Gender diversity discourse in the communications of major Bulgarian companies with stakeholders
Dayana Horozova 11105275 Master’s Thesis
Graduate School of Communication Master’s programme Communication Science
Supervisor: mw. dr. I.R. Hellsten 30-01-2017
Word count: 7,970 University of Amsterdam
Abstract
This study aimed to investigate the gender diversity discourse trends in twenty Bulgarian companies’ stakeholder communications over the past decade (2006 – 2016). Specifically, the units of analysis were PR releases, annual and corporate social responsibility (CSR) reports, website statements and official online policies, Facebook posts and visual materials. A
quantitative content analysis was carried out to explore the differences between industries on two types of framing tools – textual and visual. The results shed light on generalizable trends within industrial sectors and publication types. Contrary to expectations, the heavy industry exhibited a notable gender diversity awareness, while the consumer-oriented sectors displayed a higher degree of stereotyping. All industries, however, discussed gender diversity openly very rarely for the ten-year period. Practical implications and future directions are discussed as the observed trends allude to the desirable image traditionally “male” and “female” companies strive to perpetuate. Moreover, gender diversity awareness could not be predicted solely by time period, indicating an ambiguous social and cultural context.
Gender diversity discourse in the communications of major Bulgarian companies with stakeholders
The twenty-first century is hailed as the time for an abolishment of stereotyping. In recent years, more emphasis receive theories exploring the notion of gender as culturally determined (Morris, 2006) and reproduced via social norms and institutions, such as organizations (Singh & Point, 2006). Social discourse encompasses the view of language “as an element of social life” (Fairclough, 2005). Studying the construction of gender in discourse uncovers social
development trends as some theorists postulate that the concept of masculinity is valued more than femininity worldwide (Morris, 2006).
The effects of media messages in particular have been widely explored, especially concerning gender and advertising (Schroeder & Borgerson, 1998). However, organizational texts also aid the establishment of personal and social identities (Fairclough, 1995). Accordingly, modern-day research on business ethics emphasizes the value of examining how gender is
constructed in discourse (Kelan, 2008). By communicating the desired corporate image, organizational diversity statements exhibit how diversity should be perceived by external stakeholders and inside the organization (Singh & Point, 2006).
Today, production moves towards more service goods and flexible working environments (Kelan, 2008). Nevertheless, women still face “traditional” problems such as the pay gap (Kelan, 2008) and underrepresentation, especially on managerial and executive level (Ryan & Haslam, 2007). Moreover, exposure to even subtle visual cues might perpetuate gendered stereotypes (Klassen, Jasper, & Schwartz, 1993).
Overall, diversity management has only recently formed a noteworthy issue (Merilainen, Tienari, Katila, & Benschop, 2009) and approaches differ between corporations and national
contexts (Singh & Point, 2004). Research on diversity discourse has chiefly focused inside organizations and largely employs reactive methods such as interviews (Grosser & Moon, 2010) and surveys. Research is mainly exploratory and of inductive nature, prompting the need for more case studies (Merilainen et al., 2009). Moreover, little exploration of European
organizational diversity discourse has been done and a useful tool is online corporate communications with stakeholders (Point & Singh, 2004).
Research on organizational gender diversity discourse often focuses on stakeholder communication channels such as corporate reports and website statements. More targeted studies have considered the United States (Gröschl, 2011; Mills, 2005) and the United Kingdom (Grosser & Moon, 2010) as social arenas for diversity discourse. Others have also explored corporate diversity statements in Nordic countries (e.g., Benschop & Meihuizen, 2002; Merilainen et al., 2009; Risberg & Søderberg, 2008; Sund, 2015). Those countries, however, are already more familiar with a nuanced gender diversity discourse than ex-socialist states (Chakarova, 2003; Watson, 1993). Understanding of diversity varies contextually and between cultures mostly due to the construction of power relations (Grosser & Moon, 2010).
Few works have focused on gender diversity discourse in Eastern Europe and even fewer on Bulgaria as an exemplary case study of the region’s characteristics. Most research has
regarded the ethnic aspects of diversity in Bulgaria (e.g., Elchinova, 2001; Ghodsee, 2010; Rechel, 2008; Zhelyazkova, Kosseva & Hajdinjak, 2010). Targeted exploration of gender discourse has mostly concerned the overall societal and economic challenges in front of Bulgarian women in and after the transition period (e.g. Panova, Gavrilova, & Merdzanska, 1993; Sarwoko & Nurdiana, 2013; Stoilova, 2010). Studies have also looked into gender
representations in the field of advertising (Ibroscheva, 2007, 2013) or in a political media context (Danova, 2006).
Thus, the aim of the present study is to examine gender diversity discourse in a Bulgarian corporate context as an exemplary case study of a society “between the East and the West” (Ibroscheva, 2007). To offer a comprehensive view, this research focuses on the discourse in multiple channels of online organizational communication with stakeholders – website
statements, annual/CSR reports, PR releases and social media (Facebook) over the last decade (2006-2016). Hence, the underlying research question is formulated as follows:
RQ: How is gender diversity represented in the communications of major Bulgarian companies with stakeholders between 2006 and 2016?
Literature Review Gender Equality/Diversity
Gender equality is a fundamental aspect of diversity. Workplace equality between the sexes is still far from desirable and is an issue even for developed countries (European Commission, 2015). A lower percent of the eligible female population takes part in the global workforce, compared to male participation (“The Global Gender Gap”, n.d). Gender equality is a human rights issue and is a part of a “sustainable people-centered development” (OSAGI, 2002). The inclusion of female employees on an executive level serves as inspiration for current and potential workers through their leadership and mentor roles, personal and professional network and their
“symbolic” status of opportunity for other women (Terjesen, Sealy, & Singh, 2009).
The terms “gender equality” and “gender diversity” are often used interchangeably in discourse (Sund, 2015). However, gender equality refers semantically to the norm of equal
representation of men and women. It signifies “equal visibility, empowerment, responsibility and participation for both women and men in all spheres of public and private life” (Council of Europe, 2013, p. 5). In contrast, gender diversity represents the outcome of the equality norm (Sund, 2015), such as the multitude of viewpoints that equality should bring. Diversity discourses move beyond quotas set by regulatory bodies and give rise to internalized value in representation (Bear, Rahman, & Post, 2010).
Gender Diversity Discourse
Language has effects outside of the textual boundaries (Burman & Parker, 1993; Fowler, 1996). Discourses are embedded in texts; they represent and shape social reality through concept and object salience (Singh & Point, 2004). Discourse analysis views an object and its meaning as constructed and “(re)produced” in communication (Brandth, 2002; Singh & Point, 2004),
including visuals. Accordingly, the foundation of the present study’s conceptual framework is the critical discourse analysis (CDA), which links discursive practices with the wider social and cultural setting (Fairclough, 1995).
Power relations in society ideologically form discourses (Fairclough, 2005). According to Fairclough, analysis of organizational discourse should be viewed as an integral part of
organizational studies. Discourse analysis can clarify how an organization changes or resists to change, how it deals with external pressures and what outcomes could be expected (Fairclough, 2005).
Moreover, research has uncovered that women would conform to or challenge gendered segregation in an organizational setting depending on their consciousness of discrimination, understanding of corporate politics and eagerness to take action (Ledwith & Colgan, 1996).
