Capacity for Empathy and Emotional Contagion in Those With Psychopathic Personalities

People with psychopathic traits are sometimes adept at recognizing the emotions of others and using this knowledge in anti-social ways. However, data from incarcerated psychopaths suggest that they are incapable of true empathy. In this paper, we describe three studies that link psychopathic personality to emotional contagion and empathy, and we offer suggestions for reconciling the seemingly conflicting data. While most studies of psychopathic personality assess incarcerated respondents, the resulting data may not be generalizable to noncriminals; participants in these studies were recruited from the general population. The research confirms that empathy and emotional contagion are positively correlated and that each is negatively correlated with psychopathy, as expected. Unique to these studies is the finding that, when instructed, those with psychopathic traits can easily “catch” the emotions of others via the steps of the emotional contagion pathway, thus implying their capacity for empathy. However, without instruction, those with psychopathic traits did not automatically catch others’ emotions.

of the PCL-R, but suggest that further research is needed with forensic and correctional samples. Nonetheless, they propose "the Hare SRP could be a good choice of measure to capture psychopathy in a broad range of individuals" (p. 251).

Continuum or Taxon
While the "psychopaths" described in the popular press are radically different from normal people, the vast body of current research contradicts this notion (Clark, 2007). Most analyses describe psychopathy in a manner that is consistent with general personality disorders-they are dimensional constructs that occur along a continuum (see, for example, the research of Hare & Neumann, 2006;Marcus, John, & Edens, 2004;Marcus, Lilienfeld, Edens, & Poythress, 2006;Marcus, Ruscio, Lilienfeld, & Hughes, 2008;Walters, Brinkley, Magaletta, & Diamond, 2008). However, it is important to note that some scholars disagree and describe psychopathy as a discrete taxon-more in line with its TV and literary depictions. The Hare Psychopathy Checklists (PCLs) measure psychopathic traits on a continuum, not discrete categories.

Assessing Empathy and Contagion
In a comprehensive study, Luckhurst (2012Luckhurst ( , 2014 argued that two personality trait deficits are pivotal to the psychopathic personality: a profound lack of empathy and a failure to share others' emotions (i.e. a lack of emotional contagion [EC]). She argued that differences in the possession of these traits should be apparent both when we compare those with psychopathic traits to a normal population and when we consider variations within a normal population. Let us now consider these two constructs, empathy and contagion, in more detail.

Empathy
Scholars have defined empathy in a bewildering array of ways. For example: "A sense of similarity in feelings experienced by the self and the other…." (Decety & Lamm, 2009, p. 199).
Some scholars consider empathy to be a thought process; others describe it as a feeling. Empathy might be an antecedent to interaction or an outcome of interaction. It might come from understanding another's perspective.
Or, empathy might be facilitated by an unconscious, automatic response. To address this confusion, Davis (1996) created a model that integrates the definitions described by contemporary theorists. The model is organized by the four components of an empathic event, in rough chronological order: antecedents, processes, intrapersonal outcomes, and interpersonal outcomes. Within any component would be the characteristics that impacted the empathic event; for example, the "antecedent" component might include the person involved and his or her personal biology, history, or knowledge. The antecedent step might also contain the situation and its characteristics, such as the physical setting or emotional climate. Each empathic event would have a different process and outcomes, depending on the characteristics that influenced it. Thus, Davis's model represents the Three Studies of Empathy, EC, and Psychopathy integration of many perceptions of empathy. Davis and other scholars have called this a multidimensional concept, which he defines simply as "a set of constructs having to do with the responses of one individual to the experiences of another" (p. 12).
In keeping with Davis's inclusive perspective, the authors chose to assess empathy with the Toronto Empathy Questionnaire (TEQ). The TEQ was developed by Spreng and his colleagues (2009), to capture many types of empathic responding (see Appendix B). The TEQ consists of 16 questions that use Likert-style scales to indicate frequency of empathic behaviors (never, rarely, sometimes, often, and always). The 16 items were extracted from existing self-report measures, including the IRI (Davis, 1996), the QMEE (Mehrabian & Epstein, 1972), the BEES (Mehrabian, 2000), the Hogan Empathy Scale (Hogan, 1969), as well as a few less-known scales, such as the Nursing Empathy Scale (Reynolds, 2000), the Scale of Ethnocultural Empathy (Wang et al., 2003), and the Measure of Emotional Intelligence (Schutte et al., 1998). The resulting 16 questions capture many characteristics that are traditionally associated with empathy, as well as a few newer concepts known to be related or predictive of empathy. The TEQ shows good test-retest reliability, strong construct validity, and high internal consistency. In short, this measure captures the variety of behaviors that are described in the current empathy literature. It is the psychopath's ability to experience empathy-specifically to respond to the experiences of another-that was the principal concern of this paper.

