Abstract
Although most women assume they would confront sexism, assertive responses are rare. We test whether women’s preference for respect or liking during interpersonal interactions explains this surprising tendency. Women report preferring respect relative to liking after being asked sexist, compared with inappropriate, questions during a virtual job interview (Study 1, n = 149). Women’s responses to sexism increase in assertiveness along with their preference for being respected, and a respect-preference mediates the relation between the type of questions and response assertiveness (Studies 1 and 2). In Study 2 (n = 105), women’s responses to sexist questions are more assertive when the sense of belonging is enhanced with a belonging manipulation. Moreover, preference for respect mediates the effect of the type of questions on response assertiveness, but only when belonging needs are met. Thus the likelihood of confrontation depends on the goal to be respected outweighing the goal to be liked.
During a town hall meeting, Secretary of State Hillary Clinton publicly responded to what she understood to be a sexist question. Clinton believed she was asked what her husband thought about a political issue, to which she responded “You want me to tell you what my husband thinks? My husband is not the Secretary of State, I am. You ask my opinion. I will tell you my opinion; I’m not going to channel my husband” (Sisk, 2009), p. 7.
Confrontations such as Secretary Clinton’s reduce prejudicial attitudes and discriminatory behaviors (Czopp, Monteith, & Mark, 2006; Mallett & Wagner, 2011). Although many women imagine they would act as Secretary Clinton did and assertively confront sexism, they rarely do so (Swim & Hyers, 1999). For example, Woodzicka and LaFrance (2001) found that 68% of women imagined they would refuse to answer sexist questions during a job interview and 28% said they would either rudely confront the interviewer or leave the interview. Yet when actually asked sexist questions during an interview, all of the women answered the sexist questions and none rudely confronted or left. Similarly, when faced with sexual harassment in the workplace, women typically try to avoid or appease the harasser (e.g., smile and offer a polite response) rather than assertively confronting (e.g., asking the harasser to stop, telling him that his behavior is unwelcome, or filing a formal report; Fitzgerald, Swan, & Fischer, 1995). In sum, research that studies women’s actual responses to a sexist situation finds that assertive confrontation is the exception rather than the rule. We conduct two experiments to test whether women’s preference for respect or liking goals explains this surprising tendency with regard to responses to sexism.
One aim of the present research is to test whether preferring the goal to be respected predicts women’s use of assertive confrontation in response to sexism. Research on impression management shows an association between seeking respect and assertive behavior. When striving to be perceived as competent, people attempt to control the interaction and focus on communicating their own opinions and abilities rather than attempting to understand those of the interaction partner. When instructed to gain respect, people use a host of agentic behaviors such as directing the conversation toward their own skills and abilities, taking charge of the conversation, and disconfirming their partner’s statements (Godfrey, Jones, & Lord, 1986; Jones & Pittman, 1982). Moreover, when responding to discrimination, a daily diary study found that people report using assertive confrontation to gain respect for one’s self or social group (Hyers, 2007). Therefore we predict that as one’s preference for the goal to be respected increases, so will assertive confrontations of sexism.
Importantly, one’s desire to be respected does not exist in a vacuum. Research on confronting prejudice shows that concern with social rejection is one of the major barriers to confrontation. Perceived discrimination poses a unique threat to one’s sense of social acceptance and belonging (Carvallo & Pelham, 2006). Both witnesses to confrontation and people who are confronted typically dislike the confronter (Czopp et al., 2006). Confrontation also carries the well-documented risk of interpersonal (e.g., disbelief, humiliation) and institutional (e.g., lower job evaluations, denial of promotion, termination) retaliation (Fitzgerald et al., 1995). The costs associated with confronting prejudice raise the possibility that the goal to be liked may also be pursued in sexist situations.
Thus, a second aim of the present research is to test whether the need to be respected outweighs the need to be liked when women experience sexist situations. Being respected and liked are both essential ingredients in receiving lucrative rewards including job offers, training opportunities, and promotions (Rudman, Moss-Racusin, Glick, & Phelan, 2012). The fact that consensually held gender stereotypes portray working women as competent but not very likeable places women in a precarious position in the workplace. Women must work harder than their male colleagues and self-promote to ensure that others notice their accomplishments (Eagly, 1987; Eagly & Karau, 2002). Yet when women highlight their skills and abilities to demonstrate competence and gain respect, they are disliked more than male peers who engage in the same behavior (Rudman & Glick, 1999, 2001). Indeed, women who disconfirm gender stereotypes by demonstrating competence are uniquely susceptible to backlash in the form of social (e.g., being disliked, excluded) and economic (e.g., being denied a job, passed over for promotion or training) sanctions for their atypical behavior (Phelan, Moss-Racusin, & Rudman, 2008; Rudman et al., 2012). These competing concerns create a “double bind” for women in that they must often choose between being respected and liked.
