Abstract
Stereotypes have important social consequences, such as promoting female discrimination in the workplace, which depends on how women are categorized. Extending prior work, here we analyze how two important female subgroups, women who are categorized as professional or sexy women, are evaluated on key dimensions of stereotype content (morality, sociability, and competence), positive and negative emotions, and facilitation behavioral tendencies (active and passive). To this end, we adapted a previous scale of facilitation tendencies to the working environment. Furthermore, we aim to explore the mechanism involved in carrying out helping behaviors towards each subgroup of women in the workplace. In order to fulfill these goals, 201 participants (Mage = 28.88, SD = 12.25; 66.2% women) were randomly assigned to evaluate a woman categorized as either sexy or professional on the mentioned variables. Results show that women categorized as sexy are devalued compared to those categorized as professionals. We also found that competence has a driving role in predicting more active facilitation tendencies towards a woman categorized as professional than a woman categorized as sexy via positive emotions. These findings have implications for the career development of women.
Our evaluations of others are often influenced by the stereotypes ascribed to social groups and categories. Stereotypes, in turn, influence the emotions we experience towards members of these groups and, ultimately, our willingness to help them. Indeed, a growing body of work has shown that stereotypes have important social consequences, such as reinforcing prejudiced attitudes or promoting discrimination (see Dovidio & Gaertner, 2010).
The employment discrimination faced by women in most societies is undeniable (e.g., Adachi, 2013; Cuadrado et al., 2015). However, the way they are perceived and categorized influences the stereotypes, emotions, and behavioral tendencies towards them. Importantly, it has been shown that perceivers process group-related information at multiple levels, ranging from the broader (i.e., overall category) to more specific (i.e., subgroups) levels (Richards & Hewstone, 2001). Indeed, stereotypes ascribed to the overall category might be distinct to those ascribed to the nested subgroups (Brewer et al., 1981; Richards & Hewstone, 2001). For instance, perceivers may hold a global image of women, and several differentiated representations of specific women subgroups, such as homemaker, professional woman, and sexy woman (Altermatt et al., 2003; Deaux et al., 1985; Eckes, 2002; Glick et al., 2005).
Extending prior work, the present research aims to analyze how two important female subgroups, professional women and sexy women, are evaluated on key dimensions of stereotype content, emotions, and behavioral tendencies. Furthermore, we aim to explore the mechanism involved in carrying out helping behaviors towards both types of women in the workplace.
Stereotypes, Emotions, and Behaviors Towards Professional and Sexy Women Subgroups
Over the past few decades, numerous studies have shown the existence of two fundamental content dimensions which underlie person and group impressions (see Abele & Wojciszke, 2019). According to the stereotype content model (SCM; Fiske et al., 2002), these dimensions are called warmth and competence. Warmth reflects the intentions of different groups and its members towards us (e.g., they are sincere and friendly), while competence refers to whether they are capable of achieving their intentions (e.g., they are skillful and intelligent). Research has shown that information about warmth is more predictive of affective and behavioral reactions (see Cuddy et al., 2008).
In this vein, Leach et al. (2007) suggested that warmth might be defined by two distinct facets, morality and sociability. According to Brambilla and Leach (2014, p. 398), “sociability pertains to being benevolent to people in ways that facilitate affective relations with them (e.g., friendliness, likeability, kindness), morality refers to being benevolent to people in ways that facilitate correct and principled relations with them (e.g., honesty, trustworthiness, sincerity).” A number of studies have demonstrated that perceived morality plays a primary role over sociability and competence on person and group perception (Cuadrado et al., 2020; Riva et al., 2016; for a review, Brambilla et al., 2021).
Although warmth judgments are made more quickly than competence judgments, sometimes competence takes primacy and represents the most salient dimension in working environments (see Cuddy et al., 2011). We aimed at confirming this primacy of competence (versus other stereotype dimensions) when predicting facilitation behavioral tendencies towards two distinct subgroups of women (women categorized as sexy and women categorized as professionals).
Previous research has showed that sexy women are evaluated with a lack of competence-related traits (Altermatt et al., 2003; Deaux et al., 1985; Six & Eckes, 1991), whereas professional women are evaluated as high in competence (Eckes, 2002; Fiske et al., 2002).
Glick et al. (2005) revealed that women managers dressed in a sexy manner were perceived as less competent and elicited fewer positive and more negative emotions than women managers dressed in a neutral manner. Wookey et al. (2009) also showed that high-status sexually dressed women received lower ratings in competence than professionally dressed women.
