Accurately detecting happy facial expressions associates with perceptions of flirtatiousness

Abstract

Emotional facial expressions are relevant to flirtation because they provide information on an individual’s intentions or motivations. Individual differences in the ability to accurately detect and discriminate between normative facially expressed emotions could lead to misperceptions of the level of sexual interest being conveyed, which has been linked to sexual assault and harassment. To explore this notion, we recruited a national sample of college aged male and female participants (N = 219) who completed a novel facial expression recognition task used to detect accuracy in processing facial emotions of happiness, surprise, anger, and disgust. Participants also viewed multiple video clips of blind dates between two different-sex participants and rated each partner on their degree of flirtatiousness. Consistent with predictions, we found that individuals who misidentified other facial emotions for happiness appeared to overestimate flirtation. Though not predicted, participants who failed to accurately identify happy faces also overestimated flirtation, whereas individuals who took longer to respond to emotional facial expressions and misidentified an emotion as conveying happiness made greater errors in perceptions of flirtatiousness. Overall, these findings suggest that individual differences in the ability to detect and discriminate happiness through facial expressions are relevant to misperceptions of flirtatious behavior, and more broadly illuminates the role of basic emotion recognition on perceptions of flirtatiousness.

Keywords

flirtation sexual perception facial emotion recognition misperception of sexual interest sex differences

Millions of women are sexually assaulted or harassed each year (Smith et al., 2018) – most by men they know – with college-aged women in particular being three to four times as likely to experience sexual violence than women in general (Sinozich & Langton, 2014). Although there are factors that differentiate sexual assaults during college from those in other settings (e.g., likelihood that alcohol was present; Fedina et al., 2018), the negative mental and physical health consequences of sexual victimization remain the same (e.g., anxiety and depression; Carey et al., 2018).

Several theoretical models of sexual misconduct (e.g., the extended confluence model of sexual aggression; Malamuth et al., 1991) have specified a role for the overestimation of sexual interest as a precursor to such sexual offenses. Misappraisals of sexual interest occur, in part, because the communication of sexual interest is more often veiled than direct (Gersick & Kurzban, 2014), and certain indicators do not exclusively signal flirtation (e.g., smiling). Thus, flirtatiousness must be discriminated from ancillary information, as well as behaviors better characterized as politeness, friendliness, or mere happiness (Farris et al., 2010). There may also be internal processes that influence perceptions of sexual interest (Koenig et al., 2007; Perilloux et al., 2012; Ranganath et al., 2009; Treat et al., 2019), such as individuals projecting their own attraction when gauging the sexual interest of another. While misconstrued sexual intent in many situations may be easily corrected; in some situations, it may lead to inappropriate comments, sexual overreach, or the discounting of even overt signs of rejection, thereby facilitating coercive or assaultive behavior (Muehlenhard et al., 2016). Therefore, it is important to evaluate the factors that uniquely influence perceptions of sexual interest, especially in settings with elevated risk of misconduct.

This line of research is incompletely characterized without minding specific attention to sex differences in perceptions of sexual interest. Studies find that male participants are more prone to misperceive platonic friendliness as sexual interest, especially from female targets (La France et al., 2009), and appear to conflate perceptions of female target’s sexual interest with their own (Koenig et al., 2007). Sexually coercive male participants have been found to overperceive sexual interest in laboratory studies, as well as report histories of overestimation of sexual interest in their dating histories (Farris et al., 2006; Treat et al., 2001). These findings depict a broad frame whereby sex-based asymmetries in sexual interest and corresponding appraisals of it may partially explain behavioral asymmetries in sexual misconduct. Although sex-based differences do not always emerge in the context of sexual communication (e.g., performative understanding of affirmative consent policy; Mattson et al., 2022), the effect that male observers tend to perceive higher degrees of flirtatiousness than female observers is stable across converging operations (Abbey, 1982; Cahoon & Edmonds, 1989; Farris et al., 2008; Koukounas & Letch, 2001). This robust effect has underwritten lines of research aimed at identifying the mechanisms that perpetuate this perceptual bias. Following a social-information processing model (Crick & Dodge, 1996), we hope to contribute to this literature by examining whether individual differences in the ability to accurately identify facial affects influences perceptions of flirtation in reliable ways, both per se and with attention to sex differences.

Facial expression and perceived flirtation

Although sexual interest can be conveyed (and thus misconstrued) across a myriad of verbal and nonverbal channels, growing empirical attention is being paid to communication through facial expressions. Facial expressions are particularly relevant to flirtation given that these interactions are often face-to-face and that facial expressions convey affective information nonverbally. Facial expressions have adaptive significance during courtship (Schmidt & Cohn, 2001) specifically because they provide rich information about a potential sexual partner’s emotional or motivational state (Matsumoto et al., 2008). Facial expressions are also shown to shape observers’ judgments about various courtship-relevant emotional experiences such as happiness or anger, mental states like boredom or confusion, and personality traits including trustworthiness or dominance (Chen et al., 2015; Krumhuber et al., 2007; Montepare & Dobish, 2003; Rozin & Cohen, 2003). There is also research highlighting a specific expression as a generally agreed-upon indicator of female flirting (Haj-Mohamadi et al., 2021).

Ultimately, communication through facial expressions may be even more diagnostic of flirtation than the actual words exchanged, as language may be veiled to provide the speaker plausible deniability that protects against explicit rejection (Moore, 2010). It is this degree of ambiguity, however, that makes communication through facial expressions susceptible to misappraisal. This notion is consistent with previous work indicating that individuals closely scrutinize facial expressions to determine flirting, but certain expressions – ones that potentially convey flirtatiousness – are prone to detection errors (e.g., relatively high rates of false-positives and false-negatives; Benda & Scherf, 2020; Hall et al., 2014; Ranganath et al., 2009).

Facially expressed emotion and flirtation

Humans are highly sensitive to emotional faces (Calvo & Esteves, 2005) and automatically process the information contained therein, which permits dynamic evaluation and decision-making during social interactions. With respect to flirtation, individuals often weigh multiple, tentative hypotheses about a candidate partner’s sexual intentions (Treat et al., 2019), and there is evidence that both males and females use emotional cues to update appraisals of sexual interest (Perilloux et al., 2012). Presumably, in the context of a dating situation, expressed positive affect (e.g., happiness) will be more aligned with flirtatiousness than negative affects (e.g., anger, disgust), which may be antagonistic to perceptions of sexual interest. However, this information is only useful insofar as it is accurately identified.

From a social-information processing standpoint (Crick & Dodge, 1996), the encoding of social cues (e.g., positive affect) and their subsequent interpretation (e.g., as a sign of flirtation) influences the type of social response that ensues (e.g., moving in for a kiss). Difficulty encoding can lead to the misinterpretation of a situation and a problematic social response in tandem. Babcock et al. (2008) applied this model to an investigation of intimate partner violence, finding that violent/antisocial male participants showed deficits in coding emotional faces - particularly anger, happy, neutral, and surprise - which they linked theoretically to social misappraisals and correspondingly aggressive behavioral responses. Analogously, in the context of a dating situation, communicated emotions that are either overlooked or misidentified may lead to miscalibrations of flirtation perceptions, which in turn may influence unwanted sexual behaviors.

More specifically, if an expressed negative emotional display is missed, a person may overestimate the other party’s sexual interest, which aligns with findings that men prone to sexual misconduct tend to be less sensitive to women’s negative affective cues (Treat et al., 2019). Overestimations of sexual intentions can also result from ambiguously valenced emotions, such as surprise. This is consistent with previous evidence supporting the tendency of individuals to identify images of surprise as signifying happiness (Benda & Scherf, 2020). Expressions of surprise are not always positive, and so conflating them with expressions of happiness should erroneously inflate perceptions of sexual interest. Finally, it is also possible that even the accurate identification of happiness during an interpersonal exchange can inflate estimates of flirtation. The expression of happiness can occur in both friendly and flirtatious interactions, but the context of a dating scenario could incline individuals to appraise positive affect as more flirtatious. Thus, those with greater true positive identifications of happiness may be subject to errors at the interpretation stage, presuming greater flirtatiousness on the basis of higher frequencies of observed positive affect.

Facially expressed emotion, flirtation, and sex differences

Collectively, prior research and theory highlight the role of accurate processing of facial emotion in the overestimation of flirtatiousness. A remaining question, however, is whether sex differences in the ability to decode specific emotional facial expressions influences overestimation of sexual interest, across situations, by men relative to women. Although some research suggests that men (and women) who endorse rape-supportive attitudes may be less influenced by a woman’s affective cues (Treat et al., 2017), it is unclear if this represents a lowered capacity for accurately identifying emotion (i.e., an encoding problem) or other internal biases (e.g., perceptions of one’s own mate value; Haselton, 2003), or if their attention is on areas other than the face of the person. Indeed, men tend to strongly endorse traditional sexual scripts (Gagnon & Simon, 2017; Newstrom et al., 2021), and may accurately encode negative affects when present but view this as the woman playing “hard to get,” in accord with their prescribed role as sexual gatekeeper. Differentiating between an encoding versus an interpretation error may be relevant to explaining the reliable effect that males – relative to females – tend to overestimate the level of flirtation in cross-sex interactions (see La France et al., 2009).

