PERCEIVED RESISTANCE OF WOMEN IN ACADEMIA

PERCEIVED RESISTANCE OF WOMEN IN ACADEMIA

Master Thesis, MSc Human Resource Management University of Groningen, Faculty of Economics and Business

January 17, 2021

Laura Wörgartner Student Number: S4140842

Alois Norer Straße 3 6130 Schwaz, Tyrol

Austria

Tel.: +43 69910401414

Email: l.worgartner@student.rug.nl

Supervisor Dr. Susanne Täuber

s.tauber@rug.nl

Abstract

Gender inequality is a worldwide issue and women are still underrepresented in many organizations although companies claim to promote gender equality. A gap between policy and reality of gender equality is also evident in higher educational institutions such as universities.

Women are still underrepresented in academic positions which especially holds for women of different ethnic backgrounds. Although universities promote gender equality, the lack of meaningful progress suggests that resistance against women’s equality still exists. This thesis tests the hypothesis that women who perceive such resistance will be less inclined to let their voices be heard. Additionally, I predict that intersectionality strengthens the effects of perceived resistance on psychological safety. Psychological safety will be provided as a mediating variable. To test the hypotheses, 91 women working in academia filled out a survey about perceived resistance to gender equality. Results indicate that particularly intersectional women experience the most forms of perceived resistance and claim they would speak up about gender inequality. However, when women with higher levels of intersectionality experience discrimination and resistance, it affects their psychological safety feeling more negatively than women on less intersectional levels; and therefore, those women are more likely to leave the organizations.

Keywords: Gender Inequality; Intersectionality; Perceived Resistance; Psychological Safety;

Voice

PERCEIVED RESISTANCE AGAINST EQUALITY OF WOMEN IN ACADEMIA

INTRODUCTION

Workforce diversity can increase productivity, improve problem-solving competencies, and enhance talent acquisition, making diversity essential for organizations. The term workforce diversity refers to employees that differ such as in age, ethnicity, sexuality, religion, or gender (Saxena, 2015). According to the Deloitte University Press (2017), awareness of the importance of diversity has increased over the past years. However, awareness has not led to substantial organizational change towards more equal, diverse, and inclusive workplaces. The World Economic Forum (2020); for instance, notes in their Gender Gap Report that we will not achieve parity for another 100 years.

Workforce diversity can increase productivity, improve problem-solving competencies, and enhance talent acquisition, making diversity essential for organizations. The term workforce diversity refers to employees that differ such as in age, ethnicity, sexuality, religion, or gender (Saxena, 2015). According to the Deloitte University Press (2017), awareness of the importance of diversity has increased over the past years. However, awareness has not led to substantial organizational change towards more equal, diverse, and inclusive workplaces. The World Economic Forum (2020); for instance, notes in their Gender Gap Report that we will not achieve parity for another 100 years.

The gap between policy and reality is evident also in higher educational institutions such as universities. Universities all around the world claim to promote and to be committed to diversity;

however, it seems that they have not implemented diversity successfully (Sensoy & Diangelo,

2017). Illustrating this, a study conducted by Hofstra, Kulkarni, Galvez, He, Jurafsky, and

until 2015 to find out if diverse groups are delivering more creative ideas than non-diverse groups and if their career paths differ. Their results showed that although young academics from diverse backgrounds have more creative ideas, these are systematically discounted and undervalued, leading to less successful academic careers of minorities. This shows that inequality and biases are still present in academia.

Additionally, a gap particularly regarding gender equality has been observed in academia.

Although policies exist concerning equal opportunities, implementation of those policies does not seem to be successful yet (Salinas & Bagni, 2017). Women are still a minority in senior academic positions in Europe with percentages of only 18.7% in the Netherlands and 19.4% in Germany (European Commission, 2019). According to van den Brink and Benschop (2012), reasons why gender equality measures might fail include the primary focus on women and not on the university as a system. However, institutions are mainly influenced by male management which is stereotypical perceived as tough and focus-driven and women are perceived as generous and relationship-focused. Consequently, women are seen as deviant and in need of ‘fixing’. (Heilman, 2001; van den Brink & Benschop, 2012). Furthermore, research mentions that people who do not fulfil certain stereotypes, tend to be judged more critically (Bettencourt, Dill, Greathouse, Charlton, & Mulholland, 1997). Additionally, by virtue of the male dominated workplace and the stereotypes for women, mismatches can occur, which lead to a perception that women are not competent for the male dominated workplace (Heilman, 2001).

The above pressures make women feel that they need to execute male stereotypical

adjectives to succeed in the male dominated working environment (Kloot, 2004). As a result,

Heilman (2012) states that women often find themselves forced to change; however, this can have

disadvantages for their development and organization. For instance, women often censor

themselves which can affect “women’s maximization of their talents and wreak havoc with their emotional well-being” (Heilman, 2012, p. 125). Additionally, stereotypes can lead to discrimination of women which can have an adverse impact on equality measures (Humbert, Kelan, & van den Brink, 2015). In sum, the above research shows that ‘fixing’ the women is not the solution. Rather, to achieve gender equality, the entire organization needs to change which also includes the focus on recruitment, selection and other sectors (Adams, 2018).

Universities’ attempts to promote gender equality with special programs, however, seem to be ineffective because those programs might bring along prejudices that a woman might have been only chosen for a specific academic position because she is a woman and not because of her competencies. This affirmative action could; therefore, have negative consequences for women (van den Brink & Benschop, 2012). Krook (2016), too, argues in her research about the underrepresentation of women in politics that resistance exists regarding gender equality measures such as quotas. “This is because, if fully implemented, quotas threaten the reigning rules, practices, and norms of political life, long premised upon women’s exclusion” (Krook, 2016, p. 279). In addition, research shows that most people think that formulating policies is the same as exerting policies, leading to non-performativity once a policy has been introduced (Ahmed, 2012; Dover, Kaiser, &, Major, 2019). Accordingly, Kloot (2004) suggests that “the writing of policies as a substitute for overt action is in itself a masculine trait” (p. 481).

In sum, the literature on the reality versus policy gap concerning diversity and specifically gender equality in academia shows that equality measures do not have the desired effect (Salinas

& Bagni, 2017; Sensoy & Diangelo, 2017). Indeed, England, Levine, and Mishel (2020) argue that

the process of gender equality in academia has been slowing down and that there is a need for

universities to change. This is important because academia can benefit from gender diversity and

opportunities can arise such as innovation, diverse knowledge, and expertise (Nielsen et al., 2017).