Emotional Contagion
Primitive emotional contagion is a process which is relatively automatic, unintentional, uncontrollable, and largely inaccessible to conversant awareness. Hatfield and her colleagues (1994) defined emotional contagion as: The tendency to automatically mimic and synchronize facial expressions, vocalizations, postures, and movements with those of another person's and, consequently, to converge emotionally (p. 5).
Scholars from a variety of disciplines provide voluminous evidence that people do in fact frequently catch one another's emotions (see Hatfield, Bensman, Thornton, & Rapson, 2014;Hatfield, Cacioppo, & Rapson, 1994 for a review of this research). When people mimic expressions of fear, anger, sadness, joy, love, or disgust, they tend to feel not just any emotion (be it positive or negative) but a pale reflection of the specific emotions they have mirrored. There is considerable evidence that the process of primitive emotional contagion occurs in three stages: Mimicry -> Feedback -> Convergence.
Mimicry: People tend to automatically synchronize facial expressions, vocal productions, postures, and movements with those of others.
Feedback: Subjective emotional experiences are generated by the afferent feedback from such mimicry.
The authors of the three studies in this paper measured contagion with the Emotional Contagion Scale (ECS).
The instrument was designed by Doherty (1997) to assess people's susceptibility to catching others' emotions (see Appendix C). Responders with high overall scores on the ECS have been shown to possess a high susceptibility to contagion. Scholars in many countries have used the scale to test a variety of hypotheses about the nature of emotion (see, for example : Hietanen, Surakka, & Linnankoski, 1998, who used the ECS Finnish version.) The ECS is the only current measure of primary contagion. Luckhurst, Hatfield, & Gelvin-Smith Several researchers have assessed the reliability and validity of the scale. Originally, R. William Doherty (1997) tested the ECS in a three-part study with a large sample of U.S. participants. He established its reliability and explored its factor structure. (Doherty intended a unidimensional structure that would reflect parsimony, but later research established that five factors were the best fit for the data [Lundqvist & Kevrekidis, 2008].) He compared the results of the ECS to the outcomes of a variety of psychological measures to confirm the validity of the EC Scale's construct. In 2006, Lars-Olov Lundqvist tested the factor structure and psychometric properties of the Swedish adaptation of the ECS. He found good test-retest reliability and described the multifaceted structure of the scale. He later (2008) found a strong correlation between results of the ECS and outcomes of the Temperament and Character inventory (TCI) of Cloninger, Svrakic, and Przybeck, (1993). The TCI measures four dimensions of temperament: novelty seeking, harm avoidance, reward dependence, and persistence, and three dimensions of character: self-directedness, cooperativeness, and self-transcendence.
The strong correlation found by Lundqvist between the two measures further confirms the validity of the ECS; Hatfield and her colleagues (1994) predicted that those who sense they are part of community, rather than alone in the world, should be highly susceptible to emotional contagion; Lundqvist (2008) showed this to be true.

