Mapping the Dark Factor (D) and Agreeableness-Antagonism (A-A) on the Interpersonal Circumplex

A total of 600 participants were recruited for the study from the participant pool at a large midwestern university, and IRB approval was obtained prior to the data collection. Informed consent was obtained from all participants before they provided any data. All participants completed the study on Qualtrics. Responses were excluded if they: 1) failed to provide valid answers to EPA Infrequency and/or Virtue scales, 2) incorrectly answered more than two out of eight attention-check items, 3) answered the questions in a single-response style (e.g., all 1s), 4) completed the survey in less than 500 seconds. The final sample included 516 individuals: 62.4% Women, 35.9% Men, 0.6% Non-Binary or Other; 88.4% Straight or Heterosexual, 1.2% Gay or Lesbian, 9.4% Bisexual, 1% Other; 5.9% Hispanic or Latino, 72% White, 19.8% Asian, 4.7% more than one racial identity, and 3.3% Black or African American.

This data set has been used previously in Rose, Miller, & Lynam (2023) which examined the validity of the newly developed Five Factor Model Antagonistic Trait Measure (FFM ATM). The study reported data from the current A-A and IPC measures but only in the form of criterion variables in relation to the FFM ATM domains. The data set was also used in Rose et al. (2022) which compared profiles of D and A-A but only included the two separate scores of interpersonal warmth and dominance. Thus, these data have never been used to compare D and A-A to one another on the interstitial space of IPC. All the analyses were preregistered before being conducted (https://osf.io/vsxdw).

Agreeableness-Antagonism. Trait antagonism was assessed by reversing the agreeableness items from the International Personality Item Pool Representation of the NEO PI-R short form – Agreeableness subscale (IPIP 120; Maples et al., 2014), which is a 120-item measure of the 30 facets of the five domains of the FFM. Participants rated how accurately statements such as “[I] am not interested in other people’s problems” describe them on a Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). The Agreeableness/Antagonism subscale includes six facets measured by 24 items (four items per facet). Cronbach α’s for the facets ranged from 0.65 to 0.86 with a median of 0.68.

Dark factor of personality. The dark factor of personality was measured by The Dark Factor of Personality Scale (D; Hilbig et al., 2020), which is a 22-item self-report measure of the core features of social aversive psychopathology (e.g., psychopathy, narcissism, Machiavellianism). The Cronbach α for the scale is 0.88.

Interpersonal traits. Interpersonal traits were measured by the Revised Interpersonal Adjective Scales (IAS-R; Wiggins et al., 1988), which includes 64 adjectives that describe people’s interpersonal characteristics (e.g., “Tender”, “Ruthless”), and each octant of the IPC is measured by eight adjectives from the measure. Participants rated how accurately these adjectives describe them on an 8-point scale ranging from 1 (extremely inaccurate) to 8 (extremely accurate). Cronbach’s α’s of the octants ranged from 0.73 (warm-submissive) to 0.91 (warm-dominant), with a median of 0.86.

Validity checks. Validity items from the Elemental Psychopathy Assessment (EPA; Lynam et al., 2011) was used in this study. The EPA is 178-item self-report measure of psychopathy, and here only the 16 items that comprise two validity scales – Infrequency and Unlikely Virtue – were used.

The structural summary method (SSM; Gurtman, 1994; Zimmermann & Wright, 2017) was used to examine the interpersonal profiles of Antagonism, as well as its six facets, and D. The SSM uses four parameters to summarize correlations between external constructs (i.e., D and A-A) and the eight octants of the IPC, each representing a unique combination of warmth and dominance (see Figure 1). Elevation describes the overall endorsement of items (e.g., do individuals with higher antagonism give higher ratings of their interpersonal traits across the board?). Amplitude describes the extent to which the constructs present a specific interpersonal theme (e.g., how much does antagonism relate to specific types of interpersonal traits?). R², also known as the goodness-of-fit statistic, describes the extent to which the profile fits into a circular structure, and thus illustrates interpersonal prototypicality (e.g., a prototypically cold-dominant profile would be associated most strongly with the cold-dominant octant, then with cold and dominant octants, and increasingly less with octants further away from cold-dominance, and thus least strongly associated with the warm-submissive octant). Angular Displacement describes the specific theme of the interpersonal profile in terms of where it can be placed on the IPC (e.g., what types of interpersonal traits does antagonism relate to the most strongly?). 95% confidence intervals of Elevation, Amplitude, and Angular Displacement will also be reported in the results.

For interpretations of SSM results, based on recommendations from existing publications (Gurtman & Pincus, 2003; Hopwood et al., 2013; Wright et al., 2012), Elevation and

Amplitude values greater than .15 are considered notably elevated and differentiated, and R² values greater than .80 are considered to fit an expected circular structure moderately well. Amplitude is only interpreted with adequate R², and Angular displacement is interpreted when Amplitude and R² meet their corresponding thresholds. The SSM parameters and their associated confidence intervals were analyzed using the ‘circumplex’ package in R (v0.3.8; Girard et al., 2021). For the D and A-A profiles to be considered different, the profiles need to have interpretable angular displacements with non-overlapping 95% confidence intervals, which suggests a meaningful difference. Data and codes used in this study are publicly available at https://osf.io/up3z9/.