Some reasons for why gender equality measurements fail have been introduced above, such as the lack of organizational change regarding stereotypes and resistance to quotas, but these might not be the only factors responsible for a gap between words and actions (Humbert et al., 2015; van den Brink & Benschop, 2012). Previous research has already revealed the existence of resistance and the associated tools such as non-performativity or resistance to gender quotas (Ahmed, 2012;

Krook, 2016).

However, how perceived resistance is experienced by women and how it affects their voices, has not been subject to systematic research yet. For instance, research shows that women who experience forms of harassment are afraid to speak up because of eventual consequences (Naezer, van den Brink, & Benschop, 2019). However, this is crucial because the term voice relates to empowerment, especially when connected to speaking up about an issue (Kabeer, 1999).

Furthermore, voice can be defined as “the attempt to change rather than escape from an objectionable situation—contains the potential for transformation by bringing the self into connection with others” (Gilligan, 1988, p. 154). Additionally, the present work aims to contribute to the literature on resistance to gender equality. These insights also help to overcome resistance and to empower and support minorities in academia, which is important, because minorities are more innovative than the majority (Hofstra et al., 2020).

Furthermore, another under-researched issue concerns intersectionality. Research typically

treats women as if they were a homogenous group, although they differ in various factors such as

ethnicity and social status. These differences cause stratification within the minority group that is

commonly neglected by gender equality programs (Moughalian & Täuber, 2020; National

Academies of Sciences, Engineering, and Medicine, 2007). For instance, a survey from the

National Center for Education Statistics (2019) found that especially women of color are underrepresented in higher academic positions in the United States. Moreover, only limited research on intersectionality and about how different groups experience events differently exists (Amis, Munir, Lawrence, Hirsch, & Mcgahan, 2018; Wanelik, Griffin, Head, Ingleby, & Lewis, 2020). Addressing this gap, the current research examines the impact of intersectionality on women’s perception of resistance more systematically. Intersectionality will be operationalized as the number of dimensions on which academics occupy a minority status within their institution.

Together, the goal of this thesis is to investigate the prevalence of different resistance tools as perceived by women in academia and how resistance affects them. To shed further light on gender inequality in academia and to investigate this topic from the women’s perspective, this thesis aims to answer the following questions: Which forms of resistance do women in academia perceive and how does this affect them? Which role does psychological safety play? Does intersectionality affect the interrelations between these variables?

THEORY AND HYPOTHESES

Perceived Resistance to Gender Equality

As already mentioned in the introduction, diversity in academia still lags behind diversity in society, and especially the equal treatment of women and men in academia needs to be improved (Salinas & Bagni, 2017; Sensoy & Diangelo, 2017). However, an ignorance of the problem and resistance to change remain (Medina, 2013). People in key positions who fear that they might lose power, tend to display self-serving and corrupt behavior (Wisse, Rus, Keller, & Sleebos, 2019).

This appears especially in highly competitive environments such as academia. In order to explain

and understand why progress is slow or stalling despite universities’ commitment to gender

committed to (Mergaert & Lombardo, 2014). Thus, I build on the definition by Pardo del Val and Martínez Fuentes (2003) that resistance to change concerns “any phenomenon that hinders the process at its beginning or its development, aiming to keep the current situation” (p. 153).

To better understand the term resistance, the different responding strategies of institutions to processes need to be explained first. According to institutional theory, five different institutional strategies exist which summarize how institutions react to “pressures toward conformity with the institutional environment” (Oliver, 1991, p. 151). Those five strategies can be divided into acquiesce, compromise, avoid, deny, and manipulate and are applied by organizations when they feel pressured (Oliver, 1991). In the context of gender equality, the VicHealth report (2018) refers to eight different types of institutional resistance strategies which include “denial, disavowal, inaction, appeasement, appropriation, co-option, repression, and backlash”, furthermore, they vary from passive to active resistance (VicHealth, 2018, p. 4). The term passive resistance is connected to avoidance to change something (EIGE, 2019). Also, Oliver (1991) mentions avoidance in her institutional response strategies. Ahmed (2012) refers to this as non-performativity which relates to saying the right things but not acting upon them. Although diversity has been stated as being important, its essential implementation is missing. Universities have mission statements where they express being committed to diversity and equality, but those statements of commitments are not performances of equality themselves. Additionally, those commitments can fail to be implemented and often old patterns overrule and are reasons why new ones cannot be implemented. Thus, universities can create new policies; nevertheless, nothing happens (Ahmed, 2012).

Another response to equality measures is backfiring (Leslie, 2019), meaning that rather

than simply being inconsequential and ineffective, diversity measures can actually harm

underrepresented groups by creating backlash. An example for such backlash is the implementation of quotas in French universities: after their introduction, the number of women hired actually decreased (Deschamps, 2018). Other forms of resistance include the denial of the existence of a problem, the refusal to take responsibility, and appeasement (VicHealth, 2018). All of these resistance strategies are difficult to detect. When resistance becomes more active, it can also be detected more easily. These so-called active resistance strategies show actively that change is not accepted, for instance by not taking seriously what someone said or even actively attacking the person who asks for change (EIGE, 2019). Relatedly, Oliver (1991) mentions defying and manipulation as strategies that constitute active institutional response strategies.

Another example of active resistance is harassment (National Academies of Sciences, Engineering, and Medicine, 2018). Bowling and Beehr (2006) define the term workplace harassment “as interpersonal behavior aimed at intentionally harming another employee in the workplace” (p. 998). With this method, the harasser gives the affected person the feeling that he or she does not belong to the institution. Many different forms of harassment have been reported (National Academies of Sciences, Engineering, and Medicine, 2018). For instance, in a study conducted by Naezer et al. (2019) which deals with harassment in Dutch Academia, six types of harassment have been identified with the help of the 53 personal stories of women working in academia. These concern scientific sabotage, sexual harassment, physical and verbal threats, criticizing someone unfairly, exclusion, and the attribution of negative attributes to women.

However, many more forms of harassment which have not been mentioned in the report exist.