A third aim of the present research is to test whether satisfying the need to be liked increases assertive responses to sexism. Fear of rejection causes people to change their behavior to appease others (Fiske, 2004; Williams & Carter-Sowell, 2009). When attempting to be liked, people engage in socially acceptable behavior including agreeing with their interaction partner (Godfrey et al., 1986; Jones & Pittman, 1982). The need to be liked may therefore motivate women to avoid assertive confrontations and either ignore the offensive behavior or make a polite response. In fact, the more women fear being ridiculed or disliked the less they report confronting sexism (Good, Moss-Racusin, & Sanchez, 2012). In addition, Shelton and Stewart (2004) found that when women were more motivated to do well in an interview, they were less likely to assertively confront sexist questions. We test whether reducing concerns over being liked by, for example, affirming the sense that one belongs promotes assertive confrontations of sexism.
Overview of the Present Research
Two experiments test whether attempting to balance the need for respect and liking explains why assertive confrontation of sexist behavior is the exception rather than the rule. Although women may wish to be respected, their fear of the social costs of confronting may temper an assertive response to sexism. Study 1 examines women’s responses to sexist and non-sexist interview questions. It tests whether women’s self-reported preference to be respected versus liked differs following sexist versus non-sexist questions and whether preferring respect shapes actual responses to the questions. We predict that women will assertively respond to sexism to the extent that they prefer respect, despite social pressures to be liked. Study 2 provides a second test of these questions and also tests whether attenuating, relative to enhancing, concerns about being liked increases assertive responses to sexist questions.
Study 1
Study 1 examines women’s responses to sexist or merely inappropriate questions during a mock job interview. We use Woodzicka and LaFrance’s (2001) interview paradigm to examine women’s reactions to being asked sexist versus inappropriate (but non-sexist) interview questions. Previous research has demonstrated that the questions used in this paradigm are rated as equally surprising and offensive, but that the sexist questions are rated as more sexist than the inappropriate questions (Shelton & Stewart, 2004; Woodzicka & LaFrance, 2001).
We examine whether women report a preference to be respected or liked depending on the type of questions they are asked. We do so by assessing the relative preference for each goal using a bipolar response scale that forces participants to rate one goal in relation to the other. Participants have the option of rating the goals as equally important by choosing the scale midpoint. Using this response format enables us to determine whether participants identify one goal as taking priority over the other during the interview. Past research on goal pursuit during intergroup interactions has used the same question wording and response options and found that this self-report item corresponds to verbal and nonverbal pursuit of respect and liking goals (Bergsieker, Shelton, & Richeson, 2010).
We also test whether women’s self-reported goal preference is related to the assertiveness of their responses to the interview questions. We hypothesize that the assertiveness of women’s responses to sexist questions will increase along with their preference to be respected and that the preference to be respected will promote assertive confrontation. That is, a preference for respect should mediate the association between the type of questions and the assertiveness of responses. Supporting this hypothesis, when faced with discrimination, people who are concerned with gaining respect for themselves or their group are motivated to assertively respond by clearly telling a harasser that a behavior is inappropriate (Hyers, 2007).
Although being liked is an essential outcome of a job interview, liking may not necessarily be the goal that women consciously report preferring. People do not like to acknowledge their attempts to ingratiate themselves to others. In fact, Jones (1965) argues that ingratiation is a mildly pejorative term because it suggests that people deliberately manipulate their behavior to gain another person’s approval. This makes people hesitant to explicitly state that they wish to be liked or acknowledge that they are acting a certain way to gain approval. Moreover, women may try to balance concerns with being respected by the interviewer with concerns about being liked (Rudman et al., 2012). Walking the tightrope of demonstrating competence and maintaining liking means that leaning too far toward either goal could have disastrous consequences. Thus we test whether women report a stronger preference for being respected relative to being liked after they are asked sexist (vs. inappropriate) questions, and the influence that this preference has on their responses.
Method
Participants
One hundred fifty-four women participated to partially fulfill a course requirement. Participants ranged in age from 18 to 26 years old (M = 18.66; SD = 1.14). They were White (n = 105), Asian (n = 21), Latina (n = 17), multiple ethnicities (n = 10), and African American (n = 1).