Practices such as the overwhelming use of the male pronoun “he” in earlier management literature implicitly speak of underlying sexist assumptions that the standard employee is male, thus transferring them into corporate reality (Alvesson & Billing, 1997).
Hence, close exploration of companies’ diversity discourses ultimately uncovers
fundamental elements of how diversity is perceived internally. However, research has primarily looked into the more regulatory aspects of gender equality in an organizational context (e.g. Casey, Skibnes, & Pringle, 2011; Lewis & Campbell, 2007) or has explored direct and
quantifiable diversity outcomes such as corporate social responsibility (CSR) ratings (Bear et al., 2010).
Frames in gender diversity discourse. Discourse requires a multitude of frames (Hope, 2010; Singh & Point, 2004) through which an individual assigns meaning (Minsky, 1975). According to Goffman (1974), individuals “organize” experiences through “natural” and “social” frameworks which provide crucial information. The “social” frame subjects the individual to rules in society (Goffman, 1974). In a similar vein, according to Minsky (1974), discourse encompasses a hierarchical “network” of frames and detailed interrelated sub-frames. Van Dijk (1977) also sees discourse as a “macro structure” that is processed by multiple frames. The present study utilizes this structured conceptualization.
Hence, frame analysis focuses on concrete fragments of issue definition (Hope, 2010). Framing theory is rooted in the selection of some aspects of reality that are made “more
noticeable, meaningful or memorable” in a text (Entman, 1993, p. 53). Frames are employed to recognize, evaluate and define a problem or an event (Entman, 1993).
Expanding the CDA by Vaara, Kleymann and Seristö (2004), Singh and Point (2006) identified the most salient frames in corporate diversity discourse. They classified the most widely used framing tools as problematization, rationalization, fixation, reframing and
naturalization of diversity, accompanied by symbols such as photos. The fixation frame regards the liberal notion that those most deserving will reach their goals regardless of diversity practices. Problematization refers to the framing of diversity as a “problem”, presenting it as something to be protected or as a part of an anti-discrimination policy. In turn, reframing views the issue “in
perspective”, includes solutions or new strategies and connects it to practice. Rationalization offers justifications or explanations of diversity or links it with business opportunities. Finally,
naturalization views diversity as natural and to be expected.
The operationalization of gender diversity discourse in this study translates the above presented framework into quantifiable measures. Elements of this conceptualization occur in other studies on diversity discourse such as the emphasis on individual qualities (Gröschl, 2011), the meritocracy discourse (Risberg & Søderberg, 2008) or the “discrimination and fairness perspective” in diversity management (Dass & Parker, 1999).
Gender diversity discourse and industry. Women’s perceived inequality differs across industries (Aitchison, 2000). In the service sector, most employees are women but men occupy the higher levels of management (Aitchison, 2000). The chemical industry is also considered not “sensitive” to gender issues (Benschop & Meihuizen, 2002). Overall, most industries
underrepresent women in annual reports photos, while traditionally male industries such as construction and technology often do not feature them at all (Benschop & Meihuizen, 2002). Workers in the clothing industry are predominantly female (Hesmondhalgh & Baker, 2015) because of the “feminization” of the workforce (Hossain, Mathbor, & Semenza, 2013). The
phenomenon results from a gendered culture where women are considered docile, less aggressive than men, easier to manage and less likely to join trade unions (Hossain et al., 2013). Other “feminized” professions include primary teaching, nursing and the beauty industry
(Hesmondhalgh & Baker, 2015). Sectors connected with men are driving, plumbing, car sales and the heavy industry (Hesmondhalgh & Baker, 2015).
As predominantly “male” sectors rarely emphasize gender diversity in stakeholder communications, the first hypothesis states:
H1: Mentions of gender diversity in the stakeholder communications of Bulgarian organizations differ in occurrence, depending on the industrial sector where the organization operates.
a) Organizations operating in the consumer-oriented goods and services sectors mention gender diversity more often than companies in the heavy industry do.
Around 50% of major European companies mention diversity on their websites and most of the statements refer to anti-discrimination practices (Point & Singh, 2003; Singh & Point, 2004). Thus, the emphasis most often falls on the individual merits of employees (fixation frame) and diversity is often presented in normative terms (problematization). Following from the aforementioned “feminization” issues, this trend is likely to be stronger in the heavy industry. Sectors with more diverse personnel are assumed to have implemented a less straightforward diversity strategy:
H2: Organizations frame gender diversity in stakeholder communications differently, depending on the industrial sector where the organization operates.
a) Organizations operating in the heavy industry employ the fixation and
problematization frames regarding gender diversity more often than companies operating in the consumer-oriented goods and services sectors do.
b) Organizations operating in the consumer-oriented goods and services sectors employ the reframing and rationalization frames of gender diversity more often than
companies in the heavy industry do.
Visual materials. As complimentary to the text discourse, this study also examines all visuals. Photographs reinforce the framing tools used in discourse (Singh & Point, 2006) and provide subtle cues about gendered behavior (Goffman, 1979). Multiple studies have shown that gender-stereotyped media depictions affect the audience’s attitudes and expectations towards women, including perceived professional roles in society (Kilbourne, 1990; Lindner, 2004).
In the study Gender Advertisements (1979), Erving Goffman established an influential framework for the analyzing of gender roles in advertising. Goffman focused on the implicit sexism in the industry by studying the nuances in body positions. His key observation is that the standard body postures in advertising portray women’s vulnerability to men, thus conveying status and power. Today, little has changed in media depictions since his research on gendered portrayals (Kang, 1997; Lawton, 2009; Schroeder & Borgerson, 1998). Women are still frequently objectified in visuals and shown as “withdrawn” or in subordinate positions to men (Lindner, 2004).
Goffman’s frame/category licensed withdrawal relates to depicting women as distracted from the situation. Function ranking visualizes men as instructors and women as subordinates. Portraying women as occupying less space (or not being center) is regarded by relative size. Feminine touch conceptualizes depictions of females cradling/caressing an object. Ritualization
of subordination decodes more overt images where the woman displays a physical act of subordination.
There is a lack of research on visual materials in a corporate context as most studies have focused on media messages. Of crucial importance is the context where images are “produced and consumed” (Schroeder & Borgerson, 1998). More targeted studies have demonstrated that the visual representations of gender diversity in financial reports are mainly stereotypically masculine and reinforce traditional gendered labor division (Benschop & Meihuizen, 2002). Elements of Goffman’s conceptualization such as women’s depiction as physically below the male figure are present even in annual corporate reports (Benschop & Meihuizen, 2002).
Studies have proven the persistent nature of all of Goffman’s categories/frames, with some decline over time in the implementation of feminine touch and ritualization of
subordination (Kang, 1997). Relative size and function ranking are often visible in media messages featuring both men and women (Zotos & Tsichla, 2014).
Thus, the third hypothesis states:
H3: In visual materials found in organizational communications with stakeholders, the categories/frames licensed withdrawal, relative size and function ranking are most commonly observed.
Researchers have stipulated that occupations are mechanisms of power in society (Johnson, 1972), as evident in the strategic and exclusionary nature of professions (Parkin, 1979). Connell (1987) postulates that gender divisions are integral to “production itself”. Presently, gendered practices still dominate employee hiring and production to the extent that female sexualization plays a role in women’s career development (Hossain et al., 2013). Occupational fields are strongly associated with either females (e.g., services) or men
(Hesmondhalgh & Baker, 2015) and communications with stakeholders reflect such corporate identities (Benschop & Meihuizen, 2002).