Females are more accurate than males at decoding facial communication in general (McBain et al., 2009) and – while there appears to be little difference in accuracy with respect to identifying happiness – some studies suggest that female participants may be more accurate in recognizing negative facial affect cues (Forni-Santos & Osório, 2015). As such, it is possible that men’s relative insensitivity to negative affect cues, on average, may partially explain why men as a group tend to overestimate flirtation. It is also possible that men who accurately encode positive affect overestimate the degree of flirtation, as men are more generally theorized to possess a greater motivation to interpret a woman’s behavior as communicating sexual interest. Evolutionary theorists, for example, have suggested that mixed-sex interactions activate mating goals, especially in men (e.g., error management theory; Baumeister et al., 2001; Haselton & Buss, 2000), which might explain some men’s motivation to interpret facial expressions like happiness as flirtation to avoid missing potential mating opportunities. It is therefore possible that the proposed effect, whereby greater accuracy in identifying happiness in a dating context may inflate perceived flirtatiousness and help account for the effect of sex differences.

The current study

The current work seeks to evaluate the relevance of basic facial emotional recognition accuracy on appraisals of flirtatiousness in a blind date scenario, which was used to provide a surrounding context relevant to flirtatious behavior. Participants were college-aged males and females who viewed several simulated blind dating interactions and evaluated both the male and female interaction partners’ level of flirtatiousness. This study also piloted a novel emotional recognition task. Whereas other studies have relied on static pictures of varying facial expressions (Haj-Mohamadi et al., 2021) that clearly depict one emotion or another, unique to our task is that emotional expressions gradually morphed from a neutral face to one of four basic emotions relevant to flirtation – happiness, surprise, anger, and disgust. This task is more consistent with how emotions are conveyed naturalistically – over a range of intensity. Participants responded at the point they detected an emotion and identified which of the four emotions they believed was expressed. This allowed us to evaluate the effects of speed with which individuals infer an emotion, as well as its interplay with identification accuracy, on how individuals differentially appraise the level of sexual intent in a heteronormative dating scenario.

Hypotheses

We derived our hypotheses from a social-information processing frame (Crick & Dodge, 1996), whereby individual differences in the ability to encode particular facially expressed affects will correspond to specific types of misappraisals at the interpretive level. We first hypothesized that individuals who had more difficulty identifying anger or disgust (i.e., negative affect) when present would tend to overperceive flirtatiousness, and that this would at least partially account for the anticipated effect of sex. This hypothesis aligns with findings about the importance of sensitivity to negative affective cues in accurately detecting levels of sexual interest (Treat et al., 2019). Next, we hypothesized that – regardless of accuracy – those who more frequently identified facial expressions as emoting happiness would also overperceive levels of flirtation. This hypothesis was based on the pretense that even accurately perceived happiness – particularly in a dating context – would filter into perceptions of flirtatiousness despite that in some cases it was more indicative of friendliness. Thus, individuals who more frequently identify happiness by any route may be more prone to overestimate flirtatiousness in these situations relative to those who have difficulty detecting this emotion when present. Given that internal beliefs more pronounced in men might bias the interpretation of facially expressed happiness as signaling sexual interest, we predicted that this effect would also partially account for the hypothesized effect of sex on perceptions of flirtation.

We made no a priori hypotheses regarding surprise as it neither inherently conveys positive or negative affect. That is, individual differences in accurately identifying surprise may lead to overperceptions of sexual interest but may not be reliably biased in one direction (e.g., under-perceiving or over-perceiving). However, perceptions of surprise are shaped by context (Cheal & Rutherford, 2013), so perhaps within blind dating interactions that more frequently communicate interest, individual differences in the ability to detect surprise may reliably lead to one kind of perceptual error. Lastly, we tested the interaction between response time and each emotion recognition accuracy variable. Predicated on the speed-accuracy trade-off, we hypothesized that quick, inaccurate perceptions of facial emotions would be associated with greater misappraisals of flirtatiousness. However, we did not believe reaction times would reliably over- or under-predict the level of flirtation, so we examined the absolute magnitude of the errors in flirtation perception while excluding the valence.

Method

Participants

This study was conducted in two phases, with the initial phase focused on developing the stimuli for use in the second. More specifically, participants in phase one were recruited from a university sample of heterosexual males (n = 30) and females (n = 30) ranging from ages 17 to 21. These participants – who had to be United States citizens and fluent in English – were randomly assigned to a different-sex pair and participated in a “blind date” interaction. Excerpts from these interactions were then used as the stimuli for the second phase of the study (see Procedures). Participants for the second phase of the study were recruited using Qualtrics’ research panels. Eligibility criteria required all participants to be currently living in the United States, fluent in English, between 18 and 21 years old, and completing the study on a desktop or laptop computer (cf., cell phone, tablet, etc.). Of the 310 participants that matched these qualifications; 3 cases (1%) failed attention checks, and 5 (1.6%) and 7 (2.3%) cases evidenced a straight-lining or random response set, respectively. These cases (n = 15) were removed from further analysis. As the current focus was on over perceptions of flirtation within heteronormative interactions, gender non-binary individuals (n = 5) and those who did not identify as heterosexual (n = 71) were recruited but their data were removed from the forthcoming analyses, yielding a final sample of 219 participants. Demographic information for phase one and two is reported in Table 1.

Table 1.

Sociodemographic Characteristics of Dyad and Participant Observer Samples.

Characteristics	Dyad sample N = 56	Observer sample N = 219
Sex	Female: 28 (50.0%)	Female: 110 (50.2%)
	Male: 28 (50.0%)	Male: 109 (49.8%)
Age	19.20 (1.05)	19.74 (1.08)
	Median 19	Median 20
Race
White	27 (48.2%)	141 (64.4%)
African American/Black	8 (14.3%)	30 (13.7%)
Asian/Pacific Islander	13 (23.2%)	18 (8.2%)
Latino/a/Hispanic	3 (5.4%)	19 (8.7%)
Other	3 (5.4%)	7 (3.2%)
Did not report	2 (3.6%)	4 (1.8%)
School status
Freshman	21 (37.5%)	50 (22.8%)
Sophomore	24 (42.9%)	53 (24.2%)
Junior	3 (5.4%)	33 (15.1%)
Senior	8 (14.3%)	33 (15.1%)
Other	-	1 (0.5%)
Did not report	-	49 (22.4%)
Relationship status
Single	35 (62.5%)	151 (68.9%)
Casually dating	5 (8.9%)	10 (4.6%)
Seriously dating	16 (28.6%)	48 (21.9%)
Engaged	-	2 (0.9%)
Married	-	7 (3.2%)
Did not report	-	1 (0.5%)
Affiliated with Greek organization
Present member	6 (10.7%)	19 (8.7%)
Past member	1 (1.8%)	6 (2.7%)
Median household income	$70,000–89,999 (n = 5)	$30,000–49,999 (n = 38)
Region
Northeast	54 (96.4%)	53 (24.2%)
Midwest	-	59 (26.9%)
Southeast	-	54 (24.7%)
Southwest	-	25 (11.4%)
West	1 (1.8%)	24 (11.0%)
Other	-	3 (1.4%)
Did not report	1 (1.8%)	1 (0.5%)

Measures

All participants self-reported sex, race, age, and other sociodemographic characteristics. Participants also completed several survey instruments that tapped various gender role traditionalism (e.g., The Attitudes Towards Women Scale; Spence et al., 1973) and characterological traits associated with sexual misconduct (e.g., Levenson Self-Report Psychopathy Scale; Levenson et al., 1995), which were part of a larger study examining person-level determinants of misperceptions of flirtatious behavior. However, only the measures pertinent to the current analyses are discussed below.

Perceptions of flirtation

Dyads in the “blind date” interaction were asked to separately rate perceptions of their own as well as their partner’s sexual interest using the Perceptions of Sexual Interest Questionnaire (Abbey, 1982; Abbey & Harnish, 1995). This questionnaire employs a 7-point Likert-type scale on which participants rate a target individual using four descriptive terms relevant to flirtation (i.e., flirtatious, seductiveness, sexy, promiscuous), as well as additional trait terms, such as likability and intelligence, to hide the true focus of the study. Items tapping sexual interest were averaged to form an omnibus scale, with higher scores indicating higher levels of perceived flirtatiousness. Average scores by observers of the female and male targets were 2.99 (SD = 1.69; range = 1–7) and 2.86 (SD = 1.61; range = 1–7), with alpha coefficients of .91 and .88, respectively. We also averaged scores on the remaining scale items (e.g., friendly, likeable, etc.) to form a general evaluative index to be used as a control; that is, to ensure that the observer evaluations of flirtatiousness are unique to those descriptors and not an extension of a more or less positive overall appraisal of the individual. Omnibus scores on the general evaluation index for female and male targets had sample averages of 4.93 (SD = 1.15; range = 1–7) and 4.74 (SD = 1.24; range = 1–7), respectively, with alpha coefficients of .91 and .92, respectively, indicating strong internal consistency.

