is critically important, and has significant clinical implications. In fact, various clinical disorders with childhood maltreatment as a major etiological factor have impaired empathy as a core symptom (eg, Cluster B personality disorders and complex posttraumatic stress disorder).12,13


Empathy has been defined as the capacity to experience and understand another’s emotional state in a manner that is congruent with the other person’s emotions.14,15 A long-standing tradition in psychology divides empathy into 2 components: affective and cognitive.16 The affective component constitutes the sharing of another person’s emotions via internal sensory representations of their emotions. Based on a widely accepted framework, Bateson and Ahmad17 divide affective empathy into: 1) emotion matching, which is the phenomenon where one mirrors the emotional state displayed by another; and 2) empathic concern, which is essentially sympathy. The cognitive component of empathy reflects the ability to infer and understand the mental state of another.18,19 These subcomponents of empathy are partially distinct, but also highly interconnected. For example, emotion matching may make the suffering of the other more salient, which in turn may trigger empathic concern and cognitive empathy to imagine how the other feels.20

Although some subcomponents of empathy, like emotion matching, are present from birth, more mature forms of cognitive empathy, like the capacity for mentalization, develop later in childhood and adolescence.21 The capacity to mentalize, or reflectively engage with another’s thoughts and feelings, is an ability that is rooted in the child’s early attachment experiences with parents.22,23 Empathic parents have the capacity to observe the child’s mind, to understand and contain their mental state, and to view the child as an intentional being, which promotes the child’s understanding of minds.24-26

Child maltreatment and empathic development

Although impaired empathy is cited commonly in the literature as a consequence of child maltreatment,27-29 we found little empirical evidence that directly investigated this relationship. Most of the existing literature is theoretical in nature, in which the fields of social and developmental psychology,1,21 as well as neuroscience,22 describe conceptual models that explore the role of maltreatment in contributing to the development of impaired empathy. Often these theories are based on observations from clinical and case studies only.

Substantial indirect evidence points to a link between maltreatment and empathy, however. In particular, many of the factors considered necessary for the development of empathy are impaired in children who experience maltreatment. For example, maltreated children frequently 1) have insecure attachments with their caregivers22; 2) develop negative internal working models30; 3) are hypersensitive to emotional signals31; and 4) have impaired emotion regulation.32 In the following section, we explore these 4 lines of evidence that link maltreatment to impaired empathy.

The ability to empathize is nurtured from infancy, starting with positive attachment experiences with caregivers.22 Empathic parenting associated with secure attachment encourages the child to explore the parent’s mind, which in turn encourages the development of the capacity to mentalize.1 By contrast, child maltreatment typically leads to insecure attachment styles, which are associated with impaired empathy.22,30,33 Moreover, maltreated children with insecure attachments typically internalize negative working models. For example, they view themselves as inadequate and unlovable; see others as hostile and uncaring; and regard interpersonal interactions as potentially dangerous and painful.34 Because these negative internal working models are likely to foster disdain, fear, and mistrust of others, they undermine the development of empathy.35,36

Because maltreated children often are faced with continual fear, they also frequently become hypersensitive to negative emotional signals.31 As a result, maltreated children often experience intense emotional distress (devoid of sympathy) when faced with another’s emotional distress.35,37,38 For example, one study showed that maltreated children often became intensely distressed in the presence of another child crying in a playground.39 Instead of soothing or helping the other child, the maltreated children reacted with a combination of fear, anger, and violence. Such forms of emotional distress hinder empathy because they foster a self-directed focus that detracts from the suffering of others.15

Another factor necessary for the development of empathy is emotion regulation.40 Cummings32 identified 3 distinct patterns of emotion regulation in children when viewing another person in distress. These are 1) adaptively concerned emotion regulation, which is associated with moderate levels of negative affect that are well modulated, together with empathy; 2) undercontrolled/ambivalent emotion regulation, which is associated with underregulation of emotional behavior, high reactivity, and prolonged elevation of both positive and negative affect; and 3) overcontrolled/unresponsive emotion regulation, which is associated with low emotional and behavioral reactivity, and the inhibition of signs of distress. Maltreated individuals typically exhibit the latter 2 patterns, which are associated with significant impairments in empathy.40

Study overview

In the present investigation, we used the context of the South African Truth and Reconciliation Commission (TRC) hearings to elicit current and ecologically valid empathic responses in participants. The TRC was established after apartheid with the purpose of engaging with and healing the injustices of the past, mainly through exploring testimonial narratives of perpetrators and victims.

We used short, emotionally arousing clips from the TRC hearings to explore empirically the relationship between varying degrees of child maltreatment and empathic responding in an adult participant sample. In light of the literature reviewed above, we hypothesized that significant child maltreatment would be associated with impaired empathy, and that the severity of maltreatment would be related linearly to the degree of impaired empathic responding.



Forty-nine racially diverse participants (White: 25, Black: 12, Colored (people of mixed race): 12; 33 women) between the ages of 20 and 60 participated in the study. To increase the generalizability of our findings, we sourced participants from the university campus and the local environment through posters and an advertisement placed in a local newspaper. We thus made use of non-probability convenience sampling, and included all participants who responded to our study advertisement.

All participants completed the procedures described below and received 90 ZAR (approximately 10 USD) as gratuity. The study protocol was approved by the Research Ethics Committee of the University of Cape Town’s (UCT) Department of Psychology.

Experimental design and setting

We utilized a mixed methods research design that incorporated both quantitative and qualitative data analysis. This methodology enabled us to obtain rich, subjective data indicative of participants’ experiences, in addition to numerical measures of empathy and related emotions amenable to statistical analysis. Specifically, we used a within-subjects mixed-model design, where stages of data collection took place sequentially. The study took place in a laboratory at UCT’s Department of Psychology.