To our knowledge, only one study (Cuadrado & López-Turrillo, 2014) has compared the evaluations towards different female subgroups on the three stereotype dimensions (morality, sociability, and competence), as well as on positive and negative emotions. That research was carried out with an adolescent population. The authors found that sexy women received worse evaluations than professional women: professional women were perceived as more competent, moral, and sociable than sexy women. Their results also showed that competence was the dimension most attributed to professional women, while morality was the dimension least attributed to sexy women. Therefore, it can be concluded that sexy women are devalued compared to professional women on all three stereotypical dimensions, and that sexy women are characterized as having low morality, and professional women as being very competent. As a consequence, adolescents reported feeling more positive and fewer negative emotions towards professional women than towards sexy women.
Although Heflick et al. (2011) did not include subgroups of women as targets, they offered results that supported the devaluation of sexy women on all three stereotypical dimensions. These authors found that when the participants were instructed to pay attention to the physical appearance (vs. performance) of both men and women, women (but not men) were perceived as being less competent, warm, and moral. Focusing solely on morality, and in the domain of intimate partner violence, Pacilli et al. (2017) found that participants evaluated a fictitious victim of gender violence as less moral when she presented a sexualized appearance that when their appearance was non-sexualized. These authors also found a tendency of participants to express fewer helping intentions towards the sexualized compared to the non-sexualized victim. No research has directly examined how evaluations of women who are categorized as professional or sexy women affect participants’ helping intentions towards them in the work context. The present work fills this gap.
The Stereotypes-Emotions-Behaviors Link
Stereotypes have profound effects on our emotions and behaviors towards others. From the SCM (Fiske et al., 2002) and other perspectives (e.g., intergroup emotion theory), emotions derive from cognitive appraisals such as stereotypes (see Mackie & Smith, 2018, for a review). For example, in the workplace, the competence ascribed to a worker can easily elicit positive emotions towards him/her.
The literature has consistently revealed that emotions more strongly and directly relate to behavior than cognitions (e.g., Cuddy et al., 2007; Talaska et al., 2008), and that emotions mediate the stereotype-behavioral tendencies relationships (Cuddy et al., 2007; Mackie et al., 2000). As such, Cuddy et al. (2007) confirmed that positive emotions lead to facilitation behaviors. Facilitation tendencies lead to favorable outcomes or gains for groups and its members (Cuddy et al., 2007, 2008). That is, people tend to act for the benefit of others when they experience positive emotions towards them. By contrast, people are less willing to help or cooperate with others when they experience negative emotions towards them.
Active-passive is a frequent distinction that captures a wide range of behavioral tendencies. Active behaviors are direct, explicit, overt, intense, and carried out with directed effort to influence the target. By contrasts, passive behaviors are indirect, covert, less intense, require less directed effort, and reflect a less deliberate intention, but have repercussions for the target (Ayduk et al., 2003; Cuddy et al., 2007, 2008). Therefore, active facilitation explicitly aims to benefit a target (e.g., helping), while passive facilitation implies convenient cooperation with a target (e.g., acting with the target for one’s own purposes, simultaneously benefiting the target as a by-product) (Cuddy et al., 2007, 2008). These behavioral tendencies have traditionally been measured by using action verbs (e.g., help, cooperate with). In the present research, we will adapt this measure to the working context.
Although the analysis of the stereotypes-emotions-behaviors link is consistent in the literature, studies have not considered exploring this relation comparing evaluations towards subgroups of women. The present research will examine it by considering two relevant subgroups: sexy and professional women.
The Present Research
To the best of our knowledge, no studies considering the differences between sexy and professional women subgroups on the three stereotype dimensions (morality, sociability, and competence), positive and negative emotions, and behavioral tendencies in the work context have yet been conducted. This study aims to fill these gaps.
The main goal of the present research is to determine if women who are categorized as professional or sexy women are differently evaluated in stereotypes, emotions, and facilitation tendencies at work. Thus, based on previous research, we predict that
Women categorized as sexy women would be perceived more negatively than women categorized as professional women. They would be perceived as less moral, sociable, and competent than professional women.
Women categorized as professional women would be perceived as more competent than moral and sociable. Women categorized as sexy women would be perceived as less moral than sociable and competent.
Participants would experience fewer positive and more negative emotions towards women categorized as sexy women compared to women categorized as professional women.
