The Effect of a Diversity Statement in
Recruitment Materials on Job Seekers’
Organizational Attractiveness
U
NIVERSITY OF
A
MSTERDAM
Faculty of Economics & Business
Master of Science in Business Administration
Entrepreneurship (Joint Degree with Vrije Universiteit Amsterdam)
Author: Philip Bootsma
Email: Philipbootsma@hotmail.com
UvA ID: 10575146
VU Student ID: 2540500
Thesis Supervisor: Dr. Yuval Engel Application Date: 9th of January, 2018 Submission Date: 16th of August, 2018
Version: Final Version
i
STATEMENT OF ORIGINALITYThis document is written by Philip Bootsma who declares to take full responsibility for the contents of this document. I hereby declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
ii
ACKNOWLEDGEMENTI would like to express my special thanks of gratitude to my Master thesis supervisor, Dr. Y. Engel, for his assistance during this Master’s thesis project. Finalizing this paper within the deadline with this quality would not have been possible without the dedicated involvement, fast email responses, and assistance through every step of the process by Dr. Engel. The discussions between me and Dr. Engel where particularly helpful during the data analysis, which assisted me in narrowing down the pre-testing analysis and variables to use. Furthermore, any question could be asked during the supervision and would be answered elaborately. Therefore, I would like to thank Dr. Y. Engel for the support during my Master’s thesis project.
iii
ABSTRACTPurpose: Prior research revealed that entrepreneurship is commonly associated with masculine traits and women are underrepresented in this field of work. However, knowledge is lacking about the underlying processes that form the basis of these impressions and deter females from entering entrepreneurial activities. Therefore, this research combines role congruity, stereotype activation and signaling theories to propose that employee recruitment materials in start-ups are biased as they favor male–but not female–job seekers’ fit-perceptions with start-up jobs. We further suggest an intervention that could ameliorate this bias and enhance female attraction to start-up jobs.
Design/Methodology/Approach: Using an online experiment among a sample of 173 job seekers, we randomly assigned participants to either view an entrepreneurial job advertisement or an identical job advertisement that contained an additional “diversity statement”. This design examined whether a diversity statement intervention may reduce the effect of biasing signals in start-up job ads and the extent to which job seekers’ gender moderates the indirect effect of job advertisement signals through perceived proportion of females and anticipated belongingness, on organizational attractiveness.
Findings: The analysis of the moderated mediation model indicated that the model is not significant, except for the previously confirmed relationship between anticipated belongingness and organizational attractiveness. This means that the diversity statement did not influence perceived proportion of females, which in turn did not influence anticipated belongingness Originality/Value and Implications: This thesis adds support to the well-researched relationship between anticipated belongingness and organizational attractiveness (e.g., Gaucher et al., 2011; Lievens, Hoye & Anseel, 2007). It further found support for the assumption that entrepreneurship is associated with masculinity (Gupta et al., 2009; Short et al., 2010) and provided support for signaling theory (Rynes et al., 1991) as job seekers used the limited information to provide the organization’s proportion of female employees. Furthermore, investigating the model’s failure has shown that culture is important, due to its effect on mental definitions of individuals about the concept of entrepreneurship and its relation to gender, which can greatly differ between cultures (e.g., Lee & Peterson, 2000; Levenburg & Schwarz, 2008; Lim & Envick, 2013). More, job specificity has to be taken into account, as vagueness could lead to reduced attention towards the reading material or even loss of retention (e.g., Feldman, Bearden & Hardesty, 2006; Mason & Belt, 1986; Roberson, Collins & Oreg, 2005).
iv
TABLE OF CONTENTS STATEMENT OF ORIGINALITY ... I ACKNOWLEDGEMENT ... II ABSTRACT ... III INTRODUCTION ... 1THEORETICAL FRAMEWORK AND HYPOTHESES ... 3
JOB SEEKERS’PERCEPTION ON ORGANIZATIONAL PROPORTION OF FEMALE EMPLOYEES ... 4
JOB SEEKERS’DRIVE TO BELONG ... 6
THE RELATIONSHIP BETWEEN ANTICIPATED BELONGINGNESS AND ORGANIZATIONAL ATTRACTIVENESS ... 7
THE EFFECT OF A DIVERSITY STATEMENT IN JOB ADVERTS ON ORGANIZATIONAL ATTRACTIVENESS ... 9
THEORETICAL MODEL ... 9 METHODS ... 9 RESEARCH DESIGN ... 9 SAMPLE SELECTION ... 10 EXPERIMENTAL CONDITIONS ... 11 PROCEDURE ... 13 MEASUREMENTS ... 13 RESULTS ... 15
EXPLORATORY FACTOR ANALYSIS ... 15
DESCRIPTIVE STATISTICS AND INTERCORRELATIONS ... 16
RANDOMIZATION CHECK ... 16
MEASUREMENT MODEL VALIDATION ... 18
HYPOTHESIS TESTING... 18
DISCUSSION ... 21
INTERPRETATION OF FINDINGS... 22
THEORETICAL IMPLICATIONS ... 24
LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH ... 25
CONCLUSION ... 27
REFERENCES... 28
1
INTRODUCTIONEntrepreneurship is commonly related to stereotypical characteristics such as innovativeness, risk-taking, aggressiveness, and competitiveness, which are found to fit the male gender role better (Gupta & Bhawe, 2007; Gupta, Turban, Wasti & Sikdar, 2009). The stereotypical traits attributed to females and males in society are reflected in the ordering of positions, roles, organizations and occupations as feminine and masculine (Gupta, Turban & Bhawe, 2008; Gaucher, Friesen & Kay, 2011). This means that male employees working in a start-up would behave in accordance to their male gender stereotype. Females, on the other hand, have to overcome self-doubt, insecurity (Thébaud, 2010) and strong social gender stereotypes (Gupta et al., 2008; Gupta et al., 2009), which makes the choice to work in entrepreneurial environments more difficult.
Job seekers are not provided enough information about an advertising start-up so they make use of any available information or element in the job advertisement, such as lay-out, tone, and wording, to build a perception of the venture’s environment (Lievens & Slaughter, 2016; Rynes, Bretz & Gerhart, 1991). This phenomenon, called signaling theory, was originally developed by Spence (1973). Gaucher et al. (2011) had observed that gender stereotypical wording in advertisements influences job seekers expectations of the occupation’s gender majority and consequently the firm’s gender diversity. Thus, when start-ups that are recruiting new employees with entrepreneurial words in their job advertisement, such as visionary, inventor, initiator, pursuing, striving and capitalize (Short, Broberg, Cogliser & Brigham, 2010), they signal that the job is better fit to male job seekers. This ultimately leads to entrepreneurship being male dominated.
