The Relationship between Firm Attributes and Attitudes towards Diversity

Abstract

The attitudes of creative firms towards issues of equality, diversity and inclusiveness (“EDI”) can significantly affect their willingness to sponsor and implement effective measures in the domain. It is, therefore, essential to examine the readily measurable firm attributes that influence these attitudes. We have collected a wide range of data on almost 330 creative businesses. Our empirical investigation establishes a robust and unequivocal pattern. It indicates that more established companies tend not to view the underrepresentation or the discrimination of people with various protected characteristics as problematic. Young, innovative and efficient firms on the other hand are systematically more likely to consider these same issues as prevalent. These findings are in line with the conclusions from the previous literature which relied predominantly on anecdotal evidence. The patterns that we document suggest that EDI policies and recommendations must be tailored to the precise characteristics of the firm implementing them.

1. Introduction

In recent years, there has been growing recognition of the importance of equality, diversity and inclusiveness (EDI) in the creative industries. EDI is essential for the sustainable growth of the creative industries as it enhances innovation, creativity and talent development while also contributing to social and economic development. This paper seeks to explore firm-level attitudes towards EDI in creative industries and draw lessons for policy. We have conducted empirical studies that show stark and consistent patterns indicating that larger and more established firms have a lax attitude towards discrimination and the underrepresentation of various protected characteristics (disabilities, ethnicity, sexual orientation, etc.). On the other hand, more innovative and more efficient firms that also tend to be younger consider that people with such attributes face underrepresentation or discrimination in the creative industries.

The creative industries are an important driver of economic growth and development, accounting for a significant share of employment and economic output in many countries (UNCTAD, 2018 [1]). However, the sector has been criticized for its lack of diversity and inclusiveness, with many individuals from underrepresented groups, such as women, people of color, and people with disabilities, facing significant barriers to entry and advancement (Banks, 2020 [2]; Pratt and Jeffcutt, 2009 [3]: Cunningham et al., 2019 [4]; Watson, 2019 [5]). This lack of diversity not only limits opportunities for individuals but also stifles creativity and innovation, ultimately affecting the sector’s competitiveness and long-term sustainability.

Therefore, there is a need to foster EDI in the creative industries to ensure that the sector is inclusive and reflective of the wider society it serves. In recent years, there has been a growing recognition of numerous initiatives aimed at increasing diversity and inclusion in the workforce and promoting inclusive practices (Banks, 2020 [2]; Nicholls and Lee, 2019 [6]). However, there is little discussion around tailoring these initiatives to observable firm characteristics which are correlated with companies’ attitudes towards EDI policies in general. Our paper serves as a solid starting point for this discussion based on sound empirical evidence.

1.1. Literature Review

The link between the performance of creative firms and their inclusiveness has been well documented. Firms with diverse boards are more likely to have better financial results than those without (Carter et al., 2018 [7]). Diverse teams are also more innovative, perform better and attract a wider range of clients and consumers (Kochan et al., 2003 [8]). Firms with an adequate level of independence at the level of the management board to implement social responsibility practices are also more likely to simultaneously pursue innovation (Shafeeq Nimr Al-Maliki, Salehi and Kardan, 2023 [9]) Diversity can also bring less tangible but equally important value to firms’ operations, such as increased creativity, innovation and talent retention (Jackson and Ruderman, 2020 [10]; Tregaskis and Taylor, 2019 [11]).

However, these links between diversity and profits may not be clear to firms that have not yet prioritized EDI policies. This lack of understanding of the benefits of diversity and inclusiveness poses a significant barrier to fostering EDI (Shah, 2020 [12]). The correct implementation of EDI policies, therefore, requires a fundamental shift in the way firms operate.