We also obtained ratings by male and female targets of their own level of flirtatious behavior to be used as an anchor point against which to measure the accuracy of observer perceptions of flirtatiousness (see Analytic Plan). The female simulation partners on average reported 1.42 (SD = 0.59; range 1–3.5), whereas male simulation partners reported 2.25 on average (SD = 1.27; range = 1–5). Simulation partner’s self-assessments of flirtation were positively correlated, r = .38, p < .001, suggesting some degree of reciprocal flirtation within the interactions. Notably, however, observer perceptions of flirtatiousness were marginally correlated with female but not male simulation partner’s self-reported flirtatiousness, r = .08, p = .039, and r = .06, p = .127, respectively; though there was a small correlation such that observer perceptions of male target’s flirtatiousness were higher when the female target’s self-reported level of flirtation was higher, r = .12, p = .003.

Observer sexual attraction to the interaction targets

Participants were also asked to what extent they were sexually attracted to each of the interaction targets (rated on a scale from 1 “not at all” to 7 “extremely”). This was used as a control given evidence that this correlates with over-estimates of sexual intent and accounts for a large portion of the effects of sex on such perceptions (Lee et al., 2020). Female observer ratings of male targets averaged at 2.17 (SD = 1.57; range = 1–7), whereas male observer ratings of female targets averaged at 3.13 (SD = 1.59; range = 1–7).

NimStim facial expression recognition task

The NimStim Set of Facial Expressions comprises 672 images of human faces of varying affect and is validated for use in empirical research (Tottenham et al., 2009). These stimuli in particular were chosen because they represent multiple races and ethnicities, and research suggests that the race or ethnicity of the model can have impacts on how facial expressions are processed (Herrmann et al., 2007). We presently adapted these images for a computerized choice reaction time task that prompted participants to indicate the emotional valence of adapted NimStim images. Each target stimuli was adapted using photo morphing protocols (Vanhalst et al., 2017) to create a spectrum of emotional images that ranged from neutral to the targeted emotion for each face. A racially diverse set of six adult female and six adult male faces were selected as the target stimuli. Morpheus Photo Morpher, Version 3.17 (Morpheus Software, Inc.) was then used to create each short GIF of one picture morphing into another from the same target (see Supplemental Materials for more information). Every trial of this task began with presentation of a neutral target stimuli followed successively by morphed-target stimuli that gradually became representative of a full-intensity emotional expression (see Figure 1). There were four GIFs per face (i.e., neutral to happy, neutral to surprise, neutral to disgust, neutral to anger) and 12 faces, yielding 48 stimuli in total.

Figure 1.

A Schematic Representation of the Facial Expression Recognition Task. Note. During each trial of the task, participants viewed faces morph from a neural facial expression to a face of affect of either happiness, surprise, disgust, or anger over the course of 20 seconds.

Participant observers first completed 3 practice trials (consisting of 3 base target stimuli that were morphed to 3 different emotional expressions) followed by 12 test trials (consisting of 12 base target stimuli that were evenly morphed to different emotional expressions). Target stimuli and their emotional expression morphing were randomized across task administration (Vanhalst et al., 2017). Summary emotion recognition variables were created based on the accuracy of participant responses and included false and true positive detection of happiness (i.e., misconstruing another emotion as happiness and correctly predicting happiness when it is presented, respectively) and false negative detection of surprise, anger, and disgust (i.e., missing the facial expression when it was presented). These metrics were chosen, as opposed to other measures of accuracy (e.g., positive predictive power), since they align most closely with our theoretical framework. Response times were highly correlated across trials and a Principal Components Analysis (PCA) yielded a one component solution. We therefore averaged response times across trials to create an omnibus index. On average it took approximately 6.08s (SD = 3.43; range .51–19.17) for participants to respond to the GIFs. There were small but consistent correlations between response time and the accuracy variables (|r|s ranged from .10 to .20), such that faster response times were associated with less accuracy. Means, standard deviations, and intercorrelations between the emotion recognition variables can be found in Table 2.

Table 2.

Correlations Among Affect Recognition Rate Variables.

	n	M	SD	1	2	3	4	5
1. Anger false negative	218	.80	.91	–
2. Disgust false negative	218	.97	.87	.28**	–
3. Surprise false negatives	218	.40	.79	.33**	.34**	–
4. Happiness false positives	218	.38	.87	.43**	.35**	.64**	–
5. Happiness true positives	217	2.63	.79	−.23**	−.39**	−.63**	−.28**	–

Note. **p < .01.

Procedure

The blind date simulation was adapted from Corcoran’s previous protocol (Corcoran, 1996). The task is intended to replicate a blind date as closely as possible under experimental conditions in efforts to increase ecological validity. A key advantage of this paradigm is its use of two unacquainted participants rather than using a trained confederate. To enhance the external validity of the task further, the recording room in which participants completed the task was modeled and furnished like a residential space with a central wooden table and two chairs.

Two random different-sex participants were scheduled to participate in the study at specific timeslots. When participants arrived, they completed informed consent procedures in separate rooms to limit the interaction between participants prior to the task. Following the informed consent process, individual participants were led to the recording room where they were first introduced to their fellow different-sex participant. A research assistant then explained the task according to the following script: “The role you are being asked to play is that of a first date. So that this will seem like a blind date, I want you to imagine that you are out together at a cafe. You will have about 10 minutes to get to know one another. For confidentiality reasons, do not share your real name with your partner. Please act in the way you would in a similar situation.” Two hidden cameras were angled to capture the frontal profiles of the dyad members seated around the table. Following the task, the research assistant led participants to separate rooms to complete self-report measures. Participants in this phase of the study were remunerated with one course credit through SONA. Overall, 55 clips – each 2 minutes in length – were extracted from the dating simulations and were selected because these segments evidenced the greatest amount of active interaction between the partners that was representative of their entire interaction. The clips were presented in a split screen format so that the participant observers could watch each simulation partners’ behavior throughout the interaction.

For the second phase of the study, participant observers were randomly assigned to watch three clips of different dyads before completing self-report measures and the facial expression recognition task. On the last page of the survey, participants were given a debriefing statement which included information about the study along with counseling resources and hotline numbers if necessary. Per Qualtrics requirements, participants were paid a variable amount based on multiple factors, such as their selected payment type (i.e., gift card, check, etc.) and the subject pool from which they hailed. Remuneration ranged from $1.20 and $2.40 in e-currency. All procedures were approved by the Institutional Review Board.

Analytic plan

We estimated mixed models in SPSS 25 to evaluate our hypotheses. The outcome variables for each model were the participant observers’ perceptions of the female and male simulation partner’s level of flirtation with the other. Participants’ ratings of male and female targets were conducted in separate models to evaluate whether the pattern of effects differed across the sexes. The model intercepts were allowed to vary randomly, which segregated the variance in flirtation ratings into (a) each participant’s mean interest rating across the three randomly assigned videos, and (b) a residual term representing within-person variance in the outcome variable across videos. Fixed predictors were added in at both levels, with person-level variables (e.g., facially expressed emotion recognition) explaining variance in the random intercepts, and factors that varied across the videos (e.g., level of physical attraction to the target) correlating with the residual-level variance.

Variables were entered into the model in blocks, beginning with the interaction target’s self-rating of flirtatiousness alongside observer sex. Controlling for each simulation partner’s self-reported level of flirtation, which acted as a proxy for the actual amount of flirtation, allowed us to ask whether the viewer watching the interaction reported the level of flirtation being higher or lower than the target reported. Although not a perfect match, self-reported flirtatiousness is considered an acceptable substitute for assessing accuracy of perceptions of sexual interest (Koenig et al., 2007). Thus, we frame our outcome as representing estimates of flirtation controlling for the actual amount of flirtatiousness. Although a straight difference between reported and perceived flirtation could function as a reasonable alternative, we opted for a residualized difference score approach recommended by Cohen et al. (2014). To establish a baseline effect for sex, it was entered into the model prior to the control or primary predictor variables hypothesized to capture some of the variance of this effect. Participants’ sex was effect coded such that male = −1 and female = 1. In the next block we added each participant’s general evaluation of the male and female in the video and sexual attraction to the target person in the dyad. In the final block, we entered the emotion recognition variables of interest, which were hypothesized to capture random (i.e., between-subject) variance in estimates of flirtation across the three videos, correcting for the effect of the previously modeled control variables.

