“They Are Such an Asshole”: Describing the Targets of a Common Insult Among English-Speakers in the United States

Some investigators have examined insults in a psychosocial context. Cross-cultural patterns in the content of insults suggest that they communicate information regarding important values. For example, in Spain—a collectivist society—insults typically reference family and relationships (De Raad et al., 2005). In a multimethod exploration of insults commonly used by English-speaking adults from the United States (Hyatt, Maples-Keller, et al., 2019) ² participants were randomly assigned to rate an individual in their life (i.e., target) matching the description of a “dick”, “bitch” or “asshole”—common insults for men, women, and both men and women, respectively—in terms of the FFM. The greatest degree of consensus was observed for facets of Agreeableness (especially Tender-Mindedness and Compliance), such that participants tended to report that targets of these insults were characterized by (low levels of) these traits. The consensus insult profiles were similar to one another with strong interclass correlations across insult and target gender (r_ICC range = .88 to .96), showing that, on average, these insults communicate a similar perception or judgment: “They [i.e., the target] are so antagonistic [i.e., disagreeable].”

Hyatt and colleagues (2019) extended the nomological network of insults beyond basic personality by examining associations of insult endorsement with measures of maladaptive traits, psychosocial functioning. externalizing behaviors, Diagnostic and Statistical Manual of Mental Disorders (5^th ed.; DSM–5; American Psychiatric Association, 2013) Section II personality disorders (PDs), attachment style, self-esteem, informant reports, thin-slice ratings from strangers, social discounting, and social cognition. Self-endorsement of insults was positively associated with DSM-5 Section II PDs (antisocial, borderline, and narcissistic), DSM-5 Section III pathological personality traits (antagonism and disinhibition), so-called “dark” traits (e.g., Machiavellianism, psychopathy; Paulhus & Williams, 2002), externalizing behaviors (proactive and reactive aggression, antisocial behavior, bullying), social discounting, and rudeness and yelling in response to hypothetical provocative social situations. In contrast, insults bore small negative correlations with self-esteem and perceived adaptive functioning across various psychosocial domains.

The present study focuses on the insult “asshole” in a sample of English-speaking adults in the United States recruited from Amazon’s Mechanical Turk (MTurk) platform. We collected measures of self-reported personality as well as perceptions of the personality, beliefs, and behaviors of the “biggest asshole” (i.e., target) in the participant’s life to answer two primary questions, respectively replicating and extending Hyatt, Maples-Keller, et al. (2019): 1) what information about a target’s perceived personality is communicated when people employ insults and 2) what behaviors do people associate with this common insult? Hypotheses and planned analyses were pre-registered on the Open Science Framework (https://osf.io/9pafc). Study materials, analytic code, and de-identified data sufficient for reproduction of our quantitative analyses are also available on the OSF.³

The mean FFM profile for nominated “assholes” served as the prototype for the perceived personality of targets of the insult “asshole.” This prototype was compared to mean self-ratings of personality, other insult prototypes (Hyatt, Maples-Keller, et al., 2019), and expert-rated prototypes of relevant PDs, namely psychopathic (Miller et al., 2001), narcissistic, and antisocial PDs (Lynam & Widiger, 2001). Previous research suggests that insults describe both basic personality traits as well as configurations of traits consistent with narcissistic, antisocial, and psychopathic PDs (e.g., average r_ICC between lay-rated insult and expert-rated PD prototypes = .52 to .72; Hyatt, Maples-Keller, et al., 2019). Following Hyatt and colleagues (2019), we hypothesized that the FFM profile of the prototypical perception of an “asshole” would be characterized by low Agreeableness and high Anger (a facet of Neuroticism) and would be similar to expert rated FFM profiles of PD prototypes for psychopathy, antisocial, and narcissistic PDs. Additional pre-registered analyses comparing participants’ self-reported personality with perceptions of targets’ personality were conducted as well (see Appendix A of Supplemental Materials).