Procedure and materials
We conducted mock job interviews using a computerized instant message format similar to that used in past research on confrontation (Czopp et al., 2006; Rattan & Dweck, 2010). One of several female experimenters who were White (n = 6) or Indian (n = 1) led each participant individually through the experiment. Unbeknownst to the participant, the experimenter played the role of both experimenter (in person) and interviewer (via instant message on the computer). Participants first completed a mock job interview during which they were either asked sexist or inappropriate questions. Participants received the interview questions and submitted their answers via an instant message program. After the interview, they reported their goal preference during the interview and whether they used several specific responses using Inquisit 3.0.
Type of questions
Participants were provided with the following cover story based on Woodzicka and LaFrance’s (2001) interview paradigm:
You are about to take part in a study of online job interviews for telecommuters. With the current economy and the changing workforce, more people are choosing to telecommute to work. That is, they choose to work from home for an office that is in another location. Often the interviews for telecommuter positions are conducted over the Internet, either through Skype or simply through a formalized chat. Little research has investigated the effectiveness of these interviews or what types of strategies work well for such interviews. The present research investigates those questions.
Participants were then told that a 32-year-old male who worked in the University’s student resource center would conduct a mock job interview with them via a computerized instant message program. Participants were told that the interviewer had an MBA and extensive experience conducting online interviews.
Everyone received the same first set of questions: “What is your year in school?” “What is your major or intended major?” and “Why are you interested in this research position?” After they submitted their answers to the first set of questions, participants were randomly assigned to receive either the sexist questions (“Do you have a boyfriend?” “Do people find you desirable?” “Do you think it is important for women to wear bras to work?”) or the inappropriate questions (“Do you have a best friend?” “Do people find you morbid?” “Do you think it is important for people to believe in God?”) used by Woodzicka and LaFrance (2001). After they submitted their answers to the second set of (sexist or inappropriate) questions via instant message, the experimenter returned to the room and announced that due to technical difficulties the interview portion of the study was over.
Goal preference
We test whether women prefer one goal relative to the other, or prefer to equally balance concerns with being respected and liked by the interviewer. Bergsieker and colleagues (2010) used a single-item measure to accomplish this task. Based on that item, we asked participants to report their goal preference by answering this question: “If you had to choose between being liked and being respected by the interviewer, which do you regard as more important?” with 1 = most important to be liked, 4 = equally important, 7 = most important to be respected. Participants may rate the two goals as equally important; however, using this format increases the likelihood that participants will make a difficult choice between endorsing two favorable characteristics (Rosenthal & Rosnow, 2008). Higher scores reflect a stronger preference for being respected over being liked by the interviewer.
Assertiveness of responses
We measured responses to the interview questions in two ways: coding written responses and analyzing self-reported use of specific strategies.
Coded responses
Two coders, who were unaware of study hypotheses and the condition to which the participant was assigned, evaluated each written answer to determine whether it resembled one or more of six behaviors that Woodzicka and LaFrance (2001) found were commonly used: simply answer the question (i.e., yes or no), ask why the question was asked, refuse to answer, state the question is irrelevant to the position, tell the interviewer it is “none of his business,” and stop the interview. In response to emerging themes in the answers we also coded whether participants clarified the question (e.g., “if you mean desirable as a research assistant . . . ”) or explained their answer (e.g., “I don’t think that people find me morbid, no one has ever indicated that they feel that way about me”). Coders recorded 1 if the behavior was present and 0 if it was absent. An answer could be coded in multiple categories if more than one behavior was present. Percent agreement for coders ranged from 90% to 100%.
We summed across answers to the three questions (e.g., in the sexist condition, “Do you have a boyfriend?” “Do people find you desirable?” “Do you think it is important for women to wear bras to work?”) to create a variable for each behavior that ranged from 0 = did not do this for any question to 3 = did this for all three questions. We created one variable to represent compliant responses by computing the average of “simply answer” and “explain.” We created another variable to represent the assertiveness of responses by averaging the remaining behaviors 1 (α = .60, skew = 1.37, SE = 0.20).
Self-reported assertiveness
Participants also used a scale from 1 = not at all to 7 = very much to respond to the item, “During the interview, to what extent did you actually act in the following manner?” Participants reported how much they used each of Woodzicka and LaFrance’s (2001) six behaviors listed above over the course of the interview. We averaged the items to create an index of self-reported response assertiveness (α = .89, skew = 1.62, SE = 0.20). 2
Results
Goal preference
An independent-samples t test showed that women preferred to be respected by the interviewer, relative to being liked, after they were asked sexist (M = 5.63, SD = 1.34), compared with inappropriate (M = 4.78, SD = 1.37) questions, t(152) = 3.84, p < .001.
