Socioaffective Dynamics of Psychopathy in Daily Life

Abstract

Affective and interpersonal features of psychopathy are considered hallmarks of the disorder. Ecological momentary assessment (EMA), well-suited to examine dynamic processes in day-to-day life, has not been used to study how psychopathy influences emotional experiences and interpersonal behavior in adults. This preregistered study examined how psychopathy relates to socio-affective processes in daily life. Two samples enriched for traits related to psychopathy (Sample 1 N=142; Sample 2 N=159) completed EMA protocols focused on a variety of interpersonal and affective experiences (Observation N=8,137 to 16,460). The samples differed in sex and socioeconomic, age, and ethnic diversity, which allowed us to examine the replicability and generalizability of results. Results showed that psychopathy was related to distinct affective experiences in both samples (e.g., increased hostile affect), but was unrelated to diversity in affective experiences, and rarely moderated within-person socio-affective processes. Future directions for research on the affective and interpersonal processes of psychopathy are discussed.

Psychopathy is a longstanding focus of clinical research due to its strong ties to externalizing behaviors (e.g., aggression and antisocial behavior; Neumann et al., 2015). Psychopathy is multidimensional, characterized by impulsivity-related traits (e.g., lack of planning, sensation seeking), a manipulative, grandiose interpersonal style, and affective deficits including low fear and lack of remorse (Patrick, 2022). The affective and interpersonal components of psychopathy are considered to be central (e.g., Verschuere et al., 2018), and are thought to partially explain chronic antisocial behavior (Frick, 2009). These deficits have primarily been studied with cross-sectional self-report instruments or experimentally, using various laboratory-based paradigms (e.g., startle-response, Patrick et al., 1993; facial affect recognition, Marsh & Blair, 2008). Results from these studies have motivated multiple models to account for the distinctive affective and interpersonal features of psychopathy (e.g., Nentjes et al., 2022). Yet no research has examined the links between psychopathic traits and affective and interpersonal processes in adults using ecological momentary assessment (EMA), despite the promise of EMA in helping understand various socio-affective dynamics relevant to psychopathy. The current preregistered study was conducted to address this gap, focusing on the role of psychopathic traits in various socio-affective processes in day-to-day life.

Conceptualizing the Affective and Interpersonal Features of Psychopathy

Theorists have emphasized the distinctive affective profile of psychopathy, and historical accounts of psychopathy consistently highlight affective deficits, particularly related to lack of guilt or remorse (Horley, 2014). In The Mask of Sanity, Cleckley (1941/1988) developed sixteen criteria thought to be hallmark features of psychopathy, with four focused on affective functioning (general poverty in affective reactions, absence of nervousness, lack of guilt and remorse, incapacity for deep affectional bonds). Cleckley (1941/1988) further clarified that the general deficit in affective experiences referred to the richness and strength of affective responses as opposed to a lack of any affective experiences at all. Lykken (1995) argued for the centrality of deficient fear responses in psychopathy (termed the “low-fear hypothesis”), stemming from his early research showing that individuals seen as prototypical Clecklian psychopaths showed lower fear responses to a conditioned stimulus that had been paired with an electric shock. The work of Cleckley and Lykken was particularly influential for later conceptualizations of psychopathy that emphasized low fearfulness and anxiety, social assertiveness, and general emotional resilience, which collectively fall under the umbrella of traits termed boldness or fearless-dominance (Benning et al., 2003; Patrick, 2022).¹ Forensic theorists emphasized that while psychopathic individuals showed deficiencies in remorse and fear, there were positive relations between psychopathy and other forms of negative affect—most notably anger and hostility (McCord & McCord, 1964).

Interpersonally, psychopathy is characterized by a glib, manipulative, and exploitative disposition towards others, resulting in psychopathic individuals being more likely to disregard the feelings or needs of others in favor of their own goals or pursuits. While the interpersonal and affective features of psychopathy are generally seen as central to the disorder, affective deficits occupy a more central role in developmental models of psychopathy seeking to explain psychopathy’s robust links with serious antisocial behavior (Blair, 2017; Frick et al., 2003). Developmental models of psychopathy generally emphasize the temporal precedence of affective deficits compared to interpersonal deficits, with a focus on temperamental dimensions of low emotional sensitivity and low fearlessness. Researchers have argued that these temperamental dimensions, alongside other developmental factors (e.g., harsh parenting), contribute to the full-fledged interpersonal and affective profile of psychopathy and corresponding antisocial behavior later in life (Waller & Hyde, 2018).

Nonetheless, due in part to their bidirectional development (Viding & McCory, 2018), the affective and interpersonal features of psychopathy are closely connected. Empirically, the interpersonal features of psychopathy are strongly related to affective features in cross-sectional studies. For example, in the widely used Psychopathy Checklist-Revised (PCL-R; Hare, 2003), the interpersonal and affective facet scales of the PCL-R are strongly related (e.g., r=.70; Hare & Neumann, 2008), and coalesce to form a single factor (termed Factor 1 or the affective-interpersonal factor). Other research has shown that that the strong empirical relations observed between the interpersonal and affective factors/subscales of psychopathy instruments can be accounted for by low agreeableness or antagonism (e.g., Sherman et al., 2014; Van Til et al., 2022), suggesting that the affective features of psychopathy are saturated with interpersonally related variance (e.g., though typically framed as an emotional deficit, lack of empathy is also interpersonal in nature). However, existing measures differ in how the interpersonal and affective features of psychopathy are represented.

Divergent Nomological Networks Between and Within Measures of Psychopathy

Despite the centrality of affective-interpersonal features of psychopathy, they are not represented consistently across measures. The inconsistency primarily revolves around the inclusion or exclusion of features tied to boldness or fearless-dominance (i.e., low fearfulness, low anxiety, social assertiveness, and general emotional resilience). Several popular self-report instruments explicitly include scales to assess low anxiety, low fear, and interpersonal dominance (e.g., the Triarchic Psychopathy Measure; Patrick et al. 2009; the Psychopathic Personality Inventory-Revised; PPI-R; Lilienfeld & Widows, 1994; the Elemental Psychopathy Assessment; Lynam et al., 2011) while other instruments assess such features with a few items (e.g., the PCL-R) or not at all (e.g., the Levenson Self-report Psychopathy scale; LSRP; Levenson et al., 1995). The inclusion or exclusion of fearless-dominance or boldness-related content leads to important empirical consequences regarding socio-affective outcomes.

For example, the meta-analytic relations between affective-interpersonal features of psychopathy and affective outcomes differ depending on the instrument under investigation. Meta-analytic evidence shows that affective-interpersonal features of psychopathy assessed by PCL-based instruments are unrelated to indices of negative affect, but positively related to lack of empathy (e.g., Derefinko, 2015; Campos et al., 2022). Instruments that include boldness-related content within the affective-interpersonal dimension, such as the PPI-R, show large differences. For instance, Fearless-dominance is strongly negatively related to trait negative affectivity (meta-analytic r of −.50; Miller & Lynam, 2012). Thus, differences in operationalization across psychopathy instruments pose a challenge to identifying a consistent nomological network between psychopathy and socio-affective functioning.

Empirical Evidence for Psychopathy and Socio-affective Functioning

Affective functioning spans multiple domains subject to different levels of analysis. In their review of psychopathy and emotional functioning, Nentjes et al. (2022) highlight research spanning perceptions and appraisals of stimuli, the experience, expression, and regulation of affect, and physiological components of affect. Interpersonal behavior can be similarly examined at different levels of analysis including motives, affect, behavior, and perceptions of self and other (Hopwood, 2018). From the perspective of interpersonal theory, affective and interpersonal functioning are intimately linked, and the study of their dynamics has important implications for understanding personality pathology (Wright et al., 2023). Cross-sectional and laboratory-based research has provided valuable data to motivate more dynamic approaches to studying socio-affective functioning in psychopathy. We first provide a brief overview of research on the affective correlates of psychopathy, before discussing research on psychopathy and more general outcomes tied to socio-affective functioning.