Empathy-eliciting stimuli were film clips taken from a short documentary (A Long Night’s Journey into Day, Iris Films; 2000) depicting different encounters between victims and perpetrators who testified at the TRC public hearings. The hearings were of a case in which 7 young men from the Gugulethu Township in Cape Town were killed. The amnesty applicants were a white police officer, Captain Bellingan (the commander of the operation), and a black police collaborator, Mr. Mbelo. They divulge 2 very different stories about their involvement in the killings.

The TRC uncovered the following details: The black officer, Mbelo, was sent to Cape Town to infiltrate a group of young men and train them under the pretext that they were going to become soldiers of the anti-apartheid struggle. He then lured them into a trap where a team of policemen, including Captain Bellingan, were waiting to kill them. The press characterized the operation as a huge success of the apartheid campaign against terrorists. During the TRC hearings of Bellingan and Mbelo’s amnesty application, the mothers of the 7 victims heard the horrifying truth about what happened to their sons for the first time.

The clips for our study were taken from scenes of this hearing, as well as from a meeting between the mothers and the black officer (Mbelo). During this meeting, Mbelo asked the mothers for forgiveness. The stimuli consisted of 4 short clips (1 to 2 minutes each) and 1 long clip (12 minutes). The 4 short clips depicted: 1) a forgiving mother, 2) an unforgiving mother, 3) a mother in distress (sobbing), and 4) an unrepentant perpetrator (Captain Bellingan). These 4 clips were contained within the longer clip, which depicted the events of the hearing in more detail, including the meeting between the victims’ mothers and the remorseful perpetrator.


Self-reported emotion scales. Participants indicated their current emotional state after viewing each of the 4 short clips by rating 4 emotional qualities (empathy, shame, anger, and sadness) on a 9-point Likert-type scale, ranging from 1 (not at all) to 9 (very strongly/extremely).

Qualitative questionnaire. Participants completed the qualitative questionnaire after they had viewed all the clips. It consisted of 13 open-ended questions that explored emotions and thoughts evoked by the footage. For example, participants were asked to describe: the overall feeling they experienced when watching the film; personal memories that were triggered by the film; and any emotional detachment they felt.

Childhood maltreatment. We used the Childhood Trauma Questionnaire Short-Form (CTQ-SF)41 to assess the severity of different types of child maltreatment. CTQ-SF is a retrospective measure that consists of 25 items that measure the frequency with which different events took place when participants “were growing up”, ranging from 1 (never true) to 5 (very often true). The CTQ-SF consists of 5 subscales: physical neglect, emotional neglect, physical abuse, sexual abuse, and emotional abuse. The CTQ-SF has high sensitivity and specificity in nonclinical samples,42 and also shows high test-retest reliability (.66 to .94) and internal consistency with alphas ranging between .70 and .93.43

Participants were divided into 3 groups based on their total CTQ-SF scores. The control group had scores <36, which indicated no maltreatment or mild-moderate maltreatment on 1 subscale. The moderate maltreatment group had scores between 36 and 54, which indicated mild-moderate and moderate-severe maltreatment on multiple subscales. Finally, the severe maltreatment group had scores >55, which indicated moderate-severe and severe-extreme maltreatment on multiple scales.


On arrival, participants provided informed consent and received general instructions regarding the procedures. Each participant was seated in front of a 17-inch computer monitor and given headphones and answer booklets. To contextualize the clips, participants viewed a PowerPoint presentation providing information about the specific TRC case they were about to watch. They then viewed the 4 short clips (1 to 2 minutes each) and completed the self-reported emotion scales after each clip. Clips were shown in 3 different random sequences to avoid order effects. Finally, participants viewed the longer (12 minutes) clip, and then completed the qualitative questionnaire and the CTQ-SF. All participants were thoroughly debriefed after the study procedure and given the option to see a professional counsellor, if necessary.

Data analysis

We analyzed each of the emotions rated on the quantitative self-report scales using a 3 (group: control, moderate, and severe maltreatment) × 4 (clip type: forgiving mother, unforgiving mother, distressed mother, and unrepentant perpetrator) mixed factorial ANOVA. In instances where the assumption of sphericity was violated, the degrees of freedom were corrected using Greenhouse-Geisser epsilon corrections. These correction factors are reported. We also performed zero-order correlations between participants’ total scores on the CTQ-SF and their mean emotion ratings for each emotion. These mean emotion ratings were calculated by averaging the 4 emotion ratings after each film clip for each emotion.

Two researchers (SCL and LB), who were blind to maltreatment status, analyzed the qualitative data using thematic analysis in accordance with guidelines set out by Braun and Clarke.44 Accordingly, patterns within the data were coded, and extracts from the data were collated into themes and subthemes. Themes were compared against the data and further refined until they were internally consistent and distinctive, and offered a convincing representation of the data. In general, we approached the analysis in an inductive fashion. The themes empathy and impaired empathy were approached in a deductive fashion, however, based on the framework of empathy suggested by Bateson and Ahmad.17 Our approach varied between a semantic one, where themes were formed from explicit or surface meanings of the data, and an interpretive one, where the context of statements was taken into account.45

To integrate the quantitative and qualitative data sets, we used a multi-stage data analysis process that has been suggested to be appropriate for mixed methods research.46 Briefly, we quantified the qualitative data by summing the number of occurrences of each theme in each participant and then divided this score by the number of participants in each group. The frequency of each theme first was rated independently by 2 researchers (SCL and LB), who remained blinded to maltreatment status, and then averaged to form the final score. The level of agreement between these researchers was >85%. The thematic frequency estimates per group were then compared with the quantitative results to determine similarities and differences between the data sets.



Based on participants’ CTQ-SF scores, 18 were classified as having experienced no, or mild levels, of maltreatment (control group), 21 were classified as having experienced moderate levels of maltreatment (moderate maltreatment group), and 10 were classified as having experienced severe levels of maltreatment (severe maltreatment group). Chi-square analyses indicated that these study groups did not differ significantly in terms of the distribution of age, sex, or racial representation (P values > .39)

Quantitative results

Emotion ratings in response to the TRC clips.FIGURE illustrates subjective emotion ratings obtained after each short clip.