Assertiveness of responses
Consistent with past research on confronting, coders found that the most common response for both types of questions was to comply by simply answering the question and explaining one’s answer. However, coders found that women were significantly less likely to comply in response to the sexist questions than the inappropriate questions, t(151) = 4.65, p = .001. Moreover, coders found that women made more assertive responses to sexist than inappropriate questions, t(151) = −3.91, p < .001 (see Figure 1 for means). Self-assessed assertiveness showed the same pattern with more assertive responses to sexist (M = 2.50, SD = 1.96) than inappropriate (M = 1.49, SD = 1.23) questions, t(151) = −3.83, p < .001.
Across both studies, coders found that the most common response was to comply by answering the interview questions. However, women were more likely to assertively respond to sexist questions than inappropriate questions.
Mediation
We calculated the indirect effect of the type of question on the assertiveness of women’s responses through goal preference using the Preacher and Hayes (2004) bootstrapping method. Bootstrapping can detect effects in small samples while maintaining control over the Type I error rate (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002; J. Williams & MacKinnon, 2008). We used Preacher and Hayes (2008) INDIRECT macro for SPSS to generate 5,000 samples from the original data set using sampling with replacement. The type of question was coded so that −1 = inappropriate questions and 1 = sexist questions. Higher numbers for the goal preference variable indicates a greater preference for respect, relative to liking. Table 1 shows that the type of questions and goal preference are both positively associated with self-assessed and coded assertive responses. Overall, women reported preferring more respect, relative to liking, following sexist questions than following inappropriate questions, and as the preference for being respected increases, so does the assertiveness of responses. The confidence intervals (CIs) for the indirect effect do not include zero; therefore, we conclude that the indirect effect of the type of questions on response assertiveness through goal preference was reliable.
Results of Bootstrapping Analyses Testing for an Indirect Effect of the Type of Question on the Assertiveness of Responses That Operates Through Goal Preference (Study 1).
Note. †p < .10. *p < .05. **p < .01. ***p < .001.
Discussion
Self-assessed and coded responses were more assertive following sexist than inappropriate questions. Similar to the past research (Shelton & Stewart, 2004; Woodzicka & LaFrance, 2001), the most common response was to simply answer the questions—regardless of their content. This was, by far, the most typical way that participants chose to respond to the interview questions. Although women answered most of the sexist questions, they were more likely to assertively respond to one or more of the sexist questions than the inappropriate questions. Almost 10% (9.3%) of women even refused to answer one or more of the sexist questions. This percentage is slightly higher than that given by Shelton and Stewart (2004) who found that 5.9% of women refused to answer a question; no one refused to answer a question in Woodzicka and LaFrance’s (2001) study. We may have found slightly more assertive responses than previous studies because we used an instant message format rather than a face-to-face interview, and being physically removed from the interaction partner reduces inhibitions (Milgram, 1974; Short, Williams, & Christie, 1976).
We also found evidence that the preference for a respect goal motivated more assertive responses to the sexist, but not inappropriate, questions. When faced with sexist interview questions, women reported preferring respect relative to liking compared with when faced with merely inappropriate questions. The item we used forced women to indicate which goal held primacy during the interview—to be respected or liked. 3 Therefore, women did not say that they were only seeking respect, but they did indicate it was their preferred goal after being asked sexist questions. Moreover, preferring respect explained the association between the type of questions and the assertiveness of women’s responses. Liking was still clearly a concern, however, as even most of the assertive responses were fairly restrained. In sum, assertive confrontation is not the default response to sexism—it depends on one’s goal in the situation. If women believe it is more important to be liked at that time, their responses should be less assertive than if they believe it is more important to be respected.
One limitation of this study is that we collected reports of goal preference after the interview. That is, we assessed the proposed mediator after the dependent measure. We did so to minimize the likelihood that participants would satisfy the goal by answering the goal preference item rather than via their responses to the interviewer. However, doing so creates the possibility that the assertiveness of responses influenced goal preference. We tested the feasibility of the opposite mediation model by entering question type as the independent variable, self-reported or coded assertiveness as the mediator, and goal preference as the dependent variable. This alternative model did not provide as good of a fit to the data (CIs = [−0.04, 0.13]; CIs = [−0.003, 0.19]; coded, self-report, respectively). Thus, there is some evidence for our preferred theoretical interpretation that goal preference predicts the assertiveness of responses.