Multiple meta-analyses focused on cross-sectional studies have summarized the relations between psychopathy and theoretically relevant affective experiences (e.g., Hoppenbrouwers et al., 2016; Derefinko, 2015; Campos et al., 2022). Collectively, results have shown a constellation of positive and negative relations with different affective outcomes, and mixed findings for specific kinds of affect (e.g., fear). Hoppenbrouwers and colleagues (2016) found support for deficits in threat detection and responsivity, but no evidence for hypothesized negative relations with subjective reports of fear. Similarly, Derefinko (2015) found that total psychopathy scores were unrelated to anxiety and fear, but the interpersonal-affective components of psychopathy showed small negative relations with anxiety, fear, and constraint, and the antisocial-lifestyle components were positively related to anxiety, unrelated to fear, and strongly negatively related to constraint. Importantly, research has shown that the divergent findings across psychopathy components are also observed for other forms of negative affect (anger-hostility, depression), and these divergent relations are enhanced when controlling for overlap among the interpersonal-affective and antisocial-lifestyle features of psychopathy (e.g., Hicks & Patrick, 2006).

Other work has further emphasized that the affective profile of psychopathy is complex (see Garofolo & Neumann, 2017 for a review). For instance, Garofalo and colleagues (2019) found that psychopathy was strongly related to anger, spitefulness, and contempt, the latter two being complex affective states comprised of more basic affect components. However, consistent with other work (Hicks & Patrick, 2006), these associations varied when controlling for scale overlap and also varied across specific subscales. Boldness from the TriPM (Patrick et al., 2009) was unrelated to contempt and showed significant negative relations with various state and trait measures of anger and negative emotionality. Yet, all scales from the Self-report Psychopathy Scale (SRP; Paulhus et al., 2006) were significantly positively related to these affective outcomes.

Considering outcomes tied to socio-affective functioning more broadly, the relation between psychopathy and aggression has garnered significant research attention. Though this research does not tend to study the dynamic interplay between emotion and behavior involved in acts of aggression (i.e., aggression is most commonly assessed with self-report instruments), research has distinguished between reactive aggression (impulsive, enraged attack triggered by provocation or withdrawal of expected reward) and proactive aggression (reward-motivated and planful). Though reactive and proactive aggression show important divergences in their empirical relations (Crick & Dodge, 1996), meta-analytic evidence has found that psychopathy is similarly related to both forms of aggression (Blais et al., 2014), and there are not sizeable differences in the relations with reactive and proactive aggression when examining more specific features of psychopathy. More recent work using both offender and community samples (Garofalo et al., 2021) has found that the relations between psychopathy and different forms of aggression can be partially accounted for by psychopathy’s relation with impaired emotion regulation, highlighting the potential benefit of examining how affective experiences dynamically interface with interpersonal behavior.

Aggressive behavior is particularly indicative of impaired socio-affective functioning, but psychopathy has also been linked to stronger negative affective responses to less severe forms of interpersonal conflict. For example, Weiss and colleagues (2018) used a sample of newlywed heterosexual couples to examine how psychopathy traits related to affect expression when couples were instructed to discuss sources of tension in their relationship. For both women and men, psychopathy was related to stronger negative affect expression during the interaction. Lobbestael and colleagues (2018) found that in a sample of male undergraduates, individuals higher in Fearless-dominance (as measured by the PPI-R) responded with greater dominant behaviors when interacting with a more dominant interviewer (compared to a submissive interviewer), reflecting a deviation from the normative pattern of perceived dominance eliciting submissive behavior (i.e., complementarity; Tracey, 1994). While the results of these studies point toward heightened reactivity to interpersonal conflict and atypical responses to perceived dominance, other evidence highlights attenuated reactivity to indices of social affiliation.

Waller and colleagues (2021) explored how psychopathy related to sensitivity to affiliation, which refers to the degree of responsiveness to positive social cues that facilitate a sense of closeness with others (e.g., responding to and sharing others’ positive emotion). Results showed that psychopathy scores were significantly negatively related to sensitivity to affiliation for both men and women, and after accounting for the overlap between Factor 1 and Factor 2 of the Self-report Psychopathy Scale, only Factor 1 (interpersonal-affective features) remained negatively related to sensitivity to affiliation. Other research using a social discounting laboratory task (Sherman & Lynam, 2017) found that psychopathy was related to a lack of motivation to create or maintain positive social bonds with others. Specifically, participants were asked to identify relationships at varying levels of closeness, and then completed a series of trials where they had to choose between receiving a hypothetical monetary amount for themselves or splitting a hypothetical monetary amount with an individual at a given level of closeness. Sherman and Lynam (2017) found that those higher in psychopathy were quicker to disregard social closeness in favor of hypothetical monetary rewards, suggesting that social closeness was not as highly valued (relative to monetary gain) for those high in psychopathy. Only one study has examined the relations between psychopathy and socio-affective functioning in daily life through ecological momentary assessment, using a small sample of institutionalized adolescents. De Ridder and colleagues (2016) found that elevations on interpersonal-affective traits of psychopathy (assessed by collapsing the interpersonal and affective subscales of the Youth Psychopathic Traits Inventory; YPI; Andershed et al., 2002) were related to a greater degree of misbehavior over the course of the 8-day EMA protocol (including both rule-breaking and aggression), as well as a greater number of interpersonal conflicts with adult staff members.

This past research has laid the groundwork to explore the links between psychopathy and socio-affective functioning in more dynamic ways. For example, there is little research on how psychopathic traits relate to diversity in emotional experiences (i.e., the variety and relative amount of discrete emotion experiences; Benson et al., 2018), which is a separable component of affective functioning beyond average levels of negative or positive affect. Most focus on psychopathy and affective functioning is on between-person differences in emotional experiences, whereas emotional diversity is a within-person outcome involving fluctuations (or lack thereof) between discrete emotional states. While most conceptual work on psychopathy suggests a lack of emotional diversity (e.g., Cleckley, 1941/1988), there is little within-person focused research on emotional diversity in psychopathy.

In sum, available evidence points to psychopathic traits demonstrating a complex pattern of relations with socio-affective processes. Though dynamic approaches to studying psychopathy and socio-affective processes are relatively rare, existing research suggests that psychopathic traits may be related to the frequency of interpersonal conflicts individuals experience in daily life. In addition, psychopathic traits may enhance the degree of negative affect experienced during interpersonal conflict (particularly anger or hostility) while dampening the degree of positive affect or closeness felt during positive interactions with others. The present study sought to build on this past literature to better understand the relations between psychopathy traits and socio-affective dynamics, taking into account the oft-divergent relations between separate features of psychopathy and socio-affective functioning.

The Current Study

A noteworthy feature of research on psychopathy and socio-affective functioning reviewed above is that this work has been conducted in the laboratory using experimental paradigms, or in cross-sectional studies with retrospective self-report measures. To date, no research has examined the socio-affective dynamics of psychopathy in daily life using adult samples. This is an important gap to address—the problems and dysfunction associated with psychopathy occur in the ebb and flow of day-to-day life and in interactions with others. Thus, using methods that can assess interpersonal dynamics that unfold over time has significant promise for studying the affective and interpersonal features of psychopathy. EMA methods can also serve to establish the generalizability of laboratory-based research on socio-affective functioning in psychopathy.

EMA studies involve the repeated collection of real-time data focused on participants’ experiences and behaviors in the natural environments where they spend their lives (Shiffman et al., 2008). For those with personality pathology, EMA can provide important insights into the contexts where personality pathology is most disruptive. The present study makes use of different sampling time frames to gain insight into the socio-affective processes related to psychopathy (three days of intensive pseudo-random sampling; three weeks of moderate pseudo-random sampling). The EMA methods in the current study can also help to establish the ecological validity and boundary conditions of laboratory results (Trull & Ebner-Priemer, 2013). Calls to study personality pathology in ways that synchronize theory and method have been made recently (e.g., Wright & Zimmermann, 2019), but have not been incorporated into psychopathy research. The current study sought to address this important gap in research on socio-affective dynamics of psychopathy.

Moreover, we used two samples that allowed us to examine the consistency of results across sex as well as other important factors of diversity including socioeconomic, age, and ethnic diversity. Thus, the present study was able to evaluate both the replicability and generalizability of findings across two samples enriched for traits relevant to the socio-affective dynamics of psychopathy (i.e., high hostility and aggressiveness; see sample descriptions below). Hypotheses are detailed at the end of the Method section after introducing study-specific constructs and measures.