FIGURE 1: Mean ratings for (A) empathy, (B) sadness, (C) anger, and (D) shame for the control, moderate maltreatment, and severe maltreatment groups in response to each clip. Error bars represent standard errors
Forgive+: forgiving mother, Forgive- : unforgiving mother, Distress+: mother in distress, Repent-: unrepentant perpetrator.

The mixed factorial ANOVA for empathy ratings detected a significant main effect for the type of clip, F (2.29, 103.08) = 106.73, P < .001, ηp2 = .70, ε = .76. Post-hoc contrasts indicated that empathy was significantly higher following the distressed mother clip than the forgiving mother, the unforgiving mother, and the unrepentant perpetrator clips (P values < .001, r coefficients > .60). Empathy following the unrepentant perpetrator clip was also significantly lower than the forgiving mother and the unforgiving mother clips (P values < .001, r coefficients > .84), while empathy responses following the latter 2 clips did not differ.

Empathy ratings also differed significantly between the study groups, F (2, 45) = 6.75, P = .003, ηp2 = .23. Bonferroni post-hoc testing revealed that empathy ratings for the moderate maltreatment group were significantly lower than that of the control group (P < .05, r = .41) and the severe maltreatment group (P < .01, r = .42). However, empathy ratings did not differ significantly between the control and severe maltreatment groups (P = 1.00, r = .09). Finally, there was no significant interaction between group and type of clip, (P = .17).

Because empathy was the main focus of our study, we summarize the results for sadness, shame, and anger briefly. Mixed factorial ANOVA results indicated that these analyses were all significant for the main effect of type of clip (P values < .001). Post-hoc contrasts revealed that sadness was significantly higher following the distressed mother clip than the other clips (P values < .001, r coefficients > .60), and significantly lower following the unrepentant perpetrator clip than the other clips (P values < .001, r coefficients > .73). In terms of shame, participants’ ratings were significantly higher following the distressed mother clip than the other three clips (P values < .01, r coefficients > .44). Finally, ratings of anger were significantly higher following the distressed mother clip than all the other clips (P values < .05, r coefficients > .36). Anger was also significantly higher following the unremorseful perpetrator clip than following the forgiving mother clip (P < .001, r = .63), and significantly higher following the unforgiving mother clip than following the forgiving mother clip (P < .001, r = .62).

In terms of group effects, ratings of shame differed significantly between the study groups, F (2, 41) = 5.17, P = .01, ηp2 = .20, with significantly higher ratings of shame for the severe maltreatment group than for the control group (P < .05, r = .39) and the moderate maltreatment group (P = .01, r = .43). The main effects of group did not reach significance for sadness, (P = .07, ηp2 = .12), or anger, (P = .08, ηp2 = .11).

Associations between maltreatment and elicited emotions.TABLE 1 shows the mean self-reported emotion-ratings for each emotion in each study group when ratings for all 4 clips were averaged. This table shows a general trend, such that mean emotion ratings for all measured emotions were highest in the severe maltreatment group, whereas they were lowest in the moderate maltreatment group.


Mean self-reported emotion ratings from 1 (not at all) to 9 (very much) for each study group

Emotion Control Moderate maltreatment Severe maltreatment
(n = 18) (n = 21) (n = 10)
Mean (SD) Mean (SD) Mean (SD)
Empathy 5.94 (1.16) 4.69 (1.48) 6.25 (1.05)
Sadness 5.74 (1.25) 4.99 (1.56) 6.15 (0.98)
Anger 4.82 (1.96) 4.40 (1.95) 6.08 (1.50)
Shame 3.53 (1.88) 3.32 (1.73) 5.53 (1.89)
Means represent the mean self-reported emotion rating for each group across all 4 film clips.

To explore further the relationship between subjective emotion ratings and the degree of maltreatment, we performed zero-order correlations between total scores on the CTQ-SF and mean emotion ratings across all groups. Significant positive correlations were detected between CTQ-SF scores and ratings of shame (P < .05; r = .33) as well as anger (P < .05; r = .26). Ratings of empathy and sadness did not correlate significantly with CTQ-SF scores.

Qualitative results

Thematic analysis. We identified 7 themes (which are italicized) from our analysis of the qualitative data, which differed markedly between groups. The themes empathy, positive world view, and emotional awareness were more common in the control group, whereas the themes impaired empathy, malignant world view, and poor emotional awareness, were more common in the 2 maltreatment groups. The theme emotional distress was infrequent in the control and moderate maltreatment groups, but was expressed commonly in the severe maltreatment group.

Because empathy is the focus of this study, we discuss the themes empathy and impaired empathy in detail. The other themes and subthemes are summarized in TABLE 2 (see TABLE 3 for examples of quotes from each theme).


Summary of themes and subthemes from the qualitative thematic analysis

  Theme Subthemes
1. Empathy Imagine-self perspective; imagine-other perspective; emotional contagion; empathic concern/sympathy
2. Impaired empathy Poor understanding of others’ mental states; absence of emotional contagion; absence of sympathy; bizarre attribution of another’s behavior
3. Positive world view Positive attitude towards forgiveness; faith in goodness of humanity; sense that life has meaning
4. Malignant world view Despair; absence of hope; distrust towards others; attitude that perpetrators are inherently evil
5. Emotional distress Intense feelings of anger and hatred; fear and horror; guilt and shame; triggering of traumatic memories; somatic distress
6. Emotional awareness Focus on emotional signals; rich and complex understanding of others’ emotions; proficiency in understanding personal emotions
7. Poor emotional awareness Poor understanding of emotions in self and others; emotional blunting; avoidance tendencies

THEME 1: EMPATHY. Participants who displayed highly empathic responses (mostly those in the control group) tended to have numerous examples of each empathy subtheme, which illustrates how the component processes of empathy combine in highly empathic responses.