Our final set of analyses targeted the response time variable and its potential interaction with each of the emotion recognition accuracy variables. We hypothesized that faster reaction times would correspond to greater perceptual errors in flirtation vis-à-vis the simulation targets’ self reports. However, we did not anticipate that reaction times would reliably over- or under-predict the level of flirtation. To capture these potential effects, we first regressed each observer’s perceptions on the simulation targets’ self-reported level of flirtation and retained the absolute value of the residuals, which represented the magnitude of errors absent of their valence. Predictors were centered on their respective means to avoid non-essential collinearity between the simple effect and interaction terms. Across all models, we tested for whether estimates of effects significantly varied depending on whether the target was male or female.

Results

Data cleaning and model assumptions

Missing data for survey items were addressed using mean imputation (Schafer & Graham, 2002). There were several cases where responses to the outcome questions were completely absent or too limited to permit mean imputation (n = 12). In these cases, the offending rows of data were removed from analyses, but other rows of complete data from the same participant were retained. Where participants did not report their level of sexual attraction to the target, we imputed the mean rating of the target’s level of sexual attractiveness as reported by others of the participant observer’s self-identified sex (e.g., female observers’ mean rating of a given male target was imputed for female observers with missing data). Cases where all values for a given NIMSTIM domain were missing were deleted (n = 2). Univariate outliers (z > 3.29) were reduced on the happiness false positive variables by taking the square root, with remaining outliers being reduced in absolute value but retaining their rank in the distribution. Mahalanobis distance and Cook’s distance were calculated to detect any multivariate outliers and influential rows of data, but offending cases were ultimately retained because their removal did not substantively change the findings. All the remaining variables were linearly associated with the outcome and normally distributed.

Primary analyses

Emotion recognition accuracy

We hypothesized that emotion recognition would be associated with perceptions of sexual intentions for both male and female video targets. Specifically, we evaluated whether the tendency to detect happiness facial expressions when it was or was not present, as well as the tendency to miss surprise, anger, and disgust when they were present, would be related to perceptions of flirtatiousness. We also modeled in blocks the observer’s sex, the interaction partner’s own self-rated flirtatiousness, and the observer’s general evaluation of the dyad member and sexual attraction level to the target. Unstandardized regression estimates and 95% CIs for all effects at each step in the model are displayed in Tables 3 and 4.

Table 3.

Observers’ Evaluation of Female Interaction Targets’ Flirtatiousness with Interaction Partner.

	Block 1			Block 2			Block 3
	B(SE)	95% CI		B(SE)	95% CI		B(SE)	95% CI
	B(SE)	LL	UL	B(SE)	LL	UL	B(SE)	LL	UL
Intercept	2.90 (.14)***	2.62	3.17	−.10 (.24)	−.57	.38	−.10 (.27)	−.62	.43
Female Target’s self-reported flirtatiousness	.07 (.07)	−.07	.20	−.02 (.06)	−.14	.11	−.02 (.06)	−.14	.11
Observer sex	−.43 (.10)***	−.62	−.23	−.27 (.09)**	−.45	−.10	−.28 (.09)**	−.45	−.11
Observer’s sexual attraction to female target				.13 (.03)***	.07	.20	.13 (.03)***	.06	.19
Observer’s general evaluation of female target				.57 (.04)***	.48	.66	.57 (.04)***	.49	.66
Anger false negatives							−.04 (.10)	−.23	.16
Disgust false negatives							−.10 (.11)	−.31	.11
Surprise false negatives							−.39 (.17)*	−.73	−.05
Happiness false positives							.98 (.40)*	.19	1.78
Happiness true positives							−.41 (.14)**	−.69	−.12

Note. *p < .05. **p < .01. ***p < .001.

Table 4.

Observers’ Evaluation of Male Interaction Targets’ Flirtatiousness with Interaction Partner.

	Block 1			Block 2			Block 3
	B(SE)	95% CI		B(SE)	95% CI		B(SE)	95% CI
	B(SE)	LL	UL	B(SE)	LL	UL	B(SE)	LL	UL
Intercept	2.81 (.12)***	2.57	3.05	−.03 (.22)	−.45	.40	.02 (.24)	−.45	.48
Male Target’s self-reported flirtatiousness	.03 (.03)	−.04	.09	.04 (.03)	−.02	.10	.04 (.03)	−.02	.10
Observer sex	−.23 (.10)*	−.42	−.05	−.31 (.07)***	−.45	−.17	−.31 (.07)***	−.45	−.17
Observer’s sexual attraction to male target				.26 (.03)***	.19	.33	.26 (.03)***	.19	.32
Observer’s general evaluation of male target				.49 (.04)***	.41	.57	.49 (.04)***	.41	.57
Anger false negatives							−.02 (.09)	−.20	.15
Disgust false negatives							−.13 (.09)	−.31	.06
Surprise false negatives							−.24 (.15)	−.54	.06
Happiness false positives							.80 (.36)*	.10	1.50
Happiness true positives							−.19 (.13)	−.44	.06

Note. *p < .05. ***p < .001.

As anticipated, observer sex was significantly associated with perceptions of flirtation for both female, t(215.39) = −4.29, p < .001, r_effect = −.28, and male, t(215.95) = −2.44, p = .015, r_effect = −.28, interaction targets. General evaluations of and sexual attraction towards the targets were entered into the model. Results indicate that more positive general evaluations of the target corresponded with overestimates of flirtatiousness for male targets, t(639.33) = 11.79, p < .001, r_effect = .42, and female targets, t(621.93) = 12.93, p < .001, r_effect = .46; as did greater endorsements of sexual attraction to the targets, t(628.64) = 7.61, p < .001, r_effect = .29, and t(562.22) = 3.87, p < .001, r_effect = .16, respectively.

The emotion recognition accuracy variables were entered next, and the associated findings provided partial support for our hypotheses. Specifically, we found that greater false positive identification errors of happiness were associated with overestimates of flirtation between the interaction partners, which was consistent across both female and male interaction targets, t(260.95) = 2.44, p = .015, r_effect = .15, and t(238.51) = 2.25, p = .026, r_effect = .14, respectively. Follow-up analyses revealed that individuals misidentified surprise, disgust, and anger as happiness in 44.8%, 33.5%, and 21.6% of the cases, respectively, and that it was the negative affect categories that were driving the effect on overperceptions of flirtatiousness for female targets. Parsing out either of these categories attenuated the effects on overperceptions of male target’s flirtatiousness, though more so when removing the negative affect categories. These findings taken together suggest that the misidentification of happiness is associated with overperceptions of flirtatiousness, and this effect was more pronounced for those inclined to confuse negatively for positively valenced facial emotions.

We also found that people who accurately identified happy faces when they were presented (i.e., true positive) perceived female targets as less flirty overall, t(217.07) = −2.83, p = .005, r_effect = −.19; however, this ran counter to our hypothesis. Finally, there was a small effect for surprise, with more false negatives corresponding to perceptions of lesser sexual interest from female targets, t(231.25) = −2.27, p = .024, r_effect = −.15. However, this effect is misaligned with the trend level positive bivariate association found between surprise false negatives and perceptions of female target’s flirtatiousness, r = .07, p = .066, indicating that the regression estimate likely reflects some form of suppressor effect. These findings overall suggest that individual differences in the ability to detect happiness through facial expressions was relevant to overperceptions of flirtatious behavior, though consistent with expectations only for the effect of false positives on perceptions of both male and female targets. Contrary to predictions, participants with more true positive identifications of happiness perceived less flirtatiousness in female targets. Notably, however, the difference in effects found for happiness true positives and surprise false negatives across male and female targets were not significantly different; and a model collapsing across targets yielded the same pattern of results. In any case, adding the emotion recognition variables into the models did not significantly adjust the regression effect for observer sex, suggesting that these effects are for the most part orthogonal.

Effect of response times and interactions with accuracy

Our next set of analyses examined the effect of response times on perceptions of flirtatiousness, wherein we hypothesized an interaction effect with the accuracy variable such that individuals making quick, erroneous decisions would have the greatest misperceptions of flirtatiousness. Although longer response times on the emotion recognition task were not associated with perceptions of flirtatiousness for male, t(213.73) = 1.28, p = .203, r_effect =.09, or female targets, t(210.05) = .669, p = 0.504, r_effect = .05; there was a significant interaction between response times and false positive identifications of happiness when examining perceptions of male targets, t(248.83) = 2.65, p = .009, r_effect = .17. As displayed in Figure 2, for those with a low false positive detection rate of happiness, those with higher reaction times appeared to make smaller errors in perceptions of flirtatiousness. By comparison, and contrary to expectations; those who also took longer to respond but were relatively inaccurate made larger errors in their perceptions of flirtatiousness, relative to those who were similarly inaccurate but responded more quickly. Notably, of the control predictors; observer sex was the only one to appreciably associate with the magnitude of the errors in perceptions of flirtatiousness, particularly in female targets, t(293.82) = 2.31, p = .021, r_effect = .13, though a trend-level effect was also observed for male targets, t(215.16) = 1.71, p = .089, r_effect = 12; with both suggesting that female participants on average had higher magnitude error scores than male participants. Estimates for the full model for male and female targets can be found in Supplemental Tables 1 and 2. Estimates of effects did not differ across target sex.