Hypotheses
Given the nature of psychopathy, empathy, and emotional contagion, scholars have wondered how they are connected. The popular media promotes the idea that psychopaths have no empathy. As a result, many people in the general public believe this to be true. However, the authors found no empirical evidence to support this.
Anecdotes appear widely, especially concerning so-called "successful psychopaths," who live free in the community. Those with psychopathic traits are often portrayed as monsters.
The authors wondered why is this belief is so pervasive, and if it has blocked the progress of social science to finding the truth. Authors set out to find, first, if empathy and psychopathy correlated, either negatively or positively. Next, they considered whether emotional contagion, which appears critical to such emotional responses as empathy and compassion, was correlated with psychopathy. While answers to these questions might contradict "common knowledge" about psychopathy, researchers set up three inquiries to find out. They proposed two hypotheses to test this common knowledge: Hypothesis 1: emotional contagion and empathy are positively correlated.
Hypothesis 2: psychopathic personality, measured on a continuum, correlates negatively with both empathy and emotional contagion.
Researchers conducted two studies designed to address these two hypotheses. Based on the findings from the two studies, a third study was added to clarify results.

Study 1. Correlation of Empathy and Emotional Contagion (Luckhurst, 2012)
With permission of the University of Hawaii IRB and the consent of participants, Luckhurst (2012) asked students at a University in Hawaii to complete two questionnaires: the Toronto Empathy Questionnaire (TEQ) Three Studies of Empathy, EC, and Psychopathy and the Emotional Contagion Scale (ECS). Students completed a short demographic measure as well. As is typical of Hawaii's multicultural population, respondents came from a variety of cultural and ethnic backgrounds. They identified their ancestry as African, Chinese, European, Filipino, Hawaiian, Japanese, Korean, Pacific Islander, Samoan, Spanish/Mexican, and Vietnamese. The participants also identified with an array of religious groups and varied markedly in socioeconomic status and educational attainment.
She collected responses from a total of 268 undergraduate students, with 57% being women. Luckhurst found that, as anticipated, EC and Empathy were strongly correlated (r = .543, p ≤ .01), as shown in Figure 1. The r 2 value was .295.

Smith, 2015)
The Hawaii study was replicated with a student population in Alaska, where Gelvin-Smith (2015) surveyed students for their propensity for emotional contagion and their ability for empathy. In addition, a third variable was added: psychopathy. The author used the SRP-III to investigate the influence of psychopathy on previously-established correlative findings. She surveyed 86 students who were attending a University in Alaska. Rural Alaska students who were involved in distance education were also allowed to participate in the study. As is typical of the University, participants came from a variety of ethnic and religious groups.
Participants were informed of their rights and gave their consent to participate. Protocols were approved by both the University in Alaska IRB and the IRB of the University in Hawaii.
The Questionnaire included a demographic measure, which asked questions about gender, age, cultural identification, duration of U.S. residency, religious or spiritual affiliation, degree of religiosity, highest level of education, and family of origin's educational level. Also included were the Self Report Psychopathy Scale III (SRPIII) , the Toronto Empathy Questionnaire , and the Doherty Emotional Contagion Scale (Doherty, 1997). The three measures were formatted into a survey, which could be accessed through SurveyMonkey.
The results of this study provided strong support for Hypothesis 1; as previously shown by Luckhurst, Empathy and the EC measures correlated strongly with one another; in this case r = .67, p = .01. The SRPIII instrument was also completed by this set of students. As expected in Hypothesis 2, scores on the psychopathy measure were negatively correlated with both the Empathy Scale (r = -.29) and the EC Scale (r = -.30). p = .01 in both cases (Table 1). While 244 (53%) of participants had earned a college degree, 398 participants (86%) had attended some college.
The author first attempted to replicate the findings of Gelvin-Smith (2015) in Alaska. She had subjects complete the Emotional Contagion Scale (Doherty, 1997), the Toronto Empathy Questionnaire , and the Self-Report Psychopathy Scale . Her findings (see Table 2) were similar to those in the Alaska study. Empathy and the EC measures correlated strongly with one another (r = .60, p = .01). The scores for the SRPIII psychopathy measure were negatively correlated with both the TEQ (r = -.40) and the ECS (r = -.26); p = .01 in both cases.
Three Studies of Empathy, EC, and Psychopathy Hypothesis 2. Those with high-psychopathy personalities will possess less propensity for overall emotional contagion than do normal people.
Both of these hypotheses received strong support. As expected, the high-psychopathy group showed lower empathy scores than did the low-psychopathy group. This difference was significant at p < .001 level, t(144) = 6.83. Similarly, those in the high psychopathy group scored lower on emotional contagion than did those in the low psychopathy group. The difference was significant at the p < .001 level, t(144) = 4.34.

