We (1) identify a tactic of social exclusion considered much more common among girls and women: facial expressions of disgust directed toward targets the expresser intends to exclude (Brown and Gilligan, 1993; Owens et al., 2000; Underwood et al., 2004). Intrasexual social exclusion itself is arguably more common among girls and women than boys and men (Feshbach, 1969; Feshbach and Sones, 1971; Baker, 1994; Campbell, 2002; Underwood, 2003; Underwood et al., 2004; Benenson et al., 2008a,b,c, 2011, 2013; Benenson, 2013, 2014; Krems et al., 2015; Arnocky, 2016; Williams et al., 2016, 2022; Ayers et al., 2023). There also seem to be sex/gender differences in the preferred tactics to facilitate such intrasexual exclusion (Underwood et al., 2004; Benenson, 2013). That is, whereas there are multiple ways to exclude a peer—such as shoving a person while yelling, “Scram; we don't want you here!”—girls and women tend to favor subtler tactics, such as non-verbal expressions of exclusionary intent (e.g., Björkqvist, 1994; Bjorkqvist et al., 1994; Campbell, 1999; Underwood, 2004; Vaillancourt, 2013). Here, we zero in on one tactic of social exclusion seemingly much more frequent among women: dirty looks—and, in particular, facial expressions of disgust to communicate the expresser's desire to keep that peer socially distanced from oneself and one's group (Hines and Fry, 1994; Shute et al., 2002; Underwood, 2004; Vaillancourt and Sharma, 2011).

We (2) derive and test predictions about women targets' reactivity to this potential cue of impending exclusion, which women might recurrently—and uniquely—face. Again, our key argument is that, insofar as women (vs. men) are more likely to recurrently encounter disgust expressions from same-sex/gender excluders, women might possess distinct interpretations of and reactions to other women's expressions of disgust.

Thus, here, we test whether (a) women possess distinct interpretations and (b) affective reactions to other women's facial expressions of disgust. We predict that, compared to men, women will be more likely to interpret ambiguously directed facial expressions of disgust from same-gender expressers as cues of expresser intent to exclude them and also that women will report greater hurt in reaction to these cues. We also test (c) if these reactions are consistent with such expressivity being a cue of social exclusion in this interpersonal context (i.e., among women). Specifically, if the above logic is correct, then those women who are most dispositionally concerned with being socially excluded would report being especially hurt by cues they interpret as communicating exclusionary intent (i.e., by other women's disgust expressions). We test these predictions, their specificity [i.e., women's greater hurt reactivity to other women's (but not men's) disgust expressions], and alternative possibilities for the predicted patterns of findings. Data and code are available on Open Science Framework: https://osf.io/4hbkp/.

We conducted a 2 (Participant sex/gender) × 2 (Expression: Disgust, Anger) × 2 (Interpretation: Avoidance, Harm) mixed-factors ANOVA, which yielded significant main effects of Expression, F_{(1, 212)} = 584.30, p < 0.001, ηp2 = 0.734, and Interpretation, F_{(1, 212)} = 20.61, p < 0.001, ηp2 = 0.089, and interactions of Expression and Participant sex/gender, F_{(1, 212)} = 4.00, p = 0.047, ηp2 = 0.019, and Expression and Interpretation, F_{(1, 212)} = 37.19, p < 0.001, ηp2 = 0.149, all qualified by a three-way interaction, F_{(1, 212)} = 17.51, p < 0.001, ηp2 = 0.076.

As predicted, compared to men, women were more likely to interpret same-gender disgust expressions as meaning that the expresser intends to exclude them, F_{(1, 212)} = 17.75, p < 0.001, ηp2 = 0.077, 95% CI = (0.41, 1.14). See Table 1 for means (SEs). There were no other differences between female and male participants (ps > 0.495). Additionally, both women and men interpreted disgust expressions as cuing greater exclusionary intent than anger expressions: women, F_{(1, 212)} = 490.16, p < 0.001, ηp2 = 0.698, 95% CI = (2.77, 3.31), and men, F_{(1, 212)} = 156.41, p < 0.001, ηp2 = 0.425, 95% CI = (1.80, 2.47).

To examine whether women's emotional reactions differed from men's reactions across the six measured emotions in response to women's or men's disgust expressions, we nested the six emotions within individuals and used linear mixed modeling (using the MIXED model function in SPSS 28) to regress within-person levels of each emotional reaction onto Participant sex/gender and Expresser gender (both dummy coded: women = 0, men = 1) as well as a categorical variable for Emotion that we included as a fixed effect. In this model, we allowed the categorical fixed-factor variable for Emotion to interact with both Participant sex/gender and Expresser gender, and we entered the additional two-way Participant sex/gender × Expresser gender and three-way Emotion × Participant sex/gender × Expresser gender interactions. Finally, we allowed the intercept to vary randomly across participants, and we specified a variance components covariance structure. This linear mixed modeling approach allowed us to simultaneously estimate all within-person associations for each emotional expression in a single model, reducing the likelihood of making a Type I error. Results revealed that the three-way interaction emerged as significant, F_{(5, 1950)} = 2.28, p = 0.045 (see the Supplementary Table S3 for results from the full linear mixed model).