However, pursuing diversity is crucial for organizations, including start-ups, as studies suggest that gender diverse workforces increase employee motivation and customer satisfaction (Desvaux, Devillard-Hoellinger & Baumgarten, 2007) and creates higher levels of creativity, problem solving and innovation (Campbell & Mínguez-Vera, 2008). Fortunately, research has shown that different approaches can increase gender diversity in order to profit from diversity-related benefits (Flory, Leibbrandt, Rott & Stoddard, 2018). Evidence suggests that changes in remuneration structures (Niederle & Vesterlund, 2007), mentoring and advice institutions (Rodriguez-Planas, 2012), and preferential treatments such as quotas (Niederle, Segal & Vesterlund, 2013) can influence job applicants that will have a lasting impact on a firm’s gender and ethnic diversity. However, these methods have been costly as Fortune 500 companies have gone as far as spending $200 million or more each year to enhance the diversity among
2
employees in their firm (Flory et al., 2018). In total, estimations suggest that companies have been spending close to $10 billion each year on inclusion and diversity trainings (Hansen, 2003).Start-ups have to rely on different diversity approaches in their recruitment as they do not have the same level of financial power as Fortune 500 firms. Luckily, research has shown a more cost-effective strategy in the implementation of a diversity statement in a job advertisement to highlight the firm’s valuation on employee diversity (Flory et al., 2018). In the study of Flory et al. (2018) they show that a diversity statement can be used to raise interest and application rates by ethnic minorities in the applicant pool. Their study mainly focused on ethnic minority but they do acknowledge females were influenced too. One important thing to highlight is that the diversity statement did not only make more minorities interested, but the proportion of the minorities that were actually selected for the job shows that more strong candidates applied.
In summary, gender stereotypes and signaling theory explains that the entrepreneurial wording in start-ups job advertisements tends to lead to more male than female applicants, while it would be more beneficial to evenly recruit male and female employees. Strategies employed by Fortune 500 companies have been rather expensive, which is not an option for start-ups. Therefore, the diversity statement could be a cost-effective instrument to increase workforce diversity. As a consequence, this study investigates the effect of a diversity statement in an entrepreneurial job advertisement on job seeker’s organizational attractiveness. The following research question will be leading the remainder of this thesis:
To what extent does a diversity statement in an entrepreneurially worded job advertisement influence male and female job seekers’ organizational attractiveness? To address this question, a conceptual model has been developed with the assistance of signaling theory (Spence, 1973), role congruity theory (Eagly & Karau, 2002), and stereotype activation theory (Gupta, Turban & Pareek, 2013). This has been tested using an online experiment in which participants (N = 173; 73 males and 100 females) were randomized and asked to read and rate one of two otherwise identical job advertisements, which only differ in the inclusion or exclusion of a diversity statement.
This thesis will have a variety of theoretical and practical contributions. First off, this experiment aims to uncover job seekers’ reactions and perceptions towards an advertising company, produced by reading the base-entrepreneurial or intervention-diversity job advert and has been tested through multiple established measures, which could aid to improve start-up hiring practices. Hereby contributing to the literature in the field of recruitment and human resources (HR) in entrepreneurship by focusing on the start-up sector (e.g. Moser, Tumasjan,
3
& Welpe, 2017). Secondly, by only providing the raw job advert, participants had limited amount of information available and thus had to use signaling theory to formulate perceptions about the recruiting organization (i.e. perceived proportion of females, organizational attractiveness, anticipated belongingness). Additionally, knowing if a signal that shows job seekers the organization is an equal opportunity employer could change the dynamics of otherwise male or female dominated fields and could become a future alternative to reduce the existing barriers for females to enter the field of entrepreneurship. Therefore, this thesis also adds to the existing literature of signaling theory. Thirdly, this study investigates the existence of gender role stereotypes in entrepreneurship and confirms the assumption that entrepreneurship is stereotypically associated with masculine traits (Gupta et al., 2009; Short et al., 2010). Finally, when the hypothesized model is confirmed, this thesis will contribute the literature by providing insights in the mental step by step process of a job seeker from reading the recruitment material until perceiving the organization as attractive. This could identify various potential future directions of research to better understand and control these mental stages during the recruitment process.THEORETICAL FRAMEWORK AND HYPOTHESES
This study builds on the findings and contributions presented in the study by Gaucher et al. (2011) and applies them with a focus on the start-up sector. Their research, with four main discoveries, has shown the influence of gendered worded job advertisements on the perception of job seekers regarding those job advertisements and advertising companies. Their first discovery is that the job advertisements of male-dominated jobs are more masculine worded, compared to female-dominated jobs. Secondly, Gaucher et al. (2011) discovered that job-seekers are able to associate the wording of the job advertisement with the expected gender majority of that occupation. In other words, job-seekers expect a male (female) majority when a job advertisement is masculine (feminine) worded. Further, the gendered wording of the job advertisements also affects the perception of job seekers about the appropriate gender fit to that job. Thus, males (females) perceive a greater fit with the advertised job or company when the advertisement was masculine (feminine) worded. Finally, Gaucher et al. (2011) uncovered that job seekers were attracted to jobs when they perceived a congruent relationship between the gendered wording in the job advertisement and their own gender. This final relationship has been explained by Gaucher et al. (2011) as the mediating effect of belongingness. Thus, females (males) perceive lower levels of job appeal when reading a masculine (feminine) worded job advertisement as it signals they do not belong. The abovementioned findings have already made
4
an important contribution in building the understanding of the effects of gendered wording in job advertisements, which will be extended to the start-up sector in this study. Further, this study investigates the effect of a diversity statement, rather than gendered wording, in an entrepreneurial job advertisement on job seeker’s organizational attractiveness.Job Seekers’ Perception on Organizational Proportion of Female Employees
Attracting and maintaining skilled and talented employees is important for start-ups as this is a big influencer in regard to the success of a venture (Tarique & Schuler, 2010). However, start-ups are less visible for job seekers compared to more established firms due to their lack of reputation and prior history (Kraus, Harms & Fink, 2010), lack of funds to offer attractive benefit packages, or limited options for career development (Celani & Singh, 2011; Tumasjan, Strobel & Welpe, 2011). This makes it complicated for start-ups to attract adequately skilled and experienced employees. Potential employees make their first contact with start-ups through job advertisements and this contact lays down the foundation for the development of interest (Barber, 1998; Leung, 2003). When trying to develop an organizational image of an advertising start-up, potential job seekers are faced with an information gap and depend on the information that is shared through the recruitment methods (Walker & Hinojosa, 2014). Thus, a company’s belief about the importance of equal work opportunities for men and women shared through a diversity statement is used to develop the organizational image. Job seekers further use this image to evaluate the advertising organization on its attractiveness and its characteristics (Rynes et al., 1991), or its prevailing working conditions (Ehrhart & Ziegert, 2005). Thus, the composition of a job advertisement plays a crucial part in the recruitment success of start-ups (Mayson & Barrett, 2006).