Method

All study procedures were approved by the Institutional Review Board of the University of Pittsburgh (IRB Protocol #: 000515 and 13050549). All data and code to reproduce our results are available at the Open Science Foundation page for the project: https://osf.io/quhgm. The preregistration is available at https://osf.io/zwg5q. All deviations from the preregistration are noted explicitly in the manuscript. The preregistration also includes a detailed overview of past publications using these data and what information was known prior to the registration of our analytical plan. This study involved secondary analyses of existing data, thus sample size determination was not a part of the preregistration.

Sample 1

Sample 1 (S1) comprised 145 women drawn from the larger Pittsburgh Girls Study (PGS). The PGS involves a community sample of 2,450 women who were initially recruited in 1999 and 2000 when they were ages 5 to 8 years old. An in-depth overview of the PGS is described elsewhere (Keenan et al., 2010). Our secondary data analysis focused on a subsample of the PGS, selected using a two-stage process. First, women from the PGS who reported verbal, physical, or relational aggression or self-harm/suicidality within the past year were identified via items administered annually in the PGS assessment battery over the course of three years (2014–2017). These participants were then contacted by telephone for further screening to determine whether they had engaged in externalized aggressive behavior in the past month (e.g., yelling/screaming, insulting or calling someone names, hitting or punching someone, threatening to beat someone up) or self-harming behaviors or suicidal ideation. Women who endorsed any of these behaviors in the past month were included. The final sample size of N=142 are those women who were selected for the study, completed the Psychopathic Personality Inventory-Short Form at baseline, and completed at least 10 EMA assessments (protocol described below). We reported in the preregistration that a sample size of 145 would be used for our analyses. However, for three women, full PPI data needed to compute factor scores were not available and thus our final sample size was N=142 for analyses.

The sample was ethnically and socioeconomically diverse (69.7% African American, 28.1% Caucasian, 2.1% multi-racial; 2.1% identified as Hispanic; 52.1% reported receiving public assistance). In terms of education attainment, 51.4% did not complete high school while 41.5% completed high school or GED only, and 34% were unemployed. The mean age of the sample was 21.54 years old (SD=1.58).

S1 Procedure

After completing baseline assessments, participants were trained to use study-provided, touch-screen smartphones to complete the EMA protocol. The protocol involved three weeks of assessments, including a daily morning assessment and six daily pseudo-random prompts. For pseudo-random prompts, participants provided their typical waking hours, which were split into six bins within which prompts were delivered randomly. EMA items were rated using a Likert-scale ranging from 1 (not at all) to 5 (extremely). Pseudo-random prompts were the focus of the present analyses, since they assessed specific interpersonal experiences and allowed for examination of relevant socio-affective processes. For the pseudo-random prompts, participants completed a mean of 115.9 out of a maximum of 126 (6 prompts x 21 days; 91.9% compliance).

Sample 2

Sample 2 (S2) included 159 middle-aged adults (M=40.67; SD=5.83; 46.3% male) selected for high scores on standard self-report measures of trait hostility. In the preregistration, we reported that a final sample size of N=160 would be used for analyses. However, one participant had daily diary data available but no NEO-PI-R data available and this participant was excluded from subsequent analyses. Participants were predominantly white (80.6% white; 12.5% African American; 1.9% Asian; 1.9% Hispanic) and had higher educational attainment compared to S1 (12.8% of the sample completed high school or GED only). Approximately 41.8% of the sample reported income greater than $50,000 per year.

Full details of the study procedure are available in Kamarck et al. (2009). Briefly, potential participants were screened over the telephone on two standard measures of hostility (Buss-Durkee Motor Aggression subscale; Buss & Durkee, 1957; 10 items from the Cook-Medley Hostility scale; Cook & Medley, 1954). Participants who passed the telephone screen were invited to complete further in-person screening which included a more complete set of hostility items from the Buss–Durkee Motor Aggression subscale (43 items) and the Cook–Medley subscales (27 items). Individuals meeting inclusion criteria completed a larger battery of self-report and laboratory assessments, before completing a four-day EMA protocol that was preceded by a practice day.

S2 Procedure

The EMA protocol consisted of a 47-item self-report questionnaire developed for repeated assessment of participants’ mood and social experiences and was administered using palm pilots (Palm Pilot Professional, Palm, Santa Clara, California) provided to participants. An auditory prompt was delivered every 45 minutes during waking hours, alerting participants to complete the survey. Thus, participants could complete a total of 21 prompts per day. All EMA items were rated using a 0-100 point sliding visual analogue scale, with “No” and “Yes” displayed at either pole. Practice day data were not included in our analyses, consistent with past use of these data (Vella et al., 2012). The mean number of completed entries per participant was 59.44, giving an average compliance rate of approximately 70.7%.

Measures

Descriptive information for the primary self-report scales and EMA variables are presented in Table 1.

Table 1.

Descriptive Information for Between- and Within-person Primary Variables

	Between-person (Level 2)			Within-person (Level-1)
	Mean	SD	ω	N	Min.	Max.	Mean	SD	ICC	ω
Sample 1 (N=142)
PPI-SF-Total	118.28	15.69	--	--	--	--	--	--	--	--
PPI-SF-FD	47.7	7.97	.23	--	--	--	--	--	--	--
PPI-SF-IA	56.2	11.40	.64	--	--	--	--	--	--	--
PPI-SF-Cold	14.38	4.09	.73	--	--	--	--	--	--	--
Positive Affect	7.26	2.58	.97	16,460	4	20	7.36	4.10	.41	.90
Hostile Affect	4.04	1.14	.92	16,458	3	15	4.03	2.06	.31	.76
Sad Affect	3.85	0.97	.81	16,455	3	15	3.84	1.72	.32	.67
Remorseful Affect	2.40	0.56	.40*	16,462	2	10	2.39	1.06	.28	.23*
Negative Interaction	2.24	0.51	.79*	14,219	2	10	2.24	.93	.31	.18*
Pos. Feedback	1.22	0.33	--	14,206	1	5	1.21	.66	.25	--
Sample 2 (N=159)
NEO-PPI-Total	98.55	9.74	--	--	--	--	--	--	--	--
NEO-PPI-FD	50.91	8.82	.80	--	--	--	--	--	--	--
NEO-PPI-IA	47.64	6.90	.70	--	--	--	--	--	--	--
NEO-Tender (r)	21.06	4.34	.70	--	--	--	--	--	--	--
Positive Affect	178.51	55.57	.92	9,451	0	400	180.30	67.48	.40	.82
Hostile Affect	80.50	41.43	.94	9,451	0	300	79.51	53.58	.37	.80
Sad Affect	53.86	40.31	.89	9,451	0	300	52.33	51.48	.60	.58
Anxious Affect	48.82	35.02	.77*	9,451	0	200	47.73	46.51	.55	.25*
Negative Interaction	51.37	33.41	.97	8,137	0	300	50.89	60.18	.31	.82
Pos. Feedback	48.13	16.33	--	8,137	0	100	48.50	30.93	.29	--
Agreeable Interaction	205.03	33.72	.98	8,137	0	300	206.11	65.31	.27	.92
Closeness Interaction	110.20	39.63	.83	8,137	0	300	109.10	75.46	.27	.80

Note: PPI-SF=Psychopathic Personality-Short Form; NEO-PPI=NEO-PI-R Proxy measure of the Psychopathic Personality Inventory; IA=Impulsive Antisociality factor; FD= Fearless-dominance factor; PPI-SF-Cold=Coldheartedness subscale; NEO-Tender (r)=reverse scored Tendermindedness facet; *=value represents the squared correlation between the two items used for the composite; ω is not estimated for total score scales as they are not intended to be unidimensional scales.