To determine the simple effects for the significant three-way interaction, we next re-estimated the previous model but (a) suppressed the intercept to produce separate but simultaneous intercepts for each combination of emotion and expresser and participant sex and (b) allowed those intercepts to vary randomly across participants. We next used the TEST subcommand to create custom hypothesis tests specifying the simple effects. See Table 2 for results (means and SEs) from this model. In line with predictions, compared to men, women reported greater hurt in reaction to disgust expressions on women's faces, t_(1842.61) = 3.39, p < 0.001, ηp2 = 0.017, 95% CI = (4.92, 18.43) (see Figure 1). There were no participant sex/gender differences for reported hurt in reaction to disgust expressions on men's faces (p = 0.979). Additionally, women reported much greater hurt in reaction to disgust expressions from women than men expressers, t_(1842.61) = 4.76, p < 0.001, 95% CI = (8.81, 21.16), whereas men reported similar levels of hurt in reaction to disgust expressions from women or from men expressers (p = 0.357). The sex/gender difference in reaction to disgust expressions on women's faces did not extend to anger (p = 0.160), pride (p = 0.458), joy (p = 0.599), sadness (p = 0.383), or fearfulness (p = 0.924).

A 2 (Participant sex/gender) × 2 (Recalled expression) ANOVA yielded significant main effects of Participant sex/gender, F_{(1, 89)} = 12.00, p < 0.001, ηp2 = 0.119 and Recalled expression, F_{(1, 89)} = 5.96, p = 0.017, ηp2 = 0.063. The interaction was not statistically significant (p = 0.058).

In line with predictions made explicitly about simple effects, women recalled greater hurt in reaction to same-sex/gender disgust expressions than men, F_{(1, 89)} = 14.42, p < 0.001, ηp2 = 0.139, 95% CI = (0.92, 2.93). See Table 3 for means (SEs). Women (but not men, p = 0.720) also recalled greater hurt in reaction to disgust than to anger expressions, F_{(1, 89)} = 10.06, p = 0.002, ηp2 = 0.102, 95% CI = (0.44, 1.90).

We regressed reported hurt onto Participant s = Sex/gender (dummy coded; female = 0, male = 1), NTB (centered), PVD (centered), and the two Sex/Gender × Individual difference interaction terms. See Table 4 for results. In line with predictions, this yielded significant effects of Sex, such that women (vs. men) reported greater hurt in reaction to disgust expressions, and of NTB, such that greater NTB predicted greater hurt reported by women (but not men, p = 0.687).

We additionally found a significant Sex/Gender × NTB interaction (see Figure 2). Exploring the simple slopes revealed that, among participants higher in NTB (+1 SD), women reported greater hurt than men, b = −1.80, t₍₂₅₉₎ = −4.19, p < 0.001, 95% CI = (−2.64, −0.95). Among those lower in NTB (−1 SD), this difference was not significant (p = 0.980).

One might wonder why men failed to interpret and react to disgust expressions as women did. We would certainly not argue that boys and men are wholly insensitive to detecting or reacting to cues of impending exclusion. Rather, we raise two possibilities. First, it is possible that at least intrasexual social exclusion is more frequent and/or consequential among women (e.g., Benenson, 2013, 2014), leading women to be more likely to err on the side of inferring that other women mean to exclude them (Haselton and Buss, 2000). Second, men (vs. women) may be more likely to face comparatively overt tactics of exclusion (e.g., shoving, yells of “We don't want you here!”). If male actors use more overt cues of impending exclusion, then male targets might not need to be so attuned to such subtler ones. If these expressions are not cues of exclusion among men, then men have no need to interpret and react to them as such. Indeed, others have asserted that non-verbal expressions of exclusion may very well “have special meanings for girls” and women (see also, Brown and Gilligan, 1993; Reader, 1999; LaFrance, 2002; Underwood, 2004). If this second possibility is correct, perhaps men would react similarly as women did here to those more overt cues of impending exclusion. One might also explore how various fundamental social motivations (Kenrick et al., 2010; Cook et al., 2021), such as strong desires to affiliate, might affect men and women's reactions to cues perceived as exclusionary.