Participants would be less willing to help or facilitate at work women categorized as sexy women compared to women categorized as professional women.
1
Since the research has shown that the level of hostile and benevolent sexism of the participants is differently associated to the subgroup of women to whom it is targeted (e.g., Becker, 2010; Fowers & Fowers, 2010; Glick et al., 1997; Sibley & Wilson, 2004), we will also test whether the sexism of the participants moderates the hypothesized effects of the subgroup of woman on stereotypes, emotions, and facilitation tendencies. We will also explore whether the sex and age of the participants moderate such effects. Additionally, we will explore if the subgroup of woman could have an indirect effect on facilitation behavioral tendencies through stereotypes and emotions. In other words, we will test whether there are differences between women who are categorized as professional or sexy women when predicting facilitation tendencies at work via stereotypes and emotions. In order to achieve our aims, we adapt the scale of facilitation tendencies (Cuddy et al., 2007) to the working context, and we analyze its psychometric properties (descriptive statistics, evidence of validity based on internal structure, estimation of reliability, and evidence of validity based on their relations with other variables—stereotypes, emotions, perceived status and competition of the target, and participants’ sexism).
Method
Pretest Study
In order to conduct our study, we used photos, which were selected from a major collection based on the representativeness of the specific category of woman. Initially, we obtained several photos through Google image searches using the keywords “professional woman” and “sexy woman.” Later, four expert judges classified these photographs of unknown (non-famous) women into one of two categories of interest: professional woman or sexy woman. Subsequently, they selected the 12 photographs (six for each category) more representative of the corresponding category.
Finally, 30 undergraduates (15 males and 15 females; Mage = 20.67, SD = 3.63) classified the 12 photographs into one of the two categories, and then estimated on a scale of 1 (not at all) to 5 (very much) the degree of representativeness of each photo in the corresponding category. Based on these results, we selected one photo from each category: those that were correctly classified by 100% of the participants. The participants considered these photographs to be fairly representative of their category: sexy woman (M = 4.43, SD = 0.50), and professional woman (M = 4.47, SD = 0.63).
Participants
Two-hundred and nine people volunteered to participate in this study. The study was designed in an online platform for online forms. We followed a convenience sampling and recruited participants online by publishing the study in social networks. Participants were adequately informed of their rights and provided their informed consent. Their participation was voluntary, anonymous, and confidential. The study was approved by the authors’ University Ethics Committee and was accordance with Ethical Guidelines for Research. According to pre-established criteria, we excluded eight participants (two under the age of 18, one who did not provide their informed consent, and five who failed the manipulation check—who did not correctly identify the subgroup of woman). The final sample consisted of 201 participants (66.2% women) who ranged from 18 to 67 years old (Mage = 28.88, SD = 12.25). Most of them had Spanish nationality (97%). We conducted a sensitivity analysis using G*Power 3 program (Faul et al., 2009) to determine the effect size that the current study could detect. With a sample size of 201, an error probability of .05, and a power of .80, the minimum effect size that we could detect for an ANOVA (repeated measures, within-between interaction) with two groups and three measures was ƒ(U) = .156 (ηp2 = .024), and for two measures was ƒ(U) = .199 (ηp2 = .038). That is, lower values of ηp2 than these, would indicate that we must be cautious with the results obtained. In our study, all values of ηp2 of the statistically significant results were over ηp2 = .024.
Design
Participants were redirected to two different conditions following an intergroup design: professional versus sexy woman targets. In each condition, participants were presented with a picture depicting either a professional woman (102 participants, 50.7% of the sample) or a sexy woman (99 participants, 49.3% of the sample). Both pictures showed half body photos of women looking directly to the camera.
The women in the photographs were similar in age and hair color, but varied in their dressing and posture, and included some clues to capture the cognitive schema associated with each category. The sexy woman wore a fitted black dress with a low neckline, whereas the professional woman wore a white striped t-shirt and a fitted black cardigan. The professional woman wore an architect hat and held a blueprint and a pencil in her hands. Photos are not shown due to image copyright, but they will be available from the corresponding author upon request.
The sex of the participants was distributed equally across conditions, χ2 (1) = 0.55, p = .457, as well as their age, t (199) = 1.09, p = .276, and their level of hostile sexism, t (199) = 0.29, p = .775, and benevolent sexism, t (199) = 0.78, p = .438.