We extended previous exploration of insults by building a taxonomy of interpersonally aversive, “asshole” behaviors. We examined participants’ descriptions of targets’ most bothersome “asshole-related” behaviors, both content-coding for common themes and classifying individual behaviors in terms of FFM trait domains. Analyses of nominated bothersome behaviors were not driven by a priori hypotheses, but we did expect that some number, perhaps the majority, of these reported behaviors would be classified as indicative of antagonism, whereas others might be better described as indicators of high or low scores on other basic domains (e.g., carelessness – low Conscientiousness; anger – high Neuroticism). Additionally, we examined basic descriptive statistics regarding participants’ perceptions of nominated assholes’ appreciation of and concern for the consequences of their behaviors, as well as participants’ perceptions of targets’ ability to change these behaviors.

Potential participants (n = 1,106) were recruited via MTurk to participate in a brief demographic screening questionnaire to determine their eligibility for a psychology study. Enrollment was limited to individuals 18 or older and living in the United States. Participants who were deemed eligible (n = 589; i.e., completed captcha, provided a coherent response to a brief writing prompt) were invited to participate in a study of people’s perceptions of their own and others’ personalities. 465 individuals (78.95%) enrolled in the study. We removed 65 responses (13.98% of sample) for one or more of the following: invalid response style on the embedded Elemental Psychopathy Assessment (EPA; Lynam et al., 2011) Infrequency or Virtue scales (i.e., EPA Infrequency ≥ 4 or EPA Virtue ≥ 3; n = 48); total lack of response (n = 11, 2.4%); failure of two or more (out of three) attention checks (n = 22), and nonsensical answers to free-response items (n = 1). Three additional participants were removed for providing ratings of invalid targets (i.e., the participant themselves, Donald Trump, and “someone famous”).⁴

The remaining 397 participants had a mean age of 36.92 (SD = 11.48)⁵. Participants reported gender as female (n = 182, 45.84%), male (n = 212, 53.40%), or non-binary/third gender (n = 3, 0.76%). Four (1.01%) participants identified as transgender. Participants reported race as one or more of the following: White (n = 339, 85.39%); Asian (n = 37, 9.32%); Black or African American (n = 24, 6.05%); American Indian or Alaskan Native (n = 12, 3.02%); Other (n = 5, 1.26%; e.g., “Dominican,” “Latino”). One participant did not report race. Participants also reported on their ethnicity (i.e., “Do you consider yourself to be Hispanic or Latino?”; Yes, n = 30, 7.56%) and sexual orientation (heterosexual: n = 331, 83.38%; gay: n = 8, 2.02%; lesbian: n = 7, 1.76%; bisexual: n = 42, 10.58%; questioning or unsure: n = 4, 1.01%; other [i.e., “asexual,” “pansexual,” “queer”]: n = 5, 1.26%).

All study procedures were approved by the relevant institutional review boards and all participants provided informed consent and had the chance to email the researchers to ask any clarifying questions. Participants used Qualtrics to complete questionnaires at the location of their choosing. These surveys addressed demographic information, personality, personal beliefs, and behaviors. Additionally, participants were asked to describe the “biggest asshole” (target) that they personally know and whose negative behaviors they have directly experienced and rate their perception of that person’s personality, beliefs, and behaviors. Participants received $3.00 or $4.00⁶ for completion of the full questionnaire battery. Order of measures was counterbalanced between participant self-report and ratings of the target “asshole.”

Participants completed the IPIP-NEO-120 (Maples et al., 2014), an open-source alternative to the Revised NEO Personality Inventory (NEO PI-R; Costa & McCrae, 1992). The IPIP-NEO-120 uses 120 items from the International Personality Item Pool (Goldberg, 1999) to provide estimates of FFM domains and facets (six per domain). Items (e.g., “Get angry easily”) were rated on a five-point Likert-style scale from 1 (Disagree Strongly) to 5 (Agree Strongly). Reliability for the IPIP-NEO-120 was acceptable (domain-level ⍺ range = .90 to .96⁷; facet-level ⍺ range = .72 to .92). Participants also provided basic demographic information.