Participants demonstrated the subtheme imagine-self perspective when they imagined how they would feel if they were in the position of 1 of the people portrayed in the TRC footage. For example, 1 participant stated:


Examples of quotes for each theme and subtheme, and relation to maltreatment status

Theme Subtheme Quotes
Empathy Imagine-self perspective “I absolutely understood her need to know [the mother, who was asking Mbelo why he killed her son]. In the face of such brutality, you have to know, especially if it is your loved one… I could relate.”a
Imagine-other perspective “I perceived [Mbelo] differently in that context with the mothers. I perceived him to be truly remorseful. I was heartened by his genuineness, the fact that he took responsibility.”a
Emotional contagion “The despair of the mothers was heart-breaking, I cried uncontrollably, I empathized.”a
Empathic concern “Seeing Mbelo’s face in response to the unforgiving mother. I felt deeply sorry for him as his expression was accepting of anything that the mothers may say.”a
Impaired empathy Poor understanding of others’ mental states “I felt detached from the mother who screamed/performed in front of all the people. One doesn’t have to ‘perform’ to show one’s grief. I keep it inside.”b
Absence of emotional contagion “[I felt] very little as I’m not someone who experiences emotions easily. I felt a little sad but that’s all.”b
Absence of sympathy “I didn’t feel any pain personally…”b
Bizarre misattributions “I felt anger when Mbelo was talking about staying drunk to mask pain. [It was] as though he was having a good time forgetting.”c
Positive world view Positive attitude towards forgiveness “[When the] mother offered forgiveness—I was amazed. I felt a sense of healing and an ability to move forward. Her wisdom and character made me feel she had done the right thing.”a
Faith in the goodness of humanity “I keep on being in awe with us humans, [with our] inner power and strength.”a
Malignant world view Absence of hope “[My past experiences] made me think that people can and will do anything to each other and sometimes there’s nothing to do but be in pain.”c
Despair “I feel guilt, shame, and defeat for my situation and those on the video. Pity. Pity. Pity.”b
Distrust of others, attitude that all perpetrators are inherently evil “I would scream and run away from them [the perpetrators] in fear. They have killed before and will again. Something is mentally wrong with them.”c
Emotional distress Intense feelings of fear and horror “Seeing the footage was heart-stopping… My brain was pounding, my breathing was heavier.”c
Guilt and shame “I am feeling the guilt and shame that Bellingan and Mbelo should be feeling.”c
Extreme somatic distress “I experienced… chills down my spine, felt as if my hair was being electrocuted….”c
Emotional awareness Focus on emotional signals “I was desperately searching [Bellingan] for any hint of emotion or remorse, but I never found it.”a
Rich and complex understanding of others’ emotions “I felt her [the unforgiving mother] battle was the hardest as she was in conflict between how she truly felt and what a transcendent moral code demanded of her.”a
Proficiency in understanding personal emotions “[My detachment from the mother in distress] was related not to the events unfolding in the court but rather my own issues.”a
Poor emotional awareness Poor understanding of emotions in self and others “[I felt] no guilt or shame – [I was] not there at the time.”b
“…Mbelo described smilingly how it was just another day on the job” [referring to a scene where Mbelo was clearly ashamed].b
Emotional blunting “I feel a little detached because it hasn’t affected me personally.”b
“I find it difficult to love and sometimes I can’t [feel] sorry for other people.”b
aControl group.
bModerate maltreatment group.
cSevere maltreatment group.

“I absolutely understood her need to know [the mother, who was asking Mbelo why he killed her son]. In the face of such brutality, you have to know, especially if it is your loved one… I could relate.”

Similarly, the imagine-other perspective reflects the ability to imagine how another thinks or feels given their situation. In relation to the scene where Mbelo appealed to the mothers for forgiveness, 1 participant stated:

“I perceived [Mbelo] differently in that context with the mothers. I perceived him to be truly remorseful. I was heartened by his genuineness, the fact that he took responsibility.”

Emotional contagion (“emotion catching”) could be described as instances in which an individual feels the same emotion that they observe in another. For example:

“The despair of the mothers was heart-breaking, I cried uncontrollably, I empathized.”

Finally, the subtheme empathic concern/sympathy was detected in statements that demonstrated sympathy and compassion toward the protagonists of the film footage, for example:

“Seeing Mbelo’s face in response to the unforgiving mother. I felt deeply sorry for him as his expression was accepting of anything that the mothers may say.”

THEME 2: IMPAIRED EMPATHY. This theme was marked by responses that revealed a lack of empathy, including statements showing difficulty understanding the mental states of others. We observed impaired empathy most frequently in the responses of the maltreatment groups.

The subtheme poor understanding of others’ mental states reflects an impairment of the imagine-other perspective. For example, in response to a scene showing one of the mothers crying out in pain after viewing footage of her dead son, 1 participant in the moderate maltreatment group wrote:

“I felt detached from the mother who screamed/performed in front of all the people. One doesn’t have to ‘perform’ to show one’s grief. I keep it inside.”

Some participants demonstrated an absence of emotional contagion in statements showing emotional blunting and difficulty feeling emotions in response to others’ pain. For example, when asked to describe “your overall feeling after watching the film” a participant in the moderate maltreatment group stated:

“[I felt] very little as I’m not someone who experiences emotions easily. I felt a little sad but that’s all.”

Absence of sympathy was shown by remarks indicating a lack of concern for others’ pain.

Finally, we observed extreme impairments in empathy in bizarre misattributions, where the participant displayed strange misattributions of emotions, as in this example from a participant in the severe maltreatment group:

“I felt anger when Mbelo was talking about staying drunk to mask pain. [It was] as though he was having a good time forgetting.”

The footage the participant referred to, in fact, showed a guilt-stricken Mbelo explaining how he drank as a form of escapism to forget the (unacceptable) actions that he had committed.

Links between qualitative and quantitative results

TABLE 4 presents the quantification of themes across the study groups. We combined the quantified qualitative results with the quantitative results in order to detect a pattern that differentiated between the 3 study groups. We found that the qualitative data overlapped clearly with the quantitative data, yet we also observed some significant differences between the 2 sets of data. We describe the most important similarities and differences below.