Measures
After being exposed to the specific target (professional woman versus sexy woman), participants reported their perceptions, emotions, facilitation tendencies at work, and perceived status and competition of the target woman. Unless stated otherwise, all measures used a five-point response scale ranging from 1 (nothing) to 5 (very much). At the end, participants filled out a questionnaire of sexism and reported sociodemographic information (sex, age, and nationality).
Manipulation Check
Participants were asked to indicate the category that best fit the subgroup of woman depicted in the photo: professional woman or sexy woman. Those who failed to correctly identify the subgroup of woman to evaluate were excluded from the analyses.
Stereotypes
This variable was assessed with nine items extracted from Leach et al. (2007) and adapted to Spanish by López-Rodríguez et al. (2013, 2014). Participants were asked to think about that woman and indicate to what extent they thought that the following characteristics described her: honest, sincere, and trustworthy (morality, α = .92); kind, friendly, and warm (sociability, α = .83); competent, intelligent, and skillful (competence, α = .89). The items were randomly presented, and they were averaged so that higher scores indicate a higher perception of morality, sociability, and competence.
Emotions
We used 20 items extracted from Fiske et al. (2002) and adapted to Spanish by Cuadrado et al. (2016). The participants were asked to think about that woman and indicate the extent they feel or have felt the following emotions about her (randomly presented): admiration, respect, understanding, security, comfort, fondness, pride, and inspiration (eight items; positive emotions; α = .87); contempt, disappointment, disgust, tension, fear, anxiety, anger, resentment, shame, hate, unease, and frustration (12 items; negative emotions; α = .93). Items were averaged so that higher scores indicate higher positive and negative emotions. To make sure that positive and negative emotions items were clearly distinguishable, we conducted two factorial analyses including the 20 items (one for each subgroup of woman) and fixing the number of factors to two. Principal components analysis with varimax rotation (to avoid the factors to covary) extracted for the professional woman two factors with eigenvalues larger than one that explained 57.85% of variance. Positive emotions (loadings 0.76–0.25) and negative emotions items (loadings 0.93–0.70) loaded on the separate factors. In the case of sexy women, the two factors explained 49.74% of variance. Positive emotions (loadings 0.78–0.55) and negative emotions items (loadings 0.75–0.50) loaded on the separate factors. We conducted the same analyses with the total sample: the two factors explained 56.40% of variance. Positive emotions (loadings 0.84–0.45) and negative emotions items (loadings 0.85–0.69) loaded on the separate factors.
Facilitation Tendencies at Work
The participants were asked to report the extent to which they would be willing to carry out the following actions towards a woman like her: recommending her for a job position, facilitating her professional training, promoting her at work, facilitating her promotion at work, if possible (active facilitation); working with her on a team project, cooperating with her at work, carpooling to go to work in order to reduce expenses, and partnering with her professionally (passive facilitation). Items were averaged so that higher scores indicate higher willingness to facilitate women at work actively or passively. This scale was adapted to the working context for this study (psychometric properties of this scale are analyzed in the section of preliminary analysis).
Perceived Status
Participants also reported the perceived status of the target woman with three items adapted by Cuadrado and López-Turrillo (2014) from Fiske et al. (2002): (1) To what extent are the jobs usually held by women like her prestigious? (2) In general, how economically successful are women like her? and (3) What educational level do women like her have? The participants responded using a 5-point Likert scale with alternatives ranging from 1 (not at all) to 5 (a lot) in the first two items, and from 1 (very low) to 5 (very high) in the last item. The Cronbach’s alpha of this measurement was .39. This variable was not used for the analyses because of its low reliability.
Perceived Competition
Participants also reported the perceived competition with three items adapted by Cuadrado and López-Turrillo (2014) from Fiske et al. (2002): (1) Women like her have privileges that make things more difficult for people like me, (2) The more power women like her have, the less power people like me have, and (3) The resources earmarked for women like her are resources that are taken away from people like me (α = .84). The participants had to indicate their degree of agreement with each item using a 5-point Likert scale (1 = totally disagree, 5= totally agree).
Ambivalent Sexism
Ambivalent sexism was measured using the Ambivalent Sexism Inventory (Glick & Fiske, 1996; Spanish version by Expósito et al., 1998), which showed acceptable reliability for both scales: hostile sexism (11 items: α = .88) and benevolent sexism (11 items: α = .78). Items were averaged so that higher scores indicate stronger endorsement of sexism.
Results
Manipulation Check
Most participants correctly identify the subgroup of woman showed in the photo (102; 98% for the professional woman condition; 99; 97% for the sexy woman condition).