At the beginning of the target rating block, we asked participants to “Think of the ‘biggest asshole’ you know personally where you have directly experienced the person’s negative behaviors” and to report how difficult it was to identify this individual (1 = Not at All Difficult to 5 = Very Difficult). Participants then described their perception of the target in terms of gender, age, relationship to the participant, closeness to the participant (1 = Not at All Close to 5 = Very Close), and the extent to which the insult “asshole” characterized the target (“asshole”-ness; 1 = Not at All to 5 = Very Much).⁸ Next, participants described three of the target’s bothersome, “asshole-related” behaviors in an open-ended format. For each nominated behavior, participants additionally provided ratings of the degree to which the participants perceived that the target knows that the behavior is bothersome (“Do you think this person knows this behavior is bothersome for others?”; i.e., knowledge), cares that the behavior is bothersome (”Do you think this person cares this behavior is bothersome for others?“; i.e., caring), and would be able to change the behavior if they so desired (”If this person was sufficiently motivated (i.e., knew and cared sufficiently), do you think this person could change this behavior?“; i.e., ability to change) from 1 (Not at All) to 5 (Very Much). Finally, participants rated their perception of the target’s personality on the FFM using the IPIP-NEO-120. Items were modified to refer to the target (e.g., ”This person gets angry easily”). Reliability for IPIP-NEO-120 informant reports was mostly adequate (domains = .88 to .95; facets = .58 to .88).

First, we created an FFM profile for the prototypical perception of targets of the insult “asshole” using sample means for perceptions of targets’ IPIP-NEO-120 facet scores. Next, we compared the “asshole” profile generated in this study with the FFM prototypes for insults from Hyatt and colleagues (2019; i.e., “asshole,” “dick,” and “bitch”) and with expert-rated prototypes of psychopathic, narcissistic, and antisocial PDs (Lynam & Widiger, 2001; Miller et al., 2001). We also compared this profile with the average profile obtained from participants’ self-reported personality on the IPIP-NEO-120. To measure absolute profile similarity, we employed the double-entry intraclass correlation (r_ICC) which is sensitive to both differences in score elevation (i.e., score means) and profile shape (i.e., pattern of scores; McCrae, 2008). A zero-order correlation (r) captures only similarity in profile shape (i.e., pattern of scores). In contrast, r_ICC captures similarity in both shape and elevation. Double-entry correlations are calculated by entering two rows of data for each pair of values (e.g., FFM ratings from “asshole” profile and expert-rated prototype) such that for a pair of values (X, Y), one row includes X in the first column and Y in the second, while the order is reversed in the second row (i.e., Y in the first column and X in the second). As an example, if column A had values of 1, 2, 3, 4, 5, 6, 7 and column B had values of 2, 4, 6, 8, 10, 12, and 14 – these two columns would have a perfect degree of relative similarity (r = 1.00) but only modest to moderate absolute similarity (r_ICC = .29) as the size of the elevations are reasonably different from one another.

Next, we conducted an exploratory qualitative analysis of interpersonally aversive behaviors using two methods: 1) thematic coding and 2) rating each behavior in terms of the FFM. For the first method, the first author initially reviewed all free-response descriptions of “asshole-related” behaviors (n = 1,203⁹) to generate a list of unique categories among the nominated bothersome behaviors. This list was then reviewed by two undergraduate student raters from the laboratory of the fourth author. Finally, the third and fourth authors reviewed these categories to generate overarching themes. An iterative approach was used until the authors agreed on the sufficiency of the final categorization. For the second method, the first author reviewed the responses once again, splitting multi-barreled descriptions into discrete behavioral units (n = 1,453; e.g., “Aggressive and intimidating” became “Aggressive” and “intimidating”). Undergraduate and post-baccalaureate student raters (k = 12) then determined which single FFM domain and pole (high versus low) each behavior best exemplified. Raters were provided with descriptions of each domain (used previously in Sleep et al., 2021; see Appendix B in Supplemental Materials on OSF). An additional option (“Cannot Rate”) was made available for behavioral examples not adequately covered by the FFM (e.g., “breathes like an asshole”). We used R (Version 4.1.1; R Core Team, 2021)¹⁰ and RStudio (Version 1.4.1717; RStudio Team, 2021) for initial data manipulation and recoding as well as all substantive analyses.