Frequencies of themes in each study group

Association with empathy Theme Control Moderate maltreatment Severe maltreatment
Positive Empathy +++ ++ +
Positive world view +++ ++ ++
Emotional awareness +++ ++ +
Negative Impaired empathy + ++ +++
Emotional distress ++ + ++++
Malignant world view + ++ +++
Poor emotional awareness + +++ ++++
Frequencies were determined by calculating the mean number of occurrences of each theme in each group: + very infrequent; ++ infrequent; +++ frequent; ++++ very frequent.

With regard to empathy, the control group displayed higher levels than the moderate maltreatment group across both data sets. However, for the severe maltreatment group, empathic responses differed markedly; whereas the thematic analysis indicated that this group had the lowest ratings for empathy and the highest for impaired empathy, their quantitative empathy ratings were high (similar to that of the control group).

In terms of emotional distress, both data sets indicated that the control and moderate maltreatment groups experienced relatively low levels of emotional distress. However, the severe maltreatment group displayed high frequencies of the emotional distress theme, and also had significantly higher ratings of shame (and to a lesser extent anger) than the other 2 groups on the quantitative emotion scales. This pattern of increased emotional distress in the severe maltreatment group was corroborated by the significant positive correlations between ratings of shame and anger, and total CTQ-SF scores. Thus greater levels of maltreatment were associated with higher ratings of shame and anger in general.


In this study, we examined the relationship between child maltreatment and empathy in a cohort of adult individuals. Participants were divided into 3 study groups based on their scores on the CTQ-SF: 1) control (little to no maltreatment), 2) moderate maltreatment, and 3) severe maltreatment. Consistent with our predictions, the control group displayed high empathy on both the quantitative and qualitative measures, and the moderate maltreatment group displayed low levels of empathy on both measures. However, we did not expect to see the results we observed in the severe maltreatment group. This group displayed low levels of empathy on the qualitative measure, but high levels of empathy on the quantitative measure. We contend that distinct patterns of empathic responding in the 2 maltreatment groups may have resulted from differences in factors such as emotion regulation, attachment, and emotional avoidance.

Empathic responses in the maltreatment groups

In terms of the quantitative emotion data, the severe maltreatment group displayed significantly higher empathy than the moderate maltreatment group. However, they also displayed high levels of anger and sadness, and significantly higher levels of shame than both the moderate maltreatment and control groups. These elevated levels of negative affect could be argued to reflect increased emotional distress. Our qualitative thematic analysis supported this interpretation, showing that the severe maltreatment group had the highest ratings of the theme emotional distress. The single self-rating of empathy we employed may therefore not have been as specific to empathy as we intended, and also may have been influenced by feelings of emotional distress.

The elicitation of an isomorphic emotional response to that observed in another is an integral component of affective empathy, termed emotion matching.47 But an empathic response also involves sympathy and cognitive empathy.17 Our qualitative results indicated that the severe maltreatment group was characterized by extraordinary affective responses to the clips, indicative of emotional over-arousal or distress, but also by low levels of cognitive empathy and sympathy (ie, low ratings for empathy and emotional awareness, and high ratings for impaired empathy and poor emotional awareness). The severe maltreatment group’s data thus suggest a pattern of decreased cognitive empathy and sympathy, in addition to increased emotion contagion.

In contrast to the severe maltreatment group, participants in the moderate maltreatment group displayed impairments in empathy across both the qualitative and quantitative data sets. In particular, the moderate maltreatment group showed significantly lower empathy ratings than the other 2 groups. They also displayed low empathy and high impaired empathy themes, and often displayed poor understanding of others’ mental states, possibly reflecting poor cognitive empathy. The moderate maltreatment group had very low levels of the theme emotional distress, and did not have high ratings of shame, anger, or sadness (as seen in the severe maltreatment group). Therefore, the moderately maltreated group displayed blunted affect as well as impairments in both affective and cognitive empathy.

We found it intriguing that the moderate and severe maltreatment groups showed such contrasting affective responses, yet both groups displayed impaired cognitive empathy and sympathy. Below, we first consider how these different affective response patterns could potentially inhibit empathic responding. We then explore the potential mechanisms that may underlie these patterns. Finally, we describe factors that may contribute to impaired empathy that may be common across both maltreatment groups.

Patterns of affective responding and impaired empathy

When an individual matches the emotion they observe in another in a well-modulated manner, it is likely that this emotion matching will act as a trigger for cognitive empathy and sympathy.17 We detected this “ideal” response frequently in the qualitative responses of our control group: participants’ moderate distress in response to the footage appeared to trigger both sympathy and an attempt to understand the other’s mental state (ie, cognitive empathy).

In contrast to the control group, we observed blunted affect in response to the empathy-eliciting stimuli for the moderate maltreatment group. Blunting and/or avoidance is likely to prevent the affective synchrony that may develop when an individual is exposed to another’s distress, because blunting removes the emotional stimulus that triggers an empathic response.14,22 It is not surprising therefore that participants in the moderate maltreatment group showed little sympathy and cognitive empathy.

Participants in the severe maltreatment group tended to display emotional distress in response to the footage, but with low levels of empathy and sympathy. It has been proposed that uncontrollable distress in response to another’s suffering may overwhelm the individual to the extent that it detracts from the other’s experience of pain.15 Because intense distress is an uncomfortable sensation, it is likely to lead to a self-directed focus, which then undermines the salience of the other person’s suffering. Decety and Lamm15 rightly notes that, “in the case of empathy, the best response to another’s distress may not be distress, but efforts to soothe that distress.” Therefore the inability to regulate negative emotions in the severe maltreatment group may have prevented them from experiencing cognitive empathy and sympathy.