Hence, the fourth hypothesis states:
H4: There is a lesser degree of stereotyping in visuals posted by companies operating in the consumer-oriented goods and services sectors than in visuals posted by companies in the heavy industry.
Organizational Communication with Stakeholders
Organizational communication aids the forming of corporate-stakeholder relationships (Rybalko & Seltzer, 2010). An adequate communication often incorporates CSR messages (Dawkins, 2005), seeing that organizations should attend to their stakeholders’ needs to build lasting relationships with them (Kent, Taylor, & White, 2003). CSR could be defined as “actions that […] further some social good, beyond the interests of the firm” and regulations
(McWilliams & Siegel, 2001, p.117). Hence, informing about gender diversity can be considered going beyond non-discrimination laws and a part of CSR strategies. Not discussing diversity is a “message in itself”, signaling that diversity is not important enough for the organization (Singh & Point, 2006).
To facilitate the forming of stakeholder relationships, companies have to provide online information of worth to the audience (Kent & Taylor, 1998). Stakeholders, especially
prospective employees, progressively value information on CSR activities (Gröschl, 2011). The “manner” of communicating about social goals is also in flux due to the evolving social
expectations (Ashforth & Gibbs, 1990).
H5: Mentions of gender diversity in Bulgarian organizations’ communications with stakeholders increase over the past decade (2006-2016).
Social media and the most influential medium Facebook. The stakeholders’ attention and expectations have shifted “from rationalizing product attributes and service qualities to emotional engagements and para-social experience” (Shen & Bissell, 2013, p. 633). Research on online
discourse has lagged behind the “explosive growth” of social media (Kushin & Kitchener, 2009), especially in an organizational context (Treem & Leonardi, 2012). Social networking sites (SNSs) like Facebook allow customers to engage in dialogue with the organization, as well as with other consumers (Shen & Bissell, 2013) and are utilized to foster brand loyalty (Hermida, Fletcher, Korell, & Logan, 2012). Social media simplifies the audience’s involvement and users are now able to select, interpret and disseminate messages easier than ever before (Hermida et al., 2012). With the rise of SNSs, corporate communication is “democratized” and has the potential of going “viral” (Kietzmann, Hermkens, McCarthy, & Silvestre, 2011).
Facebook is the most extensively used social network, daily connecting users from different age and backgrounds (Shen & Bissell, 2013). As such, it provides excellent means for companies to be more personally involved with stakeholders (Shen & Bissell, 2013). The ubiquitous utilization of Facebook offers more opportunities to communicate with a brand or express personal opinions and criticism. As an overwhelming number of businesses use social media (Shen & Bissell, 2013), CSR demands increase daily. Hence, hypothesis six states:
H6: Over the past decade (2006-2016), more mentions of gender diversity and more visual materials featuring women are found on companies’ Facebook posts than in website statements, PR releases and CSR or annual reports.
The Social and Economic Landscape in Eastern Europe
The social contexts that form discourses and that discourses in turn help shape, vary between countries (Merilainen et al., 2009), regions and markets. Eastern Europe is an arena of complex political and social processes, representing a unique combination of opposing traditions between socialism and capitalism, class struggles and Western democratic ideals.
Demands for political changes under the Soviet regimes throughout Eastern Europe escalated in the late 1980’s. The revolutions led to the dissolution of the Soviet Union and to a process of democratization (transition) throughout the region. The democratization has offered a new space for women’s professional identification but also opportunities for men to distance themselves from women socially (Watson, 1993). The earlier transitional period reflects the “rise in masculinism” due to the lack of leveling in gender relations (including in law), which
socialism had imposed (Watson, 1993). However, socialism also set the definitions of gender in neo-traditional boundaries; Hence, in the region “deep-seated notions of gender difference often go hand in hand with a lack of any real sense of gender inequality” (Watson, 1993, p. 71).
In the earlier days of the transition, the issue of women’s rights was secondary to the establishment of a stable political and economic foundation in the region (Watson, 1993). Moreover, with the opening of the economy, Eastern Europe faced the patterns which the Western female employment model offered, including the “gender-specific division” of professions (Watson, 1993). Thus, without the socialist equality policies, gendered power relations re-emerged in society, displaying an ambiguous mix of old traditions and new Western values (Metcalfe & Afanassieva, 2005).
Bulgaria as a case study. In Bulgaria, at the turn of the new century, socialist notions were still formative of women’s rights discourses (Daskalova, 2000). For the biggest part of the transition period, organizations dedicated to social issues had a politically limited range for autonomous action (Daskalova, 2000). A large part of the gender equality discourses issued by such organizations reflects the patriarchal communist tradition with the difference that women are no longer “required” but encouraged to work (Daskalova, 2000).
In Bulgaria, men presently constitute 66% of the workforce, while the proportion of male population is 49% (NSI, 2016). Moreover, male employers are over three times more than female employers (NSI, 2016). More women than men are represented in the education and financial sector, as well as in service-oriented sectors such as tourism and leisure (NSI, 2016).
Researchers have postulated that the contemporary Bulgarian woman has internalized three distinct viewpoints of women’s role: the Oriental tradition (through the centuries-long Turkish rule until 1878), patriarchal Eastern Orthodox, and the totalitarian–socialist model (Panova et al., 1993). These intricate socially hereditary concepts result in a difficulty for women to “identify” and “accept” themselves (Panova et al., 1993). Thus, leading to internalization of the new forms of sexism in gender discourse (Ibroscheva, 2007).
Method Design
For the purposes of this study, I employed a quantitative content analysis. This unobtrusive method allows for the data collection from each sampled company as all bigger organizations have an online presence. It is applicable for uncovering and quantifying the presence of gender diversity framing in organizational texts via manageable categories in a codebook.1 Content analysis has an
advantage over surveys when a more socially-sensitive topic is concerned due to the lack of social desirability effects and it is not influenced by low response rates (Singh & Point, 2006). Content analysis is suitable for this study’s aim for a descriptive exploration of gender discourse in
organizational communication as the researcher is part of the audience who has to deconstruct the message (Singh & Point, 2006).
Sample
Companies. The present sample comprises the first twenty biggest Bulgarian companies regarding revenue, growth and employment, based on the leading Bulgarian economic
newspaper’s latest annual Corporate Top 100 list (Capital, 2016).2
It includes private and two publicly listed organizations. The online edition of the weekly newspaper Capital is viewed by 400,000 unique visitors monthly (“Economedia”, n.d.).
Organizational communication with stakeholders. To account for a full capture of gender diversity discourse over time, a census sample of corporate stakeholder communications was drawn. Where data are archived, I explored the timeframe of the last decade (2006-2016). The time of the data collection was December 2016 and I included in the analysis all data going back to December 2006.
The organizations’ publications were examined for statements about gender diversity.3 Only documents with relevant mentions were included in the analysis. Table 1 in Appendix A presents an overview of the sample.
Online diversity statements/corporate diversity policies. I reviewed all twenty
companies’ websites for any official statements regarding gender diversity, resulting into n = 42 statements.