Study 3, Part 2: Comparisons of High and Low Scorers on the Psychopathy Measure (Luckhurst, 2014).
Psychopaths Lack Empathic Ability-but Why?
"Common knowledge" indicates that psychopaths have little empathy. Based on our correlation values, "common knowledge" seems to be true; the higher the level of psychopathy in men and women, the less empathetic and the less likely they are to share others' emotions. Luckhurst (2014) set out to find out why this is so. Primitive emotional contagion has been shown to occur in three stages: Mimicry -> Feedback -> Contagion.
Luckhurst (2014) conducted a second phase of her study to find out which stage was problematic in the highpsychopathy sample, relative to its low-psychopathy peers. Using the population previously recruited, she artificially broke down the steps of emotional contagion (mimicry, feedback, convergence) into three independent tasks. She tested the following hypotheses: Hypothesis 1. High-psychopathy respondents will be less able to mimic facial expressions than their normal peers.
Hypothesis 2. High-psychopathy respondents will be less able to experience afferent feedback than their normal peers.
Hypothesis 3. High-psychopathy respondents will experience actual emotional contagion less often than their low-psychopathy counterparts.
Luckhurst continued to work with the same subset of 146 participants derived from the larger sample of 461 recruited from Mechanical Turk. However, during this longer and more effortful portion of the study, several participants withdrew or were released from the study due to technology problems such as software incompatibility or audio limitations. This reduced the number of subjects to 104, with 53 in the high-psychopathy group and 51 in the low-psychopathy group.

Stage 1: Mimicry
To test whether psychopaths were able to identify and mimic others' emotions, she showed them pictures of people expressing one of the six universal emotions: anger, fear, surprise, disgust, happiness, or sadness Participants were asked to imitate the targets' facial expressions and hold those expressions for several seconds, following the general protocol established in 1990 by Levenson, Ekman, and Friesen (Appendix E).
(We have no way of knowing if participants actually followed the directions or how closely they adhered to them.) Next participants answered several questions, including their degree of success in mimicking, how they went about mimicking, and the degree of ease or difficulty in mimicking. Luckhurst found that her highpsychopathy subjects were every bit as good at mimicking as were low-psychopathy subjects; a t-test indicated no significant difference between high-and low-psychopathy groups, p > .05. Therefore, the first hypothesis was rejected; those who scored in the high-psychopathy group can respond normally, if instructed to mimic others' emotional displays. Participants also correctly identified emotional expressions on the faces of the targets (a test of Facial Affect Recognition [FAR]) at the same rate as their normal peers, p > .05.

Stage 2: Feedback
Hypothesis two tested the afferent feedback mechanism of the high-scoring vs. low-scoring participants. To test this, Luckhurst used the classic Directed Facial Action instructions of Levenson, Ekman, and Friesen (1990).
These instructions guided participants through muscle-by-muscle manipulation of their faces. They narrowed their eyebrows, pulled back the corners of their mouths, gritted their teeth, pursed their lips, and so on, to express each of the six universal emotions (see Appendix E). They held an expression for several seconds, Three Studies of Empathy, EC, and Psychopathy and then they were asked to identify their feeling from a list of emotions. ("I felt nothing" was also a possibility.) The authors have no way of knowing how closely the participants followed the directions in this step.
Once again, high-psychopathy participants showed no significant differences from their low-psychopathy peers in identifying their own emotions (a t-test indicated p > .05). Therefore, the second hypothesis was also rejected; high-scoring participants do have a normal afferent response, if directed to manually manipulate their own facial displays.