Potential mechanisms underlying impaired empathy in maltreatment

Emotion regulation. Research has shown that the majority of maltreated individuals have impaired emotion regulation,32,40 which has been linked to impaired empathy.14 The behavioral patterns that are characteristic of the forms of emotion regulation typically exhibited by maltreated individuals show considerable overlap with the patterns of responses we observed in the present study. For example, the blunted emotionality and unempathic responses of the moderate maltreatment group may be associated with the overcontrolled/unresponsive emotion regulation pattern described by Cummings.32 By contrast, the intense, uncontrollable, and unempathic responses of the severe maltreatment group could be likened to the undercontrolled emotion regulation pattern described by Cummings.32

Attachment styles. Insecure attachment styles are present in the majority of maltreated individuals, and also have been linked to impaired empathy.30,48 Although we did not measure attachment directly in the present study, we offer some interpretations here about the nature of attachment styles in our 2 maltreatment groups based on previous research and speculation.

An avoidant attachment style is associated with blunted affect, and a tendency toward being emotionally distant from others.49 This affective pattern resembles the blunted affect we observed for participants in our moderate maltreatment group. By contrast, the anxious and disorganized attachment styles are associated with elevated negative affect and unregulated (disorganized) emotional responses, respectively. Both of the latter affective patterns are consistent with the intense affect we observed in our severe maltreatment group.

Emotional avoidance. It is well documented that maltreated individuals have a tendency to experience intense, uncontrollable negative affect.31 However, various authors also describe manifestations of emotional avoidance in maltreated individuals, including emotional suppression,50 and avoidant coping strategies.51 Although intense emotionality and emotion avoidance appear to be diametrically opposed, they may be related closely. For example, because distressing stimuli often elicit intense negative affect and feelings of shame in maltreated individuals,52 these individuals may be motivated to adopt an avoidant emotional stance. This behavior is consistent with our findings for the moderate maltreatment group, but does not explain the heightened distress observed for the severe maltreatment group.

One possible explanation for the above discrepancy is that, although distress may motivate avoidance in maltreated individuals, some may be incapable of implementing this defense. In particular, severely maltreated individuals may be unable to inhibit negative affect (even when they are highly motivated to do so) because of undercontrolled emotion regulation, anxious/disorganized attachment, or the post-traumatic phenomenon of triggering, which is observed more readily following high levels of trauma.34

The tendency toward emotional avoidance in maltreated individuals also has conceptual overlap with Freud’s53 theory of repression. This defense mechanism may remain intact in moderate maltreatment, but may become overwhelmed with severe maltreatment, leading to overwhelming negative affect due to the emergence of disturbing unconscious material.54,55

Common features across the maltreatment groups

Besides the differences highlighted above, participants in our maltreatment groups also may share features associated with impaired empathy. One such feature is impaired mentalization, which has been linked to poor cognitive empathy.56 Fonagy1 argued that maltreatment often results in impaired mentalization because abusive or neglectful parents typically fail to engage with the thoughts and feelings of the child. The child’s internal experience therefore remains unlabeled and confusing, which, in turn, hinders the child from forming accurate representations of others’ mental states.57,58

Another shared feature may be impaired sympathy, which has been linked to maltreatment because abused children often internalize the unsympathetic behavior of their parents.59,60 We observed a lack of sympathy in both maltreatment groups, which may be linked to the theme malignant world view. For example, when a mother in the footage forgave her son’s killer, she was described as weak; when a mother cried, she was seen as “performing”; and when a perpetrator begged for forgiveness, he was viewed as a “cold-hearted liar.”

Limitations and directions for future research

The current study contributed significantly to our understanding of impaired empathy as one of the enduring effects of child maltreatment. A number of methodological limitations should be noted, however. First, retrospective methods of assessing child maltreatment have important limitations61; corroborative evidence should thus be used whenever possible. Second, our relatively small sample size may have reduced the power of our analyses. Third, because of the exploratory nature of this study, we did not employ a randomized sampling method with groups matched for third variables that may influence empathy (eg, age, sex, race, or developmental timing and chronicity of maltreatment).19,62,63 It should be noted, however, that our groups did not differ statistically in terms of age, sex, or race. We also did not group participants according to the type of maltreatment they experienced, which may be a confounding factor because different types of maltreatment may affect empathic impairment differently.9 However, most of our participants had experienced >1 type of maltreatment; it was thus impossible to group them according to type of maltreatment.

A fourth limitation was our use of a single, global, self-rating for empathy, which may have been influenced by emotional distress in some subjects. Our rationale for employing this measure was to save time, that is, we did not want measures of state affect between clips to detract from participants’ emotional engagement with the footage. A more accurate way to assess empathic responding may be to include several adjectives describing different emotional states that load onto empathy and personal distress, respectively.64

Finally, because we did not assess participants’ attachment styles and emotion regulation patterns, some of our conclusions may be speculative. Future research may benefit from assessing the contribution of these factors that potentially may mediate the relationship between empathy and maltreatment.


In this study, we found empirical evidence that child maltreatment is associated with impaired empathy. Furthermore, our results suggest that such impairments manifest differently in moderately and severely maltreated individuals. Specifically, moderately maltreated participants displayed blunted affect in response to the empathy-eliciting clips, which could result from overcontrolled emotion regulation, avoidant attachment, or a proclivity toward avoiding distressing psychological content. By contrast, severely maltreated participants displayed intense, uncontrollable emotional distress in response to the clips, which could result from lack of emotion regulation, anxious or disorganized attachment, or “unsuccessful” emotional avoidance.

The findings of the present study, though exploratory in nature, may be of practical value in understanding psychopathology and informing current treatment programs. Our findings reinforce the notion that empathy training should form an integral component of treatment programs following maltreatment.65 In addition, the therapeutic process may benefit from a better understanding of the nature of impaired empathy in maltreated individuals. For example, a vital component of psychotherapy involves enabling patients to develop an awareness of their own and others’ mental states, thereby finding meaning in their behavior.1 The present study’s findings may help therapists gain deeper insight into the different behavioral patterns associated with impaired empathy (eg, emotional overarousal or blunted affect), thereby enabling them to better assist patients in identifying problematic patterns of cognition and behavior. Our research also suggests that, rather than using a single intervention, empathy treatment programs may need to be tailored according to the particular kind of empathic deficit.