Annual and CSR reports. Annual/CSR reports were accessed through the websites of the sampled companies where available as downloads (n = 67). Any mentions of gender diversity policies are voluntary statements as there are no legal restrictions on non-financial matters disclosure (Accountancy Act, 2001).
PR releases. A census sample of all published organizational PR releases for the period 2006-2016 was drawn, utilizing a popular national PR website -Dnevnik.Bg/PR_Sfera. It is a part of a daily business and political newspaper with over 40,000 daily readers (Raykovska, 2010). The final sample was n = 76.
Facebook communications. A census sample was drawn of all organizations’ Facebook publications dating from their first post, until the point of data collection. Facepager was used as a tool in the extraction and filtering of all relevant Facebook posts, along with photos, resulting in a final sample of n = 239.
Where available, I analyzed all visual materials across the aforementioned stakeholder communication channels. Of interest were any visuals depicting (presumed) employees or the working environment where at least one person was discernible. The total sample comprising Facebook images and visuals in all types of publications was n = 322.
All units of analysis were downloaded, archived and coded on the basis of a previously formulated codebook (see Appendix C). The resulting dataset was analyzed with SPSS software.
Measures
Industry sector. I operationalized this variable based on The Industrial Classification Benchmark (ICB), utilized by stock exchanges globally (Hawkins, n.d.).4 The industries “Consumer goods” and “Technology” were split by supersectors to include activities in the heavy industry
(e.g., Automobiles) and in the end-customer industry (e.g., Personal & Household). Table 1 in Appendix A provides full representation.
Gender diversity discourse – frames. Extending Vaara et al.'s (2004) typology of strategic discourse practices, Singh and Point (2006) identified six frames present regarding gender and ethnicity in corporate websites. Namely, fixation, problematization, rationalization, reframing and naturalization. I operationalized each of the frames with four binary variables (questions) coded with “0” No and “1” Yes. If at least one of the items is coded as “Yes”, I considered the frame present. For the full operationalization, refer to Appendix C.
Visual materials. Along with the textual discourse, I examined all visual materials via Goffman’s (1979) framework for analyzing gender roles in advertisements. He found that more subtle cues in gender representation occur in six categories/frames: relative size, feminine touch, function ranking, the family, ritualization of subordination and licensed withdrawal. Following later research applications of Goffman’s findings, I added two more categories – location
(Umiker-Sebeok, 1996; Lindner, 2004) and body display (as added to Goffman’s model by Kang in 1997).
Generally, quantitative studies replicating Goffman’s observations operationalize his six categories on a dichotomous level, indicating whether elements are observable in the photo or not (e.g, Kang, 1997; Zotos & Tsichla, 2014). The present research also operationalizes the categories by a varied number of interrelated Yes/No questions. However, to allow for a more nuanced examination of stereotypical representations, a middle category was included when both male and female figures demonstrate the behavior. For each category, a weighted index of stereotyping scores was computed (see Results section). A higher score indicated a higher stereotyping level. For the full operationalization, refer to Appendix C.
Inter-Coder Reliability
The materials were analyzed by the researcher, who is a native Bulgarian speaker. For the inter-coder reliability procedure, a second Bulgarian-speaking coder joined in. Before the
reliability analysis, the second coder received a training of approximately ten hours to familiarize themselves with this project’s operationalizations. The coders discussed the coding instruments and reviewed the literature they are based on. Afterward, a pilot test was conducted with 50 publications by ten companies outside the final sample. It preceded the reliability analysis and tested for any lingering disagreements or ambiguities concerning the codebook. All discrepancies were reviewed and discussed.
For the inter-coder reliability analysis, a 10% random sample of the final sample was drawn as per Neuendorf’s (2002) recommendations when the expected reliability level is not low. The random sample resulted into 54 cases and each coder worked on 27 cases individually. To test for the inter-coder reliability, Krippendorff’s alpha was calculated for all 44 variables. The statistic is suitable for any number of coders and applicable for variables at different levels of measurement (Hayes & Krippendorff, 2007; Krippendorff, 2004).5 The lowest value was .75 for the variable “”Is the subject of the photo shown in a domestic setting or in a unidentified environment?” Six other variables displayed a value under .80; Appendix B provides a full overview.
While .80 is considered acceptable in most situations, the value of .70 is agreeable in exploratory research (Neuendorf, 2002). Four of the lowest values regarded the
operationalization of Vaara et al.’s (2004) strategic discourse practices. All other variables ranged between .83 and 1 (see Appendix B).
Data Analysis
To test the first three hypotheses, which regarded frequencies of gender diversity
mentions and frames presence, I utilized the chi-square test of independence.6 Where more than 20% of cells displayed expected observations less than 5 and if there were cells with expected values less than 1, I employed Fisher’s exact test.7
For crosstabulating variables with more than 2x2 categories, the Fisher-Freeman-Haltonextension of the Fisher’s exact test was used as it is applicable for larger contingency tables (Conover, 1980; McDonald, 2014). To test for a
directional relationship between industry sector and frames, I employed Goodman and Kruskall’s tau statistic.8
For hypothesis four and six, the Wilcoxon-Mann-Whitney test was performed. I used this rank-based nonparametric test to determine any differences between two industry sectors
(consumer-oriented and heavy) and type of publication (Facebook post and all other publications) on degree of stereotyping and number of women pictured, respectively.
Ultimately, I ran a logistic regression to test hypothesis five with year of publication as independent variable and mentions of gender diversity as a binary dependent variable.
Results
The first hypothesis predicted that mentions of gender diversity differed in occurrence, depending on the industrial sector.9 To gauge whether there was an association between these nominal variables, I employed a chi-square test of independence. However, the assumptions were not met as 61% of the cells in the resulting contingency table had expected values less than 5 (Cochran, 1954).
The two variables were not independent, as assessed by a Fisher-Freeman-Haltontest, p < .001. To see precisely where this association occurs, I conducted a 2x2 Fisher's exact test for each of the 36 possible pairwise comparisons between gender diversity mentions and industry sector (9 categories *(9-1)/2 = 36). I did a Bonferroni correction for the multiple comparisons. Thus, for the pair to be statistically significant at the .05 level, the p value must be less than .05/36 or .0014.
Only two comparisons were significant, namely Oil and Gas vs. Industrials (p = .001, Fisher’s exact test) and Oil and Gas vs. Telecommunications (p < .001). Companies in the Oil and Gas sector did not refer to gender diversity explicitly, whereas 55% (n = 10) of all 18
explicit mentions were made in the Telecommunications industry (see Table 1). Another 22% (n = 4) of all distinct mentions were found in the Industrials sector. Hence, companies operating in the Telecommunications and the Industrials sectors regard gender diversity explicitly more often than those in the Oil and Gas sector.
Additionally, to test for directional effects, Goodman and Kruskal’s tau was run to determine if explicit mentions of gender diversity (dependent variable) could be better predicted by knowledge of the industry sector (independent variable). Taking industry sector into account, prediction of explicit mentions was improved by 17% (τ = .17, p < .001).
Hypothesis 1a predicted that the consumer-oriented goods and services sectors mentioned gender diversity more often than the heavy industry.10 To test for the hypothesized assumption, I ran a 2x2 Fisher’s exact test because 25% of the cells had expected frequencies < 5. There was a statistically significant association between industry and gender diversity mentions, p = .039. Both industries equally referred to gender diversity explicitly - 4 times. Hence, the first hypothesis was partially upheld as only hypothesis 1a was rejected.