Our effective sample size of 397 exceeded the minimum recommendations for stable estimates of correlations (n = 250; Schönbrodt & Perugini, 2013) and, according to G*Power (Faul et al., 2007), provided 91.93% power to detect correlations as small as .20 at α = .01. For our comparisons between self and target ratings on the IPIP-NEO-120, we conducted a power analyses for our matched samples t-test using an alpha of .01 (two-tailed ) and an effect size of .207. The effect size was arrived at using the average of the SDs across facets as the group SDs (i.e., .90), a small correlation between groups (i.e., .01), and a difference of .25 on the five-point IPIP scale (i.e., 2.75 vs 3.0). Under these parameters, with 397 participants, we have 93.7% power to detect an effect size of 2.07 or larger.

Participants found it relatively easy to think of the “biggest asshole” in their lives (M = 1.78, SD = 1.15; 1 = Not at All Difficult to 5 = Very Difficult) and generally considered this individual (target) to be well-described by the insult (M = 4.33, SD = 0.76; 1 = Not at All to 5 = Very Much). Targets were typically male (79.35%) with an average age of 42.71 years (SD = 13.72).¹¹ Approximately one-third of targets (35.26%) were identified as romantic partners, co-workers, bosses, family members, or friends of participants while half (50.13%) formerly held such a role (e.g., ex-partners, estranged family members). On average, participants did not consider themselves to be close to targets (M = 1.65, SD = 1.06; 1 = Not at All Close to 5 = Very Close). Full information on target characteristics is provided in Table 1.

On the IPIP-NEO-120 (Table 2), participants described targets as angry (facet M = 4.34, SD = 0.83; 1 - Disagree Strongly to 5 - Agree Strongly) and broadly disagreeable (i.e., low Agreeableness; domain M = 1.95, SD = 0.55).

Table 3 presents double-entry correlations for comparisons among the FFM profiles for both targets and participants as well as prototypes of insults (Hyatt, Maples-Keller, et al., 2019), and psychopathic, antisocial, and narcissistic PDs (Lynam & Widiger, 2001; Miller et al., 2001). The FFM profile for the perception of the prototypical “biggest asshole” in the present study was strongly¹² similar to all external profiles with effects ranging from .50 (psychopathic PD; Miller et al., 2001) to .83 (male “asshole”; Hyatt, Maples-Keller, et al., 2019). In contrast, participants’ self-reported personality was moderately to strongly dissimilar to these profiles (r_ICC range = -.69 to -.36).

Of the 1,203 responses collected, 1,196 were unique (99.42%). Initial qualitative analysis of these unique nominated “asshole” behaviors generated 315 categories (e.g., poor work behavior, rude, lying, insults others, manipulative), 216 of which were observed across multiple responses (range = 1 to 83). These categories were then organized into 14 overarching themes (Table 4): Aggression/Antisociality (e.g., drinks too much, physically violent), Anger/Emotional Dysregulation (e.g., yelling at others, angry), Arrogance/Self-Centeredness/Entitlement (e.g., know-it-all, selfish), Bigotry (e.g., racist, sexist), Callousness (e.g., doesn’t care about others, infidelity), Combativeness (e.g., argumentative, litigious), Domineering (e.g., controlling, demanding), Externalization of Blame (e.g., blames others, doesn’t take responsibility), Immaturity (e.g., acts like a child, whiny), Inconsiderateness/Boundary Violation/Passive Rudeness (e.g., doesn’t listen to others, loud), Irresponsibility (e.g., doesn’t follow through on commitments, refuses to do work), Manipulativeness/Lying/Exploitation (e.g., takes advantage of others, lying to get something), Rudeness (e.g., belittles others, talks behind others’ backs), and Other (e.g., hypocritical, plays favorites). On average, each unique response touched on 1.66 categories (SD = 0.89, range = 1 to 7) and 1.39 themes (SD = 0.64, range = 1 to 5).