Consequently, it can be said that the continuum from passive to active resistance

culminates in backlash, which refers to actively attacking the conversation partner (VicHealth,

2018). For instance, Fernando and Prasad (2018) state that when people make organizations aware

of potential problems and unfair treatments, organizational actors like managers and HR often silence them and refer to them as whistleblowers. Ahern (2018) mentions in her research that although people do the right thing by speaking up about inequality and harassment, they are often traumatized because the institution blames them and gives them the feeling that he or she did something wrong. Institutions’ reactions to reporters of inequality and harassment, or misconduct in general, can be summarized the DARVO model, which is an abbreviation of Deny, Attack, and Reverse Victim and Offender. This model suggests that the harasser silences the victim by using psychological techniques like victim blaming and victim silencing (Harsey, Zurbriggen, & Freyd, 2017). If people complain about an issue, they often get silenced and referred to as whistleblowers (Fernando & Prasad, 2018). Therefore, if individuals are afraid that they are perceived as unprofessional and not worthy, low psychological safety could be a consequence (Edmondson, 2003).

Psychological Safety

The term psychological safety means that the individual has positive assumptions about how the other party might react when asking something or reporting a problem. If people experience an open environment, they feel more psychologically safe (Edmondson, 2004).

Furthermore, Roussin, and Webber (2011) mention that “psychologically safe work environments are rich in trust, they encourage risk-taking (i.e., vulnerability) among employees without concern for the negative impact on job standing or workplace reputation” (p. 325).

Research suggests that in institutions that do not ‘walk the talk’; for instance, when they talk about being committed to diversity and equality but do not deliver, employees feel abandoned.

The psychological contract between the employee and the organization gets violated when

Morrison & Robinson, 1997). If someone experiences unfair treatment such as sexual harassment, this can negatively affect the body and the mind (van Roosmalen & McDaniel, 1999). Relatedly, Barker (2017) mentions that if, for instance, someone gets mistreated, this can harm the person’s psychological safety. Discrimination, too, has been related to less trust and a lower feeling of safety, according to a Finnish study about discrimination and well-being (Castaneda et al., 2015).

Furthermore, research has found that perceiving discrimination can lead to more depression (Wei, Ku, Russell, Mallinckrodt, & Liao, 2008).

In sum, the above considerations suggest that experiencing unequal treatment and harassment and perceiving resistance to equality will negatively affect women’s psychological safety. Hence, I propose that:

Hypothesis 1: Perceived resistance has a negative effect on the psychological safety of women.

Psychological Safety’s Effects on Women’s Voices

To raise her voice and report eventual misconduct, it is essential that an individual experiences high psychological safety (Singh, Winkel, & Selvarajan, 2013). Additionally, it is salient that the individual has trust in the institution to feel psychological safety (Zhang, Fang, Wei, & Chen, 2010). Walumbwa and Schaubroeck (2009) suggest that feeling psychologically safe increases individuals’ voice because if the individual feels less exposed to risks, she is more likely to raise her voice and speak up. Furthermore, employees who feel psychologically safe are more likely to report institutional misconduct such as sexual harassment, unfair treatment or other unethical incidents (Edmondson, 2020; Singletary Walker, Ruggs, Taylor, & Frazier, 2019).

However, if people experience psychological unsafety, they will be afraid of telling their point of

view (Gao, Janssen, & Shi, 2011).

Liang, Farh, and Farh (2012) distinguish between two different types of voice: promotive voice and prohibitive voice, where the promotive voice is referred to as raising ideas, whereas prohibitive voice refers to speaking up about problems and issues. Yet, many times, people feel afraid to speak up and; therefore, stay silent because they are afraid of punishment from the organization (Lee, Heilmann, & Near, 2004). However, speaking up when feeling treated unequally is important, so that actions against discrimination can be undertaken (Sbrocchi, 2019).

Consequently, to be able to raise one’s voice, the experience of psychological safety is essential. This might become challenging for organizations; however, it is important for a positive work climate and the organization’s development (Edmondson & Lei, 2014). In addition, if employees feel good about the organization, they are more likely to want to stay with the company (Ghosh, Satyawadi, Joshi, & Shadman, 2013). Therefore, following up previous arguments about the importance of psychological safety for voice and taking women’s voices into account, the next hypothesis suggests that:

Hypothesis 2: Psychological safety has a positive effect on women’s voices.

Perceived Resistance and its Effects on Women’s Voices

As mentioned above, women who experience forms of harassment are scared to raise their voices because they fear possible effects of punishments (Naezer et al., 2019). This is particularly salient when the one who mistreats has more power than the one who gets mistreated. The one who gets mistreated might give in and not raise his or her voice (Harlos, 2010). Furthermore, McLaughly, Uggen, and Blackstone (2017) state that women who experience unfair treatment such as harassment are more likely to leave their career and look for different jobs elsewhere.

However, if people complain and raise their voices, the institution often retaliates in order

Genugten, 2013). Relatedly, the DARVO model which was introduced by Harsey, Zurbriggen, and Freyd (2017) suggests that institutions take advantage by using psychological tricks to silence the victims. Fernando and Prasad (2018) investigated this in their research about how women are silenced after they complained about sexual harassment. They propose that women in academia might not speak up because this could harm their career. Additionally, Hirschman (1970) mentions that often people are not speaking up because they hope things will get better. However, when people do not speak up when they experience unfair treatment, this can negatively impact their mental health (Cortina & Magley, 2003). Speaking up is thus important because it means standing up for what we think is right. Additionally, it can have positive effects on others, as well (Kelly, 1996).

With the #MeToo campaign, women have been invited to speak up about forms of harassment and abuse in the media industry; additionally, people also start making others aware of gender inequality in other work fields (Khubchandani, Kumar &, Bowman 2019; Palczewski, DeFrancisco &, McGeough, 2019). However, movements attempting to give women voice are often met with different silencing techniques which are met with different resistance tools including passive form such as silencing as well as active backlash (Daly, 1978; VicHealth, 2018).

Observing and experiencing such resistance decreases women’s willingness to speak up (Naezer et al., 2019). Building on the above, the next hypothesis suggests:

Hypothesis 3: Perceived resistance has a negative influence on women’s voices.