Figure 2.

Response Time by False Positive Happiness Interaction Effects on Misperceptions of Flirtatiousness. Note. A significant interaction was present between participant observers’ response time on the facial expression recognition task and their happiness false positive rate on misperceptions of sexual interest in male targets (p = .009).

Discussion

This study evaluated observers’ task-level emotion recognition accuracy and speed using a novel facial emotion morphing paradigm, and examined whether individual differences on this task correspond to over perceptions of flirtation between simulation blind-date partners. Our theoretical frame posited that individual differences in the ability to accurately encode facial emotions will predict errors in estimates of flirtation because an individual’s threshold for perceiving facial expressions is relevant to appraising flirtation. Our findings, in general, support this contention, but converged on the identification of happiness as being particularly relevant to how individuals appraised flirtatiousness in these scenarios. These effects remained even after accounting for multiple factors that have been shown to influence perceptions of flirtatiousness in previous research, such as sex, sexual attractiveness, and general evaluation of the target individual.

The tendency for humans to quickly and accurately identify happiness has been previously described as a superiority effect or cognitive bias (Zsido et al., 2021), and it may even influence determinations of interpersonal flirtatiousness. Indeed, in the context of the present study, the tendency to make false positive identifications of happiness was most predictive of increased perceptions of flirtatiousness for male and female targets. Surprisingly, however, this tendency appeared to be heavily driven by individuals who more frequently mistook some form of negative facial emotion as indicating happiness, as opposed to surprise. One straightforward explanation is that mistaking a negative for a positive facial expression is the larger error, so those liable to perceive negative facial expressions as conferring that the interaction is going smoothly will, to a larger degree, overestimate flirtatious intentions across scenarios.

We also found that the effect of false positives identifications of happiness on overall misjudgments of flirtation were stronger for those who took longer to respond. Contrasting predictions, those responding quickly but erroneously were more similar in their perceptions of flirtation to those with low false positive rates. It was those making errors when the morphed facial expressions became more pronounced that tended to misperceive levels of flirtation. Given the nature of these errors, our effects for happiness false positives errors may be circumscribing individuals with broader deficits in social cognition (e.g., emotional intelligence; Mayer et al., 2001), especially given that happiness is more readily recognized and categorized than other facial expressions for most people (Leppänen & Hietanen, 2004; Wells et al., 2016). Such deficits are indeed associated with performance on emotion recognition tasks (Connolly et al., 2020), as well as implicated in cases of sexual misconduct (Gillespie et al., 2015). If so, the tendency to misidentify negative emotional facial expressions as conveying happiness, and thus misjudge flirtation in dating interactions, may be one mechanism linking these phenomena. This would be consistent with social-information processing models positing that deficits in perceptions of negative facial expressions lead to overestimates of sexual interest, thereby influencing one’s unwanted sexual behavior in response to their inaccurate perceptions of the interaction.

We also found that those with fewer true positive identifications actually tended to perceive greater flirtation across the blind date interactions. Once again, it is plausible that difficulty identifying happiness, in particular, is a proxy for other social cognitive deficits. For example, there is research suggesting that schizotypal symptoms are associated with a tendency to view facial emotions more negatively (Brown & Cohen, 2010), whereas meta-analytic work shows a tendency for those on the Autism spectrum to have difficulty with recognizing happiness specifically (Uljarevic & Hamilton, 2013). Future research should collect self-reported mental health and disability information, a limitation of our study, to be able to assess some of these subgroups who may have difficulties perceiving certain affects. It is alternatively plausible that these difficulties are focally related to emotion recognition; that is, maybe some individuals are not very good at recognizing facially expressed emotions, but their other social faculties are intact. Additional research will be needed to tease out these possibilities.

In any case, why difficulty identifying true positive signals of happiness would be biased towards overestimating flirtation, as currently found, is not entirely clear. One possibility is that a dearth of accurate information regarding the communication of positive affect made the interactions more ambiguous, leaving room for any number of internal factors to have a greater influence in perceptions of flirtation. Research highlights a number of different characteristics that reliably associate with over perceiving sexual intentions, such as sociosexuality (Howell et al., 2012; Lee et al., 2020), particular sexual scripts (Lenton & Bryan, 2005), and so forth. Correspondingly, the inability to accurately identify happiness could act as a modulator, strengthening the influence of other mechanisms that tilt towards overperceptions of sexual intent.

Although we found the typical effect that male participants tended to rate dyadic interactions as having greater sexual content than female participants, the effects for the emotion recognition variables were fairly orthogonal to those found for sex. Therefore, emotion recognition accuracy did not explain the effect for participant sex but did predict flirtatiousness regardless of sex. However, we found that female participants made higher magnitude errors (i.e., inclusive of both over and underestimates) than male participants in their appraisals of flirtatiousness. This finding is inconsistent with some previous work (Farris et al., 2008), but dovetails with others that have highlighted greater misappraisals of sexual interest by female participants in particular situations, such as when examined in multivariate mediation models (Koenig et al., 2007). Men are also shown to be more accurate at detecting flirting when it is present (Hall et al., 2014). Lastly, consistent with previous studies (La France et al., 2009), we found that male participants in general rated both male and female targets higher on flirtatiousness than female participants. This finding aligns with Harnish et al.’s (1990) idea of “sexual schema” which suggests that men view the world through a sexualized lens decoding all cues as sexually relevant regardless of the sex of the target. As a set, these findings suggest that the effect of sex on perceptual errors in flirtatiousness are more complicated and may require alternative design and data analytic approaches to fully explain. For example, future research may want to test if men’s accuracy in detecting flirtation varies based on their base rates of flirtatious interactions due to their “flirt-assume bias”, similar to how some probability models posit in the deception literature (Levine et al., 2006). Additionally, the body of literature on misperceptions of sexual interest has overwhelmingly assessed biological sex differences, however these differences are more likely driven by gender socialization, for example, as opposed to biological sex differences. Therefore, future work in the field should consider both biological sex and gender identity.

Methodological considerations

A strength of the study is that we used ecologically valid blind date interactions as stimuli, which also allowed us to use the interaction partners’ self-reports of flirtatiousness as an anchor point against which to evaluate third-party appraisals. However, the naturalistic approach relinquishes experimental standardization, which can create statistical noise and thus attenuate power, or otherwise overlook existent effects that are detectable only under controlled conditions. This may be particularly relevant to the role of negative affect recognition. Our analyses did not indicate that difficulties in identifying negative affect when present were relevant to perceptions of flirtation in the interactions, but this may reflect that negative affect during these interactions was less frequent. Using standardized stimuli that depict negative affect in response to flirtation may identify that difficulties in detecting negative affect may be pertinent for how these types of interaction are viewed. Additionally, we used four affect categories we deemed most relevant to dating interactions, presuming that other basic emotions such as displays of sadness and fear may be less likely to emerge in these scenarios outside of extreme cases. Nevertheless, further research may consider expanding the set of stimuli to include these, as well as other potentially relevant emotive facial expressions (e.g., compound expression; Du et al., 2014).

We used a performance-based assessment of emotion recognition rather than a subjective self-reported measure and expanded on existing operations using facial morphs that not only better mimic actual facial expression, but also allowed us to consider reaction times as a potentially relevant variable. However, this task had a limited number of trials relative to other studies in this domain of research (e.g., Vanhalst et al., 2017), which could have restricted response variance and to some extent yoked the recognition variables (e.g., misidentifying anger as happiness would increase anger false negatives and happiness false positives). Although these operations were adequate for detecting the effects of happiness recognition, they nevertheless could have inflated Type II error and thus alternatively explain why effects for the remaining emotion recognition variables did not emerge. Additionally, while the use of true and false positives to measure accuracy aligned best with our proposed mechanisms of perceptions of sexual interest, the setup of our emotion task did not clearly map onto other used metrics for detecting emotional accuracy, such as those offered from Signal Detection Theory. Although our hypotheses focused on the type of errors and accuracy in labeling facial emotion stimuli, we were nevertheless unable to discriminate between response biases (i.e., tendency to respond a certain way to a particular stimuli) and sensitivity (i.e., discrimination between signal and noise in stimuli detection), which may be useful in the broader context of flirtation literature (Stanislaw & Todorov, 1999).