Study 2
Study 1 showed that when asked sexist (compared with merely inappropriate) interview questions, women reported that respect was more important than liking. Furthermore, preferring respect relative to liking predicted the assertiveness of women’s responses to sexist interview questions. In Study 2, we attempt to replicate the effect of preferring respect on assertive responses to sexist behavior. Specifically, we test whether preferring respect mediates the relation between the type of questions and the assertiveness of responses.
We also test whether reducing the need to be liked moderates the impact that the type of questions has on the strength of the respect goal. To do so, we manipulate whether the need to be liked is reduced or enhanced before the interview by asking participants to recall a time they were accepted or rejected (Knowles & Gardner, 2008). Recalling a time that one was accepted should enhance the sense of belonging and allow respect concerns to come to the fore. In comparison, recalling a time that one was rejected should reduce the sense of belonging and emphasize the need to be liked.
We also test whether reducing versus enhancing the sense of belonging affects the assertiveness of responses. We expect that responses to sexist questions should be more assertive when the sense of belonging has been enhanced compared with when it has been reduced. The social situation is more powerful than we think (Jones, 1990) and priming rejection, and therefore reducing the sense of belonging, should mimic the power that social pressures to be liked exert on our behavior. In comparison, satisfying the need to be liked by enhancing the sense of belonging may allow the need to gain respect to take the lead, thereby increasing the assertiveness of responses. Indeed, when faced with discrimination, people who are primarily concerned with being liked are motivated to avoid confrontation by making a joke or ignoring the offensive behavior whereas people who are concerned with gaining respect use confrontation (Hyers, 2007). Finally, we test for moderated mediation—that is, whether attenuating liking-related concerns via enhancing the sense of belonging moderates the impact that the type of questions has on assertive responses through goal preference.
Method
Participants
One hundred ten women participated to partially fulfill a course requirement. Participants ranged in age from 18 to 23 years old (M = 18.96; SD = 1.21). Participants were White (n = 68), Asian (n = 23), African American (n = 9), Latina (n = 5), multiple ethnicities (n = 4), and American Indian/Alaska Native (n = 1). We excluded 3 participants due to procedural errors (e.g., computer malfunction) and 2 participants who suspected the interview was not real.
Procedure
One of several female (3 White, 1 Indian, 1 Latina) or male (1 White) experimenters conducted the experimental session. Participants were randomly assigned to condition using a 2 (sense of belonging: enhanced, reduced) × 2 (type of questions: sexist, inappropriate) between-participants factorial design. The Inquisit 4.0 computer program presented all study materials. Participants were first instructed to complete the belonging manipulation, which they were told we were pilot testing for a future study. Participants were then randomly assigned to receive sexist or inappropriate questions and wrote responses to the questions. The experimenter cited technical difficulties to end the interview after the answers were submitted and then asked the participant “to tell us your thoughts about the interview experience.” At this point, participants completed a number of self-report items about the interview.
Materials
Belonging manipulation
Participants were randomly assigned to receive Knowles and Gardner’s (2008) belonging manipulation to either reduce (via social rejection) or enhance (via social acceptance) the sense of belonging. The following instructions were used to reduce the sense of belonging; instructions used to enhance the sense of belonging appear in brackets.
Please write about a time in which you felt intensely rejected [accepted] in some way, a time that you felt as if you did not belong [you belonged]. This rejection [acceptance] can be interpersonal in nature (e.g., a time in which someone broke up with you, or no longer wanted to be your friend [a time in which someone wished to date you or wanted to be your friend]) or can be a rejection [acceptance] from a group (e.g., a time in which you were chosen last for a team or excluded from a clique [a time in which you were chosen for a team or included in a clique]).
Participants used a scale from 1 = not at all to 7 = very much to answer the following manipulation check item, “I felt socially accepted during this event.”
Goal preference
As in Study 1, participants indicated their goal preference using the item, “If you had to choose between being liked and being respected by the interviewer, which do you regard as more important?” with 1 = most important to be liked, 4 = equally important, 7 = most important to be respected.
Type of questions
Participants experienced the same interview and received the same interview questions described in Study 1. After the participant submitted her answers to the questions, the experimenter notified her that there were technical difficulties and they could not continue the interview.