The present work does not address why women and men sometimes differently enact intrasexual social exclusion in the first place. Some have reasoned that perpetrating physical aggression or other tactics that evoke physical retaliation might be costlier for women and/or that women's indirect tactics are especially well-tailored to harm other women (e.g., Bjorkqvist et al., 1994; Campbell, 2002; Benenson et al., 2013). On this view, perhaps women's disgust expressions facilitate the exclusion of disfavored others while also allowing expressers some plausible deniability of aggressive intent, attenuating the likelihood that aggressor-expressers face retaliation from targets and their allies. On this view, expressions of disgust among women might be an implicature—inferred content from a communication (Sperber and Wilson, 1998)—as is indirect speech (e.g., “Officer, is there a way we can make this ticket go away?”; Pinker et al., 2008). For example, compared to bald propositions, such insinuations allow speakers to enjoy the greatest possible benefits (e.g., getting out of the ticket) without facing the full possible costs (e.g., getting arrested for attempted bribery). Here, women expressers might enjoy the benefits of excluding same-sex/gender others without risking the costs of being deemed an overt aggressor by targets, their allies, and other would-be third-party punishers.

This raises the question of whether third parties—notably, men—recognize what one woman's look of disgust toward another can mean. Indeed, the potentially encrypted nature of this message—primarily made by and for women—could imply not only that women expressers are unlikely to get tagged as aggressors by (male) onlookers, but also that women targets might be less believed and supported by (male) allies. Future work should examine the extent to which women's tactics of aggression allow them to avoid retaliation from targets as well as from others (e.g., targets' allies, audiences of gossip)—and the extent to which this differs by sex/gender.

The present work used U.S. community samples, in large part because the work detailing women's disgust expression use was primarily conducted in the U.S. However, women seem to eschew physical tactics of aggression across cultures, especially when aggressing against other women (Burbank, 1987). To the extent that women across cultures use non-physical tactics, then women across cultures might possess distinct interpretations of and reaction to such cues—be they facial expressions of disgust or others. Future work might seek to identify culturally distinct male- and female-typical tactics of aggression to explore the broader hypothesis that people are attuned to that tactics of aggression that they recurrently face. One might also wish to explore how factors even within nations, such as ecology and race in Western nations (Campbell, 2002; Williams et al., 2016), might shape the intrasexual aggression that women experience, and thus their reactivity. For example, faces used here appeared White; given racialized perceptions of emotion in the U.S. (e.g., Becker et al., 2010), future work might examine the role of race.

Although data in each studied largely supported a priori predictions, we note that some studies did not employ attention or bot checks, whereas others did (as explicitly noted in the respective “Methods” sections). Note that the inclusion of inattentive participants would likely work against our ability to find a signal through the noise.

Universally, disgust expressions are made in response to potentially noxious stimuli, with social perceivers recognizing and often benefitting from what others' disgust connotes (Shariff and Tracy, 2011). The present work suggests that such disgust expressions might also have social functionality (e.g., Kupfer and Giner-Sorolla, 2017)—and not just in dyadic situations, wherein, say, Akeelah's disgust at Betty connotes Akeelah's desire to avoid Betty—but also in triadic and n-person situations. Consider, for example, Akeelah and her friend Carla are chatting when Carla catches sight of Dee, a woman she finds distasteful and wants to avoid. In such a context, Carla's disgust face “at” Akeelah would be read by Akeelah as connoting not Carla's desire for distance from Akeelah, but rather Carla's desire for both her and Akeelah to maintain their social distance from Dee. This illustration is consistent with the notion that, just as people's internal feelings of disgust help them maintain safe distances from potentially noxious objects, interpersonal disgust expressions might help people and groups maintain “safe” distances from socially undesirable people. Perhaps this is why the 2015 Pixar film “Inside Out” introduced its anthropomorphized Disgust character as the being responsible for keeping the 11-year-old girl she inhabited “from being poisoned—physically and socially” (Rivera and Docter, 2015).

Women's and men's tactics of intrasexual aggression sometimes diverge. We reasoned that women targets of intrasexual aggression would be uniquely attuned to recognizing and reacting to the tactics that they uniquely and recurrently face. We found evidence for this proposition with respect to facial expressions of disgust, a tactic of indirect aggression—and specifically of social exclusion—more likely among women (Underwood, 2004): More than men, (1) women participants interpreted other women's disgust expressions as cuing expresser intent to exclude them, and—as such exclusion hurts (Eisenberger et al., 2003)—(2) reported greater hurt in reaction. (3) Further, women more dispositionally concerned with social inclusion reported greater hurt in reaction to other women's disgust expressions. Together, these data suggest value in examining women's responses to the somewhat distinct tactics of aggression they recurrently face, particularly in the context of their same-sex/gender social ecologies.