Resistance to Gender Equality as an Intersectional Problem

Finally, as already mentioned in the introduction, it is essential to note that women are not

a homogenous group and differ in ethnicity or social class (National Academies of Sciences,

intersectionality, which explains an intersection of different identities such as gender and ethnicity.

Furthermore the term intersectionality has been defined as “the interconnected nature of social categorizations such as race, class, and gender as they apply to a given individual or group, regarded as creating overlapping and interdependent systems of discrimination or disadvantage”

(Oxford English Dictionary, n.d., Definition 2). A study conducted by Miller and Roksa (2019) found that white men are seen as most privileged, followed by men with different backgrounds, followed by white women, and lastly accompanied by women with different backgrounds. In academia, this effect has recently been demonstrated in a Dutch business school in which white Dutch men earned more and were higher in rank than non-Dutch females (Bago d'Uva & Garcia- Gomez, 2020). This is reflected in the fact that women with underrepresented minority backgrounds, such as Afro-American or Hispanic descent, are less represented in academia (National Center for Education Statistics, 2019; Wyche & Graves, 1992).

Furthermore, women with different backgrounds tend to be less likely to receive tenure or become full professors (Wilds, 2000). Reasons for that include ineffective equality measures and intersecting experiences of discrimination based on sex and race (Lazos, 2012; Menges & Exum, 1983). Thus, women with different backgrounds face more difficulties than white women or their male peers. Yet, besides inequality often not being perceived by the privileged group (Townsend- Bell, 2020), the intersectional aspect of gender inequality (Crenshaw, 1991; Moughalian & Täuber, 2020) is under-researched, too.

The intersectional aspect; however, is important to consider because different groups

experience resistance differently (Townsend-Bell, 2020). For instance, Simons, Friedman, Liu,

and McLean Parks (2007) find that non-white people perceive a mismatch between words and

actions stronger than white people. Furthermore, a study conducted by Singh et al. (2013) found

out that non-white people tend to feel psychologically more unsafe than white people in a predominantly white settings when the organization does not value diversity. Additionally, this can have negative influences on psychological safety and performance. In a similar vein, Silverschanz, Cortina, Konik, & Magley (2007) state that sexual minorities feel less psychologically safe, often because of harassment. Beyond that, Hall, Hall, Galinsky, and Phillips (2019) report that women of color are seen as masculine and stereotyped as being aggressive.

Therefore, women of color are often labeled as the “angry black women”, which can affect them in the way they behave and psychologically. They might feel that they need to be more careful in how they act, so they do not confirm the negative stereotype. Furthermore, they often feel psychologically unsafe in a predominantly white environment (Ashley, 2013). Related to this, the term minority stress refers to “excess stress to which individuals from stigmatized social categories are exposed as a result of their social, often a minority, position” (Meyer, 2003, p. 675). Finally, when it comes to harassment in the workplace, women who belong to minority groups are more affected than women who do belong to the majority (Berdahl & Moore, 2006).

However, despite these insights, only limited research on this topic exists (Wanelik et al., 2020). Therefore, taking the role of intersectionality into account, I suggest that when minority women experience resistance, this also affects their psychological safety more negatively because literature suggests that minority groups experience more forms of harassment and in general feel less psychological safe than the majority (Ashley, 2013; Berdahl & Moore, 2006; Singh et al.

2013). Consequently, hypothesis 4 proposes that:

Hypothesis 4: Intersectionality accentuates the effect of perceived resistance on psychological

safety, such that experiences of psychological safety decrease more among employees with more

Linking together all hypotheses, the following overall hypothesis and conceptual framework is guiding my research:

Hypothesis 5 (complete model): The interactive effect of perceived resistance to equality and intersectionality on women’s voices is mediated by psychological safety.

Figure 1. Theoretical Framework

METHOD Participants and Procedures

To test the underlying hypotheses, I conducted a survey research combining quantitative and qualitative elements because as Driscoll, Appiah-Yeboah, Salib, and Rupert (2007) say

“qualitative data provide a deep understanding of survey responses, and statistical analysis can provide detailed assessment of patterns of responses” (p. 21). Additionally, to receive more information on this topic, I asked several open questions so that the participants could write more about their individual experiences with perceived resistance.

I pre-registered my research at Open Science Framework (OSF) under the link

https://osf.io/87sn9/?view_only=6e6f75c504ff4c2b907d65537e09fd00. Since I did not find a

namely discrimination, harassment, complaint management, and resistance. I choose different components to measure my independent variable because I assume that perceived resistance can take many forms (e.g., passive and active forms), and I wanted to measure many different aspects of this variable. Data collection started only after the pre-registration was set. Women who work in various academic fields were invited to participate in a survey about perceived resistance to gender equality. A priori power analysis conducted with G*Power (Faul, Erdfelder, Buchner, &

Lang, 2009), revealed that a minimum of 54 participants was required, given three predictors (resistance, intersectionality, and psychological safety) to test the hypotheses. We thus set 54 respondents as the minimum number to collect, but aimed to collect as many responses as possible within a timeframe of 6 weeks. Women from various international universities were contacted, including women from the Netherlands, Germany, Austria, and other countries through mine and my supervisor’s personal and professional network. Potential participants were contacted via email and were asked to participate in the survey. Before filling out the survey, participants needed to fill out a consent form where their rights were explained. Only participants who clicked “yes” on the consent form were forwarded to the questions. If participants clicked “no” they were forwarded to the last page without filling out the questions.

Furthermore, the anonymity of the participants was ensured. However, if the participants

wanted to receive the study results, they could voluntarily share their email addresses at the end of

the survey. Any personal data was deleted after receiving the data, and the email addresses were

stored in a separate file to ensure anonymity. In total, 204 people started the survey, but only 100

filled it out completely. From the 100 completed questionnaires, we only used responses from

women, resulting in a final data set of 91 respondents. These 91 respondents belonged to 28

different nationalities, with most from the Netherlands (28.7%), followed by Austria (13.2%), and

Germany (12.1%). As a current place of residence, respondents named the Netherlands (47.3%), followed by Austria (14.3%), and the United Kingdom and Northern Ireland (6.6%). Respondents worked in various universities and scientific fields, with most of them from Social and Behavioral Sciences (45.1%), followed by Business and Economics (22%), and Arts and Humanities (13.2%).

Additionally, 45% of the respondents indicated having tenure, 50.5% indicated not having tenure, and 3.3% did not wish to disclose this information. Additional information about the demographics can be found in the Tables 1 and 2 in Appendix A.