Of the 1,453 separate behaviors nominated across 1,203 responses, 1,414 (97.32%) were unique. Raters achieved fair to moderate agreement (overall κ = .42; category-wise κ = .02 to .46) with the highest concordance when classifying the low pole of Conscientiousness and the lowest for high Agreeableness. Most nominated asshole behaviors were categorized as being consistent with low Agreeableness (n = 1005, 71.07%). This was followed by low Conscientiousness (n = 180, 12.73%), high Neuroticism (n = 60, 4.24%), and low Openness (n = 45, 3.18%). “Cannot Rate” (n = 42, 2.97%), high Extraversion (n = 11, 0.78%), low Extraversion (n = 5, 0.35%), low Neuroticism (n = 1, 0.07%) and high Openness (n = 1, 0.07%) were infrequent. The remaining nominated behaviors (n = 64, 4.53%) were classified into multiple categories (i.e., multiple modal responses across raters) with three-quarters (n = 48) including low Agreeableness and approximately half (n = 33) involving low Conscientiousness. High Agreeableness did not appear as the modal rating for any behavior, further supporting the primary role of disagreeableness.

For each target, scores were derived for perceptions of the target’s knowledge, care, and ability to change by taking the mean of the corresponding items across all three nominated “asshole-related” behaviors. Participants perceived targets to have some awareness of the bothersome nature of their behavior for others (M = 3.41, SD = 1.24; 1 - Not at All to 5 - Very Much) but did not care that the behaviors were off-putting to others (M = 1.34, SD = 0.64), suggesting more of a deficit in motivation rather than insight. Targets were perceived to have a small degree of ability to change their behaviors if sufficiently motivated (M = 2.41, SD = 1.16). Exploratory post-hoc analyses examining the relations between social perceptions and perceptions of target personality suggested that perceived target Agreeableness was negatively associated with perception of knowledge (r = -.25, p < .001) and positively associated with perceptions of caring (r = .36, p < .001) and ability to change (r = .16, p = .001). Results for overall perceived “asshole”-ness were consistent with these findings as targets with higher ratings on the insult were perceived as caring less about (r = -.39, p < .001) and having less ability to change (r = -.17, p = .006) their behavior. Perceived “asshole”-ness was not significantly related to perception of knowledge (r = .11, p = .065).

The present study examined information about the perceived personality of targets as communicated via a common insult. Participants perceived the “biggest assholes” in their lives—mostly middle-aged men—as chiefly characterized by a combination of antagonism (i.e., low Agreeableness; immodesty, uncooperativeness, lack of empathy) and hostility (i.e., Anger facet of Neuroticism), replicating Hyatt, Maples-Keller, et al. (2019). Overall, the perceived FFM profile for “asshole” in the present study was similar to prototypes of psychopathic, antisocial, and narcissistic PDs (Lynam & Widiger, 2001; Miller et al., 2001). This is not particularly surprising as Agreeableness tends to account for approximately one-third to three-fourths of overall variance in these constructs explained by the FFM (Lynam & Miller, 2019). Given antagonism’s links to aggression (Hyatt, Zeichner, et al., 2019), antisocial behavior (Vize et al., 2019), relational dissatisfaction (Malouff et al., 2010), and counterproductive work behavior (Mount et al., 2006), popular insults like “asshole” convey important information about the speaker’s perception of a target’s interpersonal functioning that is relevant across a wide array of contexts.