Preliminary Analysis: Psychometric Properties of the Scale of Facilitation Tendencies at Work
Descriptive Statistics for the Items of Facilitation Tendencies
Note. Scores ranged from 1 (nothing) to 5 (very much).
In the active facilitation factor, modification indices recommended co-variating the errors of the item “recommending her for a job position” and the item “promoting her at work” in the sexy woman sample. The two-factor model yielded an adequate fit for the professional woman sample, χ2 (19) = 29.488, p = .059, CFI = .962, TLI = .943, RMSEA = .074, SRMR = .050, and an acceptable fit in the sexy woman sample, χ2 (18) = 28.250, p = .058, CFI = .970, TLI = .953, RMSEA = .076, SRMR = .046. All factor loadings were statistically significant and over .56.
Internal consistency of each factor in each subsample was also adequate. Active Facilitation: α = .75/ω = .75 (professional woman) and α = .78/ω = .78 (sexy woman); Passive Facilitation: α = .77/ω = .78 (professional woman) and α = .83/ω = .84 (sexy woman).
Descriptive Statistic and Bivariate Correlations Between Variables for Sexy and Professional Women Targets.
Note. Descriptive statistic and bivariate correlations between variables in the group that evaluated a sexy woman (above diagonal, n = 99) and a professional woman (below diagonal, n = 102). *p < .05; ** p < .01.
Stereotypes of Women Categorized as Professional and Sexy Women
We conducted a mixed-MANOVA to test the hypotheses that participants can ascribe different attributes to a professional and a sexy woman, with subgroup of woman as a between-factor variable, and stereotype dimensions (i.e., perceived morality, sociability, and competence) as a within-factor variable. The multivariate effect of subgroup of woman on stereotypes was significant: Wilk’s λ = .55, F (2, 198) = 81.05, p < .001, η2p = .450. Univariate analyses confirmed that a sexy woman received worse evaluations than a professional woman: F (1,199) = 72.02, p < .001, η2p = .266. As shown in Figure 1, the sexy woman was perceived as less moral, F (1,199) = 50.74, p < .001, η2
p
= .203, less competent, F (1,199) = 163.00, p < .001, η2
p
= .450, and even less sociable, F (1,199) = 8.94, p = .003, η2
p
= .043, compared to the evaluations received by a professional woman. The multivariate analysis also revealed a two-way interaction effect between subgroup of woman and stereotype dimensions: Wilk’s λ = .69, F (2, 198) = 44.29, p < .001, η2
p
= .309. Stereotypes, emotions, and facilitation tendencies at work for sexy and professional women. 
A closer inspection of the within-group differences confirmed that the pattern of stereotypes ascribed for a professional woman and a sexy woman was clearly different. Pairwise comparisons with Bonferroni tests showed that the professional woman was perceived as more competent than moral and sociable (ps < .001), and more moral than sociable (p < .001). However, the sexy woman was perceived as less moral than sociable and competent (ps < .001), and as sociable as competent (p = .209), which have medium values. To summarize, while a professional woman was perceived as highly competent, a little less moral, and much less sociable (see darker bars in Figure 1: Stereotypes), a sexy woman was considered less moral than sociable and competent (see brighter bars in Figure 1: Stereotypes).
Emotions Towards Women Categorized as Professional and Sexy Women
We conducted a different mixed-MANOVA to test the hypotheses that participants can feel different emotions towards professional and sexy women, with the subgroup of woman as a between-factor variable, and positive and negative emotions as within-factor variable. The multivariate effect of subgroup of woman on emotions was significant: Wilk’s λ = .26, F (1, 199) = 556.20, p < .001, η2p = .736. The analysis also yielded a two-way interaction effect between subgroup of woman and emotions: Wilk’s λ = .67, F (1, 199) = 100.42, p < .001, η2 p = .335. The univariate analysis on emotions yielded a significant effect of subgroup of woman: F (1,199) = 70.66, p < .001, η2p = .262. As shown in Figure 1, participants felt less positive emotions towards a sexy woman compared to a professional woman: F (1,199) = 131.57, p < .001, η2 p = .398. However, as previously advanced by the interaction effect, there were no differences in negative emotions between both subgroups of women: F (1,199) = 1.31, p = .254, η2 p = .007.