Measures

Unless indicated otherwise, all items were measured on a 5-point Likert scale where 1 means very uncommon or strongly disagree, and 5 means very common or strongly agree. The complete questionnaire can be found in the Appendix C.

Demographic factors were measured with questions at the beginning of the survey.

Participants were asked to indicate their gender, nationality, and current place of residence.

Furthermore, participants were asked to fill in the university they were currently affiliated with, their position, their academic field, and the years they were already working in academia. If participants did not wish to disclose their demographic information, they could also choose the option “do not wish to disclose” for each demographic question.

Intersectionality. Participants were asked to tick off boxes containing factors on which

they differ from the majority of the people they work with. These factors included, for instance,

race, ethnicity, sexual orientation, disability status, and religious affiliation. Furthermore, they

could select other options and mention other dimensions themselves. The options were adopted

from Athena Survey of Science, Engineering and Technology (2016). Participants indicating

differing from the majority on more factors were considered as more intersectional compared to participants differing from the majority on fewer factors.

Perceived Resistance consists of four different components namely Discrimination,

Harassment, Complaint Management, and Resistance based on questions adapted from the Athena Survey of Science, Engineering and Technology (2016), Naezer et al. (2019), Svensson and Genugten (2013), and VicHealth (2018). An exploratory factor analysis (EFA) showed that, in line with my assumption that resistance is reflected in different ways, the items of the respective scales did not load on a single factor. Consequently, I treated them as separate components.

The first component that was used to measure perceived resistance is called Discrimination (15 items, α = .90). Furthermore, an open question addressed the attributes that are needed to be privileged in resource allocation. However, this open question could be answered voluntarily.

The second component, Harassment (6 items, α = .90) was adopted from Naezer et al.

(2019). Additionally, participants could mention other forms of harassment and rate them accordingly. In order to facilitate a shared understanding of the terms correctly, I provided definitions before presenting the items.

To measure Complaint Management, the items of Svensson and Genugten (2013) on retaliation were adapted to the context of (gender) equality. Example questions were “Being criticized for complaining about gender inequality” or “Being denied an opportunity for a deserved training” (12 items, α = .95).

Furthermore, the component Resistance was used to measure perceived resistance more explicitly. Therefore, descriptions of resistant behaviors of the institution and examples were provided (8 items, α = .88). Those descriptions and examples were taken from VicHealth (2018).

Additionally, an open question was provided which encouraged participants to mention other

forms of perceived resistance, harassment, discrimination, or other examples and experiences.

High values on this component mean that participants experienced resistance when complaining about inequality.

Psychological Safety was measured with seven items adopted from Edmondson (1999; α

= .82). Example questions are “I am able to bring up problems and tough issues.” and “It is safe to take a risk on this team.” Three items were reverse coded. High values on this scale indicate that participants feel psychologically safe in their institutions.

Voice was divided into two components, namely Voice Attitudes and Voice Behavior. I also

conducted an exploratory factor analysis and the components did not load together and therefore, I decided to operationalize the two components separately in the model.

Voice Attitudes was measured with 10 items adapted from Liang et al. (2012; α = .95). The items tapped into a promotive and prohibitive voice, referring to bringing up new ideas and speaking up about problems and issues, respectively. Example questions were “I proactively develop and make suggestions for issues that may influence (gender) equality in my working environment” and “I dare to voice out opinions on (gender) inequality, even if that would embarrass others.” In the following paragraphs this was called Voice Attitudes. High values on this scale mean that participants speak up about inequality.

Furthermore, Voice Behavior was measured with four self-developed items concerning

voice and career (α = .80). Example questions are “I consider changing careers” and “I would

recommend other women to work in academia.” Two items were reverse coded. High values on

this scale indicate that women like to work in their institutions and in academia and would like to

recommend this to other women too.

Open Questions. Effects on career were adapted from the Athena Survey of Science,

Engineering and Technology (2016). The question dealt with possible effects of various social identities on their careers. Furthermore, participants could add dimensions that have effects on their careers and rate them accordingly. Participants had the opportunity to add other forms of harassment that have not been mentioned yet, that they have experienced themselves or witnessed.

To introduce the fourth component complaint management, the question “Have you ever complained about harassment or discrimination, or do you know of other who have done so?” was asked, where participants could choose the options that they have either complained themselves, know others, or know no one who complained about (gender) equality. In addition, at the end of the component complaint management, an open question gave the participants the opportunity to elaborate more on their personal experience with retaliation. If answers were not completed in English, I translated them accordingly.

Statistical analyses. First, I calculated the descriptives of my components. Second, I ran

the correlational analyses, to show the correlations for my variables and third, I ran the moderated

mediation model using Hayes Process Model 7 (2018; 5000 bootstrap samples; predefined) eight

times in total, to test all of my components separately. The independent variable is Perceived

Resistance with its components discrimination, harassment, complaint management, and

resistance, Psychological Safety is the mediator, the dependent variable is Voice with its

components voice attitudes and voice behavior. Furthermore, Intersectionality serves as the

moderator.

RESULTS

Preliminary Results

Descriptive Statistics and Correlations

Regarding intersectionality, 21 respondents indicated they do not differ from the majority they work with, 32 respondents indicated differing on one dimension from the majority they work with, and 38 participants indicated differing from the majority on more than one dimension. Table 3 and 4 provide an overview of means, standard deviations, deviations from the respective scale midpoint of each construct, and intercorrelations for the variables used in this study. Table 3 shows the correlations of women who ticked off one or more dimensions they differ from the majority referred as intersectional. Table 4 shows this for women who ticked none of the dimensions;

therefore, they are referred to as non-intersectional.

For women, who mentioned, they differ from the majority (table 3) on one or more dimensions, one finding is that the components of perceived resistance (Discrimination, harassment, Complaint Management, and Resistance) each correlate negatively with Psychological Safety. This could be interpreted as support for the notion that experiences of discrimination, harassment, resistance, and retaliation following complaints about inequality is associated with lower levels of Psychological Safety.

Furthermore, an interesting finding is that for intersectional women (table 3) Psychological

Safety positively correlates with Voice Behavior, suggesting that feeling psychologically safe

makes women less likely to change careers, more likely to stay in their institution, and to

recommend working in academia to others. A correlation for non-intersectional women (table 4)

has not been found.