This reaction of job seekers to hiring materials, such as building a perception of an organization based on a job advertisement (Lievens & Slaughter, 2016), has been explained through signaling theory. This theory, originally developed by Spence (1973), has been used to describe the behavior of both the receiver (e.g. job-seeker), who interprets and reacts to information, and the signaler (e.g. start-up), who gives information while trying to send positive characteristics (Connelly, Certo, Ireland & Reutzel, 2011). Job advertisement signal elements such as lay-out, tone, and wording to the job seekers to attract potential job applicants (Goldberg & Allen, 2008; Rynes, 1989). Job seekers assess their perceived fit with the advertising company based on these elements (Backhaus, 2004). Research has discovered that the applicant’s impression of the advertised company is influenced by the wording of the job advertisement (Lievens & Chapman, 2010; Rynes & Cable, 2003). Further, Gaucher et al. (2011)
5
had observed that gender stereotypical wording in advertisements influences job seekers expectations of the occupation’s gender majority and consequently the firm’s gender diversity. However, as we know that job seekers base their expectation about the company on recruitment signals, a different image about the firm’s proportion of female employees can be created with a message in the recruitment material that shows the firm’s importance of equal work opportunities for men and women. Subsequently moderating the expected occupation’s gender majority based on the gender stereotypical wording in the advertisement.Prior research has agreed that the entrepreneurial orientation of start-ups has attributes that can be associated with masculine traits (Gupta et al., 2009; Short et al., 2010). The stereotypical representation of an entrepreneur is that of someone as being competent, risk tolerant, competitive, independent, and agentic, which display comparable attributes to the stereotype of a man (Ahl, 2006; Thébaud, 2015). On the other hand, traits such as cooperative and social have been associated with females (Bem, 1985). In addition, job seekers have been argued to be attracted to a company when the advertisement signals similar characteristics or traits to those the job seekers possesses (Celani & Singh, 2011). This occurrence, seeking congruency, can be explained through role congruity theory. The advertising company can be seen as a group that possesses certain traits and characteristics that make them fit to occupy the organization's role (Hoffman & Hurst, 1990). Job seekers actively search for congruency between their attributes and those of an organization. Members of a group with congruent characteristics have been found to facilitate performance (Eagly & Diekman, 2005). Stereotype activation theory (SAT) explains how job seekers are able to automatically assess their congruence with the gender stereotypical characteristics of a job advertisement. Prior research has collected evidence that individuals automatically access trait concepts and stereotypes from their memory when perceiving and making judgments about others (Banaji, Hardin & Rothman, 1993; Macrae, Milne & Bodenhausen, 1994). The stereotype eventually comes to be activated reflexively upon the mere presence of those features, without the perceiver’s intending it or being aware of the trait concept or stereotypic influence on perceptual interpretation or judgment (Chen & Bargh, 1997).
Based on these concepts, entrepreneurial worded job advertisements that are recruiting for start-ups might signal masculinity about the advertised job and activate the male stereotype of entrepreneurship. This would lead to job seekers to perceive the job and the organization as male dominated. However, signaling theory shows that people base their perception about the job and organization on the available recruitment materials. Thus, when including an intervention in the job advertisement in the form of a diversity statement, this will be used to
6
make an image about the company’s dynamics. Further, a diversity statement could show the importance the company lies on including employees from both genders, which collides with the male stereotype of entrepreneurship and could result in more females being interested in the advertised job. The following hypothesis illustrates these arguments:Hypothesis 1: Job Seekers’ gender moderates the relationship between start-up job advertisements signaling and perceived proportion of females, such that the base entrepreneurial version will lead to lower perceived proportion of females among women than among men, while the diversity intervention version will lead to higher perceived proportion of females among women than among men.
Job Seekers’ Drive to Belong
As stated before, Gaucher et al. (2011) had uncovered that job seekers were attracted to jobs when they perceived a congruent relationship between the gendered wording in the job advertisement and their own gender, which can be explained through role congruity theory. However, in line with this thinking, we can also argue that job seekers are attracted or anticipate belongingness to a firm when they perceive the gender majority, rather than gendered wording, to be congruent with their own gender. The following sections will elaborate on the arguments for this statement.
A job applicant’s belongingness, or need to belong, has been explained as the drive for “a minimum quantity of lasting, positive, and significant interpersonal relationships” (Leary & Baumeister, 2017, p. 57) and refers to an individual’s perception to be part or suited to a certain group (Gaucher et al., 2011). In short, belongingness is the attempt to be a part of in order to prevent feelings of alienation and loneliness (Lee & Robbins, 1995). Consequently, to satisfy the drive, individuals, such as job seekers, pursue frequent positive interactions to build and preserve high quality interpersonal relationships (Watt & Badger, 2009). Knowing that job seekers are able to stereotype the advertised job on its signals, a job seeker’s feeling of belongingness could be higher (lower) when having the same gender as the gender majority (minority). Thus, belongingness could be affected by the feeling of congruency or similarity between the job seeker and the individuals working at a start-up.
This argument can be further supported by similarity-attraction theory, which states that job seekers are looking for teams and ventures with similar attributes (Jackson et al., 1991). In line with role congruity and SAT, this means that job applicants compare their own demographic attributes with the members of different groups to assess their degree of similarity (Tsui, Egan & O’Reilly, 1992). Based on the perceived congruency or similarity, job seekers
7
are able to assess whether they could belong or fit among a certain group of individuals. This feeling of belongingness is strengthened, when job seekers perceived the norms and values to be similar between them and employees working at an organization (Von Walter, Wentzel & Tomczak, 2012). Further, we can argue that job seekers develop increased feelings of belongingness and fit when they perceive the predominate gender in a venture to be the same gender as theirs.This argument can be supported through the person-organization-fit construct, developed by Cable and Judge (1994), which concerns the compatibility between employer and employee. In order to achieve compatibility, Kristof (1996) defined that either both sides should be sharing the same norms, or at least one side should provide the desires of the other side. An important predictor of the attractiveness of an advertising start-up is the perception of job seekers on fair treatments, whereby job seekers desire to experience fair practices during their employment period (Hom, Griffith, Palich & Bracker, 1999). Job seekers are able to make this judgement of fair treatment by a prospective venture based on the signals that have been send through job advertisements (Breaugh & Starke, 2000). Feelings of anxiety and reduced belongingness have been found when job seekers perceived a miss-fit of norms and values (Murphy, Steele & Gross, 2007; Osborne, 2007). A better match between job seekers and a start-up venture’s norms and values leads to increased attraction and commitment (Cable & Judge, 1997; Kristof, 1996). Job seekers with demographic similarities have been found to have a higher chance of liking the advertised job and thus increased feelings of belongingness and perceive higher P-O fit when interactions take place with demographically similar individuals (Jackson et al., 1991). Therefore, the following hypothesis illustrates these arguments:
Hypothesis 2: Job seekers’ gender moderates the relationship between perceived proportion of females and anticipated belongingness, such that this relationship is negative for men and positive for women.