Table 1 presents the correlations between D, A-A total score, and A-A facets, and Table 2 presents the structural summary parameters of D, A-A total score, and A-A facets, as well as the 95% confidence intervals of Elevation, Amplitude, and Angular Displacement. The Amplitude, and Angular Displacement confidence intervals are also plotted on the interpersonal circumplex for visualization of the results (see Figures 2 and 3). Because the Antagonism facets were calculated by reversing Agreeableness facets, antonyms of the original Agreeable facets were used in the tables and figure to describe these facets more accurately (e.g., reverse of the “Trust” facet was renamed as “Distrust”).

The results showed that the interpersonal profiles of D and A-A factors were both highly specific (Amplitude = .58 and .66, with a threshold of .15) and interpersonally prototypical (R² = 1.00 and .99, with a threshold of .80), and the Angular Displacements of the two profiles are nearly identical (160.6° and 162.1°), both demonstrating a robust, well-differentiated cold-dominant theme. In other words, from an interpersonal perspective, there is no distinction between D and A-A.

At the facet level, all the facets showed highly differentiated, prototypical, and thus interpretable profiles, though the specific interpersonal themes varied across the facets: Immorality, Uncooperativeness, and Callousness facets showed similar cold-dominant profiles as D and Antagonism, with a higher emphasis on coldness. Grandiosity showed a slightly more dominant profile, while Distrust and Selfishness displayed almost purely cold profiles. It is particularly worth highlighting that the profiles of D and A-A are closer to each other on the figure than many of the A-A facets are to A-A total score or to other A-A facets.

In an analysis that was not preregistered, we applied the structural summary method to the individual items from D (20 items) and A-A (24 items); all items were coded to run in the same direction. Figure 4 depicts the frequency of the angular displacements of the D and A-A items on the interpersonal circumplex. With the exception of a single D item at 81.7°, “having a lot of money is not one of my goals in life” (reversed), all D items fell within the range of A-A items. The angular displacements of the D items ranged from 81.7°.to 181.2° with a median of 159.4°; A-A items ranged from 102.7° to 201.3° with a median of 161.25°. Even at this most granular level, D and A-A are distributed virtually identically across interpersonal space.

The present study tested whether the dark factor and Agreeableness-Antagonism, at both the domain, facet, and item levels, could be distinguished from one another using the interpersonal circumplex. The IPC provides a well-established theoretical and empirical model for mapping the interpersonal characteristics of other constructs along two major axes—Agency and Communion, which are the fundamental dimensions of interpersonal behaviors. The IPC has been used previously to distinguish the domains and facets of personality disorder traits (Wright et al., 2012), the three components of the Dark Triad (i.e., psychopathy, narcissism, and Machiavellianism; Dowgillo & Pincus, 2017), and grandiose and vulnerable narcissism (Miller et al., 2012). Importantly, the Dark Triad covers constructs that some consider “interpersonal disorders (Wright et al., 2022; cf., Widiger et al., 2023) and the proponents/creators of the D factor invoke definitions that are almost entirely interpersonal in nature (e.g., Moshagen et al., 2018). As such, the IPC seems like an ideal paradigm through which to interrogate possible differences in these constructs.

Results from the current study indicate that the dark factor and Agreeableness-Antagonism are functionally identical in interpersonal terms. Both fall solidly within the cold-dominant octant. This octant typically describes individuals who are competitive, arrogant, and calculating (Leary, 1957), which fits well descriptions of both the dark factor and Antagonism. The facets of Agreeableness-Antagonism showed some differentiation among themselves, ranging from across the cold and cold-dominant octants, with the dark factor falling solidly in the middle.

Clarifying the degree of overlap/distinction between similar seeming constructs like D and A-A is a valuable endeavor, as scarce resources such as money, time, and attention can be invested in studying these so-called novel constructs when relatively little new is discovered (e.g., see grit vs. conscientiousness literatures as an example). When highly similar constructs can be identified early in the process, scientists can push the field forward more effectively by focusing on more constructive research. The current study highlights concerns that a new, siloed literature on the dark factor is unnecessary given that these findings can be comfortably organized with the general and pathological personality literatures on antagonism. That is, the existing D-focused literature can be integrated relatively seamlessly into the much larger and more well-established literature on A-A (e.g., Lynam & Miller, 2019).

This study has several limitations worth noting. It used a convenience sample with limited demographic diversity (e.g., age; socioeconomic status, education), which may affect the generalizability of our findings. In addition, we relied exclusively on self-report measures of D, A-A, and the IPC, which may introduce some bias and overestimation of relations given shared method variance. Additionally, although the interpersonal circumplex is a compelling model, it is possible that differences between the dark factor and A-A lie outside of interpersonal space and thus was not fully captured in the study. Future researchers are encouraged to compare the dark factor and A-A further by examining how they associate with other aspects of human behaviors and functioning. Nonetheless, the current findings add to an existing literature that calls into question the viability of the D factor as a standalone, new construct deserving of intensive investigation.

Contributed to conception and design: TVD, DRL

Contributed to acquisition of data: LR, DRL

Contributed to analysis and interpretation of data: TVD

Drafted and/or revised the article: TVD, JDM, DRL

Approved the submitted version for publication: TVD, LR, JDM, DRL

The authors received no financial support for the research, authorship, and/or publication of this article.

The authors declared no potential conflicts of interest respect to the research, authorship, and/or publication of this article.

All data files and analysis scripts used for this study can be found on the OSF project page: https://osf.io/up3z9/