Another limitation was our use of NimStim stimuli which may represent more stereotypical and sometimes exaggerated expressions as opposed to the wide range of expressions of an emotion (Barrett, 2011). Stereotyped facial expressions are oversimplified expressions that do not capture the variety in which people move their faces to express emotions, which may account for some of the errors in detection (Barrett et al., 2019). An alternative perspective, however, may be that if individuals cannot identify these commonly agreed upon facial expressions, it may mean that they are even less likely to be able to identify ambiguous facial expressions that are common in the communication of sexual interest. Future research should incorporate recent innovations of measuring emotion perception such as reverse correlation (Jack et al., 2018) or using virtual humans (Rickel et al., 2002), which some researchers recommend for improving emotion research’s reliability and specificity (Barrett et al., 2019), to see if our findings hold true.

Our study - in line with the majority of literature of facial expressions - follows basic emotion theory, which suggests that a small number of categorical facial displays exist (i.e., happiness, sadness, disgust, anger, surprise, and fear) and are universal across many cultural backgrounds (Ekman, 1999; Ekman & Friesen, 1971). Although these perspectives remain influential (Cordaro et al., 2018), there is a growing body of new research suggesting that emotion categories are a conceptual act of categorizing core affects that vary in expression across different cultures and situations (Barrett et al., 2019). This notion may be especially relevant to the detection of surprise, as variability in responding may be better accounted for by inconsistencies in how this emotion is facially expressed, as opposed to individual differences in detection accuracy. Unlike other emotive facial expressions (e.g., happiness), surprise is more ambiguous and some of its initial facial movements (e.g., raised eyebrows) are non-descript and could be confused with other facial expressions (Kosonogov & Titova, 2019). Additional research is needed to determine which theoretical perspective holds true. In any case, it is important to acknowledge that our findings across facial emotions may to some extent be culturally contextualized, and thus may not generalize beyond English speaking western college age students. Additionally, other relevant situations and facial expressions of these emotions should be evaluated to determine how these effects are contextually bound.

There are other limits on the generalizability of our findings. Although we sampled individuals who are in an age range where dating is prevalent, and risk of sexual assault and harassment are high; these phenomena are not isolated to these contexts. Whether or not the effects of facial emotion recognition extend to other venues – particularly those that are not inherently sexual (e.g., the workplace) – is also uncertain. In addition, we used third party evaluations of flirtatiousness, as opposed to appraisals from first-person encounters. Therefore, whether or not the current findings would generalize to actual flirtatious interactions, as opposed to third-party perceptions, is not certain. This critique should be somewhat tempered, however, as findings across methods frequently parallel with the mode of observation acting as a moderator that functions to influence the size of flirtation estimates, rather than the valence (La France et al., 2009; Tomich & Schuster, 1996). Moreover, in addition to facilitating the recruitment of a nationally representative sample; the use of third-party observers provided other benefits. More specifically, the use of multiple interactions allowed us to identify individual proneness to misestimate flirtation across varied interactions, suggesting a general tendency rather than effects tied to a particular instance or standardized interaction. Third-party observations also help rule out a host of other potential causal factors inherent to flirting in naturalistic first-person encounters, such as one’s own motivations for perceiving flirtation (Henningsen, 2004) and perceptions of mate value (Haselton, 2003), that may interface with the processing of facial emotions.

Finally, our analytic approach anchored level of attraction present in the dyadic interaction early in the statistical model by inputting self-reports of the dyad members’ attraction level towards their partner. Using interaction member attraction as representative of the level of attraction present may be problematic because post-hoc introspection of an experience can redirect attention and moderate previous judgements (Wilson et al., 1993; Wilson & Schooler, 1991) and the relative novelty of the dyadic interaction may have led the participants to confound overall arousal for romantic attraction (Dutton & Aron, 1974). It also should be noted, that the targets’ reported level of flirtation was not highly correlated with perceptions of their behavior. It is possible that this was because we showed participants 2-minute clips from the overall 10-minute interactions, so the target’s self report was based on a broader range of time. This is a limitation, as stronger correlations may have emerged had we shown participants the full interaction. However, we presented 2-minute clips that were selected because they encompassed the overall tone of the interaction; with the goal to reduce observer drift, remove parts of the videos that violated participant anonymity, and allowed us to get multiple assessments of each participant without creating fatigue. Regardless, removing this self-perception control did not substantially alter the observed estimates, which suggests that our effects may largely be predicting relatively higher versus lower levels of perceived flirtatiousness without reference to accuracy. Nevertheless, higher versus lower perceptions of flirtation should correspond to be more false positives and negatives, on average, when played out across situations.

Conclusion

Considered overall, our findings indicate that basic emotion recognition of happiness is linked to third-party perceptions of flirtatiousness in both males and females. Although these findings may reflect individual differences across the normative continuum, it is possible they identify those who struggle with discriminating basic emotional expressions. Future research should seek to identify individual differences that account for the detected emotional deficits, as well as other person-level constructs that are relevant to perceptions of flirtatiousness. Our findings lend validation for our novel approach to measuring emotion recognition, which includes response times, and we provide recommendations for expanding this paradigm to include more basic emotions, additional trials, and different stimuli that pull for the relevance of different emotions. Extending this approach to different populations and scenarios will likely prove fruitful for enhancing our understanding of misperceptions of flirtatiousness, as well as how they may relate to sexual misconduct.

Supplemental Material

Supplemental Material - Accurately detecting happy facial expressions associates with perceptions of flirtatiousness

Supplemental Material for Accurately detecting happy facial expressions associates with perceptions of flirtatiousness by Emily S Bibby, Allison M McKinnon, Michael Shaw and Richard E Mattson in Journal of Social and Personal Relationships

Footnotes

Author’s note

Preliminary results using this dataset and related research questions were previously presented at a symposium in an IARR conference.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Open research statement

As part of IARR’s encouragement of open research practices, the authors have provided the following information: This research was not pre-registered.

The data used in the research can be publicly posted. The data can be obtained at: or by emailing: emily.bibby@stonybrook.edu. The materials used in the research can be publicly posted. However, IRB restrictions preclude the dissemination of the videotaped interactions in an open access format. Specific requests for these materials will be considered on a case-by-case basis and in consultation with this institution’s IRB. The available materials can be obtained at: https://osf.io/8bk7f/ or by emailing: emily.bibby@stonybrook.edu.

Supplemental material

Supplement material for this article is available in online.

ORCID iDs

Emily S. Bibby

Richard E. Mattson

References

Abbey

(1982). Sex differences in attributions for friendly behavior: Do males misperceive females' friendliness? Journal of Personality and Social Psychology, 42(5), 830–838. https://doi.org/10.1037/0022-3514.42.5.830

Abbey

Harnish

R. J.

(1995). Perception of sexual intent: The role of gender, alcohol consumption, and rape supportive attitudes. Sex Roles: A Journal of Research, 32(5–6), 297–313. https://doi.org/10.1007/BF01544599

Babcock

J. C.

Green

C. E.

Webb

S. A.

(2008). Decoding deficits of different types of batterers during presentation of facial affect slides. Journal of Family Violence, 23(5), 295–302. https://doi.org/10.1007/s10896-008-9151-1

Barrett

L. F.

(2011). Was Darwin wrong about emotional expressions? Current Directions in Psychological Science, 20(6), 400–406. https://doi.org/10.1177/0963721411429125

Barrett

L. F.

Adolphs

Marsella

Martinez

A. M.

Pollak

S. D.

(2019). Emotional expressions reconsidered: Challenges to inferring emotion from human facial movements. Psychological Science in the Public Interest, 20(1), 1–68. https://doi.org/10.1177/1529100619832930

Baumeister

R. F.

Catanese

K. R.

Vohs

K. D.

(2001). Is there a gender difference in strength of sex drive? Theoretical views, conceptual distinctions, and a review of relevant evidence. Personality and Social Psychology Review, 5(3), 242–273. https://doi.org/10.1207/S15327957PSPR0503_5

Benda

M. S.

Scherf

K. S.

(2020). The complex emotion expression database: A validated stimulus set of trained actors. PLOS ONE, 15(2), Article e0228248. https://doi.org/10.1371/journal.pone.0228248

Brown

L. A.

Cohen

A. S.

(2010). Facial emotion recognition in schizotypy: The role of accuracy and social cognitive bias. Journal of the International Neuropsychological Society: JINS, 16(3), 474–483. https://doi.org/10.1017/s135561771000007x

Cahoon

D. D.

Edmonds

E. M.

(1989). Male-female estimates of opposite-sex first impressions concerning females’ clothing styles. Bulletin of the Psychonomic Society, 27(3), 280–281. https://doi.org/10.3758/BF03334607

10.

Calvo

Esteves

(2005). Detection of emotional faces: Low perceptual threshold and wide attentional span. Visual Cognition, 12(1), 13–27. https://doi.org/10.1080/13506280444000094

11.