The Relationship Between Anticipated Belongingness and Organizational Attractiveness Job seekers affective and attitudinal thought about a specific venture to be a future employer can be seen as company attractiveness (Barber, 1998). This attitudinal thought of job seekers has been defined as passive in nature as it does not necessarily imply that any actual behavior will be taken towards the company (Barber, 1998). This passivity allows individuals to be attracted to many companies simultaneously; more active pursuit of the company would require resource conservation and a more limited set of feasible possibilities (Barber, 1998). That type of behavior has been defined as intention to pursue, which refers to thoughts about a company
8
that specifically imply further action (Highhouse, Lievens & Sinar, 2003). Job seekers feeling of anticipated belongingness could lead to feelings of attractiveness to the advertising company. Social identity theory, which posits that individuals, such as job seekers, categorize themselves in groups on their self-concept (Tajfel & Turner, 1985), can further explain the anticipated feelings of belongingness. At the center of social identity theory lies the concept of social roles, which is a socially shared expectation that characterizes a person who posits a certain social position or is a member of a particular social category (Biddle, 1979; Sarbin & Allen, 1968). Job seekers are able to characterize an advertising start-up through the signals in the job advertisement, which are stereotyped through SAT. With social identification, job seekers are able to further classify individuals in their environment into different groups based on age, gender, or religion (Ashforth & Mael, 1989). Further, job seekers are able to define themselves in their environment based on these classifications (Ashfort & Mael, 1989).Due to social identification, job seekers are able to identify certain groups they perceive themselves as fitting or belonging to (Gaucher et al., 2011; Hogg, 2001). Further, applicants are likely to become more attracted to a start-up when they perceive the job advertisement to signal traits that are similar to those in their possession (Celani & Singh, 2011). Male job seekers (women) have been perceived a greater fit with advertisements that signal masculine (feminine) through wording (Eagly & Karau, 2002). The greater the extent to which job seekers perceive similar characteristics between themselves and the recruiting firm the greater the extent to which job seekers will identify with and thus feel attracted to the recruiting firm (Celani & Singh, 2011). In other words, feelings of anticipated belongingness can be associated with organizational attractiveness. Research on role congruity theory indicates that women would violate stereotypical gender norms when they manifest in male-typical attributes (Ritter & Yoder, 2004). As a consequence, due to their socially accepted roles, women are likely to feel less confident to apply, less attracted, and feel less belonging to a position that is signaled and perceived as masculine (Gaucher et al., 2011). Thus, the groups job seekers categorize themselves with, are the groups they associate with favorable and positive feelings (Stets & Burke, 2000). Therefore, we argue that the favorable and positive feelings of job seekers anticipated belongingness is important for the decision if the company is suited to them, and therefore anticipated belongingness will lead to organizational attractiveness. This leads to the following hypothesis:
Hypothesis 3: Job seekers’ anticipated belongingness is positively related to their perceived organizational attractiveness
9
The Effect of a Diversity Statement in Job Adverts on Organizational AttractivenessThis study further argues that signals in start-up job advertisements are related to organizational attractiveness. Thus, job seekers that receive a signal through a diversity statement in the recruitment material may have different perceptions of organizational attractiveness than job seekers without that signal. However, this relationship is expected to be influenced in a serial, rather than direct, manner, as the previous hypotheses argue that the job advertisement first leads to perceived proportion of females of the firm, which in turn leads to certain feelings of anticipated belongingness and subsequently to organizational attractiveness. Therefore, job adverts influence organizational attractiveness through serial mediation. Furthermore, gender moderates the relationship (1) between job advertisement signaling and perceived proportion of females and (2) between perceived proportion of females and anticipated belongingness. The full serial moderated mediation relationship is shown in the following hypothesis:
Hypothesis 4: The relationship between start-up job advertisement signaling and organizational attractiveness is serial mediated through (1) perceived proportion of females and (2) anticipated belongingness, while job seeker’s gender moderates the path between (i) job advertisement signaling and perceived proportion of females and between (ii) perceived proportion of females and anticipated belongingness.
Theoretical Model
The literature review has resulted in the development of four different hypotheses. These hypotheses combined will be used to answer the aforementioned main research question. The questions are summarized in the theoretical model, shown in figure 1 (p. 10). The model aims to explain the effect of the entrepreneurial start-up advertisement on perceived organizational attractiveness and examine whether this effect, which we hypothesize to be moderated by job seeker’s gender and mediated by perceived proportion of females and anticipated belongingness, can be affected with the use of a diversity statement.
METHODS Research design
In the following sections, the design, sample selection, experimental conditions, and measurements of this thesis are described. To test and analyze the theoretical model, an experimental design was used in which participants were randomly assigned to one of two conditions. This way, two groups were created with the sole difference between the shown job advertisement being the inclusion or exclusion of the diversity statement. Participants of the
10
experiment were asked to read the job advertisement and subsequently fill in the remainder of the survey, which provides the necessary information for the analysis.Figure 1: Theoretical Model
Sample Selection
This study investigates how entrepreneurial start-up job advertisements will bias potential job seekers’ organizational attraction and if a diversity statement might moderate this effect. The web-based Qualtrics survey has been spread among students at the Vrije Universiteit Amsterdam and Universiteit van Amsterdam. Students were selected as we specifically sought participants who are a representation of the population that were looking for a start-up job or are expected to do so within 12 months. In total, 238 respondents have passed the eligibility check (18+) and finished the survey. The exact response rate cannot be determined as the survey has been distributed through various social media channels.
Participants have been excluded on strict criteria in order to include only the most relevant cases for further testing. The exclusions have been based on partial completion, hiding their gender, not being a relevant job seeker, not open to start-ups, or failing one of the two attention check questions. Further, participants, aged 36 or older, have been excluded as it has been decided that this custom age group of young adults is most relevant as start-up employees. Moreover, participants had to watch the job ad for at least 15 seconds, as this is the expected time to skim all the words of the job advertisements, a technique that will allow to accurately get the message of the job advertisement faster than by reading word for word (Duggan & Payne, 2009). As a further step after the initial exclusions, an in dept review of the individual
11
answers has been done and excluded participants that explained they do not understand the concept of entrepreneurship in the respective comment box.Due to these exclusions, the total sample was reduced to 173 individuals. This amount is just below the required sample size of 177 and matches 89.3% power. The required sample size of 177 participants has been estimated with the help of the G*Power analysis (Faul, Erdfelder, Lang & Buchner, 2007), selecting the option ANOVA: fixed effects, special, main effects and interactions. The following numbers have been used in the G*Power analysis to estimate the required sample size: a = .05, power = 90%, numerator df = 2, number of groups = 4, and effect size f = 0.27 (equivalent to partial h2 = 0.066; retrieved from the effect size Wording x Gender by Gaucher et al., 2011).
Of the total 173 participants, 100 were female (57.8 percent) and 73 were male (42.2 percent), with a mean age of 22.99 years old (SD = 3.19) and 117 participants (67.6 percent) identified themselves with the Dutch nationality. Further, 167 individuals (96.5 percent) are currently attending university, and of them 140 respondents (80.9 percent) have finished either a Bachelor’s or Master’s degree. The area of their current or completed study has been spread out among participants. We also asked participants about their understanding of the English language and 171 out of 172 participants indicated that they have an above average understanding of English. This indicates that participants had no issue in reading the job ad and understanding the questions. Further, the survey consisted of two experimental conditions where 34 males and 51 females completed the entrepreneurial-baseline-condition, while 39 males and 49 females have finished the diversity statement intervention condition.