Another interesting finding is that for intersectional women (table 3), all components of perceived resistance have each a positive correlation with Voice Attitudes. This could mean that if women experience unfair treatment, they are more likely to raise their voices. However, for intersectional women (table 3) Discrimination, Complaint Management and Resistance relates negatively to Voice Behavior, which suggests the opposite, namely that perceiving resistance is associated with women more likely to leave the institution and not recommending others to work in academia. Those correlations with the voice components have not been found for non- intersectional women (table 4).

Additionally, an important finding is the negative correlation between Intersectionality (table 3) and Psychological Safety, which suggests that the more intersecting disadvantages a woman reports, the less psychologically safe she feels. Furthermore, Intersectionality is negatively related to Voice Behavior, which can be interpreted as the more intersectional someone is, the more likely this person wants to leave the institution and do not want to recommend others to work there. However, this might of course be a result of these women experiencing more inequality.

Additionally, for intersectional women (table 3) all components of perceived resistance are positively correlated with each other. This could be interpreted as an interrelation between all of the components which could indicate if someone scores high on one component, this person might score high on the others.

.

Table 3. Descriptive Statistics and Correlations: Intersectionals (>=1)

Notes. ^†p < .10, ^*p < .05, ^**p < .01. Significant t-test statistics indicate that the observed mean is significantly above (positive value) or significantly below (negative value) the scale mid-point.

Table 4. Descriptive Statistics and Correlations: Non-Intersectionals (<0)

Notes. ^†p < .10, ^*p < .05, ^**p < .01. Significant t-test statistics indicate that the observed mean is significantly above (positive value) or significantly below

Variables N M SD t 1 2 3 4 5 6 7 8

1. Discrimination 2. Harassment

3. Complaint Management 4. Resistance

5. Intersectionality 6. Psychological Safety 7. Voice Attitudes 8. Voice Behavior

70 70 70 70 70 70 70 70

3.80 3.07 3.33 3.53 1.97 3.13 3.70 3.09

.62 .99 .93 .88 1.17

.85 .84 1.04

1.23 ^**

.57^**

.83^**

1.03^**

-.53^**

.63^**

1.15^**

.63^**

1 .40^**

.52^**

.55^**

.22^† -.35^**

.34^**

-.26^*

1 .73^**

.65^**

.31^**

-.64^**

.31^**

-.23 1 .69^**

.27^* -.59^**

.29^* -.37^**

1 .29^* -.55^**

.43^**

-.33^**

1 -.29^*

.25^* -.12

1 -.10 .34^**

1

.23 1

Variables N M SD t 1 2 3 4 5 6 7

1. Discrimination 2. Harassment

3. Complaint Management 4. Resistance

5. Psychological Safety 6. Voice Attitudes 7. Voice Behavior

21 21 21 21 21 21 21

3.32 2.90 2.87 3.31 3.40 3.54 3.51

.52 1.00

.97 .56 .83 .72 1.03

.82^**

.40^† .37^† .81^**

.90^**

1.04^**

1.01^**

1 .65^**

.43^† .42^† -.32 .21 -.32

1 .70^**

.30 -.42^†

.36 -.25

1 .21 -.54^* .16 -.20

1 .03 .42^† -.05

1 -.05 .56^**

1 .20 1

Main Analysis

To analyze the underlying hypotheses, I tested the predicted moderated mediation model using Hayes Process Model 7 (2018; 5000 bootstrap samples; predefined). Because the dependent variable was measured by four different components, I ran the model four times to test whether the interactive effect of resistance and intersectionality on voice attitudes is mediated by psychological safety, and four times to test whether the interactive effect of Perceived Resistance and Intersectionality on Voice Behavior is mediated by Psychological Safety. All variables were z-standardized for the purpose of the analysis. Table 5 gives an overview over the results of the analyses. Models 1-4 show the results for the relationships between the components of the independent variable Perceived Resistance and the mediator Psychological Safety. Models 5-12 show the results for the relationships between the independent variable, the mediator and the components of the dependent variable Voice.

Table 5. Regression Analyses

Mediator Variable Model: Psychological Safety

Model 1 Model 2 Model 3 Model 4

Intercept .06 (.10) .06 (.10) .05 (.90) .07 (.10)

Independent Variables

Discrimination -.29 (.10)**

Harassment

Complaint Management -.59 (.09)**

Resistance -.40 (10)**

Moderator

Intersectionality -.14 (.10) -.08 (.09) -.06 (.10) -.08 (.10)

Interaction -.17 (.09)† -.15 (.09) -.16 (.10) -.26 (.11)*

Dependent Variable Model: Voice Attitudes

Model 5 Model 6 Model 7 Model 8

Intercept .00 (.10) .00 (.10) .00 (.10) .00 (.09)

Independent Variables

Discrimination .33 (.11)**

Harassment .42 (.12)**

Complaint Management .33 (.13)*

Resistance .50 (.11)**

Mediator

Psychological Safety .03 (11) .16 (.12) .10 (.13) .14 (.11)

R

10.39% 12.06% 07.91% 20.32%

Model 9 Model 10 Model 11 Model 12

Intercept .00 (.10) .00 (.10) .00 (.10) .00 (.10)

Independent Variables

Discrimination -.19(.10)†

Harassment -.003 (.12)

Complaint Management -.18 (.12)

Resistance -.14 (.11)

Mediator

Psychological Safety .33 (.11)** .40 (.12)** .30 (.12)* .33 (.11)**

R

19.05% 15.97% 17.99% 17.58%

Notes. Standard Errors between parentheses ^†p < .10, ^*p < .05, ^**p < .01, z-standardized data, Interaction refers to the Interaction effect of the components of the independent variable with the moderator intersectionality.

Hypothesis 1 stated that perceived resistance will undermine women academics’ feelings of psychological safety. Models 1-4 in Table 5 support this prediction by revealing that the association with psychological safety is negative and significant for each component of perceived resistance (model 1: t = -2.84, CI

-.496, -.087; model 2: t = -6.59, CI

-.741, -.398; model 3: t = -6.43, CI

-.770, -.407; model 4: t = -4.16, CI

-,584, -.206, respectively). Hypothesis 1 is thus supported.