Moreover, implementing EDI policies can be challenging and requires significant resources and commitment from firms. They may be more likely to prioritize profit and short-term gain over long-term sustainability, and EDI policies may be viewed as a hindrance to these objectives (Dobbin and Kalev, 2018 [13]; Kanter, 2018 [14]). EDI policies require the development of training programs, the creation of inclusive hiring practices and the establishment of diverse leadership teams (Banks, 2020 [2]). Firms need to re-evaluate hiring policies and implement a culture of inclusivity. Some firms may be hesitant to implement such measures because of the assumption that there is a limited pool of diverse talent to recruit from. A survey by the Creative Industries Federation found that over 75% of creative industry leaders found it challenging to find and recruit talent from diverse backgrounds (Creative Industries Federation, 2021 [15]).

Another significant barrier to the implementation of EDI policies is the fear of backlash from the dominant group. When firms begin to implement EDI policies, individuals from dominant groups may feel threatened and believe that they are being unfairly disadvantaged (Dobbin and Kalev, 2018 [13]). Therefore, EDI measures may be seen as a threat to the status quo, leading to resistance from those who benefit from the current system (Hunt et al., 2015 [16]). This fear of backlash can lead to resistance from both employees and management, ultimately hindering the effectiveness of EDI policies.

Finally, there is a lack of regulatory support for EDI policies in many countries. While some countries have introduced legislation to encourage EDI in the workplace, others have not (UNCTAD, 2018 [1]). Without regulatory support, firms may not see the need to implement EDI policies or may prioritize other policy measures that are mandated by law. For example, in the UK, the Gender Pay Gap Reporting requirement was introduced in 2017 to address gender inequality in the workplace. However, only those firms that have more than 250 employees have an obligation to report on the issue, leaving many smaller companies “free” to address or ignore EDI reporting (Banks, 2020 [2]).

1.2. Objectives and the Theoretical Framework

The objective of this paper is to build on this existing literature to operationalize the firm-level attitudes that may systematically impact how creative businesses view EDI-related challenges. Their attitudes toward the challenges themselves may significantly affect their willingness to sponsor and implement EDI measures and policies. We have collected a sample of more than 300 creative businesses. Specifically, we seek to evaluate whether firm attributes, such as size, turnover, profitability and innovativeness, determine attitudes toward EDI policies.

Our work is novel in several ways. First, we focus specifically on the creative industries which have received relatively little focused attention in the literature on EDI policies and attitudes toward them. Second, we collect data on a large sample of firms, enabling us to conduct a rigorous study. Third, we examine a wide range of observable firm characteristics, allowing us to identify the factors which are most strongly associated with EDI attitudes. The need for the type of rigorous empirical evidence that we provide in this paper cannot be overstated. As observed by Baumann and Albinsson (2020) [17], the current discourse on EDI in the creative industries is often influenced by anecdotal evidence. We briefly summarize the previous pieces of the literature that have attempted similar empirical studies, resulting in conflicting outcomes.

Firm size has been found to be positively associated with the implementation of formal diversity and inclusion initiatives (Berggren and Magnusson, 2019 [18]; Banks, 2019 [19]). At the same time, other studies have suggested that larger firms are less likely to prioritize EDI policies than smaller firms (McPherson et al., 2016 [20]; Pendleton and Valizade, 2019 [21]). Other papers have found no significant relationship between firm size and EDI attitudes (Majid et al., 2020 [22]). Similarly, while some studies have suggested that more profitable firms are less likely to prioritize EDI policies (O’Regan and Ghobadian, 2006 [23]), others have found no significant relationship between profitability and EDI attitudes (Bello et al., 2020 [24]). Research shows that fostering equality, diversity and inclusiveness within an organization can contribute positively to its long-term sustainability and financial performance (Hunt et al., 2018) [25]. Companies that prioritize EDI can benefit from improved decision-making processes, a better understanding of diverse markets, and an enhanced ability to attract and retain top talent (Orlitzky et al., 2003) [26].