Carey

K. B.

Norris

A. L.

Durney

S. E.

Shepardson

R. L.

Carey

M. P.

(2018). Mental health consequences of sexual assault among first-year college women. Journal of American College Health, 66(6), 480–486. https://doi.org/10.1080/07448481.2018.1431915

12.

Cheal

J. L.

Rutherford

M. D.

(2013). Context-dependent categorical perception of surprise. Perception, 42(3), 294–301. https://doi.org/10.1068/p7130

13.

Chen

Garrod

Schyns

Jack

(2015). The face is the mirror of the cultural mind. Journal of Vision, 15(12), 928–928. https://doi.org/10.1167/15.12.928

14.

Cohen

West

S. G.

Aiken

L. S.

(2014). Applied multiple regression/correlation analysis for the behavioral sciences. Psychology press.

15.

Connolly

H. L.

Lefevre

C. E.

Young

A. W.

Lewis

G. J.

(2020). Emotion recognition ability: Evidence for a supramodal factor and its links to social cognition. Cognition, 197(1), Article 104166. https://doi.org/10.1016/j.cognition.2019.104166

16.

Corcoran

K. J.

(1996). The influence of gender, expectancy, and partner beverage selection on (meta)perceptions in a "blind" dating situation. Addictive Behaviors, 21(3), 273–282. https://doi.org/10.1016/0306-4603(95)00061-5

17.

Cordaro

D. T.

Sun

Keltner

Kamble

Huddar

McNeil

(2018). Universals and cultural variations in 22 emotional expressions across five cultures. Emotion, 18(1), 75–93. https://doi.org/10.1037/emo0000302

18.

Crick

N. R.

Dodge

K. A.

(1996). Social information‐processing mechanisms in reactive and proactive aggression. Child Development, 67(3), 993–1002. https://doi.org/10.1111/j.1467-8624.1996.tb01778.x

19.

Tao

Martinez

A. M.

(2014). Compound facial expressions of emotion. Proceedings of the National Academy of Sciences of the United States of America, 111(15), E1454–E1462. https://doi.org/10.1073/pnas.1322355111

20.

Dutton

D. G.

Aron

A. P.

(1974). Some evidence for heightened sexual attraction under conditions of high anxiety. Journal of Personality and Social Psychology, 30(4), 510–517. https://doi.org/10.1037/h0037031

21.

Ekman

(1999). Basic emotions. In Dalgleish

Power

M.J.

(Eds.), Handbook of cognition and emotion (pp. 45–60). John Wiley & Sons.

22.

Ekman

Friesen

W. V.

(1971). Constants across cultures in the face and emotion. Journal of Personality and Social Psychology, 17(2), 124–129. https://doi.org/10.1037/h0030377

23.

Farris

Treat

T. A.

Viken

R. J.

McFall

R. M.

(2008). Perceptual mechanisms that characterize gender differences in decoding women's sexual intent. Psychological Science, 19(4), 348–354. https://doi.org/10.1111/j.1467-9280.2008.02092.x

24.

Farris

Viken

R. J.

Treat

T. A.

(2010). Perceived association between diagnostic and non-diagnostic cues of women's sexual interest: General Recognition Theory predictors of risk for sexual coercion. Journal of Mathematical Psychology, 54(1), 137–149. https://doi.org/10.1016/j.jmp.2008.10.001

25.

Farris

Viken

R. J.

Treat

T. A.

McFall

R. M.

(2006). Heterosocial perceptual organization: Application of the choice model to sexual coercion. Psychological Science, 17(10), 869–875. https://doi.org/10.1111/j.1467-9280.2006.01796.x

26.

Fedina

Holmes

J. L.

Backes

B. L.

(2018). Campus sexual assault: A systematic review of prevalence research from 2000 to 2015. Trauma, Violence, & Abuse, 19(1), 76–93. https://doi.org/10.1177/1524838016631129

27.

Forni-Santos

Osório

F. L.

(2015). Influence of gender in the recognition of basic facial expressions: A critical literature review. World J Psychiatry, 5(3), 342–351. https://doi.org/10.5498/wjp.v5.i3.342

28.

Gagnon

J. H.

Simon

(2017). Sexual conduct: The social sources of human sexuality. Routledge. https://doi.org/10.4324/9781315129242

29.

Gersick

Kurzban

(2014). Covert sexual signaling: Human flirtation and implications for other social species. Evolutionary Psychology, 12(3), 549–569. https://doi.org/10.1177/147470491401200305

30.

Gillespie

S. M.

Rotshtein

Satherley

R.-M.

Beech

A. R.

Mitchell

I. J.

(2015). Emotional expression recognition and attribution bias among sexual and violent offenders: A signal detection analysis. Frontiers in Psychology, 6, 595–595. https://doi.org/10.3389/fpsyg.2015.00595

31.

Haj-Mohamadi

Gillath

Rosenberg

E. L.

(2021). Identifying a facial expression of flirtation and its effect on men. Journal of Sex Research, 58(2), 137–145. https://doi.org/10.1080/00224499.2020.1805583

32.

Hall

J. A.

Xing

Brooks

(2014). Accurately detecting flirting: Error management theory, the traditional sexual script, and flirting base rate. Communication Research, 42(7), 939–958. https://doi.org/10.1177/0093650214534972

33.

Harnish

R. J.

Abbey

DeBono

K. G.

(1990). Toward an understanding of “the sex game”: The effects of gender and self‐monitoring on perceptions of sexuality and likability in initial interactions 1. Journal of Applied Social Psychology, 20(16), 1333–1344. https://doi.org/10.1111/j.1559-1816.1990.tb01475.x

34.

Haselton

M. G.

(2003). The sexual overperception bias: Evidence of a systematic bias in men from a survey of naturally occurring events. Journal of Research in Personality, 37(1), 34–47. https://doi.org/10.1016/S0092-6566(02)00529-9

35.

Haselton

M. G.

Buss

D. M.

(2000). Error management theory: A new perspective on biases in cross-sex mind reading. Journal of Personality and Social Psychology, 78(1), 81–91. https://doi.org/10.1037/0022-3514.78.1.81

36.

Henningsen

D. D.

(2004). Flirting with meaning: An examination of miscommunication in flirting interactions. Sex Roles: A Journal of Research, 50(7-8), 481–489. https://doi.org/10.1023/B:SERS.0000023068.49352.4b

37.

Herrmann

Call

Hernández-Lloreda

Hare

Tomasello

(2007). Humans have evolved specialized skills of social cognition: The cultural intelligence hypothesis. Science (New York, N.Y.), 317(5843), 1360–1366. https://doi.org/10.1126/science.1146282

38.

Howell

E. C.

Etchells

P. J.

Penton-Voak

I. S.

(2012). The sexual overperception bias is associated with sociosexuality. Personality and Individual Differences, 53(8), 1012–1016. https://doi.org/10.1016/j.paid.2012.07.024

39.

Jack

R. E.

Crivelli

Wheatley

(2018). Data-driven methods to diversify knowledge of human psychology. Trends in Cognitive Sciences, 22(1), 1–5. https://doi.org/10.1016/j.tics.2017.10.002

40.

Koenig

B. L.

Kirkpatrick

L. A.

Ketelaar

(2007). Misperception of sexual and romantic interests in opposite-sex friendships: Four hypotheses. Personal Relationships, 14(3), 411–429. https://doi.org/10.1111/j.1475-6811.2007.00163.x

41.

Kosonogov

Titova

(2019). Recognition of all basic emotions varies in accuracy and reaction time: A new verbal method of measurement. International Journal of Psychology, 54(5), 582–588. https://doi.org/10.1002/ijop.12512

42.

Koukounas

Letch

N. M.

(2001). Psychological correlates of perception of sexual intent in women. Journal of Social Psychology, 141(4), 443–456. https://doi.org/10.1080/00224540109600564

43.

Krumhuber

Manstead

Kappas

(2007). Temporal aspects of facial displays in person and expression perception: The effects of smile dynamics, head-tilt, and gender. Journal of Nonverbal Behavior, 31(1), 39–56. https://doi.org/10.1007/s10919-006-0019-x

44.

La France

B. H.

Henningsen

D. D.

Oates

Shaw

C. M.

(2009). Social-sexual interactions? Meta-analyses of sex differences in perceptions of flirtatiousness, seductiveness, and promiscuousness. Communication Monographs, 76(3), 263–285. https://doi.org/10.1080/03637750903074701

45.

Lee

A. J.

Sidari

M. J.

Murphy

S. C.

Sherlock

J. M.

Zietsch

B. P.

(2020). Sex differences in misperceptions of sexual interest can Be explained by sociosexual orientation and men projecting their own interest onto women. Psychological Science, 31(2), 184–192. https://doi.org/10.1177/0956797619900315

46.