Stage 3: Convergence
Finally, to test hypothesis three, which asked whether high and low-psychopathy participants "catch" others' emotions at the same rate, the author showed each participant seven short video clips. The video clips were selected by undergraduate research students to represent the six universal emotions. Students demonstrated an inter-rater reliability of between 90-100% in their agreement of the emotion that was shown by each clip. The clips showed a person (the "target") or a group ("targets") in a life situation, for example eating a meal, at a sporting event, or talking with a friend (see Appendix F for a representative image from each film clip). Each clip lasted between eight and 15 seconds, and the order of the clips was randomly varied. This task was practiced once and repeated six times, so each participant saw clips of all six universal emotions.
Clips showed the facial expression of only the target(s)-such as jubilant faces in the crowd during 2011's World Series game-six home run-and did not show the facial expressions of others reacting to the target.
Hence, the participant had no cues from other responders. After watching a film clip, participants were asked to describe what emotions they felt while watching, and to quantify how strongly they felt the emotions. In addition, they were asked to identify the emotions expressed by the targets as a test of FAR.
It was here, watching emotions that occurred in a natural setting, that the differences between the highpsychopathy respondents and their normal peers emerged. Luckhurst once again found that the highpsychopathy respondents were able to identify others' emotions as well as their low-psychopathy counterparts did, p > .05. However, high-psychopathy group simply did not feel the emotions that they saw displayed in the short film clips at the same level as the low-psychopathy group, t(102) = 2.695, p < .01. These data supported the third hypothesis. In addition, MANOVA testing of three steps on two levels of psychopathy showed nonsignificant multivariate effects, p > .05. However, between-subjects effects were significant, p < .01, on the third step of EC, convergence.

Other Findings
Race and Gender Bias -Representations of race and gender were varied throughout this study to minimize effects of bias. The NimStim Facial Stimulus set contained photos representing Caucasian, African, Asian, and Figure 2). Six photos were used for each of six

Ingroup/Outgroup Bias -Protocols for Stage 1 and Stage 3 (of Study 3) contained questions from the Social
Distance Scale (Byrne, 1971) to check for in/out-group perceptions. This study found no differences between high-psychopathy and low-psychopathy respondents to questions that asked whether they could be friends with targets in the photos and videos, or whether the targets would fit into respondents' social groups. No patterns arose connecting respondents to targets of their own race or gender.

Discussion
In the second part of study three, the author artificially isolated the three steps of emotional contagion to identify which step(s) was unsuccessful for people with high-psychopathy personalities compared to low. If our methods measured the intended behaviors, then in two of the three steps, high-psychopathy respondents performed with no significant differences from their normal counterparts. However, in the third step, the high-psychopathy participants simply did not converge without prompting. This finding is consistent with "common knowledge" about psychopaths-they did not "catch" the emotions of others.
A number of explanations are possible for these interesting findings.

Susceptibility to Contagion
First, Hatfield and her colleagues (1994, p. 148) itemized factors that may determine people's susceptibility to contagion (in general).

Their level of attention to others
2. Their sense of being interrelated to others rather than being independent 3. Their ability to read others' emotions (including voice, gestures, postures)

Their level of physiological (autonomic) response
Any of these factors could influence the vulnerability of a mentally healthy person to contagion; these factors are likely to influence those with highly-psychopathy personalities as well. Of all the factors listed, the one most likely to have influenced the study three participants is factor number two: the sense that they are connected with others rather than independent. This factor was not tested in the study. All other characteristics listed here were measured, manipulated, or controlled. Thus, we suspect the five other characteristics are not responsible for the differences we found. We further hope that the sense of interrelationship felt by those with psychopathic characteristics will be investigated in future research. Social Psychologists, aware of the weakness of any one method, recommend that scholars try triangulationtesting one's theories with different populations, using slightly different assessment techniques, and employing a variety of dependent variables. In the Studies 1-3 (above) we used community and student populations, classified them as psychopathic or non-psychopathic using the SRPIII (a self report measure) and assessed our dependent variables with the Toronto Empathy Questionnaire (TEQ) and the Emotional Contagion Scale (ECS). In an effort to triangulate, we searched for other studies investigating questions similar to our own. We found many. In the following neuroscience studies, scholars studied prison populations, classified prisoners as psychopathic/non-psychopathic by a variety of techniques, and used fMRI techniques to assess psychopathsʻ ability to empathize with others and catch their emotions. Luckily, they came to the same conclusions we did.