Experimental Conditions
As highlighted before, this study conducts an experiment about the effect of a diversity statement included in a job advertisement on job-seekers perception of organizational attractiveness. In order to conduct this experiment, two different start-up job advertisements had to be created that make up for each experimental condition. Sentence fragments and relevant wording was collected from a review among various job advertising websites, such as angel.co, jobfluent.com, dutchstartupjobs.com, and tq.co, to make a generic and realistic framework of a job advertisement. This framework was kept ambiguous in nature to attract a larger population of interested job seekers by limiting the presented information about the company, its product, and job position. To make this framework more representable for a start-up, the wording of the advertisement had to be adjusted to be more entrepreneurial. The entrepreneurial dictionary containing 244 words, developed by Short et al. (2010), has been
12
used to increase the percentage of entrepreneurial words to be a realistic representation of a start-up job advertisement. In order for the experiment to only test on the effect of entrepreneurial words, the word list of Gaucher et al. (2011) was used to reduce the level of masculine and feminine words to be of an insignificant effect on the experiment. The job advertisement for the entrepreneurial condition, as seen in appendix A, has a total of 209 words of which 21 (10.05%) are entrepreneurial and 3 (1.5%) are masculine.The paper by Flory et al. (2018) has shown that a diversity statement can be an effective means to increase the levels of workforce diversity. This experiment tries to increase the attractiveness of start-ups for female job-seekers and, as a result of that, to increase workforce diversity. Therefore, the second condition has copied the job advertisement of the first condition and added a diversity statement to it. So, the sole difference between the two conditions is the inclusion or exclusion of the diversity statement. The research papers written by Flory et al. (2018), Singh and Point (2004 & 2006), and Windscheid et al. (2016) have been the first source of the development of the diversity statement, while the other source has been the aforementioned job advertising websites. Both sources have led to the development of the following diversity statement: “the equality of opportunities for men and women is an essential part of our company’s culture, values and principles. We believe a diverse company is a better company and therefore offer an environment that appreciates, respects, and accepts gender diversity”. Including this statement, the advertisement for the diversity statement condition, as seen in appendix B, has a total of 249 words of which 21 (8.43%) are entrepreneurial and 3 (1.2%) are masculine.
In order to insure the validity of the job advertisements and their manipulated wording, three recruitment/human resource experts have been asked to evaluate if the job ads are realistic, believable and attractive. The three experts had positive comments about the job advertisements in terms of perceiving the vacancy as real, typical for a startup, and are likely to be attractive for job-seekers who are open to work at a startup. Their only concern was about the information gap regarding what the company does, what specific job position is open, and where the work is being performed (e.g., while traveling, at an office, at home). We considered this feedback but chose to not change the level of specificity about the company and the job position in order to maintain the applicability of the job ads to a larger proportion of our respondents. However, we added details about the opportunity to travel abroad and the work environment (working at the office). After including the feedback from the experts, the real survey was made through Qualtrics. In order to test the experiment works as planned through the survey, a pilot study was done among students, family, and friends. The pilot test revealed minor points to address,
13
such as various small spelling mistakes and a question asking for the proportion of females working in the advertised position. This question has been removed from the survey as the advertisements do not advertise a job position and is thus irrelevant for the experiment and could lead to confusion. The correction of the spelling mistakes has led to the final version of the experiment, ready to be distributed among the sample.Procedure
Participants have either been recruited via a distributed link or by having face-to-face contact at the University of Amsterdam and Vrije Universiteit Amsterdam. Respondents that joined online have done so via the anonymous link found at either their university e-mail, LinkedIn, Facebook wall, or shared university / study groups on Facebook. The participants were invited to join and fill in the Qualtrics survey that was explained as a study to investigate the role of job advertisements in start-up’s employee recruitment. After reading a short introduction emphasizing confidentiality and voluntary participation (appendix C) at the landing page, which has been kept limited to prevent information overload and immediate dropouts, participants either declined or agreed to join the experiment. After this, participants were confronted with a first brief set of questions about their age, employment status, job seeker status, and openness to work at a start-up. These questions are placed to filter the participants and confirm their eligibility to the study. Participants continuing beyond this point have been, after reading a final instruction (appendix D), randomly assigned to one of the two experimental conditions to create a between-subjects design experiment. After filling out the remainder of the questions and completing the survey, respondents were briefly informed about the real purpose and background of the study in a debriefing (appendix E), and subsequently thanked for their participation.
Measurements
The survey has included the following measures to operationalize the concepts of the theoretical model. The responses were measured on a 7-point Likert scale, unless noted otherwise, ranging from ‘strongly agree’ to ‘strongly disagree’.
The only dependent variable of this experiment will be measured through organizational attractiveness. The organizational attractiveness of the job advertising organization to be a future employer of the survey respondent has been measured by implementing the 5-item scale developed by Highhouse et al. (2003). The five items implemented from this study are “I would accept a job offer from this company”, “I would make this company one of my first choices as an employer”, “If this company invited me for a
14
job interview, I would go”, “I would exert a great deal of effort to work for this company”, and “I would recommend this company to a friend looking for a job”.The experiment has in total two mediating variables of which perceived proportion of females has been established on the assumption that the field of entrepreneurship has been dominated by men (Kelley, Singer & Herrington, 2012), this research has made perceived workforce proportion testable through the perceived percentage of females working in a start-up. Each increment in the share of females present in the total workforce assists in an increased gender diversity in a start-up. To test this, expected female proportion of advertising company’s workforce has been measured through the following two items: “proportion of female employees” and “proportion of females in leadership positions”. Respondents were able to answer their perceived proportion 11-point Likert scale from 0% to 100% with interval increments of 10%.
Anticipated belongingness has been based on role congruity theory that explains the importance of feeling fit in a certain role, this survey has implemented the 4-item scale of anticipated belongingness from Gaucher et al. (2011). The four items that have been included are “The type of people who would apply for this job are very different from me”, “My values and this company’s values are similar”, “I’m similar to the people who would apply to this job”, and “I could fit in well at this company”.
Besides the dependent variables, the survey also included three control variables. The first measurement that has been asked in the reflection is if the participants did imagine looking for a job. Further, the participants are asked to reflect the organization’s entrepreneurial image by asking if the job advertisement, company advertising this job, and employees at this firm are entrepreneurial. The survey further reflects the organization’s masculinity image by asking if the participants thought of a female employee while answering (reversed), thought if the company is looking for male employees, and would agree that the job is more attractive for male employees.