Self-reported Psychopathy

Sample 1

Participants in S1 completed the Psychopathic Personality Inventory-Short Form (PPI-SF; Lilienfeld & Hess, 2001). The PPI-SF is a 56-item self-report measure of psychopathy comprised of eight subscales. Empirical investigations of the parent measure, the Psychopathic Personality Inventory-Revised (PPI-R; Lilienfeld & Widows, 1994), have shown that the subscales form two psychopathy dimensions termed Impulsive Antisociality (made up of the Machiavellian Egocentricity, Carefree Nonplanfulness, Blame Externalization, and Impulsive Nonconformity subscales) and Fearless-dominance (made up of the Social Potency, Fearlessness, and Stress Immunity subscales) (Benning et al., 2003). The Coldheartedness scale does not load on either factor and is considered a standalone subscale.

Based on reviewer feedback, the Coldheartedness subscale was used for non-preregistered secondary analyses. In the current study, internal consistency values were α=.82 for Impulsive Antisociality, α=.70 for Fearless-dominance, and α=.73 for the Coldheartedness subscale. Internal consistency for the PPI-SF-Total score was α=.80. Impulsive Antisociality and Fearless-dominance were significantly positively correlated (r=.19; p=.02). PPI-Coldheartedness showed a small positive correlation with Fearless-dominance (r=.14; 95% CI= −.05; .31) and a small non-significant negative relation with Impulsive Antisociality (r=−.08; 95% CI= −.26; .12).

Sample 2

Participants completed the NEO-Personality Inventory-Revised (NEO-PI-R; Costa & McCrae, 1992), a 240-item self-report measure based upon the Five-factor Model of personality. Past work has shown that a reliable and valid proxy measure of the PPI-R can be derived using items from the NEO-PI-R (Witt et al., 2009). Thus, we took the same approach used by Witt and colleagues (2009) to construct scales for Impulsive Antisociality (17 items), Fearless-dominance (17 items), and a total score (34 items) using a subset of NEO-PI-R items. Internal consistency values were α=.69 for NEO-Impulsive Antisociality, α=.80 for NEO-Fearless-dominance, and α =.64 for the NEO-PPI-Total score. NEO-Impulsive Antisociality and NEO-Fearless-dominance were significantly negatively correlated (r= −.23; p <.01).

Based on feedback during the review process, we also used the reverse-scored Tendermindedness facet scale in S2 (hereafter referred to as (low) Tendermindedness) as a focal scale for non-preregistered secondary analyses. The original derivation of the PPI proxy scales using NEO-PI-R items (Witt et al., 2009) did not include a proxy scale of Coldheartedness. (Low) Tendermindedness was selected given the content overlap it shared with the Coldheartedness scale from the PPI-R-SF. Specifically, the Tendermindedness facet, “measures attitudes of sympathy and concern for others…low scorers are more hardheaded and less moved by appeals to pity. They would consider themselves realists who make rational decisions based on cold logic.” (Costa & McCrae, 1992; p. 18), whereas Coldheartedness, “measures a propensity toward callousness, guiltlessness, and unsentimentality.” (Lilienfeld & Andrews, 1996; p. 495). Internal consistency for (Low) Tendermindedness was α=.70. (Low) Tendermindedness and NEO-PPI-Fearless-dominance were positively correlated (r=.16; 95% CI= .00; .30) while (low) Tendermindedness showed a small but non-significant positive relation with NEO-PPI-Impulsive Antisociality (r=.04; 95% CI= −.11; .21).

Ambulatory Assessment

See Table 1 for EMA variable descriptives, including ICCs and multilevel reliabilities.

Positive and Negative Affect

Three affect composite scales were created in S1 and S2. Four positive affect items (excited, happy, cheerful, joyful) were combined to create positive affect scales in S1 and S2. Negative affect composites used in both S1 and S2 included a hostile affect composite of three items (hostile, irritable, angry2) and a non-hostile negative affect composite of three items (scared, sad, lonely). In S1, two additional affect items (guilty, regretful) were combined to index remorse-related affect. In S2, an additional two affect items (nervous, jittery) were combined to index anxiety-related affect.

Interpersonal Interactions

Negative Interpersonal Interactions.

In S1, perceived negative interpersonal interactions were assessed with two EMA items. At each random prompt, participants were asked if they had 1) felt insulted or criticized and 2) felt rejected, abandoned, excluded, or left out since they had received the last prompt. The two items were combined to create a negative interpersonal interaction scale.

In S2, perceived negative interpersonal interactions were assessed with three EMA items. During each EMA prompt, participants were asked when their most recent social interaction was. If participants reported that the interaction was less than 45 minutes ago, they were asked whether 1) Someone treated you badly 2) Someone interfered with your efforts and 3) Someone was in conflict with you. The three items were combined to create a negative interpersonal interaction scale in S2.

Positive Interpersonal Interactions.

In S1, participants were asked if they had felt complimented or praised since they had received the last prompt. This single item was used to index perceived positive feedback interactions.

In S2, a larger number of EMA items assessed perceived positive interpersonal interactions. We examined a single EMA item asking participants whether Someone gave you positive feedback in order to have a comparable positive interaction outcome in both S1 and S2. Other positive interpersonal EMA items administered in S2 had participants rate whether the interaction was 1) Pleasant, 2) Agreeable, 3) Friendly, and 4) Intimate/close, and to what extent participants 5) Felt close to others and 6) Confided in others. Pleasant, agreeable, and friendly interaction items were combined to create a perceived agreeable interaction composite, and the intimate, felt close, and confided items were combined to create a perceived intimacy interaction composite.

Primary Hypotheses³

We had several primary hypotheses. First, we expected divergent relations for Fearless-dominance and Impulsive Antisociality. We hypothesized that Impulsive Antisociality would be positively related to all negative affect and perceived negative interactions, but negatively related to positive affect and perceived positive interpersonal interactions. Furthermore, Impulsive Antisociality would significantly enhance the within-person links between perceived negative interactions and different forms of negative affect, with the exception of remorseful affect. Last, we expected Impulsive Antisociality to have an enhancing effect on each path of our within-person mediation model focused on hostility and perceived negative interpersonal interactions (described below). We hypothesized that Fearless-dominance would be negatively related to all forms of negative affect as well as frequency of perceived negative interactions, and positively related to positive affect. Second, we expected FD-traits to have a dampening effect on within-person links between perceived negative interactions and negative affect. We did not have specific hypotheses for Fearless-dominance regarding its enhancing or dampening effects on perceived positive interactions and positive affect. Regarding emotional diversity (i.e., evenness and richness of affective experiences), we hypothesized that psychopathy traits would generally be linked to lower emotional diversity (i.e., negatively correlated).

Results

Average Momentary Affect and Interpersonal Outcomes

We examined the relations between psychopathy and average reported affect and interpersonal outcomes over the course of the EMA protocols in S1 and S2 using multilevel linear regression models with robust maximum likelihood estimation due to the nesting of observations within participants. All level-2 psychopathy predictors were grand-mean centered for analyses. Time of day (assessed continuously within each day) and day of the week (weekday vs. weekend) were included as within-person covariates in each model.

Results are displayed in Table 2. Overall, findings showed few replicable relations between psychopathy traits and average momentary affect in S1 and S2. The exception was for hostile affect. In both S1 (β=.28; 95% CI=.14; .42) and S2 (β=.21; 95% CI=.02; .40), Impulsive Antisociality was significantly positively related to hostile affect. Three other effects were statistically significant. In S1, there was a significant negative relation between PPI-SF-Fearless-dominance and hostile affect (β=−.20; 95% CI=−.38; −.02). In S2, while NEO-Impulsive Antisociality was positively related to sad affect (β=.20; 95% CI=.03; .37), NEO-Fearless-dominance showed a negative relation (β=−.21; 95% CI=−.35; −.06).

Table 2.