Assertiveness of responses
As in Study 1, participants wrote their response to each interview question. We coded their writing to determine whether each answer resembled one or more of eight behaviors (listed in Study 1 method). Two coders, who were unaware of study hypotheses and the condition to which the participant was assigned, evaluated each written answer. Percent agreement for coders ranged from 96% to 100%. We combined answers to the three questions to create a variable for each behavior that ranged from 0 = did not use this response to any question to 3 = used this response for all three questions. As in Study 1, we created one variable to represent compliant responses by combining “simply answer” and “explain.” We created another variable to represent the assertiveness of responses by averaging the remaining behaviors (α = .83, skew = 1.24, SE = 0.38).
As in Study 1, participants also used a scale from 1 = not at all likely to 7 = very likely to self-assess how much they used each behavior over the course of the interview. We averaged these items to create an index of self-assessed response assertiveness (α = .89, skew = 2.02, SE = 0.24).
Results
An independent-samples t test showed that participants felt more socially accepted when responding to the belonging enhancement (M = 6.47, SD = 0.82) compared with the belonging reduction (M = 2.10, SD = 1.20) prompt, t(102) = −21.69, p < .001.
Goal preference
We first tested whether manipulating the need to be liked moderates the impact that the type of questions has on goal preference. We used a 2 (type of question: sexist, inappropriate) × 2 (sense of belonging: enhanced, reduced) between-subjects ANOVA with goal preference as the dependent variable. Replicating Study 1, we found a main effect of type of question, F(1, 104) = 7.17, p = .009, η2 = .07, such that participants reported preferring respect relative to liking more in the sexist (M = 5.40, SD = 1.15), compared with the inappropriate (M = 4.63, SD = 1.39) interview condition. There was no main effect of the belonging manipulation on goal preference, F(1, 104) = 0.63, p = .43, and no interaction, F(1, 104) = 0.77, p = .38.
Self-assessed response assertiveness
We then used the same between-subjects ANOVA and examined self-assessed response assertiveness as the dependent variable. We found a main effect of the type of questions, F(1, 105) = 4.36, p = .04, η2 = .04, such that responses were self-assessed as more assertive in response to sexist (M = 1.79, SD = 0.99) than inappropriate questions (M = 1.44, SD = 0.74). There was also a main effect of the belonging manipulation, F(1, 105) = 3.94, p = .05, η2 = .04, such that responses were self-assessed as more assertive following the belonging enhancement prompt (M = 1.78, SD = 1.04) than the belonging reduction prompt (M = 1.45, SD = 0.67). Moreover, there was an interaction, F(1, 105) = 5.16, p = .025, η2 = .05. Participants reported using the most assertive responses when they received the belonging enhancement prompt and sexist questions; all other conditions were significantly lower in assertiveness (see Figure 2). Simple effects tests showed that responses to sexist questions were self-assessed as significantly more assertive following the belonging enhancement prompt than the belonging reduction prompt, F(1, 105) = 9.24, p = .003, η2 = .09. Responses to the inappropriate questions did not differ based on the belonging manipulation, F(1, 105) = 0.04, p = .84. Moreover, following the belonging enhancement prompt, responses were more assertive to the sexist questions than the inappropriate questions, F(1, 105) = 9.69, p = .002, η2 = .09. Following the belonging reduction prompt, responses did not differ depending on the type of question, F(1, 105) = 0.02, p = .90.
Self-assessed response assertiveness was highest in response to sexist questions when the sense of belonging had been enhanced via a belonging manipulation (Study 2).
Coded response assertiveness
Finally, we used the same between-subjects ANOVA to examine coded assertiveness in the written interview responses as the dependent variable. When considering compliance, as in Study 1, we found that the most common response to both types of questions was to comply by simply answering and explaining one’s answer (see Figure 1 for the means). Coders noted that women were significantly less likely to comply with the sexist questions (M = 1.45, SD = 0.69), compared with the inappropriate questions (M = 2.04, SD = 1.12), F(1, 102) = 10.25, p = .002, η2 = .10. There was no effect of the belonging manipulation on compliance and no interaction, ps > .56.
When considering the assertiveness captured in coding of written responses to the interview questions, we found a main effect of the type of question F(1, 102) = 6.01, p = .02, η2 = .06 such that responses were more assertive in response to sexist (M = 1.25, SD = 1.44) than inappropriate questions (M = 0.37, SD = 0.73). There was also a main effect of the belonging manipulation, F(1, 102) = 4.26, p = .04, η2 = .04 such that responses were coded as more assertive following the belonging enhancement prompt (M = 1.19, SD = 1.45) than the belonging reduction prompt (M = 0.44, SD = 0.81). The interaction was not significant, F(1, 102) = 2.71, p = .10, though cell means were in line with predictions and matched the pattern of self-assessed assertiveness (see Figure 3).