Neuroscience Investigations
Neuroscientists have begun to try to discover psychopaths' exact biochemical deficiencies. Several recent studies have assessed brain glucose activity using fMRI techniques, including assessing the neural activity of psychopaths. Many scientists have noted a lack of activity or disrupted activity in the amygdalae of psychopaths (see, for example Cima, Tonnaer, & Hauser, 2010;Glenn, Raine, & Schug, 2009;Marsh & Blair, 2008). Abnormal activity is especially high when psychopaths are asked to consider moral dilemmas. However, Cima and colleagues found no differences in judgments that result from this abnormal amygdala activity. As a result, those authors reject the hypothesis that emotional processes are causally necessary for judgments about moral dilemmas. Instead, they suppose that moral judgment might occur first, followed by emotional decision-making. As a result, they conclude that while psychopaths show good moral judgments, they are unable to rely on emotions to implement them. Wilson, Juodis, and Porter (2011) point out that previous studies on facial affect recognition by psychopaths have produced conflicting results. They noted that emotion-processing deficits could not always be traced to amygdala dysfunction. These authors suggest that while the amygdala itself may not be impaired, its activation might in some way contribute to a left-brain mechanism that is dysfunctional, resulting in difficulty in information processing, including cognitive processes.
According to Kosson, Suchy, Mayer, and Libby (2002), psychopaths may have difficulty with left-hemisphere resources in general. The authors noted that some psychopaths use the right brain to process language-related information, a process that usually occurs in the left brain in healthy people. Right brain processing is likely inadequate, slow, and awkward, thus rendering those with psychopathic personalities able, but inefficient, at processing such information. Dolan and Fullam (2006) also found much slower recognition time for facial expressions by criminals with antisocial personality disorder relative to non-criminal healthy men. The influence of the amygdala and the use of right-brain versus left-brain processes may well explain why the autonomic emotional contagion process does not occur in psychopaths as it does in normal subjects.
Recently, a fascinating study conducted by Meffert and colleagues (2013) revealed more precise differences between normal and psychopathic neurology. In research conducted in the Netherlands, the authors used fMRI to compare the brain activity of non-psychopathic controls with that of criminal psychopaths. Participants watched video clips of hands interacting to indicate emotions (e.g., caressing, slapping). Resulting brain activity in criminal psychopaths was much lower than the brain activity in control subjects, especially in regions of the brain known to be connected to empathy (including temporal, insular, parietal and frontal lobe regions). However, when participants were instructed to empathize when watching the videos, differences in brain activity between the two groups were reduced. Authors conclude that "psychopathy is not a simple incapacity for vicarious activations but rather reduced spontaneous vicarious activations…" (p. 2550).
The studies reviewed in our paper were conducted in a natural setting, outside of the fMRI lab, yet the results were similar; response differences between those with high-psychopathy characteristics and control subjects were reduced when participants were prompted. The studies also suggest that psychopaths at large in the community respond in a manner similar to criminal psychopaths who are incarcerated.