Table 1
Frequencies of Explicit Gender Diversity Mentions by Industry Sector
Industry Sector
Are there any explicit references to gender diversity?
Total
Yes No
Oil & Gas 0 48 48
Industrials (Construction & Materials; Industrial Goods & Services)
4 8 12
Consumer Goods - Automobiles & Parts 0 3 3
Utilities (Electricity; Gas & Water) 0 24 24
Consumer Goods (Food & Beverage; Personal & Household Goods)
2 3 5
Health Care 0 15 15
Consumer Services (Retail – Food & Drug, General Retail)
2 13 15
Travel & Leisure 0 10 10
Telecommunications 10 35 45
Total 18 159 177
The second hypothesis stated that organizations framed gender diversity mentions differently, depending on the industrial sector.11 I employed the Fisher-Freeman-Haltontest for all four 9x2 contingency tables as the overall frequencies were small. For the fixation and the problematization frames, no significant association with industry sector was found (p = .486, p = .468, respectively). However, there was a significant association between industry sector and rationalization and reframing, p = .011 and p = .013, accordingly (see Table 2). Knowledge of
the industry sector improved prediction of rationalization occurrence by 12% (τ = .12, p = .008) and reframing utilization by 11% (τ =11, p = .014).
Hypothesis 2a further specified that the heavy industry employed the fixation and problematization frames more often than the consumer-oriented sectors.I conducted a 2x2 Fisher’s exact test as the overall frequencies were small. There was no significant relationship between fixation and problematization frames occurrence and industry sector (p = .170, p = .087, respectively). Thus, hypothesis 2a was rejected.
Hypothesis 2b predicted that the consumer-oriented sectors employed the reframing and rationalization frames more often than the heavy industry. Hereby, I ran a chi-square test of independence. All expected cell frequencies were larger than 5. There was a statistically
significant association between industry sector and reframing occurrence, χ² (1) = 4.03, p = .045. The opposite result of predicted was evident – companies in the heavy industry employed the reframing frame more often (n = 40) than those in the consumer-oriented sector (n = 14).
Knowledge of the industry sector improved prediction of reframing utilization by 4% (τ = .04, p = .036). Moreover, there was a statistically significant association between industry sector and rationalization occurrence, χ² (1) = 10.81, p = .001. As above, the opposite result of expected occurred - organizations in the heavy industry employed the rationalization frame more often (n = 31) than those in the consumer-oriented sector (n = 15). Taking industry sector into account, prediction of rationalization frame presence was improved by 11% (τ = .11, p = .055).
Hypothesis 2b was thus rejected.
The third hypothesis predicted that Goffman’s categories/frames licensed withdrawal, relative size and function ranking would be the most commonly observed for all industries. To test this assumption, I performed a chi-square test of independence.12
Table 2
Frequency of Frames Presence (Vaara et al., 2004) by Industry Sector
Frames Total Industry Sector Oil & Gas Industrials Consumer Goods- Cars Utilities Consumer Goods (Food) Health Care Consumer Services (General) Travel & Leisure Telecom- muni-cations χ² p Fixation 167 44 12 3 24 5 14 11 10 44 n.a. .486 Problema 153 40 12 3 22 3 12 12 9 40 n.a. .468 Rationaleb 86 16 8 1 6 4 10 11 7 23 n.a. .011* Reframing 98 15 9 3 13 4 7 10 7 30 n.a. .013*
Note. Chi-square test of independence was not computed because more than 40% of cells had an expected count less than 5. Fisher-Freeman-Halton significance value is reported.
Variables cross-tabulated by industry sector and frame Present and Not Present. a
– Problematization frame b
– Rationalization frame * - significant at the .05 level.
There was no statistically significant association between industry sector and relative size frame occurrence, χ² (7) = 12.88, p = .075 (Table 3). Regarding the other five frames, the overall frequencies were small. Hence, I ran the Fisher-Freeman-Halton test. There was an association between industry sector and the occurrence of licensed withdrawal and ritualization of
subordination, p = .005 and p = .013, accordingly (Table 3). The third hypothesis was partially supported for the licensed withdrawal frame, which was the most common across industries (n = 287). Second most visible was ritualization of subordination (n = 283).
The fourth hypothesis predicted that there was a lesser degree of stereotyping in visuals posted by the consumer-oriented sectors than by the heavy industry. Firstly, all items used to operationalize Goffman’s (1979) categories/frames (see codebook in Appendix C) were added into aggregate variables. Some frames however had a differing number of questions. Following Kang’s (1995) approach to overcoming this weighting problem, each aggregate variable was divided by the number of items (e.g., all three relative size items were summed up and then divided by 3). The resulting ratio-level variables thus had a range of 0-2 and were comparable.13
As the assumptions for an independent samples t-test did not hold, the non-parametric Wilcoxon-Mann-Whitney test was conducted to determine if there were differences in stereotyping score between the consumer-oriented and heavy industry.
A visual inspection revealed that the distributions of the stereotyping scores for both industry sectors were similar. The feminine touch stereotyping score (ranging from 0 to 2) was statistically significantly higher in the consumer-oriented sector (Mdn = 0.67) than in the heavy industry (Mdn = 0.33), U = 2.44, z = 3.27, p = .001. Similarly, the licensed withdrawal stereotyping score was statistically significantly higher in the consumer-oriented sector (Mdn = 0.50) than in the heavy industry (Mdn = 0.38), U = 3.47, z = 3.19, p = .001.
Table 3
Frequency of Goffman’s (1979) Categories Presence by Industry Sector
Goffman Categories Total Industry Sector Oil & Gas
Industrials Utilities Consumer Goods (Food) Health Care Consumer Services (General) Travel & Leisure Telecom- muni-cations χ² p Relative Size 189 18 11 5 2 15 57 6 75 12.88 .075 Licensed Withdrawal 287 21 15 16 2 24 100 10 99 n.a.a .005* Function Ranking 41 4 3 1 2 2 14 3 12 n.a.a .061 Feminine Touch 220 26 15 10 1 15 71 5 77 9.54 .216 Ritualization of Subordination 283 24 16 11 2 23 99 10 98 n.a.a .013*
Body Display 43 0 4 1 0 5 16 2 15 n.a.a .117
Location 47 2 2 0 0 2 21 2 18 n.a.a .240
Note. a - chi-square test of independence was not computed because more than 20% of cells had an expected count less than 5. Fisher-Freeman-Halton significance value is reported.
Variables cross-tabulated by industry sector and frame Present and Not Present. * - significant at the .05 level.
There was a statistically significant difference in median relative size stereotyping scores between the heavy industry (Mdn = 1.50) and the consumer-oriented sector (Mdn = 1), U = 592.50, z = -2.32, p = .020. A full representation of the results is included in Table 4. Thus, hypothesis four was partially supported for the relative size stereotyping score.