Lenton

A. P.

Bryan

(2005). An affair to remember: The role of sexual scripts in perceptions of sexual intent. Personal Relationships, 12(4), 483–498. https://doi.org/10.1111/j.1475-6811.2005.00127.x

47.

Leppänen

J. M.

Hietanen

J. K.

(2004). Positive facial expressions are recognized faster than negative facial expressions, but why? Psychological Research, 69(1-2), 22–29. https://doi.org/10.1007/s00426-003-0157-2

48.

Levenson

Kiehl

Fitzpatrick

(1995). Assessing psychopathic attributes in a noninstitutionalized population. Journal of Personality and Social Psychology, 68(1), 151–158. https://doi.org/10.1037//0022-3514.68.1.151

49.

Levine

T. R.

Kim

R. K.

Sun Park

Hughes

(2006). Deception detection accuracy is a predictable linear function of message veracity base-rate: A formal test of park and levine’s probability model. Communication Monographs, 73(3), 243–260. https://doi.org/10.1080/03637750600873736

50.

Malamuth

N. M.

Sockloskie

R. J.

Koss

M. P.

Tanaka

J. S.

(1991). Characteristics of aggressors against women: Testing a model using a national sample of college students. Journal of Consulting and Clinical Psychology, 59(5), 670–681. https://doi.org/10.1037/0022-006X.59.5.670

51.

Matsumoto

Keltner

Shiota

M. N.

O'Sullivan

Frank

(2008). Facial expressions of emotion. In Lewis

Haviland-Jones

J. M.

Barrett

L. F.

(Eds.), Handbook of emotions (pp. 211–234). The Guilford Press.

52.

Mattson

R. E.

Parker

M. M.

McKinnon

A. M.

Massey

S. G.

Merriwether

A. M.

Hardesty

Young

S. R.

(2022). Just say yes? A cluster analytic approach to evaluating college students’ understandings of affirmative consent. Archives of Sexual Behavior, 51(5–6), 3197–3209. https://doi.org/10.1007/s10508-021-02217-z

53.

Mayer

J. D.

Salovey

Caruso

D. R.

Sitarenios

(2001). Emotional intelligence as a standard intelligence. Emotion, 1(3), 232–242. https://doi.org/10.1037/1528-3542.1.3.232

54.

McBain

Norton

Chen

(2009). Females excel at basic face perception. Acta Psychologica, 130(2), 168–173. https://doi.org/10.1016/j.actpsy.2008.12.005

55.

Montepare

Dobish

(2003). The contribution of emotion perceptions and their overgeneralizations to trait impressions. Journal of Nonverbal Behavior, 27(4), 237–254. https://doi.org/10.1023/A:1027332800296

56.

Moore

M. M.

(2010). Human nonverbal courtship behavior--a brief historical review. Journal of Sex Research, 47(2), 171–180. https://doi.org/10.1080/00224490903402520

57.

Muehlenhard

C. L.

Humphreys

T. P.

Jozkowski

K. N.

Peterson

Z. D.

(2016). The complexities of sexual consent among college students: A conceptual and empirical review. The Journal of Sex Research, 53(4-5), 457–487. https://doi.org/10.1080/00224499.2016.1146651

58.

Newstrom

N. P.

Harris

S. M.

Miner

M. H.

(2021). Sexual consent: How relationships, gender, and sexual self-disclosure affect signaling and interpreting cues for sexual consent in a hypothetical heterosexual sexual situation. Sex Roles, 84(7), 454–464. https://doi.org/10.1007/s11199-020-01178-2

59.

Perilloux

Easton

J. A.

Buss

D. M.

(2012). The misperception of sexual interest. Psychological Science, 23(2), 146–151. https://doi.org/10.1177/0956797611424162

60.

Ranganath

Jurafsky

McFarland

D. A.

(2009, August). It’s not you, it’s me: Detecting flirting and its misperception in speed-dates. In Proceedings of the 2009 conference on empirical methods in natural language processing (pp. 334–342). Suntec City.

61.

Rickel

Marsella

Gratch

Hill

Traum

Swartout

(2002). Toward a new generation of virtual humans for interactive experiences. IEEE Intelligent Systems, 17(4), 32–38. https://doi.org/10.1109/MIS.2002.1024750

62.

Rozin

Cohen

A. B.

(2003). High frequency of facial expressions corresponding to confusion, concentration, and worry in an analysis of naturally occurring facial expressions of Americans. Emotion, 3(1), 68–75. https://doi.org/10.1037/1528-3542.3.1.68

63.

Schafer

J. L.

Graham

J. W.

(2002). Missing data: Our view of the state of the art. Psychological Methods, 7(2), 147–177. https://doi.org/10.1037/1082-989X.7.2.147

64.

Schmidt

K. L.

Cohn

J. F.

(2001). Human facial expressions as adaptations: Evolutionary questions in facial expression research. American Journal of Physical Anthropology, 116(S33), 3–24. https://doi.org/10.1002/ajpa.20001

65.

Sinozich

Langton

(2014). Rape and sexual assault victimization among college-age females, 1995-2013. US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.

66.

Smith

S. G.

Zhang

Basile

K. C.

Merrick

M. T.

Wang

Kresnow

Chen

(2018). The national intimate partner and sexual violence survey: 2015 data brief – updated release. CDC.

67.

Spence

J. T.

Helmreich

Stapp

(1973). A short version of the attitudes toward women scale (AWS). Bulletin of the Psychonomic Society, 2(4), 219–220. https://doi.org/10.3758/BF03329252

68.

Stanislaw

Todorov

(1999). Calculation of signal detection theory measures. Behavior Research Methods, Instruments, & Computers, 31(1), 137–149. https://doi.org/10.3758/BF03207704

69.

Tomich

P. L.

Schuster

P. M.

(1996). Gender differences in the perception of sexuality: Methodological considerations. Sex Roles, 34(11), 865–874. https://doi.org/10.1007/BF01544320

70.

Tottenham

Tanaka

J. W.

Leon

A. C.

McCarry

Nurse

Hare

T. A.

Marcus

D. J.

Westerlund

Casey

B. J.

Nelson

(2009). The NimStim set of facial expressions: Judgments from untrained research participants. Psychiatry Research, 168(3), 242–249. https://doi.org/10.1016/j.psychres.2008.05.006

71.

Treat

T. A.

Church

E. K.

Viken

R. J.

(2017). Effects of gender, rape-supportive attitudes, and explicit instruction on perceptions of women’s momentary sexual interest. Psychonomic Bulletin & Review, 24(3), 979–986. https://doi.org/10.3758/s13423-016-1176-5

72.

Treat

T. A.

McFall

R. M.

Viken

R. J.

Kruschke

J. K.

(2001). Using cognitive science methods to assess the role of social information processing in sexually coercive behavior. Psychological Assessment, 13(4), 549–565. https://doi.org/10.1037/1040-3590.13.4.549

73.

Treat

T. A.

McMurray

Betty

J. R.

Viken

R. J.

(2019). Tracking men’s perceptions of women’s sexual interest. Current Directions in Psychological Science, 29(1), 71–79. https://doi.org/10.1177/0963721419884322

74.

Uljarevic

Hamilton

(2013). Recognition of emotions in autism: A formal meta-analysis. J Autism Dev Disord, 43(7), 1517–1526. https://doi.org/10.1007/s10803-012-1695-5

75.

Vanhalst

Gibb

B. E.

Prinstein

M. J.

(2017). Lonely adolescents exhibit heightened sensitivity for facial cues of emotion. Cognition & Emotion, 31(2), 377–383. https://doi.org/10.1080/02699931.2015.1092420

76.

Wells

L. J.

Gillespie

S. M.

Rotshtein

(2016). Identification of emotional facial expressions: Effects of expression, intensity, and sex on eye gaze. Plos One, 11(12), Article e0168307. https://doi.org/10.1371/journal.pone.0168307

77.

Wilson

T. D.

Lisle

D. J.

Schooler

J. W.

Hodges

S. D.

Klaaren

K. J.

LaFleur

S. J.

(1993). Introspecting about reasons can reduce post-choice satisfaction. Personality & Social Psychology Bulletin, 19(3), 331–339. https://doi.org/10.1177/0146167293193010

78.

Wilson

T. D.

Schooler

J. W.

(1991). Thinking too much: Introspection can reduce the quality of preferences and decisions. Journal of Personality and Social Psychology, 60(2), 181–192. https://doi.org/10.1037/0022-3514.60.2.181

79.

Zsido

A. N.

Arato

Ihasz

Basler

Matuz-Budai

Inhof

Schacht

Labadi

Coelho

C. M.

(2021). “finding an emotional face” revisited: Differences in own-age bias and the happiness superiority effect in children and young adults. Frontiers in Psychology, 12, 580565. https://doi.org/10.3389/fpsyg.2021.580565.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.10 MB