Many responses to our request to “be as specific as possible” about targets’ nominated bothersome behaviors were simple and direct (e.g., “He is rude.”), but others were lengthy and colorful descriptions of unique behaviors. Specific examples of antagonistic behavior ranged from the mundane (“rolling their eyeballs when I speak”) to the frankly antisocial and violent (“drinking excessively and abusing children”).

Consistent with our hypothesis, the low pole of Agreeableness figured heavily into both FFM-based profiles of perceptions of targets and classification of nominated characteristic behaviors. In fact, many of the themes in Table 4 overlap specifically with facets of antagonism (i.e., low Agreeableness) including grandiosity/self-centeredness, manipulation, and callousness. Other themes like aggression and blame externalization do not have specific representation in the FFM (e.g., Chester & West, 2020). Disinhibition (e.g., inconsiderateness and irresponsibility; low Conscientiousness) and high Neuroticism (e.g., emotional dysregulation) also appeared in the preliminary taxonomy, albeit to a much lesser extent. These themes also reflect DSM-5 criteria for antisocial (e.g., irresponsibility, lying, callousness) and narcissistic PDs (e.g., entitlement, exploitation), In general, violation of social norms and mores (e.g., antagonism; disinhibition) and dysregulation of affect (e.g., anger) appear to be central correlates of nominated “asshole” behaviors.

Strengths of the present study include replicating and extending previous work as well as pre-registration of methods and hypotheses. We also note some limitations to the generalizability of our results. More than three-quarters of the target “assholes” were male. This may accurately reflect patterns in the use of the insult (which is consistent with basic gender differences in trait antagonism; Schmitt et al., 2008), but may have limited the scope of our preliminary behavioral taxonomy. Additionally, our participants were predominantly White, heterosexual, and cisgender, and results may not generalize to more diverse samples. Detailed demographic data was not collected regarding targets. Our methodology also suffers from single-rater and mono-method bias. Future studies should seek to study the use of insults among dyads to determine whether characteristic patterns of self-other agreement emerge (for recent work regarding evaluative personality judgment within dyads, see Sun et al., 2021).

The findings also suggest other valuable future directions. Consistent with Hyatt and colleagues (2019), we observed variability in the perceived FFM profiles and nominated characteristic behaviors of prototypical “assholes.” Personality information conveyed by insults may be partially dependent upon individual differences between raters in personality, demographics, or sociocultural identity. Further research should examine insult-related cultural and regional norms of raters as well as other potential predictors of variability such as target gender (e.g., differences in the use of “bitch”), ethnicity, group membership, or other salient characteristics. Additionally, we established that the term “asshole” conveys information regarding perceptions of the target’s personality, but did not examine whether this is an intended, or indeed even a primary consequence of employing insults. Future studies should examine the function of insults, both in terms of motivations for and outcomes of their use. Finally, our qualitative analyses were atheoretical and exploratory. Pre-registered confirmatory analyses of the identified categories and themes among nominated (or observed) aversive interpersonal behaviors is warranted. “Asshole” is used by English speaking adults in the American lay public in reference to individuals who are perceived to exhibit traits consistent with constructs of clinical concern (e.g., psychopathy, narcissism). Thus, exploration of the behaviors which inform further research on interpersonal dysfunction in antagonism-related disorders.

Contributed to conception and design: BMS, CSH, DRL, JDM

Contributed to acquisition of data: BMS

Contributed to analysis and interpretation of data: BMS, DRL

Drafted and/or revised the article: BMS, CSH, DRL, JDM

Approved the submitted version for publication: BMS, CSH, DRL, JDM

The authors have no competing interests.

Data and syntax (https://osf.io/4psq6/?view_only=dcc34a5ee3d2463d9047e4594ca29f95) and supplemental materials and codebook (https://osf.io/n6a7f/?view_only=f666f15ac37b476a81cbd58a8eb040de) can be viewed on OSF. To protect participant anonymity, qualitative data are available only by request. Interested parties are encouraged to contact B. M. Sharpe ([email protected]) or J. D. Miller ([email protected]).