Hypothesis five predicted that mentions of gender diversity increased over the past decade (2006-2016). A binominal logistic regression was run with year as the predictor and the dichotomous dependent variable Gender Diversity Mentions. The assumptions of independence of observations, a minimum of 15 cases per independent variable and of linearity were upheld.14
The logistic regression model was not statistically significant, χ2
(1) = 1.22, p = .270. However, the Hosmer and Lemeshow goodness of fit test revealed that the model is not a poor fit to the data, χ2
(5) = 1.83, p = .872. The model only accounted for 2.4% (Nagelkerke R2) of the variance in gender diversity mentions and correctly classified 89% of the cases. The predictor Years was not a statistically significant contributor to the model (p = .308). However, the direction was contrary to the hypothesized effect (see Table 5) – the odds of not mentioning gender diversity increased with time. Thus, hypothesis five was rejected.
Ultimately, hypothesis six predicted that more mentions of gender diversity and more depictions of women would be found in Facebook posts than in other publications.15 The chi-square test assumption of no more than 20% of cells with expected counts less than 5 did not hold. There was no statistically significant association between the type of post and gender diversity mentions, as assessed by Fisher’s exact test, p > .99.
Table 4
Differences on Goffman’s Categories (1979) Stereotyping Score between Industries
Goffman categories
Industry
U p
Consumer-oriented sector Heavy industry
Median Minimum Maximum Median Minimum Maximum
Feminine Touch 0.67 0 1.33 0.33 0 0.67 2.44 .001** Licensed Withdrawal 0.5 0 1.5 0.38 0 1 3.47 .001** Relative Size 1 0 2 1.50 0 2 592.5 .020* Body Display 2 0 2 2 0 2 53.5 .275 Location 1 0 2 0.5 0 2 53.5 .409 Function Ranking 1 0 2 1.5 0 2 63.5 .976 Ritualization of Subordination 0.33 0 1 0.34 0 1.17 2.83 .305 Note. * *p < .05, **p < .01.
Table 5
Summary of Logistic Regression Predicting Likelihood of Explicit Gender Diversity Mentions by Year (N = 99)
Predictor
Gender diversity mentions
95% CI for Odds Ratio (eB) B SE B eB Year (2006 - 2016) -.18 .17 .83 .59 1.19 Constant 368.79 359.76 χ2 1.22 df 1
Note: eB = exponentiated B. Dependent variable coded as 1 for No (No Mentions) and 0 for Yes (Mention Present).
*p < .05.
To test the second part of hypothesis six, I conducted the non-parametric Wilcoxon-Mann-Whitney test as the assumptions for an independent samples t-test were violated.16 There was no statistically significant difference in median of women included in visuals between Facebook posts (Mdn = 1) and all other types of organizational publication (Mdn = 1), U = 6.92, z = -0.60, p = .547. Thus, hypothesis six was rejected.
Discussion
This research set out to explore gender diversity discourse in Bulgarian companies’ stakeholder communications over the past decade (2006 – 2016). Specifically, the project aimed to uncover the tendencies in gender diversity discourse in PR releases, CSR and annual reports, website statements and official online policies, Facebook posts and visual materials. A targeted
approach was undertaken regarding industrial differences on two types of framing devices – Vaara et al.’s (2004) and Goffman’s (1979) framework. The results shed light on generalizable trends within industrial sectors and publication types.
Firstly, explicit gender diversity mentions differed in occurrence, depending on the industrial sector where the companies operated. The Telecommunications sector regarded gender diversity the most over the decade, in line with the predicted and theoretically established trend of “feminization” of the service sector (Hossain et al., 2013). A large part of the
Telecommunications business consists of end-customer interactions and frequent product promotion. It is understandable that telecommunications operators would aim to connect with as many of their diverse clients as possible, including women, by communicating diversity
practices.
An unpredicted result occurred in the Industrials sector, which is considered part of the “heavy” and traditionally male industry (Benschop & Meihuizen, 2002). Contrary to Benschop and Meihuizen’s (2002) observations, hereby companies displayed a noticeable gender diversity awareness. Generally, both the heavy and the consumer-oriented industries exhibited an equal tendency to regard gender diversity. However, the amount of explicit references instead of broader non-discrimination discussions remained only a handful for the entire time period. This general observation is in line with Singh and Point's (2006) conclusions that only 6% of top European companies promote gender diversity online. The finding lends credence to the shared sentiment of a large part of Bulgarian women who despite facts to the contrary, believe that their problems are “undistinguishable” from those of men or the society (Ghodsee, 2004). Moreover, it corroborates that there is a continuing resistance to feminist issues in Eastern Europe
Secondly, I uncovered a significant association between industry sector and rationalization and reframing frames presence. While they were not the most frequently observed frames overall, they displayed the most non-uniform tendencies across companies. Contrary to expectations, the heavy industry employed reframing and rationalization
considerably more often than the consumer-oriented sector when broadly discussing diversity. The observations correspond to Bellard and Rüling’s (2001) findings that companies assign differing meanings to diversity according to their current situation and market scope. Firms in the heavy industry operate on a broader scale and, presumably, feel more pressure to mollify stakeholders with socially-aware statements as their actions inevitably affect climate change (Meehl et al., 2007). In line with legitimacy theory (Suchman, 1995), they provide more
diversity statements and statistics to retain and enhance their CSR reputation. The heavy industry was generally more active than the consumer-oriented sector in terms of written stakeholder communications.
The above discussed shows that end-customer sectors do not exhibit enough commitment to gender diversity, despite women being overrepresented in employment (Risberg & Søderberg, 2008). As the heavy industry is considered traditionally “male”, it might strive to branch out and include more diverse employees, thus highlighting its diversity management practices (Gröschl, 2011). However, providing rationalizations for gender diversity also corresponds to Kaptein's (2004) findings that companies would regard anti-discrimination practices more often than actually assign a value to diversity. An encouraging observation is that companies also communicate a significant amount of diversity statements in a change context (regarding the reframing frame), signifying some internalization of the presented ideas into the corporate culture (Vaara et al., 2004).
Thirdly, this study broadened the gender diversity discourse analysis by considering visual materials as per an extension of Goffman’s (1979) framework of gender roles
presentation. Overall, Bulgarian companies most commonly employed the licensed withdrawal frame. Moreover, the degree of stereotyping regarding licensed withdrawal was statistically significantly higher in the consumer-oriented sector than in the heavy industry, contrary to expectations. This finding parallels Morris et al.’s (2006) study which corroborated that women are mostly depicted as product users and not authority figures, as opposed to men’s frequent presence in a “serious” setting.
Benschop and Meihuizen (2002) conclude that firms in the construction and technology sectors rarely feature women. It appears that if they do, the visualizations have to represent a hands-on and a determined approach, associated with the “male” qualities of the sectors (Powell, Bagilhole, & Dainty, 2009). This assumption was also corroborated by the stereotyping score for the frame feminine touch. Countering expectations, it was statistically significantly higher in the consumer-oriented sector than in the heavy industry. Even though this means less stereotyping in sub-sectors such as construction, it alludes to the desired image traditionally “male” companies strive to perpetuate. It seems that it is considered more appropriate to display distancing and non-utilitarian touch in sectors mostly associated with women. Moreover, this corresponds to
Gherardi's (1994) postulation that when women are accepted in a masculine environment, they are commonly made “the object […] that typifies the community” (p. 603). Organizations in the heavy industry could be displaying this phenomenon where men treat women as equals but their defining characteristics as females are diminished.
After licensed withdrawal, the second most visible frame in images was ritualization of subordination, conceptualizing unnatural postures, shyness or subordination (Goffman, 1979).