Coded response assertiveness was highest in response to sexist questions when the sense of belonging had been enhanced via a belonging manipulation (Study 2).
Mediation
We used Hayes PROCESS macro for SPSS to estimate a conditional process model (Preacher, Rucker, & Hayes, 2007), which tests the contingent nature of the effect of the type of questions on response assertiveness through goal preference, depending on the belonging manipulation. We found evidence that enhancing the sense of belonging moderated the indirect effect as there was an interaction between the type of questions and the belonging manipulation in the model of goal preference (see Table 2; Model 5, Hayes, 2012). 4 This interaction was significant for self-assessed assertiveness and marginally significant for coded assertiveness. The indirect effect is conditional on the belonging manipulation, so we separately examine the mediational pathways for belonging enhancement and reduction (Hayes, 2012). When the sense of belonging is enhanced, the indirect effect of the type of questions on the assertiveness of responses through respect goal preference is positive and significant for both self-assessed and coded responses. Moreover, for the belonging enhancement condition, a 95% bootstrap CI for the conditional indirect effect is above zero. When the sense of belonging is reduced, the indirect effect is not significant for either self-assessment or coding, as evidenced by CIs that include zero. Therefore, preference for a respect goal mediates the effect of the type of questions on the assertiveness of responses, but only when belonging is enhanced.
Results of Conditional Effects Bootstrapping Analyses Testing for an Indirect Effect of the Type of Question on Assertive Responses That Operates Through Goal Preference and Is Dependent on the Belonging Manipulation (Study 2).
Note. ††p = .05. †p < .10. *p < .05. **p < .01. ***p < .001.
Discussion
Replicating Study 1 and past research, the most common response to the interview questions was to comply by providing an answer. This suggests that there are strong response norms for an interview context that constrain behavior to be mostly compliant, even in the face of inappropriate or biased questions. Responses were more assertive to sexist than inappropriate questions, but the assertiveness of responses was still tempered in relation to how Woodzicka and LaFrance (2001) found that women imagined they would respond to such questions. This may reflect an attempt to balance being respected and liked by the interviewer in the service of gaining a favorable evaluation (e.g., Rudman et al., 2012).
Replicating Study 1, we found that preferring a respect goal mediated the relation between the type of questions and response assertiveness. Women report a stronger preference for respect, relative to liking, following sexist questions than following inappropriate questions, and as the preference for respect increases, so does the assertiveness of responses. This research provides the first test (and replication) of the causal role that preferring respect plays in predicting assertive responses to sexism. Goal preference was unaffected by the belonging manipulation; as in Study 1, we simply found a main effect whereby the preference for respect relative to liking was greater following sexist, compared with inappropriate, questions. Recalling a time that one had been accepted or rejected in the past did not modify this self-reported preference.
Instead, we found that the belonging manipulation moderated the effect of the type of questions on the assertiveness of responses. 5 That is, when women were asked sexist questions, their responses were less assertive when the sense of belonging was reduced by considering past rejection compared with when the sense of belonging was enhanced by recalling a time they had been accepted. Less assertive responses following the reduced sense of belonging correspond to how many women actually responded to the sexist interview questions used in this study (Woodzicka & LaFrance, 2001) and to sexual harassment in the workplace (e.g., Fitzgerald et al., 1995). In comparison, enhancing belonging increases assertive responses to sexism. Perhaps belonging enhancement lowers the perceived risk of confronting. Although it does not change participants’ preference for respect over liking from the interviewer it may still affect their willingness to act on that goal and assertively confront. 6 One of the biggest barriers to confronting any type of prejudice is the fear of social repercussions (Good et al., 2012; Shelton & Stewart, 2004) and Study 2 suggests that reducing that fear increases assertive responses.
General Discussion
The present research enhances our understanding of why assertive responses to discrimination are the exception rather than the rule. The need to be liked is a fundamental human motivation and the fear of retaliation for confronting inhibits assertive responses to sexism. Study 1 shows that when faced with a sexist interviewer, preferring respect over liking explains the assertiveness of women’s responses. Study 2 shows that the indirect effect of the type of questions on the assertiveness of responses through respect goal preference only operates when liking-related concerns are attenuated by enhancing the sense that one belongs.