Additionally, there is growing evidence that diverse teams can enhance creativity and innovation within an organization which, in turn, may lead to better financial results and increased competitiveness (Lorenzo et al., 2018) [27]. Firms that actively work towards creating an inclusive environment can also bolster their corporate reputation which plays a crucial role in attracting customers, investors and employees (Brammer et al., 2007) [28].

It has also been found that companies that integrate sustainability and EDI into their core business strategies are better positioned to mitigate potential risks and adapt to changing market conditions, thus contributing to long-term financial success (Eccles et al., 2014) [29]. By addressing issues of inequality and underrepresentation, organizations can not only improve their social and environmental performance but also enhance their economic performance in the long run (Husted and de Jesus Salazar, 2006) [30].

More innovative firms are more likely to prioritize EDI policies (Lepak and Snell, 2002 [31]; Appelbaum et al., 2018 [32]). However, it is unclear whether this relationship holds in the creative industries which are characterized by high levels of innovation and creativity as a baseline. Firm characteristics that can be classified as “soft” (such as organizational culture and leadership) have been suggested by the literature as critical drivers of attitudes to EDI-related issues (e.g., Nkomo et al., 2019 [33]; McCormack and Wergin, 2018 [34]). Nevertheless, such “soft” explanatory variables are very hard (if not impossible) to measure rigorously.

In light of some of the conflicting results from the previous literature, our study aims to provide theoretical clarity and empirical evidence on the way that firm-level attributes influence EDI-related attitudes. To do so, we have developed a theoretical framework set on five pillars that allow for establishing systematic, causal and testable relationships. These pillars include insights from institutional theory, the resource-based view (RBV), innovation theory, social identity theory and from ambidexterity theory. We use this five-dimensional theoretical framework to formulate the hypotheses that we test empirically (see Section 2.1). Overall, all of the theoretical pillars suggest—both jointly and separately—that large established firms have limited incentives in engaging with EDI incentives, while smaller, more innovative ones have a vital operational incentive in fostering them.

1.2.1. Institutional Theory

Institutional theory offers a useful lens to understand how organizations operate within their social and cultural environment (DiMaggio and Powell, 1983 [35]). This theory posits that organizations conform to established norms, rules and practices to gain legitimacy and increase their chances of survival (Meyer and Rowan, 1977 [36]). Within the context of EDI, the institutional theory suggests that organizations may adopt diversity and inclusion practices to comply with societal expectations, gain legitimacy and align with the norms of their industry (Greenwood et al., 2002 [37]).

However, as organizations become more established, they may become more resistant to change due to the process of institutionalization (Selznick, 1957 [38]). This resistance to change may manifest in older, larger firms as being less likely to adopt innovative EDI practices or view underrepresentation and discrimination as problematic. This observation suggests that older and more established firms will tend to be less concerned with EDI-related issues than smaller and younger ones.

1.2.2. The Resource-Based View

The resource-based view of the firm posits that organizations can achieve a competitive advantage by leveraging their unique resources and capabilities (Barney, 1991 [39]). In the context of our study, diverse human capital can be considered as a valuable resource that enhances creativity, innovation and problem-solving (Jackson and Ruderman, 2020 [10]; Tregaskis and Taylor, 2019 [11]). Therefore, the RBV suggests that firms that prioritize EDI may benefit from a competitive advantage in the marketplace.

This theory predicts that smaller, more innovative firms which rely on and early stage competitive edge to establish themselves in their marketplace are more likely to view EDI as important, potentially due to their recognition of the value of diverse human capital. Conversely, larger, more established firms may not view EDI as essential for their competitive advantage, possibly due to their focus on other resources or capabilities that have historically contributed to their success.

1.2.3. Innovation Theory

Innovation theory highlights the importance of creativity, experimentation and risk taking in driving organizational performance and growth (Schumpeter, 1934 [40]). Organizations that prioritize innovation are more likely to adopt novel practices, experiment with new ideas and adapt to changing environments (Tidd and Bessant, 2009 [41]). Our study suggests that more innovative firms are more likely to prioritize EDI policies, possibly because they view diversity as a driver of creativity and innovation.