Even though in line with expectations, it is discouraging that this more overt stereotyping technique is still so prevalent. Contrary to the hypothesized however, the stereotyping score for ritualization did not differ significantly between industrial sectors.
Fourthly, companies in the heavy industry displayed a tendency to portray men as taller or being in the foreground, concordant with expectations. Hence, signifying the importance attributed to male figures in the image. This contradicts previous observations that relative size rarely occurs (Kang, 1997; Umiker‐ Sebeok, 1996) and lends credence to Zotos and Tsichla's (2014) findings that the category in fact makes a comeback in contemporary media messages.
Fifthly and also countering predictions, Facebook posts and all other types of
publications did not differ significantly regarding the number of women in images. It could be argued that today’s “visual culture” has extended to all communications with stakeholders, similar to all media messages (Schroeder & Borgerson, 1998), and has become part of the corporate identity expression (Benschop & Meihuizen, 2002).
Lastly, the conducted logistic regression did not reveal time to be a significant predictor of explicit gender diversity mentions, contrary to expectations. The model accounted for a small amount of variance concerning gender diversity mentions, thus indicating that there are other factors at play than the time period. It is clear that gender diversity discourse is constructed in an ambiguous and non-straightforward social context, which often results in contradictory meanings (Brandth, 2002).
Practical Implications
Across European companies, there is a large variety of framing regarding diversity statements online, despite binding EU diversity management legislation (Singh & Point, 2004).
The present study revealed that there is an overall lack and an inconsistent representation of gender diversity in Bulgarian corporate context. Practitioners should be mindful of the subtle stereotyping cues regarding gender diversity. When communicating with stakeholders, corporate departments have to consider that images’ overt or latent characteristics can perpetuate social understandings (Belknap & Leonard II, 1991) or imply existing hierarchies (Zotos & Tsichla, 2014). Moreover, gender equality is far from being attained on higher executive levels (Singh & Point, 2006). Even if it is achieved, diversity is present in multiple forms; Depicting more women in non-stereotypical postures is not enough when they do not also represent other minorities (e.g., race and disability).
Practitioners have to keep in mind that organizations are part of society and their
presentations of gender roles indicate corporate values and social responsibility. Companies can play a significant part in the professional experiences of young potential employees by
communicating more welcoming and socially-aware messages. Moreover, Bulgarian companies frequently regard individual achievements as the determining factor in career development. Even though it seems fair, it may signify overlooking diversity practices (Singh & Point, 2006) or may further establish the privileged positions of a selected few (Merilainen et al., 2009).
Limitations
This study has some limitations which have to be highlighted. Firstly, the research design (content analysis) prevents from drawing causality conclusions. Moreover, as evident by a logistic regression, explaining gender diversity discourse is not achievable in a straightforward manner. As a social phenomenon, the understanding of diversity or diversity discourse is in constant flux and evolution (Merilainen et al., 2009).
Secondly, during the data collection, it became evident that organizations consistently neglected explicit mentions of gender diversity practices. To procure a relevant sample, I considered both the rare explicit mentions and the implied references to gender diversity. It is most likely that companies include the gender aspect when promoting general diversity practices but it was still only assumed during the analysis.
Thirdly, the present research provided an overview of the companies’ voluntary diversity disclosures online. It remains unclear how truthful and complete their implementation is inside the organizations. Furthermore, I looked at corporate publications from the last decade but few of the sampled companies provided access to their older messages online. The publication of annual and CSR reports was also irregular, limiting the generalizability regarding this particular type of communication with stakeholders. Further, website statements are often updated and contributed with relevant observations only for the current year of sampling - 2016.
Fourthly, this study’s geographical field of interest was Bulgaria. Findings might be generalizable to ex-socialist Eastern European countries with similar social developments
(Ghodsee, 2004; Manning, 2004; Watson, 1993) but they are still predominant for Bulgaria only.
Finally, the inter-coder reliability test uncovered seven variables (out of 44) with Krippendorff’s alpha values lower than .80 - most commonly accepted as a threshold
(Krippendorff, 2004). The results were mostly due to the novelty of translating Vaara et al.’s (2004) strategic discourse practices into quantifiable measures and the rarity of the binary categories, leading to a lower Krippendorff’s alpha despite few disagreements. Moreover, the researcher was a coder for the final reliability test due to time constraints. The application of the codebook by a separate second coder might yield different results because of the lack of
Future Research Directions
Further work regarding gender diversity discourse in a corporate context might expand with combining a non-reactive research design such as a content analysis with employee and manager interviews, which would provide a unique perspective on the implementation of diversity practices. McKay (1997) demonstrated that employee perceptions differ largely between men and women, presumably experiencing the same corporate culture. Due to time limitations, this study did not elaborate on such personal experiences but a full overview of the social phenomenon gender diversity cannot be achieved through discourse analysis only (Fairclough, 2005) but also by close observation.
Moreover, companies in the Bulgarian heavy industry exhibited a contradictory but still notable gender diversity awareness. More research is needed to uncover the drivers behind these enhanced socially responsible practices and whether they are driven by economic factors or genuine internalization.
Further, when considering social media channels for communicating with stakeholders, I chose Facebook as a representative medium. This omits media such as LinkedIn, for instance. Facebook users typically present their “leisure-persona”, while emphasizing their professional self on LinkedIn (van Dijck, 2013). Hence, companies might also display different “personas” to seem closer to their stakeholders, depending on the channel. Further research could elaborate on diversity discourse in more types of social media, which users utilize for a slightly different way of communicating.
Conclusion
Gender diversity discourse in organizational messages is highly dynamic and its meaning fluctuates across industries, time and even specific companies within the same sector of
operation. Practitioners need to consider the latent characteristics of the language they employ when communicating with stakeholders online, along with any subtle visual cues. Such
information reveals what diversity means for the particular organization (Singh & Point, 2006), how it is integrated into the corporate culture and what qualities and values are appreciated in this environment. If the company does not display an adequate social awareness in today’s globalized world, it might face reputational damage.
References
Accountancy Act. Закон за счетоводството. (2001). Retrieved from http://lex.bg/laws/ldoc/2135213569
Economedia – Капитал – издания на Икономедия. (n.d.) Retreived from http://www.economedia.bg/izdaniainfo.php?guid=283201
The Global Gender Gap Index Results in 2015. (n.d.). Retrieved from
http://reports.weforum.org/global-gender-gap-report-2015/the-global-gender-gap-index-results-in-2015/
Industry Classification Benchmark (ICB) – Structure. (2016). Retrieved from http://www.icbenchmark.com/structure
Aitchison, C. (2000). Women in leisure services: Managing the social-cultural nexus of gender equity. Managing Leisure, 5(4), 181–191. http://doi.org/10.1080/13606710010001789 Alvesson, M., & Billing, Y. D. (1997). Understanding Gender and Organizations. London:
SAGE Publications.
Ashforth, B. E., & Gibbs, B. W. (1990). The double-edge of organizational legitimation. Organization Science, 1(2), 177–194. http://doi.org/10.1287/orsc.1.2.177
The Australian Government. Department of Innovation, Industry, Science and Research. (2011). Workplace Discrimination Harassment Policy. Retrieved from
https://industry.gov.au/AboutUs/InformationPublicationScheme/Ourpolicies/Documents/W orkplaceDiscriminationHarassmentPolicy.pdf
Bear, S., Rahman, N., & Post, C. (2010). The impact of board diversity and gender composition on corporate social responsibility and firm reputation. Journal of Business Ethics, 97(2),
207–221. http://doi.org/10.1007/810551-010-0505-2
Belknap, P., & Leonard II, W. M. (1991). A conceptual replication and extension of Erving Goffman’s study of gender advertisements. Sex Roles, 25(3–4), 103–118.