In sum, our study shows that child maltreatment can have a significant impact on an individual’s empathic responding toward others. Because empathy plays such a vital role in social relations, it is important to explore further how maltreatment impairs empathy, and how the burden of impaired empathy can be treated or prevented. Our study has made important first steps in exploring these complex relationships.

DISCLOSURES: Mr. Locher, Ms. Barenblatt, Dr. Fourie, and Dr. Gobodo-Madikizela report no financial relationship with any company whose products are mentioned in this article or with manufactuers of competing products. Dr. Stein has received research grants and/or consultancy honoraria from AMBRF, Biocodex, Lundbeck, National Responsible Gambling Foundation, Novartis, Servier, and Sun.

ACKNOWLEDGMENTS: We thank Professor Mark Solms for editorial comments as well as for his conceptual input. This article was part of a 3-year research project (2009 to 2012), “Empathy: Emotional responses to video-taped scenes from the Truth and Reconciliation Commission of South Africa,” funded by the Fetzer Institute and the National Research Foundation.

Authors S.C. Locher and L. Barenblatt contributed equally to this article.


  1. Fonagy P. Attachment and borderline personality disorder. J Am Psychoanal Assoc. 2000;48:1129–1146.
  2. Hussey JM, Chang JJ, Kotch JB. Child maltreatment in the United States: prevalence risk factors, and adolescent health consequences. Pediatrics. 2006;118:933–942.
  3. Gilbert R, Widom CS, Browne K, et al. Burden and consequences of child maltreatment in high-income countries. Lancet. 2009;373:68–81.
  4. Pierce L, Bozalek V. Child abuse in South Africa: an examination of how child abuse and neglect are defined. Child Abuse Negl. 2004;28:817–832.
  5. Hart SN, Binggeli NJ, Brassard MR. Evidence of the effects of psychological maltreatment. Journal of Emotional Abuse. 1997;1:27–58.
  6. McCord J. A forty year perspective on effects of child abuse and neglect. Child Abuse Negl. 1983;7:265–270.
  7. Watkins HD, Bradbard MR. Child maltreatment: an overview with suggestions for intervention and research. Family Relations. 1982;31:323–333.
  8. de Paul J, Guibert M. Empathy and child neglect: a theoretical model. Child Abuse Negl. 2008;32:1063–1071.
  9. Feshbach N. The construct of empathy and the phenomenon of physical maltreatment of children. In: Cicchetti D Carlson V, eds. Child maltreatment: theory and research on the causes and consequences of child abuse and neglect. New York, NY: Columbia University Press; 1989:349-373.
  10. Mash EJ, Wolfe DA. Abnormal child psychology. 4th ed. Belmont CA: Wadsworth Publishing; 2010.
  11. Kaufman J, Zigler E. Do abused children become abusive parents? Am J Orthopsychiatry. 1987;57:186–192.
  12. Gelder M, Mayou R, Cowen P. Shorter oxford textbook of psychiatry. 4th ed. Oxford United Kingdom: Oxford University Press; 2005.
  13. Kaplan H, Sadock B. Synopsis of psychiatry. 7th ed. Baltimore MD: Williams and Wilkins; 1994.
  14. Decety J. Dissecting the neural mechanisms mediating empathy. Emotion Review. 2011;3:92–108.
  15. Decety J, Lamm C. Human empathy through the lens of social neuroscience. ScientificWorldJournal. 2006;6:1146–1163.
  16. Hein G, Singer T. I feel how you feel but not always: the empathic brain and its modulation. Curr Opin Neurobiol. 2008;18:153–158.
  17. Bateson CD, Ahmad NY. Using empathy to improve intergroup attitudes and relations. Social Issues and Policy Review. 2009;3:141–177.
  18. Preston SD, de Waal FB. Empathy: its ultimate and proximate basis. Behav Brain Sci. 2002;25:1–20.
  19. Singer T. The neuronal basis and ontogeny of empathy and mind reading: review of literature and implications for future research. Neurosci Biobehav Rev. 2006;30:855–863.
  20. Zahavi D. Simulation projection and empathy. Conscious Cogn. 2008;17:514–522.
  21. Abu-Akel A. A neurobiological mapping of theory of mind. Brain Res Rev. 2003;43:29–40.
  22. Watt D. Toward a neuroscience of empathy: integrating affective and cognitive perspectives. Neuropsychoanalysis. 2007;9:119–172.
  23. Gobodo-Madikizela P. Trauma forgiveness, and the witnessing dance: Making public spaces intimate. J Anal Psychol. 2008;53:169–188.
  24. Fonagy P, Target M. Playing with reality: I. Theory of mind and the normal development of psychic reality. Int J Psychoanal. 1996;77:217–233.
  25. Fonagy P, Target M. Attachment and reflective function: their role in self-organization. Dev Psychopathol. 1997;9:679–700.
  26. Target M, Fonagy P. Playing with reality II: the development of psychic reality from a theoretical perspective. Int J Psychoanal. 1996;77:459–479.
  27. de la Vega A, de la Osa N, Ezpeleta L, et al. Differential effects of psychological maltreatment on children of mothers exposed to intimate partner violence. Child Abuse Negl. 2011;35:524–531.
  28. Horwitz AV, Widom CS, McLaughlin J, et al. The impact of childhood abuse and neglect on adult mental health: a prospective study. J Health Soc Behav. 2001;42:184–201.
  29. Chaitin J, Steinberg S. “You should know better”: expressions of empathy and disregard among victims of massive social trauma. Journal of Aggression Maltreatment, and Trauma. 2008;17:197–226.
  30. Yates TM. The developmental psychopathology of self-injurious behavior: Compensatory regulation in posttraumatic adaptation. Clin Psychol Rev. 2004;24:35–74.
  31. Heim C, Nemeroff CB. The role of childhood trauma in the neurobiology of mood and anxiety disorders: Preclinical and clinical studies. Biol Psychiatry. 2001;49:1023–1039.