Relations Between PPI Scales and Average Affect Ratings During EMA Protocols

	Hostile Affect		Sad Affect		Positive Affect		Remorseful Affect		Anxious Affect
	Estimate (SE)	95% CI	Estimate (SE)	95% CI	Estimate (SE)	95% CI	Estimate (SE)	95% CI	Estimate (SE)	95% CI
Sample 1
PPI-SF-IA	.28** (.07)	.14; .42	.14 (.08)	−.02; .30	−.06 (.08)	−.21; .09	.11 (.08)	−.05; .27	--	--
PPI-SF-FD	−.20* (.09)	−.38; −.02	−.16 (.11)	−.37; .05	.04 (.08)	−.11; .19	−.20† (.10)	−.41; .00	--	--
PPI-SF-Cold	−.09 (.08)	−.25; .09	−.18* (.08)	−.33; −.04	−.22** (.07)	−.37; −.08	−.24** (.06)	−.36; −.11	--	--
PPI-SF-Total	.08 (.07)	−.06; .22	−.03 (.09)	−.21; .15	−.08 (.07)	−.23; .06	−.09 (.08)	−.24; .07	--	--
Sample 2
NEO-PPI-IA	.21* (.10)	.02; .40	.19* (.09)	.02; .36	−.06 (.09)	−.24; .11	--	--	.08 (.08)	−.07; .24
NEO-PPI-FD	−.10 (.08)	−.26; .05	−.21* (.07)	−.35; −.07	.13 (.08)	−.03; .29	--	--	−.09 (.08)	−.26; .07
NEO-Tender (r)	.00 (.07)	−.13; .14	−.08 (.08)	−.23; .06	−.07 (.09)	−.25; .11	--	--	.02 (.08)	−.13; .08
NEO-PPI-Total	.05 (.08)	−.10; .21	−.05 (.08)	−.20; .10	.07 (.09)	−.09; .24	--	--	−.03 (.09)	−.20; .15

Note: PPI-SF=Psychopathic Personality-Short Form; NEO-PPI=NEO-PI-R Proxy measure of the Psychopathic Personality Inventory; IA=Impulsive Antisociality factor; FD= Fearless-dominance factor; PPI-SF-Cold=Coldheartedness subscale; NEO-Tender (r)=reverse scored Tendermindedness facet; All values are standardized estimates to facilitate comparisons across samples; †= p<.10; *=p<.05; **=p<.01.

For aggregated interpersonal outcomes, most results showed null relations for psychopathy dimensions in both S1 and S2 (Table 3). The lone exception was a significant positive relation between PPI-Impulsive Antisociality and perceived negative interpersonal interactions in S1 (β=.18; 95% CI=.09; .28), meaning that individuals higher in Impulsive Antisociality tended to perceive interpersonal interactions as more negative, on average.

Table 3.

Relations Between PPI Scales and Average Interpersonal Interaction Ratings During EMA Protocols

	Negative Interactions		Praise/Positive Feedback		Agreeable Interactions		Intimacy Interactions
	Estimate (SE)	95% CI	Estimate (SE)	95% CI	Estimate (SE)	95% CI	Estimate (SE)	95% CI
Sample 1
PPI-SF-IA	.18** (.05)	.09; .28	.08 (.08)	−.07; .23	--	--	--	--
PPI-SF-FD	−.07 (.08)	−.22; .08	.15 (.10)	−.04; .33	--	--	--	--
PPI-SF-Cold	−.13* (.05)	−.23; −.02	−.27** (.07)	−.41; −.13	--	--	--	--
PPI-SF-Total	.06 (.05)	−.04; .17	.06 (.08)	−.09; .22	--	--	--	--
Sample 2
NEO-PPI-IA	.13 (.08)	−.05; .30	.06 (.09)	−.11; .23	−.16† (.09)	−.33; .02	.09 (.09)	−.07; .26
NEO-PPI-FD	.04 (.08)	−.12; .20	−.02 (.08)	−.17; .14	.02 (.08)	−.13; .16	−.05 (.08)	−.21; .12
NEO-Tender (r)	−.02 (.07)	−.16; .12	−.11 (.09)	−.29; .06	−.02 (.07)	−.16; .12	−.23** (.08)	−.38; −.08
NEO-PPI-Total	.13 (.08)	−.02; .27	.02 (.08)	−.13; .18	−.10 (.08)	−.25; .06	.03 (.08)	−.13; .18

Note: PPI-SF=Psychopathic Personality-Short Form; NEO-PPI=NEO-PI-R Proxy measure of the Psychopathic Personality Inventory; IA=Impulsive Antisociality factor; FD= Fearless-dominance factor; PPI-SF-Cold=Coldheartedness subscale; NEO-Tender (r)=reverse scored Tendermindedness facet; All values are standardized estimates to facilitate comparisons across samples; †= p<.10; *=p<.05; **=p<.01.

Emotional Diversity

Individual scores for emotional diversity were calculated using metrics that quantify the evenness/distribution of emotional experiences (Gini coefficient) and the total number of emotions experienced (Simpson’s index) during the EMA protocol using R code adapted from Benson et al. (2018).⁴ Individual differences in these indices were correlated with each psychopathy scale (Fearless-dominance, Impulsive Antisociality, and total score). Scores were calculated for global, positive, and negative emotional diversity. Results for the relation between psychopathy features and emotional diversity are presented in Table S1 while Figure 1 displays emotional diversity data for select individuals in S1 and S2 high in psychopathy (i.e., one standard deviation above the sample mean).⁵ Overall, results showed that null effects were most common (r range= −.17-.15; median r = −.01). PPI Total scores in S1 showed a significant negative relation with negative Emodiversity (r = −.17, 95% CI= −.32; .00), though this was confined to Simpson’s index of negative emotional diversity.

Figure 1. Between-person Heterogeneity in Affect Ratings for Select Participants High in Psychopathy in Samples 1 and 2.

Note: PPI-SF=Psychopathic Personality Inventory-Short Form; NEO-PPI=NEO-PI-R Proxy Scale of the PPI-R; IA=Impulsive Antisociality; FD=Fearless-dominance; Sample 1 participants are displayed in the first row and Sample 2 participants are shown in the second row; All participants displayed in the figure scored at least 1 SD above their respective sample average on total psychopathy scores; The length of each “petal” indicates the number of occasions each emotion was reported by the participants; the colors of each “petal” reflect the proportion of occasions the emotion was rated at low (darker shades closer to edge) to high (lighter shades, closer to center) intensities; Positive and negative emodiversity are Gini coefficients, which range from 0 to 1, with higher values reflecting more diverse emotional experiences. Gini coefficients reflect the variety and relative abundance of emotions that individuals reported during the respective EMA protocols in Samples 1 and 2.

Dynamic Affective and Interpersonal Outcomes

We also examined whether psychopathic traits moderated within-person affective processes following different types of interpersonal interactions. Specifically, we examined how perceived negative interactions impacted feelings of hostile affect and non-hostile negative affect and how perceived positive interactions impacted positive affect. Next, we tested whether level-2 psychopathy scales enhanced or dampened the level-1 relation between perceived negative interpersonal interactions and the two negative affect scales (i.e., a cross-level interaction effect).

Additionally, we extended these basic dynamic models to test a dynamic mediational model involving hostile affect and perceived negative interpersonal interactions. Specifically, we evaluated whether the strength of the relation between lagged hostile affect (hostile affect at time t-1) and current hostile affect (hostile affect at time t) was partly mediated by perceived negative interpersonal interactions, and to what extent psychopathy traits moderated these pathways.

We note that across S1 and S2, participants reported on interpersonal interactions that had occurred since the last prompt. Thus, while participants provided information about affect and interpersonal interactions during the same EMA assessment, reports on interpersonal interactions were retrospective. This allows for the appropriate temporal sequencing between interpersonal situations and affective responses.

We used multilevel structural equation models (MSEM; Sadikaj et al., 2021) to test all dynamic models involving interpersonal interactions and affect. MSEM was carried out using Bayesian model estimation. Diffuse, default priors were used for all model parameters. Psychopathy scales were grand-mean centered for all models. Time of day and day of the week served as within-person covariates. Last, lagged hostile affect was person-mean centered in models estimating the mediational effects of hostility. All MSEM and multilevel linear regression models were implemented in MPlus (version 8.7; Muthén & Muthén, 2017).