Bellard, E. & Rüling, C. (2001, July). Reflections and projections
of boundaries in the diversity management discourses in the United States, France and Germany. In 17th EGOS Colloquium, Lyon, France.
Benschop, Y., & Meihuizen, H. E. (2002). Keeping up gendered appearances: Representations of gender in financial annual reports. Accounting, Organizations and Society, 27(7), 611–636. http://doi.org/10.1016/S0361-3682(01)00049-6
Brandth, B. (2002). Gender Identity in European Family Farming: A Literature Review. Sociologia Ruralis, 42(3), 181–200. http://doi.org/10.1111/1467-9523.00210
Burman, E., & Parker, I. (1993). Discourse analytic research: Repertoires and readings of texts in action. London: Routledge.
Busch, C., De Maret, P. S., Flynn, T., Kellum, R., Le, S., Meyers, B., ... & Palmquist, M. (1994). Content Analysis. Writing@ CSU. Colorado State University. Retrieved from
http://writing.colostate.edu/guides/guide.cfm?guideid=61
Capital. (2016). K100 | Класацията Капитал 100 - най-големите компании в България. Retrieved from http://www.capital.bg/k100/
Casey, C., Skibnes, R., & Pringle, J. (2011). Gender equality and corporate governance: Policy strategies in Norway and New Zealand. Gender, Work & Organization, 18(6), 613–630. Chakarova, V. (2003). Development of Gender Relations in the Context of Social
Transformation: The Case of Bulgaria. Centre International de Formation Européenne. Cochran, W. G. (1954). Some methods for strengthening the common χ2 tests. Biometrics, 10(4),
417-451.
Connell, R. W. (1987). Gender and power: Society, the person and sexual politics. Oxford: John Wiley & Sons.
Conover, W. J. (1980). Practical nonparametric statistics. (2nd ed.). New York, NY: John Wiley & Sons.
Council of Europe. (2013, November). Council of Europe Gender Equality Strategy (2014-2020). REF. 248813GBR2014. Retrieved from
https://edoc.coe.int/en/gender-equality/5992-council-of-europe-gender-equality-strategy-2014-2017.html
Danova, M. (2006). Women in politics in Bulgarian newspapers: Post-feminism in a post-totalitarian society. In N. M. Bamburać, T. Jusić, & A. Isanović (Eds.), Stereotyping: Representation of women in print media in South East Europe (p. 111). Sarajevo: Mediacentar.
Daskalova, K. (2000). Women’s Problems, Women’s Discourses in Bulgaria. In Gal, S., & Kligman, G. (Eds.), Reproducing gender: politics, publics, and everyday life after socialism. (pp. 337–370). Princeton, NJ: Princeton University Press..
Dass, P., & Parker, B. (1999). Strategies for managing human resource diversity: From resistance to learning. Academy of Management Perspectives, 13(2), 68–80. http://doi.org/10.5465/AME.1999.1899550
Dawkins, J. (2005). Corporate responsibility : The communication challenge. Journal of Communication Management, 9(2), 108–119. http://doi.org/10.1108/13632540510621362 Elchinova, M. (2001). Ethnic discourse and group presentation in modern Bulgarian society.
Development and Society, 30(1), 51–78.
Communication, 43(4), 51–58. http://doi.org/10.1111/j.1460-2466.1993.tb01304.x
The European Commission. (2015). Report on equality between women and men. Retrieved from http://ec.europa.eu/newsroom/just/item-detail.cfm?item_id=52696#annual_reports
Fairclough, N. (1995). Media discourse. London: Arnold Publishers. http://doi.org/10.1017/CBO9781107415324.004
Fairclough, N. (2005). Discourse analysis in organization studies: The case for critical realism. Organization Studies, 26(6), 915–939. http://doi.org/10.1177/0170840605054610
Field, A. (2013). Discovering statistics using IBM SPSS statistics. London: Sage.
Fowler, R. (1996). On critical linguistics. In Caldas-Coulthard, C. & Coulthard, M. (Eds.), Texts and practices: Readings in critical discourse analysis. London & New York, NY:
Routledge.
Gherardi, S. (1994). The gender we think, the gender we do in our everyday organizational lives. Human Relations, 47(6), 591–610. http://doi.org/10.1177/001872679404700602
Ghodsee, K. (2004). Feminism-by-design: Emerging capitalisms, cultural feminism, and
women’s nongovernmental organizations in postsocialist Eastern Europe. Signs, 29(3), 727– 753. http://doi.org/10.1086/380631
Ghodsee, K. R. (2010). Muslim lives in Eastern Europe : gender, ethnicity, and the
transformation of Islam in postsocialist Bulgaria. Princeton, NJ: Princeton University Press. Goffman, E. (1974). Frame analysis: An essay on the organization of experience. Cambridge,
MA: Harvard University Press.
Goffman, E. (1979). Gender Advertisements. Cambridge, MA: Harvard University Press.
Gröschl, S. (2011). Diversity management strategies of global hotel groups: A corporate web site based exploration. International Journal of Contemporary Hospitality Management, 23(2),
224–240. http://doi.org/10.1108/09596111111119347
Grosser, K., & Moon, J. (2010). Developments in company reporting on workplace gender equity: A corporate social responsibilty perspective. In Accounting Forum, 32(3), 179–198. Hayes, A. F., & Krippendorff, K. (2007). Answering the call for a standard reliability measure
for coding data. Communication methods and measures, 1(1), 77-89.
Hermida, A., Fletcher, F., Korell, D., & Logan, D. (2012). Share, like, recommend: Decoding the social media news consumer. Journalism Studies, 13(5–6), 815–82.
Hesmondhalgh, D., & Baker, S. (2015). Sex, gender and work segregation in the cultural industries. The Sociological Review, 63(S1), 23–36.
http://doi.org/http://dx.doi.org/10.1111/1467-954X.12238
Hope, M. (2010). Frame Analysis as a Discourse-Method: Framing ‘climate change politics. In Proceedings of Post-Graduate Conference on Discourse Analysis. Bristol, UK.
Hossain, M. I., Mathbor, G. M., & Semenza, R. (2013). Feminization and labor vulnerability in global manufacturing industries: Does gendered discourse matter? Asian Social Work and Policy Review, 7(3), 197–212. http://doi.org/10.1111/aswp.12018
Ibroscheva, E. (2007). Caught between east and west? Portrayals of gender in Bulgarian
television advertisements. Sex Roles, 57(5–6), 409–418. http://doi.org/10.1007/s11199-007-9261-x
Ibroscheva, E. (2013). Selling the Post-Communist Female Body. Feminist Media Studies, 13(3), 443–462. http://doi.org/10.1080/14680777.2012.708515
Johnson, T. (1972). Professions and Power. London: Macmillan.
Kang, M.-E. (1997). The portrayal of women’s images in magazine advertisements: Goffman’s gender analysis revisited. Sex Roles, 37(11–12), 979–996.