To finish off the information from the respondents, several demographics were asked. One of those measurements was job seeker’s gender, where we asked participants of the survey to indicate their respective gender (male, female, or rather not say). Besides that, respondents have been asked to fill in their respective age to formulate the mean age of the sample and indicate their eligibility to the experiment. Next to that, participants were asked to fill in their current employment status and nationality. Furthermore, we asked participants their highest level of completed education level and the educational orientation of their completed or current study. The following measurement is the English understanding of participants. This
15
will indicate if participants are able to fully understand the survey. Furthermore, our final two demographic measurements of participants their openness to start-ups and their job seeker status, have been included as exclusion criteria. As stated before, indicating to not be open to work in a start-up or not looking for a job currently, in the past 12 months, in the following 12 months will lead to exclusion of the experiment.RESULTS
Before the hypotheses are ready to be tested, several preliminary analyses have been done to prepare the data, which will be elaborated on below. In short, the analysis consists of an exploratory factor analysis, creation of the descriptive statistics and intercorrelations table, a randomization check, and a confirmatory factors analysis.
Exploratory Factor Analysis
The survey has implemented multiple comparable questions to test the measurements of organizational entrepreneurial image and organizational masculinity image. Therefore, an Exploratory Factor Analysis (EFA) has been done to check if the questions can be combined to use them for the randomization check. The three questions “the job advertisement was very entrepreneurial”, “the company advertising the job is very entrepreneurial”, and “employees at the company advertising the job are expected to be very entrepreneurial” have been checked to form an averaged organizational entrepreneurial image. The items “reading the job advertisement I had a picture in my mind of female employee” (reversed), “this company is looking for male rather than female employees”, and “most people would agree that this job is more attractive for male than for female employees” have been checked to formulate organizational masculinity image.
When running the EFA calculations through SPSS, the results show that the KMO statistic is .630. This shows that the model has sufficient survey size. The Bartlett’s Test of Sphericity is significant (<.001), indicating at least two items are highly significant. Also, two components are found with Initial Eigenvalue of 1 or higher (2.320 & 1.695). Further, the component correlation matrix shows that the correlation between the components is low with a value of -.161. The rotated component matrix shows that the three organizational entrepreneurial image items and the three organizational masculinity image items are both in their own component.
After this, the reliability of the components is tested. The Cronbach’s Alpha for the entrepreneurial component is .804, which shows good reliability. The Cronbach’s Alpha is lower for the masculinity component (.671), and the Chronbach’s Alpha if item is deleted is
16
checked. If “reading the job advertisement I had a picture in my mind of female employee” is deleted the Cronbach’s Alpha would increase to .790. Therefore, this item will be dropped in order to formulate a reliable organizational masculinity image component. The full output of the EFA calculations can be found in appendix F.Descriptive Statistics and Intercorrelations
All means, standard deviations, and intercorrelations of the relevant survey items are provided in table 1 (p. 17). The correlations that are stated below are either important for the experiment or show surprising correlations. First off, the condition was significantly negatively correlated with organization masculinity image (r = -.235, p <0.01). Perceived proportion of females was significantly negatively correlated with organization masculinity image (r = -.475, p <0.01). Anticipated belongingness is significantly positively correlated with organizational attractiveness (r = .669, p <0.01). Perceived proportion of females was significantly negatively correlated with age (r = -.157, p <0.05). Anticipated belongingness is significantly positively correlated with organizational nationality (r = .167, p <0.05). Gender is significantly positively correlated with educational orientation (r = .247, p <0.01). Age is significantly positively correlated with organization masculinity image (r = .172, p <0.05). Nationality is significantly positively correlated with English understanding (r = .239, p <0.01).
Randomization Check
In order to check if the groups are randomized correctly, the means of the group reading solely the entrepreneurial version have been compared with the group reading the diversity statement version. Due to the types of data, a combination of the Independent T Test, Chi-Square Test and Mann-Whitney U Test have been used to compare the groups. The test results can also be found in appendix G. The analysis has revealed that there were no significant differences between groups regarding their age (T = .071, p = .994), gender (χ2 = .331, p = .565), employment status (χ2 = 1.169, p = .760), nationality (χ2 = .233, p = .629), educational level (Z = -.476, p = .634), educational program (χ2 = 9.576, p = .792), imagined they looked for a job (T = .465, p = .642), entrepreneurial image (T = .305, p = .761), and English understanding (T = .073, p = .942 or Z = -.111, p = .911). The only significant difference between groups was found on the variable organization masculinity image (T = .332, p = .002). However, this variable was ought to differ and is a check before using the real variables to test the perception difference.
17
Table 1: Descriptive Statistics and Intercorrelations Table
N Mean SD Variables
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)
(1) Condition 173 .51 .50 1
(2) Perceived Proportion of Females 173 4.59 1.27 .078 1 (3) Anticipated Belongingness 173 4.78 1.07 -.030 .061 1 (4) Org. Attractiveness 173 5.18 1.08 -.095 .082 .669** 1 (5) Gender 173 .58 .50 -.044 .148 .025 .085 1 (6) Age 173 22.99 3.19 -.005 -.157* -.021 -.044 -.101 1 (7) Employment Status 173 1.59 1.01 -.022 -.059 .115 .084 -.104 .293** 1 (8) Nationality 173 .32 .47 -.037 .034 .167* .145 .116 .141 .037 1 (9) Educational Level 173 7.36 1.77 .033 -.148 -.002 -.009 -.109 .512** .319** .018 1 (10) Educational Orientation 173 8.37 7.74 -.083 .079 -.022 .003 .247** .056 -.143 -.022 .030 1 (11) Imagined Looking For Job 173 5.86 .91 -.036 .061 .220** .246** -.105 -.022 .001 .051 -.001 .035 1 (12) Org. Entrepreneurial Image 173 5.71 .86 -.023 .150* .118 .216** .132 .061 -.049 .058 -.026 .097 .205** 1 (13) Org. Masculinity Image 173 3.60 1.41 -.235** -.475** -.177* -.191* -.098 .172* -.057 -.180* .172* -.136 -.062 -.091 1 (14) English Understanding 172 4.75 .52 -.006 -.100 .052 .146 -.068 .206** .036 .239** .111 -.057 .111 .263** -.107 1 ** Correlation is significant at the 0.01 level (2-tailed)
* Correlation is significant at the 0.05 level (2-tailed) Notes: Explanation of the categorical variables:
Condition Job Advertisement: 0 = Entrepreneurial; 1 = Diversity Statement
Perceived Proportion of Females : 1 = 0%; 2 = 10%; 3 = 20%; 4 = 30%; 5 = 40%; 6 = 50%; 7 = 60%; 8 = 70%; 9 = 80%; 10 = 90%; 11 = 100% Gender: 0 = Male; 1 = Female
Employment Status: 0 = Unemployed not looking for work; 1 = Unemployed looking for work; 2 = Employed part time; 3 = Employed full time Nationality: 0 = Dutch; 1 = Other
Educational Level: 1 = No schooling; 2 = Some high school; 3 = High school graduate; 4 = Some college credit; 5 = Trade/technical/vocational training; 6 = associate degree; 7 = Bachelor’s degree; 8 = Pre-master program; 9 = Master’s degree; 10 = Professional degree; 11 = Doctorate degree
Educational Orientation: 1 = Business/Management; 2 = Economics/Finance; 3 = Behavioral Science; 4 = IT; 5 = Engineering; 6 = Political Science; 7 = Law; 8 = Sociology; 9 = Communications; 10 = Culture; 11 = History; 12 = Math; 13 = Philosophy; 14 = Religion; 15 = Health; 16 = Biology; 17 = Bio-Tech; 18 = Arts & Humanities; 19 = Language; 20 = Other
18
Measurement Model ValidationStata 14 has been used to analyze the hypothesized model (figure 1, p. 10) through a confirmatory factor analysis (CFA). The model has been replicated through structural equation modelling (SEM) and estimated. The analysis (detailed test results shown in appendix H) shows that the factor correlation between anticipated belongingness and organizational attractiveness is 0.72, which is lower than the required 0.85 to indicate proper discriminant validity (Fornell & Larcker, 1981). Further, the factor loadings should be higher than 0.7 to ensure face validity. The model shows good statistics except for the items OA4 (.60) and AB3 (.67).