Hypothesis 2 stated that Psychological Safety has a positive effect on Women’s Voice. The

regression equations in models 5-8 show that the relationship between Psychological Safety and

the component Voice Attitudes are not significant (model 5: t = 0.25, CI

-.188, .242; model 6: t

= 1.25, CI

-.092 .404; model 7: t = 0.79, CI

-.152, .3510; model 8: t = 1.27, CI

-.077, .350).

However, for Psychological Safety and Voice Behavior, the regression equations are significant and in the predicted direction (model 9: t = 3.22, CI

.127, .536; model 10: t = 3.26, CI

.155, .640; model 11: t = 2.48, CI

.059, .534; model 12: t = 3.06, CI

.117, .550). Summarizing this, hypothesis 2 is partly supported.

Hypothesis 3 predicted a negative direct effect of Perceived Resistance on Voice. The direct effect of the components of Perceived Resistance on Voice Attitudes was significant (model 5: t = 3.06, CI

.116, .547; model 6: t = 3.35, CI

.169, .665; model 7: t = 2.59, CI

.077, .579;

model 8: t = 4.64, CI

.284, .711). However, the direct effect of the components of Perceived Resistance on Voice Behavior was non-significant (model 9: t = -1.83, CI

-.393, .016; model 10:

t = -0.03, CI

-.246, .238; model 11: t = -1.47, CI

-.413, .062; model 12: t = -1.31, CI

-.360 .074). Hypothesis 3 proposes that perceived resistance has a negative influence on women’s voices and our results for Voice Attitudes show the opposite. Therefore, hypothesis 3 needs to be rejected for Voice Attitudes. Additionally, for Voice Behavior, hypothesis 3 needs to be rejected as no direct effect was observed. Based on these results, hypothesis 3 needs to be rejected.

Hypothesis 4 stated that Intersectionality accentuates the effect of perceived resistance on

psychological safety, such that the negative impact of perceived resistance on psychological safety

is more pronounced for employees with more minority attributes. As can be seen in Table 5, the

interaction term indicating moderation was only (marginally) significant for Discrimination and

Resistance. The interaction between Perceived resistance and Intersectionality was marginally

significant for Discrimination (t = -1.80, CI

-.352, .017). The conditional effects show that the

effect of experienced Discrimination on Psychological safety is significant only for average and

high levels of Intersectionality, with the effect being strongest for high levels of Intersectionality (β = -.23, t = -2.03, p = .045, CI

-.446, -.005; β = -.48, t = -3.38, p = .001, CI

-.762, -.197, respectively).

The interaction between Perceived Resistance and Intersectionality was marginally significant for resistance (t = -2.38, CI

-.470, -.042). Furthermore, the conditional effects show that the effect of experienced resistance on psychological safety is significant only for average and high levels of intersectionality, with the effect being strongest for high levels of intersectionality (β = -30, t = -2.75, p = .007, CI

-.508 -.082; β = -.68, t = -4.64, p < .001, CI

-.975, -.390, respectively). Based on these results, hypothesis 4 is supported, but only for the components Discrimination and Resistance.

Hypothesis 5 combines the whole model and stated that the interactive effect of perceived resistance to equality and intersectionality on women’s voices is mediated by psychological safety.

Since moderation was only found for Discrimination and Resistance with the dependent variable Voice Behavior, we only expect moderated mediation for these two components. Indeed, the moderated mediation index was significant (index = -.06, CI

-.143, -.003) for Discrimination, showing that Psychological Safety mediated the effect of Discrimination on Voice Behavior only for women academics scoring high on intersectionality (CI

-.308, -.051), but not on average (CI) and low (CI) levels of Intersectionality.

Regarding Resistance, the moderated mediation index was significant (index = -.09, CI

- ,179, -.014), too, showing that Psychological Safety mediated the effect of Resistance on Voice Behavior only for women academics scoring average and high on Intersectionality (0, CI

-.242;

-.017; +1 SD, CI

-.413, -.094, respectively), but not on low (CI) levels of Intersectionality. Based

on these results, Hypothesis 5 is supported for Discrimination and Resistance: women academics

who differ from the majority on multiple dimensions (i.e., who score average and high on intersectionality) suffer from greater impairment of their psychological safety when they are discriminated against and when they perceive resistance against equality at work. In turn, the lower psychological safety is associated with greater willingness to leave the institution and academia.

Thus, the effect of Discrimination and Resistance on Voice Behavior is fully mediated by Psychological Safety.

Open Questions

In addition to filling out the survey questions, I gave participants the opportunity to express their own experiences and thoughts to certain questions. Accordingly, this section deals with the open questions that have been asked to participants voluntarily. Furthermore, the effects on career are displayed and discussed, and people could mention other forms of harassment and retaliation they have experienced themselves or witnessed. I conducted a thematic analysis because this is a useful approach to analyze qualitative data (Braun & Clarke, 2006). I coded the qualitative data and created themes. Furthermore, the full answers can be found in the Appendix B.

Effects on Career

At the beginning of the survey, participants needed to select whether certain factors

affected their careers positively or negatively. Eleven options were given: race or ethnicity, country

of origin, legal sex, age, sexual orientation, gender identity, disability status, religious affiliation,

caring responsibilities, and marital or civil partnership status. However, the participants could

mention other dimensions that have not been listed before, as well. Participants had the opportunity

to tick off the neutral option, where the dimensions either negatively or positively affect their

career. More than half of the participants mentioned that their sex harms their career. Additionally,

more than one-third of the participants mentioned that their age harms their career. Furthermore, over one-third of the participants indicated that their gender identity influences their career negatively. Personality, pregnancy, not being willing to work full-time, being the first/only female working in the job, being a first-generation student, appearance, network (not belonging to the inner circle), speaking another language, and political orientation negatively impact some participants’ careers.

Many participants who identified as being similar to the majority in their workplace reported positive effects of the dimension on their career. For instance, almost half of the participants indicated that their race positively impacted their careers. Also, almost half of the participants mentioned their country of origin as a positive contributor to their career. Furthermore, more than one third indicated that their social class has a positive impact on their career.

Participants also mentioned their network, knowledge of the local language, international experience, working experience, political orientation, and having a senior male mentor as contributing to their career.