Facilitation Tendencies at Work Towards Women Categorized as Professional and Sexy Women
We conducted a third mixed-MANOVA with subgroup of woman as a between-factor and active, and passive facilitation tendencies at work as within-factors. The multivariate effect of subgroup of woman on facilitation intentions was significant: Wilk’s λ = .78, F (1, 199) = 55.21, p < .001, η2p = .217. The univariate analysis on facilitation tendencies yielded a significant effect of subgroup of woman: F (1,199) = 66.08, p < .001, η2p = .249. As shown in Figure 1, participants were less willing to initiate active facilitation actions, F (1,199) = 77.53, p < .001, η2p = .280, and passive facilitation actions, F (1,199) = 38.08, p < .001, η2p = .161, towards a sexy woman than towards a professional woman. The analysis revealed a two-way interaction effect between subgroup of woman and facilitation tendencies: Wilk’s λ = .95, F (1, 199) = 10.50, p = .001, η2 p = .050. Although participants always tended to more facilitation towards a professional woman than towards a sexy woman, these differences were more acute for active tendencies, such as recommending her for a job position than for passive tendencies, such as working with her in a team project (see F values and effect sizes for univariate analyses).
The Moderating Role of the Level of Ambivalent Sexism, Sex, and Age of the Participants
In order to test if hostile or benevolent sexism moderated the effect of the subgroup of woman on stereotypes, emotions, and facilitation tendencies, we conducted 14 regression analyses with subgroup of woman as predictor—one for each dependent variable: morality, sociability, competence, positive emotions, negative emotions, active facilitation, passive facilitation; and for each moderator variable: participants’ hostile sexism, benevolent sexism—using the Model 1 of the macro PROCCESS 3.0 developed by Hayes (2018). Hostile sexism only moderated the effect of subgroup of woman on negative emotions, B = 0.20, SE = .09, t = 2.13, R2 = .04, ΔR2 = .02, p = .034. The conditional effects at different levels of hostile sexism revealed that high hostile sexists felt more negative emotions towards women categorized as sexy women than towards those categorized as professional women, B = 0.24, SE = .10, t = 2.31, p = .022. However, no differences were found for those with medium or low hostile sexism, p > .263. Benevolent sexism did not moderate the effects of the subgroup of woman on any variable, p > .059.
We conducted the same analyses in order to test if the age of the participant moderated the effect of subgroup of woman on stereotypes, emotions, and facilitation tendencies (7 regression analyses). The age of the participants did not moderate the effects on any variable, p > .206.
In order to test if the sex of the participant moderated the effect of subgroup of woman on stereotypes, emotions, and facilitation tendencies, three two-factor MANOVAS were conducted with subgroup of woman and sex of the participant as factors. The sex of the participant did not play a role in any variable. There was no main effect of this factor nor interactions with the subgroup of woman, ps > .067.
Therefore, we can conclude that our results are not significatively moderated by the level of sexism, age, or sex of the participants.
Predicting Active and Passive Facilitation at Work
We conducted a serial multiple mediator model to test the hypothesis that stereotypes and emotions could serially mediate the effect of the subgroup of woman on active and passive facilitation tendencies at work. We tested whether the subgroup of woman could affect active and passive facilitation through different indirect pathways: only through stereotypes (i.e., perceived morality, sociability, and competence), only through positive emotions, and through both stereotypes and positive emotions serially, with stereotypes affecting positive emotions.
To contrast this serial multiple mediator model, we used the macro PROCESS 3.0 for SPSS developed by Hayes (2018), specifically model 80, which combines properties of parallel and serial mediation. Perceived morality, sociability, and competence were specified as parallel mediators that send a pathway to a common additional mediator, positive emotions. This model tests seven specific indirect effects—four passing through only one mediator (i.e., morality, sociability, competence and positive emotions), and three passing through two mediators (morality → positive emotions; sociability → positive emotions; and competence → positive emotions). A heteroscedasticity consistent standard error and covariance matrix estimator was used. All coefficients reported were partially standardized given that dichotomous variables such as subgroup of woman were in partially standardized form. Percentile bootstrap confidence intervals using 5000 bootstrap samples were used for the inferential approach to test indirect effects. We can interpret that there are indirect effects when the bootstrap confidence interval does not include zero.
The first analysis predicting active facilitation revealed an indirect effect of the subgroup of woman on active facilitation through competence, B = −.54, SE = .14, 95% CI = −0.788, −0.262, and through positive emotions, B = −.20, SE = .07, 95% CI = −0.366, −0.073, and through both competence and emotions serially, with competence affecting positive emotions, B = −.10, SE = .04, 95% CI = −0.202, −0.035. Neither of these bootstrap confidence intervals included zero. As shown in Figure 2, although the total effect of the subgroup of woman on active facilitation was significant, once mediators were considered, the direct effect lost its significance. Perceived morality and perceived sociability had no direct or indirect effects on active facilitation at work.