Similar to past research, we found low rates of confrontation which suggests that most women remain reluctant to confront even when motivated to gain the respect of the interviewer. Despite this fact, many people expect that a “reasonable person”—including themselves—would assertively confront a harasser. In fact, United States’s law relies on a reasonable person standard when evaluating sexual harassment cases (Ellison v. Brady, 1991). If sexual harassment occurred, then the target of harassment is expected to make an immediate and assertive attempt to stop the harassment. If such a response is not forthcoming, it is assumed that either harassment did not occur or that the target welcomed the behavior (Estrich, 1991). Given existing research, requiring such a response is an unreasonable expectation.
Our experimental paradigm involved a power differential, with participants always imagining themselves in a low-power position (i.e., the interviewee). Power affects one’s response options. The need to be liked may be stronger when the target of sexism is in a low power, compared with high power, position relative to the harasser. Future research on confrontation should manipulate the power dynamics of the situation. A person with high power should be more likely than a person with low power to risk disrupting a social exchange with confrontation (Keltner, Van Kleef, Chen, & Kraus, 2008). In support of this idea, women are more likely to report sexism if the harasser has equal or lesser status within the workplace than if the harasser is in a more powerful position (Fitzgerald et al., 1995).
Even women who hold positions of power may be unwilling to confront due to their fear of backlash. The Backlash Avoidance Model (Rudman & Glick, 1999, 2001; Rudman et al., 2012) argues that women who work in masculine-stereotyped domains (e.g., science, engineering) and women who engage in gender atypical behaviors (e.g., leadership) are particularly susceptible to backlash for agentic behavior. To avoid backlash, women in these positions use strategies such as hiding atypical behavior and increasing their conformity to norms. Therefore, women who have attained leadership positions may be unwilling to risk their status by confronting sexism and, instead, increase their conformity to the norm of polite behavior and silence themselves in the face of discrimination. Doing so ensures that they keep their position, but reinforces discrimination that ultimately puts them at a disadvantage in relation to their male peers and maintains sexist norms that impede the progress of other women.
Thus, the experience of sexism creates a double bind for all women, regardless of their power. Remaining silent can send the message that one is complicit with the behavior (Czopp, 2013). Indeed, many times the targets of sexism are blamed for their decision to not make an assertive response at the time the discrimination takes place (Cohen & Cohen, 1993; Fitzgerald et al., 1995). Yet the need to be liked by one’s colleagues is a very real concern with significant consequences; women who act agentically are disliked and penalized via sabotage, social rejection, and hiring discrimination (Heilman, Wallen, Fuchs, & Tamkins, 2004). Moreover, most people expect to be disliked if they confront discrimination (Czopp et al., 2006; Good et al., 2012). Worries that confronting sexism will jeopardize the opportunity for training, promotion, or a pay raise make the idea of questioning sexism, even in a polite manner, unappealing and therefore unlikely for most women.
Understanding how the need to be liked shapes behavior provides a more complete picture of the situation in which targets of sexism must make decisions about how to respond. When we put women in a sexist situation, preferring the interviewer’s respect over his liking increased assertive confrontation (Studies 1 and 2). Moreover, reducing liking-related concerns increased assertive responses to sexism (Study 2). The finding that the need to be liked tempers the assertiveness of responses has profound implications for how we understand reactions to sexism. People who decide the outcome of sexual harassment cases (e.g., judges, managers) should be aware that the need to be liked can overwhelm the need to be respected—even if it goes against one’s own self-interest. Procedures for reporting sexist behavior could be modified to attenuate concerns related to the need to maintain liking in the workplace. For example, conducting routine anonymous workplace climate surveys could reduce the consequences associated with filing a grievance. Keeping such surveys anonymous would protect employees who report sexism from repercussions such as being denied a promotion and allow third parties (e.g., coworkers) a chance to report sexist behavior that they observe.
If we wish to create a more just and equal society, we should identify ways to reduce the social and financial costs of confronting discrimination. Allowing biased behavior to proceed unchecked signals that such actions are normative and acceptable (Crandall, Eshleman, & O’Brien, 2002). It is only by changing the norms for confronting prejudice from being taboo to expected and even encouraged that we may reduce the consequences of confronting. If people get in the habit of confronting, the fear and anxiety that is experienced when engaging in this deviant behavior should be reduced. Not all confrontations will go well, but some may go better than expected and reduce the likelihood of future biased behavior (Mallett & Wagner, 2011). If confrontation becomes commonplace, the defensiveness of those confronted and the stigma attached to those who confront should also be reduced.
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) received the following financial support for the research, authorship, and/or publication of this article: The authors wish to thank The Gannon Center for Women and Leadership at Loyola University Chicago for supporting this research with a faculty fellowship awarded to the first author.