Results from these models are displayed in Table 4. At the within-person level, there was robust evidence that different types of interpersonal situations gave rise to specific affective responses in both S1 and S2. Specifically, relative to participants’ average levels, perceived negative interactions increased feelings of hostile affect (S1: β=.39; 95% CI=.36; .42, S2: β=.42; 95% CI=.39; .45), sad affect (S1: β=.38; 95% CI=.34; .40, S2: β=.16; 95% CI=.13; .19), remorseful affect (S1: β=.26; 95% CI=.23; .29), and anxious affect (S2: β=.18; 95% CI=.15; .21). Additionally, receiving positive feedback increased positive affect (S1: β=.26; 95% CI=.22; .28, S2: β=.28; 95% CI=.25; .30), as did agreeable interactions (S2: β=.38; 95% CI=.35; .40) and intimate interactions (S2: β=.24; 95% CI=.22; .27). However, psychopathy features did not generally moderate these within-person responses to interpersonal interactions. One cross-level moderation effect was observed in S1, where PPI-Fearless-dominance significantly dampened the relation between perceived negative interpersonal interactions and sad affect.

Table 4.

PPI Moderation ofWithin-person Affective Processes Following Interpersonal Interaction

	Neg. Int.→Hostile Affect		Neg. Int.→Sad Affect		Neg. Int.→Remorseful Affect		Neg. Int.→Anxious Affect		Pos. Feedback→Pos.Affect		Agr. Interaction→Pos. Affect		Int. Interaction→Pos. Affect
	Estimate (SD)	95% CI	Estimate (SD)	95% CI	Estimate (SD)	95% CI	Estimate (SD)	95% CI	Estimate (SD)	95% CI	Estimate (SD)	95% CI	Estimate (SD)	95% CI
Sample 1
PPI-SF-IA	.04 (.07)	−.11; .18	−.07 (.07)	−.21; .07	.03 (.07)	−.13; .15	--	--	−.05 (.07)	−.16; .10	--	--	--	--
PPI-SF-FD	.07 (.07)	−.09; .21	−.19* (.08)	−.35; −.05	−.06 (.07)	−.22; .08	--	--	.11 (.07)	−.01; .27	--	--	--	--
PPI-SF-Cold	−.01 (.07)	−.15; .11	.03 (.08)	−.13; .16	.05 (.08)	−.12; .19	--	--	.01 (.08)	−.15; .18	--	--	--	--
PPI-SF-Total	.06 (.07)	−.09; .19	−.14† (.08)	−.28; .00	−.01 (.07)	−.16; .14	--	--	.06 (.07)	−.05; .22	--	--	--	--
Sample 2
NEO-PPI-IA	−.02 (.07)	−.14; .12	−.06 (.07)	−.19; .08	--	--	.05 (.09)	−.12; .19	.05 (.08)	−.12; .18	.04 (.07)	−.10; .18	.01 (.07)	−.11; .16
NEO-PPI-FD	.01 (.08)	−.14; .14	−.01 (.07)	−.16; .13	--	--	.06 (.07)	−.08; .21	.02 (.07)	−.12; .16	.10 (.08)	−.06; .23	.06 (.07)	−.09; .21
NEO-Tender (r)	.02 (.07)	−.13; .16	−.15† (.07)	−.29; .01	--	--	−.09 (.08)	−.26; .08	−.02 (.08)	−.17; .14	.03 (.08)	−.13; .17	.04 (.08)	−.08; .21
NEO-PPI-Total	−.02 (.07)	−.15; .13	−.07 (.08)	−.20; .08	--	--	.10 (.08)	−.07; .23	.06 (.07)	−.08; .18	.12† (.07)	−.02; .24	.06 (.07)	−.08; .20

Note: PPI-SF=Psychopathic Personality-Short Form; NEO-PPI=NPO-PI-R Proxy measure of the Psychopathic Personality Inventory; IA=Impulsive Antisociality factor; FD=Fearless-dominance factor; PPI-SF-Cold=Coldheartedness subscale; NEO-Tender (r)=reverse scored Tendermindedness facet; Neg. Int.=Negative interaction; Pos. Feedback=interpersonal interaction where the participant received positive feedback; Agr. Interaction=Agreeable interaction; Int. Interaction=Intimacy interaction; Estimated effects are standardized Bayesian posterior estimates; †= p<.10; *=p<.05

The additional MSEM estimating a mediational within-person process involving hostile affect and perceived negative interpersonal interactions found support for negative interpersonal interactions mediating the relation between past and future hostile affect in both S1 and S2. Specifically, hostile affect at t-1 was related to future negative interactions (S1: β=.14; 95% CI=.12; .16, S2: β=.11; 95% CI=.09; .13), and negative interactions increased hostile affect (S1: β=.37; 95% CI=.32; .40, S2: β=.40; 95% CI=.36; .42). The mediational effects were significant in both samples (S1: B=.05; 95% CI=.03; .07, S2: B=.06; 95% CI=.04; .07), but direct paths between past hostile affect at t-1 and future hostile affect at time t remained significant (S1: β =.21; 95% CI=.20; .22, S2: β=.26; 95% CI=.23; .27). Despite replicable within-person mediation processes, tests for cross-level moderation of within-person effects showed mostly null results (Table 5). The one exception was the impact of PPI-Fearless-dominance on hostility inertia in S1, where PPI-Fearless-dominance had a dampening effect (β= −.20; 95% CI=−.35; −.04).

Table 5.

PPI Moderation of Within-person Mediation Model of Hostility

	Hostility Generation		Hostile Reactivity		Hostility Inertia
	Estimate (SD)	95% CI	Estimate (SD)	95% CI	Estimate (SD)	95% CI
Sample 1
PPI-SF-IA	.05 (.07)	−.07; 20	−.01 (.07)	−.15; .12	.07 (.07)	−.04; .21
PPI-SF-FD	−.08 (.07)	−.21; .05	.09 (.07)	−.05; .20	−.20* (.08)	−.35; −.04
PPI-SF-Cold	−.11 (.07)	−.24; .03	−.02 (.07)	−.16; .11	.07 (.07)	−.07; .20
PPI-SF-Total	−.02 (.07)	−.16; .10	.04 (.07)	−.09; .16	−.02 (.07)	−.15; .12
Sample 2
NEO-PPI-IA	−.04 (.09)	−.18; .16	−.05 (.08)	−.19; .11	.02 (.07)	−.12; .16
NEO-PPI-FD	−.02 (.07)	−.15; .12	−.02 (.08)	−.19; .13	−.05 (.07)	−.17; .07
NEO-Tender (r)	−.09 (.08)	−.20; .09	.05 (.07)	−.14; .17	−.08 (.07)	−.21; .03
NEO-PPI-Total	−.03 (.08)	−.18; .15	−.04 (.09)	−.23; .13	−.03 (.07)	−.17; .08

Note: PPI-SF=Psychopathic Personality-Short Form; NEO-PPI=NEO-PI-R Proxy measure of the Psychopathic Personality Inventory; IA=Impulsive Antisociality factor; FD= Fearless-dominance factor; PPI-SF-Cold=Coldheartedness subscale; NEO-Tender (r)=reverse scored Tendermindedness facet; Estimated effects are standardized Bayesian posterior estimates; Hostility Generation=within-person path from past hostility at time t-1 to negative interpersonal interactions; Hostile Reactivity=within-person path from negative interpersonal interaction to hostility at time t; Hostility Inertia=within-person path from past hostility at time t-1 to hostility at time t; *=p<.05.

Secondary Analyses

Based on feedback from reviewers, we conducted additional exploratory analyses focused on the Coldheartedness scale from the PPI-R in S1 and the reverse-scored Tendermindedness facet scale in S2 (hereafter referred to as (low) Tendermindedness), given the content overlap between the two scales. Results are presented separately for the Coldheartedness and (low) Tendermindedness scales in Tables 2-5. These analyses were not preregistered.

Average Momentary Affect and Interpersonal Outcomes

Coldheartedness in S1 was significantly negatively related to average momentary affect ( β range =−.18 to −.24) with the exception of hostile affect (β=−.09; 95% CI=−.25; .09). In S2, however, (low) Tendermindedness was unrelated to average momentary affect. Results for aggregated ratings of interpersonal interactions showed that Coldheartedness in S1 was related to interactions being perceived as less negative, on average (β= −.13; 95% CI=−.23; −.02) as well as interactions being perceived as having lower amounts of praise or positive feedback (β= −.27; 95% CI=−.41; −.13). In S2, (low) Tendermindedness was related to perceiving interactions as less intimate (β= −.23; 95% CI=−.38; −.08).