1
Serial mediation of stereotypes and positive emotions on active and passive facilitation at work.
The second analysis, predicting passive facilitation, revealed an indirect effect of the subgroup of woman on passive facilitation through competence, B = −.66, SE = .14, 95% CI = −0.944, −0.376. As shown in Figure 2, although the total effect of the subgroup of woman on passive facilitation was significant, once mediators were considered, the direct effect was no significant. Perceived competence, positive emotions, and active and passive facilitation tendencies showed substantial explained variances, over .41. Perceived morality and perceived sociability had no direct or indirect effects on passive facilitation at work. The same analyses, including negative emotions, instead of positive emotions, did not reveal new indirect effects.
Discussion
The present research aims to analyze how two important female subgroups, women categorized as professional and sexy, are evaluated on the fundamental dimensions of stereotype content, emotions, and behavioral tendencies. Furthermore, we aim to explore the mechanism involved in carrying out helping behaviors towards both subgroups of women in the workplace.
As predicted (H1a), women categorized as sexy were perceived more negatively than women categorized as professionals: the former is perceived as less moral, sociable, and competent than a professional woman. These findings confirm and extend previous research, according to which “sexy women” are devalued in different stereotype dimensions (Altermatt et al., 2003; Cuadrado & López-Turrillo, 2014; Deaux et al., 1985; Glick et al., 2005; Heflick et al., 2011; Pacilli et al., 2017; Six & Eckes, 1991; Wookey et al., 2009).
Moreover, according to H1b, and Cuadrado and López-Turrillo’s results (2014) with adolescent participants, a woman categorized as professional is perceived as more competent than moral and sociable, whereas a woman categorized as sexy is evaluated as less moral than sociable and competent. Therefore, competence is the defining characteristic of the woman categorized as professional, while lack of morality is key in order to define a woman categorized as sexy, both from the perspective of adolescents and adults.
These findings are consistent regardless of whether the raters are male or female. Previous studies find that both men and women subscribe to similar stereotypes of “sexy women” (Six & Eckes, 1991; Smith et al., 2018). The present work not only confirms this pattern with women categorized as sexy women, but also reveals that it also occurs when evaluating women categorized as professional women. In addition, the sexism and age of the participants also do not influence their evaluations of the women’s subgroups.
We also found that participants feel fewer positive emotions towards sexy women compared to professional women, which partially confirms H3, and the results obtained by Glick et al. (2005) with adults and by Cuadrado and López-Turrillo (2014) with adolescents. However, in contrast to these studies, we have not found that participants experience more negative emotions towards sexy women than towards professional women. This may be due to (a) the time elapsed between the performance of the studies (Glick et al., 2005 vs. the present study, 2020), which may have moderated the negative emotions aroused by sexy women, or (b) the type of participants: adolescents (Cuadrado & López-Turrillo, 2014) versus adults. Future research should further investigate the role of negative emotions generated by some female subgroups such as those categorized as sexy women.
Our participants tend to carry out more facilitation tendencies at work towards women categorized as professional women than towards those categorized as sexy women, confirming H3. There are no studies assessing both subgroups of women on this scale, so this is a contribution of the present work. Furthermore, the results reveal that the differences are more pronounced in active facilitation tendencies than passive ones, which highlights the need to measure both types of behavior separately.
These findings are not moderated by the level of sexism, age, or sex of the participants. Since the participants’ sexism was measured at the end of the questionnaire, to better capture the moderating role of this variable future studies would measure it before any manipulation.
Finally, we have tried to explain why participants are more willing to help professional women than sexy women in the work context on the basis of the stereotyped perceptions and emotions they experience towards them. Our results show that this tendency to favor professional women over sexy women in the workplace is explained by the evaluations that both types of women receive in competence and the emotions they arouse. Specifically, participants are more willing to engage in active facilitating behaviors at work (i.e., hiring and promotion) towards a woman categorized as professional than towards a woman categorized as sexy because they perceive the professional woman to be more competent, which in turn generates more positive emotions towards her than towards the sexy woman.