Several statistics have been checked in order to assess the goodness of fit for the model. The chi-square statistic, with a value of 49.103 and value of 0.004, shows poor fit as the p-value should not be significant (Lance & Vandenberg, 2009). However, the model shows better fit with a RMSEA value of 0,072, which should be lower than 0,08 (MacMallum, Browne & Sugawara, 1996), and with a P-close value higher than 0,05 cut-off with 0,119 (Browne & Cudeck, 1993). According to Byrne (1994), the CFI should exceed 0,93 for a good fit, which Stata calculated to be a value of 0.978. Thus, according to the CFI, this model can be seen as a good fit. The TLI should have a comparable value to CFI, as Hu and Bentler (1999) defined that the cut-off value of a fitting model is lower than 0,95. The calculated TFI value for the hypothesized model is 0.969, which represents a greater value than the cut-off point. The SRMR value of a model should at least be less than 0.08 (Browne & Cudeck, 1993; Hu & Bentler, 1999) and preferably under 0.05 (Stieger, 1990). The model also has good fit on this statistic as the calculated value is 0.049. Overall, it can be concluded that this model has a good fit between the hypothesized model and the observed data as all, except chi-square, goodness of fit statistics have correct values.
Hypothesis Testing
The hypothesized model has been replicated in a custom model for the conditional PROCESS analysis, for which the results are shown below. Figure 2 (p. 21) shows the unstandardized coefficients, together with the significance levels, of the predictors of organizational attractiveness, anticipated belongingness and perceived proportion of females.
Hypothesis 1 proposed that the relationship between start-up job advertisements signaling and perceived proportion of females is moderated by job seekers’ gender, such that the base entrepreneurial version will lead to lower perceived proportion of females among women than among men, while the diversity intervention version will lead to higher perceived proportion of females among women than among men. The conditional process analysis
19
indicated, through the first model with perceived proportion of females as an outcome, F(3, 169) = 1.854, p = .139, R2 = .032, that the inclusion of a diversity statement did not have a significant effect on perceived proportion of females (b = 0.06, t = .20, p = .84). The first model further indicates a non-significant main effect of gender (b = 0.25, t = .89, p = .37) and a non-significant interaction effect (condition x gender; b = 0.27, t = .70, p = .49). Moreover, graph 1 shows the separate effects of females and males for the base-entrepreneurial and intervention-diversity version regarding their perceived proportion of females. This shows that the perceived proportion of females is steady for males under both conditions while females anticipate more gender diversity in the intervention-diversity version. However, we can conclude from the conditional process analysis that hypothesis 1 is not supported.Graph 1: Conditional effects of diversity statements on perceived proportion of females among male and female job-seekers
The second hypothesis proposed that job seekers’ gender moderates the relationship between perceived proportion of females and anticipated belongingness, such that this relationship is negative for men and positive for women. The conditional process analysis indicated, through the second model with anticipated belongingness as an outcome, F(3, 169) = .401, p = .753, R2 = .007, that perceived proportion of females does not have a significant effect on anticipated belongingness (b = 0.004, t = .04, p = .97). The second model further indicates a significant main effect of gender (b = -0.40, t = -.64, p = .52) and a non-significant interaction effect (perceived proportion of females x gender; b = 0.09, t = .73, p
4 4,2 4,4 4,6 4,8 5
Entrepreneurial Version Diversity Version
Per cei ved Pr op or tio n of F em al es Male Female
20
= .47). Moreover, graph 2 shows the separate effects of females and males for the base-entrepreneurial and intervention-diversity version regarding their anticipated belongingness. This shows that for men the perceived proportion of females does not have an effect on anticipated belongingness, while for women it shows that more gender diversity (read: more females) leads to higher feelings of anticipated belongingness. However, we can conclude from the conditional process analysis that hypothesis 2 is not supported.Graph 2: Conditional effects of perceived proportion of females on anticipated belongingness among male and female job-seekers
The third hypothesis proposed that job seekers’ anticipated belongingness is positively related to their perceived organizational attractiveness. The conditional process analysis indicated, through the third model with organizational attractiveness as an outcome, F(1, 171) = 138.80, p < .001, R2 = .448, that anticipated belongingness did have a significant effect on organizational attractiveness (b = 0.67, t = 11.78, p < .001). Therefore, hypothesis 3 is supported.
The fourth hypothesis proposed that the base entrepreneurial version will lead to lower perceived proportion of females among women than among men, while the diversity intervention version will lead to higher perceived proportion of females among women than among men, which would then affect job seeker’s anticipated belongingness, based on their gender, such that this relationship is negative for males and positive for females. It further hypothesized that this heightened belongingness in turn leads to higher feelings of organizational attractiveness. The conditional process analysis indicated, through the computed
4,4 4,6 4,8 5 3 4,5 6 An tic ip at ed B el on gi ng ne ss
Perceived Proportion of Females
Male Female
21
‘index of moderated mediation’, that the indirect effect of the condition on organizational attractiveness via the two mediators, with gender moderating the first two of three relationships, was non-significant (b = .022, 95% CI -.038; .114). This trend holds when further breaking down the indirect effects on the separate genders, as neither being female (b = .022, 95% CI -.021; .106) nor male (b = .0001, 95% CI -.040 .040) significantly moderated the proposed serial moderated mediation PROCESS model. Therefore, hypothesis 4 is not supported.Figure 2: Model and Hypotheses with PROCESS conditional analysis results.
Note: * p < .05; ** p < .001
DISCUSSION
The aim of this thesis was to test whether the inclusion of a diversity statement in an entrepreneurial job advertisement would influence job seeker’s organizational attractiveness. This effect has been tested through a gender-moderated serially mediated relationship (see figures 1 & 2) that has been split through three hypotheses. These hypotheses measured whether the intervention, as a function of gender, would influence perceived proportion of females, this, as a function of gender, would affect anticipated belongingness, which in turn would affect organizational attractiveness. The goal being to confirm the gender-moderated serially mediated relationship between the advertisement signal and organizational attractiveness. However, the results did not support three out of four hypotheses and with the only exception being the significant relationship between anticipated belongingness and organizational attractiveness. At first sight, the results of the thesis might seem contradictory. However, when investigating the collected data and arguments or findings in other literature, various possible explanations for the results of this thesis can be given.