  32. Cummings EM. Coping with background anger in early childhood. Child Dev. 1987;58:976–984.
  33. Barnett D, Ganiban J, Cicchetti D. Maltreatment negative expressivity, and the development of type D attachments from 12 to 24 months of age. In: Vondra J, Barnett D, eds. Atypical attachment in infancy and early childhood among children at developmental risk. Oxford, United Kingdom: Blackwell Publishing; 1999:97-118.
  34. Herman JL. Complex PTSD: a syndrome in survivors of prolonged and repeated trauma. J Trauma Stress. 1992;5:377–391.
  35. Briere J, Spinazzola J. Phenomenology and psychological assessment of complex posttraumatic states. J Trauma Stress. 2005;18:401–412.
  36. Laub D, Auerhahn NC. Failed empathy: a central theme in the survivor’s holocaust experience. Psychoanalytic Psychology. 1989;6:377–400.
  37. Liotti G. Trauma dissociation, and disorganized attachment: Three strands of a single braid. Psychotherapy: Theory, Research, Practice, Training. 2004;41:472–486.
  38. Thompson RA. Emotion and self-regulation. In: Thompson RA ed. Socioemotional Development. Lincoln, NE: University of Nebraska Press; 1990:367-467.
  39. Klimes-Dougan B, Kistner J. Physically abused preschoolers’ responses to peers’ distress. Developmental Psychology. 1990;26:599–602.
  40. Maughan A, Cicchetti D. Impact of child maltreatment and inter adult violence on children’s emotion regulation abilities and socioemotional adjustment. Child Dev. 2002;73:1525–1542.
  41. Bernstein DP, Stein JA, Newcomb MD, et al. Development and validation of a brief screening version of the Childhood Trauma Questionnaire. Child Abuse Negl. 2003;27:169–190.
  42. Paivio SC, Cramer KM. Factor structure and reliability of the Childhood Trauma Questionnaire in a Canadian undergraduate student sample. Child Abuse Negl. 2004;28:889–904.
  43. Bernstein D, Fink L. Manual for the Childhood Trauma Questionnaire. San Antonio TX: The Psychological Corporation; 1998.
  44. Braun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology. 2006;3:77–101.
  45. Willig C. Introducing qualitative research in psychology: adventures in theory and method. Maidenhead United Kingdom: Open University Press; 2005.
  46. Johnson RB, Onwuegbuzie AJ. Mixed methods research: a research paradigm whose time has come. Educational Researcher. 2004;33:14–26.
  47. Freedberg D, Gallese V. Motion emotion and empathy in esthetic experience. Trends Cogn Sci. 2007;11:197–203.
  48. Burnette JL, Davis DE, Green JD, et al. Insecure attachment and depressive symptoms: the mediating role of rumination, empathy, and forgiveness. Personality and Individual Differences. 2009;46:276–280.
  49. Sandberg DA, Suess EA, Heaton JL. Attachment anxiety as a mediator of the relationship between interpersonal trauma and posttraumatic symptomatology among college women. J Interpers Violence. 2010;25:33–49.
  50. Gratz KL, Bornovalova MA, Delany-Brumsey A, et al. A laboratory-based study of the relationship between childhood abuse and experiential avoidance among inner-city substance users: the role of emotional nonacceptance. Behav Ther. 2007;38:256–268.
  51. Thabet AA, Tischler V, Vostanis P. Maltreatment and coping strategies among male adolescents living in the Gaza Strip. Child Abuse Negl. 2004;28:77–91.
  52. Loader P. Such a shame – a consideration of shame and shaming mechanisms in families. Child Abuse Review. 1998;7:44–57.
  53. Freud S. Three essays on the theory of sexuality and other works. vol 7. London United Kingdom: Penguin Books; 1977.
  54. Freud S. The standard edition of the complete psychological works of Sigmund Freud. Strachey J trans-ed. London, United Kingdom: Hogarth Press; 1986.
  55. Zepf S, Zepf F. Trauma and traumatic neurosis: Freud’s concepts revisited. Int J Psychoanal. 2008;89:331–353.
  56. Bateman A, Fonagy P. Psychotherapy for borderline personality disorder: mentalization-based treatment. Oxford United Kingdom: Oxford University Press; 2004.
  57. Alessandri SM. Mother-child interactional correlates of maltreated and nonmaltreated children’s play behavior. Development and Psychopathology. 1992;4:257–270.
  58. Fonagy P, Leigh T, Steele M, et al. The relation of attachment status, psychiatric classification, and response to psychotherapy. J Consult Clin Psychol. 1996;64:22–31.
  59. Widom CS. Does violence beget violence? A critical examination of the literature. Psychol Bull. 1989;106:3–28.
  60. Wiehe VR. Approaching child abuse treatment from the perspective of empathy. Child Abuse Negl. 1997;21:1191–1204.
  61. Maughan B, Rutter M. Retrospective reporting of childhood adversity: issues in assessing long-term recall. J Pers Disord. 1997;11:19–33.
  62. Aber JL, Zigler E. Developmental considerations in the definition of child maltreatment. New Directions for Child and Adolescent Development. 1981;11:1–29.
  63. Mrazek PJ, Mrazek DA. Resilience in child maltreatment victims: a conceptual exploration. Child Abuse Negl. 1987;11:357–366.
  64. Batson CD, Early S, Salvarani G. Perspective taking: Imagining how another feels versus imagining how you would feel. Personality and Social Psychology Bulletin. 1997;23:751–758.
  65. Hollingsworth J, Glass J, Heisler KW. Empathy deficits in siblings of severely scapegoated children: a conceptual model. Journal of Emotional Abuse. 2007;7:69–88.

CORRESPONDENCE: Melike M. Fourie, PhD, Department of Psychology, University of the Free State, Bloemfontein, South Africa, 9301 E-MAIL: marethem@gmail.com