These results have several implications. On the one hand, despite the primacy of warmth in social perception, especially morality, our findings indicate that the differential evaluations that both female subgroups receive in sociability and morality do not affect intentions to help them in the work context. On the contrary, the perceived competence of these women is the only dimension involved in this process. Thus, we confirm the primacy of competence evaluations in the work context (e.g., Cuddy et al., 2011), and we also extend the research by including sociability and morality and revealing that these dimensions play no role in helping intentions in the work setting towards two female subgroups.
On the other hand, we demonstrate the key role played by positive emotions (versus negative ones) towards professional and sexy women in the relationship stereotypes–helping behaviors at work, which is in line with results obtained in studies conducted in other contexts, and towards immigrant groups (e.g., Cuadrado et al., 2020; López-Rodríguez et al., 2016). Given that behaviors elicited by positive emotions have beneficial effects in the long term (as opposed to those elicited by negative emotions, whose effects are immediate; Fredrickson, 2013), the greater willingness to actively help professional women than sexy women in the workplace would also have effects that benefit the former in the long term. Therefore, it would be advisable to carry out interventions in work settings aimed at breaking the association “sexy women-incompetence.”
Finally, the relevance of considering the intensity of behavior (active versus passive) is highlighted. It allows us to have an accurate picture about what may be happening in the workplace regarding the type of help given to a greater extent to one subgroup of woman than to another. These are behavioral intentions aimed at explicitly benefiting the target (active help), as opposed to those that involve only convenient cooperation with a target (passive help).
Our study has a number of limitations. The photographs of the two women were distinct targets; therefore, there were many differences in their appearance, beyond their sexy versus professional clothing. Moreover, social clues that evidence competence and affect perceived status (i.e., the professional woman wore an architect hat and held a blueprint and a pencil in her hands) might be responsible for a different evaluation towards the woman categorized as sexy versus categorized as professional. For example, a woman implicitly perceived as highly trained might be evaluated as more competent, inspire more admiration, and motivate more behaviors to facilitate her at work.
However, our aim was not to compare the effect of the same target performing the same job, dressed differently, on stereotypes, emotions, and behavioral tendencies. We actually aimed to activate in the participants the cognitive schema of women categorized as professional versus sexy women (using pretested photographs), and to capture the different evaluation that these subgroups of women could generate. In other studies, to activate the corresponding category of women, authors used vignettes (e.g., Quiles et al., 2008) or descriptions (Cuadrado & López-Turrillo, 2014), which included different social clues to operationalize the subgroups of women. Authors have also simply used the category’s name (e.g., Eckes, 2002; Fiske et al., 2002; Glick et al., 1997) to activate it, so that each participant could imagine totally different targets for each category. Accordingly, in our study, to activate the cognitive schema of professional versus sexy woman, the manipulation of the subgroup of woman involved multiple clues to elicit a coherent inference.
Our manipulation used physical appearance, but not only focused on dressing and posture, but also on other social clues related to professionalism such as having a high-status job, all of which are informative factors in the operativization of subgroups of women. For example, the attributes associated with the career woman subtype match the traditionally masculine traits associated with high-status work roles (Glick et al., 2005). That is, as in daily-life evaluations, multiple factors conflate in the impression-formation process (e.g., when judging women categorized as sexy vs. professional women). Future studies might individualize the factors that more strongly guide the impression about women categorized as sexy and professional women. Moreover, since we did not include a control condition, future research should also compare the evaluations of these two subgroups of women with a woman without explicit attributes of professional or sexy in order to better understand the evaluations towards these two subgroups of women.
Ultimately, our results show that presenting oneself as sexy impacts women’s careers by their being perceived as less competent, their generating fewer positive emotions, and the tendency to hire or promote them less often than women who fit the professional subgroup. Consequently, sexy women may be systematically relegated to the lowest status roles and task assignments, which can influence their opportunities to demonstrate competence at valued tasks, affecting their career development.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Ministry of Science and Innovation of Spain under grant numbers PRX19/00263 and PID2019-105114GB-I00/ AEI 10.13039/501100011033, and by la Universidad de Almería, la Consejería de Economía, Conocimiento, Empresas y Universidad y el Fondo Europeo de Desarrollo Regional (FEDER) under grant number UAL18-SEJ-D007-B.
Ethical Approval
The study was approved by the authors’ University Ethics Committee (Human Research Bioethics Committee of the University of Almería; Ref: UALBIO2019/016) and was accordance with Committee on Publication Ethics’ International Standards for Authors.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author, upon reasonable request.