22
Interpretation of FindingsFirst off, the analysis showed support for the third hypothesis, which confirms the previously researched relationship between anticipated belongingness and organizational attractiveness (e.g., Gaucher et al., 2011; Lievens, Hoye & Anseel, 2007) in the presented sample. The validated measures for belongingness and attractiveness worked as expected and show that the sample can be interpreted as conforming. Moreover, the randomization check and hypothesis testing both show that participants from the base-entrepreneurial and intervention-diversity condition are not significantly different on the key variables. However, in order to have supporting results for hypothesis one and two, the stimulus material should have sparked stereotype activation (Gaucher et al., 2013; Gupta et al., 2013) and thereby influence perceived proportion of females through a strong gender-related effect. Unfortunately, this did not occur, which means that there was no significant effect found for the first, second and consequently the moderated serial mediation hypotheses. Therefore, by investigating why the stimulus materials did not work, we will likely find an answer for why the model did not work as predicted.
The first focus for the failure of this model is investigating why the diversity statement did not have an effect on the participants. The two experimental conditions are exact copies with the sole difference being the diversity statement, thus any difference between conditions would greatly be on the statement. The anticipated effect of the diversity statement is that participants would perceive more female employees in this version compared to the base entrepreneurial version. However, this is not the case as a simple t-test shows that the difference between the entrepreneurial (!̅ = 4,49) and diversity statement (!̅ = 4,69) is not significant. Further, the hypothesis testing has shown that there are little differences between male’s and female’s perception of gender diversity between conditions, which, however, are too small to be significant. Thus, the stimulus material did not work as anticipated and the following sections will provide arguments as to why it failed to.
Literature has shown that placing a diversity statement on corporate websites leads to a better environment, added stakeholder value, enhanced corporate reputation, and better performance (Singh & Point, 2004). However, less is known about the effect of the diversity statement placement in recruitment material on participants. It can be hypothesized that presenting the statement in different job advertisement positions (e.g. near-bottom, middle or near-top) influences the reader’s perception about the importance the advertiser lies on diversity and the readers ability to memorize it. Also, placing the diversity statement alone and separate from the other text might be perceived in different ways. First, the readers could interpret this
23
as the generic best interest of the company on diversity, but it could also be perceived as a red flag for participants as it is included to increase the low amount of employee diversity. Further, as there is no solid literature to draw upon the concept of diversity statements, the exact wording or sentence build-up that is most effective in such statement is not known yet. Therefore, the diversity statement of this thesis, designed through the limited available literature and online sources could be ineffective.Next to this, the mental perception and culture of a reader is important as this will result in a local stereotype towards entrepreneurship, which in turn could influence the effectiveness or need of a diversity statement to moderate this stereotype. Research has shown that gender stereotypical images tend to be similar within cultural homogenous groups, but instead differ between culturally different groups (Cuddy et al., 2015). Thus, the decision to keep our job advertisement vague and interesting for a broad group of job seekers may have backfired, as they could have different mental perceptions. More, the actions and messages from the media influences a nation’s entrepreneurial stereotype (Ahl, 2006; Hamilton, 2013), which is inevitably different between countries. Furthermore, Lim and Envick (2013) have provided support that entrepreneurial orientation dimension scores are significantly different between nations and gender, which indicates that the cultural context needs to be considered when studying the concept of entrepreneurship. Next to that, participants for this experiment have been collected among two universities in the Netherlands, which resulted in the sample with a majority of individuals being Dutch and the remainder of the individuals have lived here for at least a year due to the data being collected near end of the academic year. Therefore, perceptions of entrepreneurship and its associated stereotypical gender are likely based on the Dutch culture, which influenced the responses on the experiment. One could argue that the Dutch culture does already show a more gender-neutral image of entrepreneurship instead of a male preferred field of work, due to its progressive image that includes people being against gender discrimination (Tesch-Römer, Motel-Klingebiel & Tomasik, 2008), women entrepreneur’s representation in the media, and high social status for successful female entrepreneurs (Kelly, Bosma & Amorós, 2011). This means that the difference between the perceived proportion of females in the base-entrepreneurial and intervention-diversity version could be lower in the Netherlands, by neutralizing the gendered effect, compared to countries that are not as progressive as the Dutch. Our final explanation for the failure of the model lies in the message specificity of the job advertisements and its subsequent distribution method among participants. The expert review during the pilot testing revealed that our job advertisements were perceived as vague and did not provide specific information (e.g. the experts mentioned scarce information about
24
firm, job, ideal employee characteristics and work location). The decision was made to not adopt all the feedback and remain vague on the job and firm description in order to not exclude participants due to their backgrounds or area of study. However, the decision to do so has potentially backfired as Feldman, Bearden & Hardesty (2006) have stated that the level of detail and specificity in recruitment material’s objects, actions, or contexts will have a number of positive outcomes in job seekers. One of such outcomes is that increased specificity about the job description, compared to the vagueness of our advertisements, increases perceived job appropriateness and ad truthfulness, while also positively influencing individuals attitude towards the advertising company and job ad (Feldman et al., 2006). Additionally, the content of a job advertisement has shown to influence organizational attitudes of job seekers (Walker, Field, Giles & Bernerth, 2008). Further, taking the experiment online has provided a totally different setting from what would occur in real life situations, as job seekers, when looking for an actual job, would have a far more serious mindset. Therefore, individuals participating in the experiment could have a more relaxed attitude and paid lower attention towards the job advertisement. This can be seen from the number of participants that finished reading the job advertisement rather quick (i.e. below the standard reading time but longer that the skimming time). This means that not all participants have read the advertisement word for word and could have missed important details in the recruitment material, such as the diversity statement. Therefore, our distribution method might have evoked feelings of reduces seriousness, which could have influenced the answers.Theoretical Implications
The first theoretical contribution of this study is the added support the relationship between anticipated belongingness and organizational attractiveness. Previous literature has demonstrated the existence of this relationship in similar contexts (e.g., Gaucher et al., 2011; Lievens, Hoye & Anseel, 2007), which have been replicated and proven in this thesis. However, one does have to take into account that only this relationship was found significant and none of the other individual hypotheses and subsequently the overall model were confirmed. Therefore, this contribution and the following contributions should be seen as hypothetical reflections, which should be considered before conducting further research in the entrepreneurial stereotype.
Next to this, the data collected for this thesis has found some support for the assumption that entrepreneurship is associated with masculinity. The combined data of both the base-entrepreneurial and intervention-diversity versions show that participants perceived more men working in the hypothesized start-up, as N = 129 (74.6%) thought that the percentage of females