Emotional Diversity

The relations between Coldheartedness and (low) Tendermindedness and measures of emotional diversity (Gini coefficients and Simpson’s indices) were also examined. As before, scores were calculated for global, positive, and negative emotional diversity. In S1, results showed that while Coldheartedness was not significantly related to positive or global emotional diversity (r range= −.16 to −.02), it was significantly negatively related to negative emotional diversity as operationalized by the Gini coefficient (r= −.21; 95% CI= −.36; −.04) and Simpson’s index (r= −.27; 95% CI= −.42; −.10). In S2, no correlations between indices of emotional diversity and (low) Tendermindedness were significant (r range= −.13 to .03).

Dynamic Affective and Interpersonal Outcomes

Similar to results for Impulsive Antisociality and Fearless-dominance, neither Coldheartedness nor (low) Tendermindedness moderated any within-person processes involving interpersonal interactions and affective responses (i.e., no evidence to support cross-level moderation effects).

Discussion

We sought to investigate how psychopathic traits impacted various socio-affective processes in day-to-day life. Overall, we found limited support for our primary hypotheses. Results provided robust evidence for various within-person socio-affective processes, but little evidence of psychopathy enhancing or dampening these effects. Furthermore, replicable effects were confined to specific dimensions of psychopathy and in many cases psychopathy dimensions showed divergent results. While Coldheartness and (low) Tendermindedness showed a similar lack of moderation of within-person socio-affective processes, there was some evidence that these dimensions were linked with less social affiliation and dampened emotional experiences. Moreover, Coldheartedness was significantly negatively related to the diversity of negative emotions over the course of three weeks in S1. Other dimensions of psychopathy were unrelated to diversity in emotional experiences. Together, the findings highlight important avenues for future research on the socio-affective dynamics of psychopathy.

Making Sense of the Predominantly Null Findings for Psychopathy

While there was robust evidence for various within-person socio-affective processes, psychopathy had little impact on these processes. Impulsive Antisociality did show replicable relations with averaged hostile affect in S1 and S2, but this finding was the exception. In most cases, significant findings were only observed in one sample, if at all. For averaged affect and interpersonal outcomes, these findings highlight that psychopathic dimensions do not manifest moderate or large relations with relevant affective or interpersonal experiences in daily life. Regarding the lack of moderation of within-person effects, the findings highlight that specific socio-affective processes (e.g., increases in negative affect following a perceived negative interpersonal interaction) function similarly for individuals across the range of psychopathy scores in our two samples. Collectively, the findings suggest that accounts of broad-based affective-functioning deficits in psychopathy (c.f. Cleckley, 1941/1981) require further refinement.

However, null findings are informative when statistical power is high to detect the smallest effects of interest. Sensitivity analyses were conducted to examine the smallest effect size detectable in S1 and S2.⁶ Results are reported in supplementary Table S2. Tests focused on average reported affect, average reported interpersonal interactions, and emotional diversity outcomes had low power (.22-.25) to detect small effects (r=.10) but were adequately powered (≥.80) to detect effects of approximately r=.25 or larger. Power was very similar for tests of cross-level moderation. Thus, our primary tests were underpowered to detect small effects across our primary analyses, while they were well-powered to detect moderate to large effects. Thus, the results suggest that psychopathic traits are unlikely to demonstrate moderate to large relations with socio-affective functioning in day-to-day life (e.g., average number of interpersonal conflicts, reactivity to conflict). Nonetheless, the results highlight potential avenues for future EMA research on psychopathy.

Future work can enhance the assessment specificity of interpersonal situations. Although we examined broad-based indices of negative interpersonal interactions, psychopathic traits may exert their largest influence in specific interpersonal situations (e.g., significant dampening of remorse- or guilt-related affect after harming another person). These effects may be difficult to detect when aggregating across various negative interpersonal interactions. These points are underscored by past research using S1 data while focused on borderline personality disorder and antisocial personality disorder traits. Scott and colleagues (2017) relied on theoretical conceptualizations of BPD to identify relevant within-person processes (i.e., a mediational pathway between perceived rejection, negative affect, and aggressive urges and behavior) and found evidence that BPD traits enhanced the link between perceived rejection and broad-based negative affect, while this cross-level enhancement was not observed for antisocial personality disorder traits, nor when perceived criticism was substituted for perceived rejection. Similarly, Wright and colleagues (2017) found that narcissism amplified the link between perceptions of other dominance and negative affect. Thus, the lack of empirical support for psychopathy traits moderating the various within-person processes can be contrasted with past efforts that were successful when focusing on BPD and narcissism, though the latter efforts were closely tied to specific, theoretically relevant interpersonal contexts. Though future EMA research on psychopathy dynamics may benefit from such enhanced designs, the present results also highlight that common operationalizations of psychopathy pose more general statistical and methodological hurdles.

Divergent Effects for Psychopathy Components

Across socio-affective outcomes, Impulsive Antisociality and Fearless-dominance tended to show opposing effects, consistent with past research. For example, while Impulsive Antisociality was significantly positively related to hostile affect in S1 and S2, Fearless-dominance was negatively related to hostility in both samples. Though the PPI scales have not been previously used in EMA-based studies, the divergent findings are in line past research showing that Impulsive Antisociality is related to a wide range of negative outcomes while Fearless-dominance has been shown to largely overlap with traits related to extraversion and low neuroticism, and in turn, is mostly unrelated to maladaptive outcomes (Miller & Lynam, 2012). Researchers have argued that these opposing findings are to be expected since Fearless-dominance and Boldness are meant to index the “mask” first highlighted by Cleckley (1941/1981) which disguises or hides the maladaptive functioning related to psychopathy (e.g., Lilienfeld et al., 2012).

Yet, the opposing relations between psychopathy dimensions raise important practical issues regarding the interpretation of psychopathy total scores for specific measures like the PPI-R. Researchers have suggested alternative approaches to understand how psychopathic traits come together to produce expected relations with relevant outcomes, whether it be curvilinear relations (e.g., Blonigen, 2013) or trait-by-trait interactions (e.g., Lilienfeld et al., 2019). However, evidence for these types of relations is limited (Benning & Smith, 2019; Crowe et al., 2021).⁷ Other approaches include subgrouping methods to identify particular “types” of psychopaths and examine how the types relate to external correlates (e.g., Hicks et al., 2004). Together, these considerations suggest that future EMA-based studies on psychopathy and socio-affective functioning may benefit from alternative measurement approaches. Relatedly, future work could use more homogenous and narrow-band trait scales. Recent EMA research focused on antagonism supports this approach, as variability was observed across different facets of antagonism (e.g., callousness vs. manipulativeness) for affective and interpersonal outcomes (Vize et al., 2022b).

Psychopathy and the Experience of Emotion in Daily Life

Dominant conceptualizations of psychopathy suggest that it should be related to a lack of diversity in emotional experiences. There was some evidence to support this contention, as Coldheartedness in S1 was negatively related to the diversity of negative emotions (but not positive emotions) over the course of the three-week EMA protocol. More generally, however, results ran counter to our hypotheses and highlight significant emotional heterogeneity worthy of further study. For example, for participants high in psychopathy, positive affect was commonly reported and at various levels of intensity. In some cases, individuals high in psychopathy reported very little negative affect while reporting high levels of positive affect (Panel 2A, Figure 1). In others, the reported emotional experiences were in line with typical descriptions of psychopathy (Panel 1A, Figure 1). Importantly, the heterogeneity was not explained by factor profiles (i.e., higher standing on Fearless-dominance relative to Impulsive Antisociality, and vice versa). It will be beneficial for future research to explore this variability underlying the emotional experiences of psychopathy. To date, psychopathy research has largely been focused on between-person data (i.e., nomothetic research) while investigations of the individual variability within psychopathy is less common (i.e., idiographic research). Insights from the large body of nomothetic research on socio-affective processes of psychopathy can be further enhanced by leveraging the respective strengths of nomothetic and idiographic traditions in future research (Wright & Woods, 2020).