Attention Deficit Disorders
Two recent papers suggest yet another reason why psychopaths may be unable to converge emotionally with others. Both papers discuss a type of attention deficit in psychopaths, which makes them easily distracted from the information necessary to understand others. Dadds and his colleagues (2012) found that children with oppositional defiant disorder, who tested high for Factor 1 (callous-unemotional) characteristics of psychopathy, showed low levels of eye contact and low levels of affection toward their mothers. The authors suggest, "psychopathic disorder begins as a failure to attend to the eyes of attachment figures" (p. 191). In an earlier study, Dadds et al. (2006) found that facial affect recognition (FAR) problems could be reduced or eliminated by instructing psychopaths to focus on the eye region of the target's face. While Luckhurst's participants did not struggle with FAR, their emotional convergence might have been deterred by lack of eye contact.
Sadeh and her colleagues (2013) argued that each psychopathy factor is associated with distinct cognitive and emotional deficits. The authors found that psychopathic individuals differed in their ability to process emotions and maintain attention according to their dominant (most high-scoring) factor. For example, the affectiveinterpersonal factor (Factor 1) was associated with enhanced sensitivity to positive stimuli. Participants showed increased error rates on cognitive tasks when positive distracters (words) were present. The social deviance factor (Factor 2) was associated with increased behavioral interference to both positive and negative stimuli, which resulted in delayed reaction times on cognitive tasks. This finding supports previous research that indicates that high Factor 2 individuals are "particularly sensitive to emotional context" (p. 291). In addition, the authors found that these deficits were interactional, not merely additive, in individuals who scored high in both factors. The authors conclude that each psychopathy factor may indicate predictable dysfunction in cognitive and affective processes. Further research may clarify whether emotional contagion is affected by one factor of psychopathic personality over another.

Limitations of This Work
The authors recognize that study three has several limitations. They include our use of Mechanical Turk as a research platform, the unanticipated self-conciousness of participants in the mimicry and afferent feedback steps of Study 3, and the lack of validation of the method of isolating steps of emotional contagion.

Use of an Untested Method
In isolating the steps of emotional contagion, we created steps that we believed were reasonable replications of events that occur involuntarily, leading to emotional convergence. We also believed that participants would follow our directions. However, our method has not been validated and we have no direct evidence that either belief is true. For example, in step one, participants were asked to mimic expressions they saw on faces in photos. Although responses from the participants indicate they did so, we have no direct evidence that this is true, or that the mimicry task is reasonable replication of mimicry as it would occur naturally.

Mechanical Turk (MTurk)
Many workers ("Turkers") use MTurk as a source of income, and they are motivated to participate in studies that pay for participation; we are not sure how this motivation might have affected recruiting or responses.
Based on the demographic information that we collected, we know that this sample of Turkers differed from the general US population in their levels of education (higher) and age (lower) as well as its proportion of women participating (higher) relative to their proportion in the general population.

Self-Consciousness
We did not anticipate the influence of self-consciousness on the study data. Several participants commented that they felt "silly" or "stupid" making the faces to complete the mimicry and afferent feedback tasks (Steps 1 and 2). While most participants were not in public places, some were around people: usually family members or roommates. Their consciousness of others may have negatively influenced their participation. That is, greater privacy probably would have increased their abilities to mimic and experience feedback, thus increasing their experience of emotional contagion.

Conclusion
In conclusion, we consider these new studies to offer evidence that "common knowledge" of reduced empathy in psychopaths appears to be true. However, we believe it was not the lack of capability to complete the cascade of steps that lead to emotional contagion that leads to the inability for empathy. Rather the lack of focus or lack of connectedness felt by the psychopath that leads to the reduced feelings. In Study 3, the direction provided by Luckhurst's protocol artificially focused participants on important cues that made the difference in performance between those with psychopathic traits and those without. The authors speculate that while those with psychopathic traits are able to catch others emotions, they do not do so if undirected. This speculation is supported by studies of patients on the Autism Spectrum who are known to pay little attention to the emotions of others naturally, but who are able to attend, identify, and understand others' emotions when prompted (see Evers, Kerkhof, Steyaert, Noens, & Wagemans, 2014;Sawyer, Williamson, & Young, 2012;Zager, Alpern, & Boutot, 2009). Many published studies provide insight in to the complex interplay of abilities and disabilities that shape the psychopath's cognitions, emotions, and behaviors. This same complexity was found in our studies.
While neuroscientists explore psychopaths' biochemical characteristics, perhaps social scientists will further investigate the processes at work in the emotional responses of those with psychopathic traits to social stimuli.
We hope the popular media and, in turn, the general public will begin to recognize psychopathic personality as an emotional disorder and relinquish the easy stereotype of psychopath-as-monster.

Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.