Figure 2. Effects on Career

Attributes of People Working in Academia

Participants could fill in up to eight empty boxes, mentioning attributes that they would ascribe to people working in academia. I sorted them and created categories which are presented in figure 3. In general, 291 attributes were collected. Most of the respondents mentioned male, followed by attributes that relate to academic success and network. Furthermore, many participants ascribed attributes such as being outspoken/assertive, being white, attributes about commitment, and local nationality to working in academia.

Figure 3. Categories of Attributes to be most favored/privileged in Resource Allocation in Academia

45

36 32

27

24 24 24

14 13 10

7 7 6 5 4 3 2 2 2 2 1 1

Male

Academic Success Network

Outspoken/Assertive White Com

mitment Local Nationa

lity Age Social Class

Local Langua ge

No Caring

Responsibilities Social Skills

Egoi

stic Behavior Privilege

Authority/Independence Appe

arance Working

Time Morality

None -Morality

Old Patterns Christian

Heterosexua l

N u m b er of A tt ri b u te s

Categories of Attributes to be most

favored/privileged in Resource Allocation in

Academia

Perceived Resistance

Harassment

Twenty-five other forms of harassment have been added that were not previously mentioned. Respondents reported harassment in terms of who is targeted (e.g., racism, gender- based harassment, language-based harassment) and behavior (e.g., exclusion, being made invisible, aggressive forms of harassment). Examples for targets are “discrimination in hiring - racial bias in what constitutes hiring criteria” (racism), “gender imbalance in higher positions”

or “sexism” (gender-based harassment), and not speaking the local languages (language-based harassment). Furthermore, examples for harassment behaviors are “withholding of relevant information” or “being treated with different expectations” (exclusion), not being seen and heard (being made invisible). More aggressive forms of harassment were also reported, such as “physical harassment” or “unnecessary phone calls late at night or during holidays” (aggressive forms of harassment).

Complaint Management

Concerning the question if participants have complained about discrimination or

harassment, or know of others who have complained, 31 respondents indicated having complained

about harassment or discrimination themselves, 55 indicated knowing of others who have

complained, and 25 respondents indicated not having complained about harassment or

discrimination and not knowing people who have done so. Participants could tick off more than

one option. This means over two-third of the women academics participating in our study either

complained themselves or know someone who has complained about discrimination or

harassment.

Furthermore, one open question asked what other forms of retaliation women had experienced after complaining about harassment or discrimination. Respondents mentioned exclusion, reputation damage, advantage being taking of one's work, unfair treatment, unpleasant confrontation, and threats. Examples for exclusion are “being excluded from research work”, a denied promotion and “withholding resources (grant application, withholding of required equipment).” An example of reputation damage is “being associated with unpleasant.”

Furthermore, an example for advantage being taken of one's work is “others taking credits for my job.” An example of unfair treatment is “only one person graded me for a research job, she graded me badly because she did not like me.” An example of unpleasant confrontation is “someone complained, was confronted to face the bully, was stressful.” Different examples of threats were

“threats of sullying reputation”, “being blamed and blackmailed to apologize” and “threats to engage in dismissal procedure.”

General Perceived Resistance

At the end of the questionnaire, the participants could write down their thoughts about perceived resistance, (gender) equality and speaking up about (gender) inequality. I received in total 24 answers for this general open question. I coded the answers and then categorized the answers into five different themes, namely, unfair treatment, non-performativity, being silent about gender equality, speaking up about inequality, and encouragement and change.

Unfair Treatment

The category of unfair treatment deals with the unfairness that women have seen or

experienced. Eight respondents referred to this category explicitly. For instance, one participant

mentioned that men often get credits for the females’ work. Additionally, two other participants

another participant stated that she has “experienced age discrimination towards female staff members.” Another participant reported that she knows others who experienced discrimination and harassment. Furthermore, she reported that she is “not taken seriously” because she works in gender studies, adding “I also know that pursuing a career in academia is going to be more difficult for me because of my gender.” Another participant wrote that she was denied a Ph.D.

position because the supervisor was afraid that she would get pregnant. Furthermore, this participant referred to the questionnaire and stated: “I had to give the neutral option while sometimes I was not neutral, I was oblivious.” Another participant mentioned that someone could get discriminated against because they cannot speak the language fluently.

Non-Performativity

The category non-performativity deals with the non-performativity of the university.

Although universities know that inequality is an issue, they do not do enough against it. Six

participants gave answers that fit to this category. One participant said that there are still not

enough daycare possibilities for children and not enough part-time arrangements. She also

mentioned that “the university (and the academic world) almost feels too rigid to make any change

in this, and that's where these issues get stuck.” Furthermore, another participant said that “gender

inequality is often shifted to the 'individual' level and is not perceived or discussed as a structural

problem.” Additionally, another participant mentions that the university cares about the issue, but

no actions take place. A different participant said: “In our institution, gender equality is just a pro-

forma subject to the chancellor, he is very misogynistic; there is no useful board to control the

chancellor.” Also, another respondent mentioned that “one confidential advisor, who is not an

international, is not effective to address the needs of international students and staff.” Another

participant pointed out that a gap still exists between the saying that “women have never had it so good” and women actually working in the business.

Speaking up About Inequality and Complaint Management

This category deals with participants who raised their voice about inequality and experienced an inappropriate complaint management system of their universities. Three participants reported answers that fit this category. For instance, one participant mentioned that she had experienced a lot of sexual harassment at her university; however, there “are no protocols in place for reprimanding the perpetrators.” Furthermore, she had been bullied and harassed;

however, “the University failed to use appropriate procedures to handle the case.” Additionally, she mentioned that “a victim of bullying and harassment should, for instance, never be put in a situation of mediation with the perpetrator - that will only instill the bullying further and create a triggering situation.” This respondent said that in her university, that the incidents are handled like an “equal power relationship” is commonplace in her university, although the opposite is the case.

Another participant said that in her institution, “the head of our department did not deny the claims of sexual harassment brought up by one of the female academics so he showed to know them.” Furthermore, she complained to the equality office about it, but nothing happened. Another participant said she needed to sign a non-disclosure agreement after reporting sexual harassment;

furthermore, everything was covered up. She added that a female dean had her sign the non- disclosure agreement. She said if you open up about this inequality, you lose tenure.

Silence and Anxiety

This category deals with women keeping silent because they were afraid to speak up about