Limitations

Our results should be interpreted with important limitations in mind. First, neither sample was recruited based on elevations for an array of psychopathic traits, but rather specific features with relevance to psychopathy (aggressiveness and hostility). Thus, there were few individuals in either sample likely to represent the more severe end of the psychopathic trait continuum. Moreover, S1 was composed of young, predominantly African American women of low socioeconomic status while S2 was composed of middle-aged adults who were predominantly white and of medium to high socioeconomic status. While these features of the samples allowed us to examine the generalizability of the findings across various diversity dimensions, it is also important to carefully consider how sociodemographic differences may drive the lack of replication for some observed findings. That is, while a lack of replicable significant findings may indicate that the finding is not robust, it is also possible that sample-specific effects are worthy of future investigation. For example, in S1, there was evidence that Fearless-dominance showed stronger buffering effects on different kinds of negative affect dynamics (e.g., women higher in Fearless-dominance reported significantly less sadness following a negative interpersonal interaction [Table 4]; the carry-over effect of hostility from one assessment to the next was significantly dampened for women higher in Fearless-dominance [Table5]). While it is important to replicate this pattern of findings using similar samples, the lack of replication across S1 and S2 observed here does not preclude that sample-specific findings may inform future research on socio-affective dynamics related to psychopathy.

Relatedly, it will be important to extend our approach to other samples relevant to psychopathy research where the full constellations of psychopathic traits may be better represented (e.g., incarcerated males). To do so, there are barriers to implementing EMA-based designs in correctional settings that are mostly absent in other contexts (e.g., how to facilitate access to smartphone devices). Some initial work in juvenile forensic settings does suggest that EMA designs can be successfully implemented in forensic settings. Pihet and colleagues (2017) demonstrated the feasibility of using EMA in a sample of juvenile youth—out of the initial 149 youth recruited, roughly 75% of the sample completed the 8-day protocol involving four assessments per day. Descriptive results showed high compliance rates (84%) as well as moderate to high between- and within-person reliability for various EMA measures (e.g., affect, aggressive behavior). Future research can build on these promising results to explore effective ways to implement EMA-based designs in correctional and forensic settings which will be an important future direction for EMA studies of psychopathy.

Second, different operationalizations of psychopathy are frequently used in research. The PPI-based operationalization of psychopathy emphasizes traits related to interpersonal boldness, whereas measures like the PCL-R assess this content in a more limited fashion. Because Fearless-dominance manifested divergent relations compared to Impulsive Antisociality, it is almost certain that different results will be obtained if psychopathy measures that deemphasize boldness-related content are used. Thus, future work will benefit from using alternative operationalizations of psychopathy in EMA studies. Ideally, multiple measures can be used and compared against one another to better understand the influence of psychopathy measurement on findings regarding socio-affective processes. Relatedly, we used multiple samples to examine the replicability of findings but we were unable to use the exact same measure in both samples. While psychopathy in S2 was assessed with a previously validated proxy measure of the PPI-R (Witt et al., 2009), our results suggest differences in how psychopathic traits within these measures related to one another in S1 and S2. While Fearless-dominance and Impulsive Antisociality were positively correlated in S1, they were negatively correlated in S2. In both cases, previously published studies (e.g., Marcus et al., 2013; Witt et al., 2010) have found similar correlations using these specific measures, but these relations differ from the negligible meta-analytic correlation between the PPI-R factors (Miller & Lynam, 2012). Though these differences did not appear to influence our results (i.e., the sign and magnitude of the Fearless-dominance and Impulsive Antisociality effects were similar across samples), future work using multiple samples will benefit from the use of identical instruments to help ensure any differences across samples are not due to measurement idiosyncrasies.

Last, the EMA protocols in S1 and S2 were not explicitly designed to study socio-affective processes tied to psychopathy. Future work would benefit from EMA protocols purposefully designed to assess processes thought to be most relevant to psychopathy. For example, as described above, interpersonal situations where callousness may be most relevant can be explicitly assessed with EMA items written to capture those contextual features. Relatedly, interpersonal situations where manipulation or dishonest behavior occurs would also be relevant to the socio-affective features of psychopathy. Regarding assessments of interpersonal conflict more broadly, subsequent research will benefit from additional assessment approaches that are not solely reliant on participants’ perceptions of conflict and behavior of others (e.g., collecting partner reports, coding passively collected audio recordings; Sun & Vazire, 2019).

Future Directions for EMA-based Psychopathy Research

In the sections above, we have highlighted how future research can expand on the present findings and improve on design limitations. More broadly, EMA-based research on psychopathy is uncommon and it is important to give additional attention to EMA design considerations to help ensure that EMA research can address relevant theoretical questions related to psychopathy in day-to-day life. When considering EMA research for personality pathology more broadly, Kaurin and colleagues (2023) highlight the “three D’s” heuristic for EMA study planning: Density, Duration, and Depth. Density refers to how frequently individuals will be assessed, such as daily diaries or several samples of emotions and behavior throughout the day. Depth refers to the number of questions or items administered at each assessment. Last, duration refers to how long a protocol lasts. Kaurin and colleagues (2023) found in reviewing the literature on ambulatory assessment of personality disorder that density was negatively correlated with depth and duration, whereas duration and depth were positively correlated. Thus, when researchers increase how frequently they sample individuals each day, they typically administer briefer surveys for a shorter period of time. Long protocols typically employ only once-daily assessments, which span longer periods of time.

Consideration of Density, Duration, and Depth will be explicitly guided by the processes and behaviors researchers seek to understand. For example, a study testing whether psychopathy is related to more frequent or intense interpersonal conflicts involving aggression will likely require a much longer EMA protocol (i.e., greater Duration)—aggression is a low-base rate behavior so EMA designs lasting only a few days or a week are unlikely to capture instances of aggression. Instead, a daily diary design that asks various questions about an episode of aggression when such an episode is reported is better suited for this research question. With the decreased frequency of assessments, the protocol may be extended to a month or longer without overdue burden on participants.

Alternatively, if researchers wanted to test a hypothesis that psychopathic traits will lead to a blunted emotional response in the aftermath of a conflict, a measurement-burst design may be used (Sliwinski, 2008). In this design, participants would be instructed to complete a brief survey following any negative interpersonal interaction (termed an event-contingent survey, since the completion of the survey is dependent on an event occurring and is typically initiated by participants themselves). Once participants report a negative interaction has occurred, a burst of short, but frequent surveys are sent over a specific time frame (e.g., assessing a limited set of emotions like remorse, guilt, or anger). When considering emotional responses to a negative interaction, delivering surveys every 15 minutes until one hour has passed since the event may be sufficient to capture the emotional trajectory of interest and subsequently examine if psychopathic traits moderate the trajectory.

Last, EMA research can serve as a valuable way to validate laboratory tasks involving socio-affective functioning, since laboratory tasks have been the primary method used to probe socio-affective deficits in psychopathy. A recent example of such an approach can be found in Edershile et al. (2023). The authors examined how narcissistic reactivity in the laboratory mapped onto narcissistic reactivity in interpersonal situations during an EMA protocol, finding that the same participants who responded with decreased grandiosity in response to experimental defeat during a competitive laboratory task responded with similar decreases in grandiosity to status threat in daily life. To date, we know of no study that has sought to provide ecological validity of lab-based tasks related to psychopathy in this way, highlighting a valuable direction for future work. For example, researchers can examine whether individuals who demonstrate decreased reactivity to distress cues in the laboratory similarly show decreased affective responses to others’ reports of distress in daily life.

Conclusion

The present study offers important data regarding the role psychopathy plays in various socio-affective processes that unfold in daily life. Though we observed primarily null findings, the present project highlights promising directions for future EMA-based psychopathy research, including designing EMA studies focused on the interpersonal situations most relevant to psychopathy. Thus far, psychopathy research has taken a nomothetic approach to understanding the disorder. The present results point towards integrating the strengths of nomothetic and idiographic approaches in order to better understand dynamic processes related to psychopathy. Last, echoing calls from other researchers (Verschuere et al., 2021), preregistration and other open science practices have much to offer psychopathy research. Many of the factors that impact the robustness of psychological findings are present in psychopathy research (e.g., small sample sizes, flexible measurement and analytical frameworks, novel and eye-catching findings). Open science methods are valuable tools for researchers working to further understand the socio-affective dynamics of psychopathy.