In contrast, larger, more established firms may be less innovative and more risk-averse, leading to a reluctance to implement EDI initiatives or address issues of underrepresentation (Bartlett and Ghoshal, 1990 [42]).

1.2.4. Social Identity Theory

Another theoretical perspective that can help explain the relationship between firm attributes and attitudes toward EDI is social identity theory. Social identity theory posits that individuals categorize themselves and others into social groups based on shared characteristics, such as race, gender or nationality (Tajfel and Turner, 1979 [43]). These social categorizations can lead to in-group favoritism and out-group discrimination which may contribute to organizational attitudes and behaviors related to EDI (Tajfel and Turner, 1986 [44]).

In the context of our study, social identity theory suggests that the composition of a firm’s workforce may influence its attitudes toward EDI. Firms with a more diverse workforce are likely to have employees with a broader range of social identities which may foster a more inclusive organizational culture and promote positive attitudes towards EDI (van Knippenberg and Schippers, 2007 [45]). Conversely, firms with a more homogenous workforce may exhibit stronger in-group biases, leading to less favorable attitudes toward EDI initiatives and a reluctance to address underrepresentation and discrimination.

As smaller firms are more agile and adaptable, they are known to be able to recruit employees from diverse backgrounds and foster a more inclusive culture (Joshi and Roh, 2009 [46]). Consequently, they are more likely to prioritize EDI policies. In contrast, larger, more established firms may have a more rigid organizational structure and culture which may impede their ability to embrace diversity and foster an inclusive environment.

1.2.5. Ambidexterity Theory

Ambidexterity theory is another relevant framework for understanding the relationship between firm attributes and attitudes towards EDI. Organizational ambidexterity refers to the ability of firms to simultaneously pursue exploration (innovation, experimentation and risk-taking) and exploitation (efficiency, optimization and refinement) to achieve long-term success (March, 1991 [47]; O’Reilly and Tushman, 2008 [48]). Achieving ambidexterity requires organizations to balance the conflicting demands of exploration and exploitation which can be challenging, particularly for larger, more established firms (Gibson and Birkinshaw, 2004 [49]).

In the context of our study, ambidexterity theory suggests that firms that are more successful at balancing exploration and exploitation may be better positioned to prioritize and implement EDI policies. These firms may recognize the value of diversity as a driver of innovation (exploration) while also implementing policies and practices to ensure that diverse employees are included, and their contributions are maximized (exploitation). This suggests that smaller, more innovative firms are more likely to view EDI as important which may be related to their ability to achieve ambidexterity.

On the other hand, larger, more established firms may struggle to balance exploration and exploitation, leading to less favorable attitudes toward EDI. These firms may prioritize exploitation, focusing on efficiency and optimization, at the expense of exploration and innovation. This focus on exploitation may contribute to the reluctance of larger, more established firms to address underrepresentation and discrimination, as they may not view EDI initiatives as necessary for achieving their short-term goals.

1.3. The Research Gap

Despite existing research on the benefits of EDI for firm performance, innovation and talent retention, there is a lack of understanding of how specific firm attributes affect attitudes toward EDI. This gap is particularly pronounced in the creative industries which have been underrepresented in EDI research. The conflicting results from previous studies on firm size, profitability and innovation in relation to EDI attitudes further underscore the need for a more robust and comprehensive investigation. This study aims to provide such an analysis, rooted in sound theory and ultimately offering insights for policymakers and further research.

We find that the previous literature has shown a lack of quantitative focus on the creative industries in the EDI literature despite the unique challenges and opportunities these industries present in terms of fostering diversity and inclusiveness. We also believe that there is insufficient empirical evidence taking into account a wide range of firm attributes jointly. This results in inconsistent findings and a limited understanding of the underlying dynamics.

Filling this research gap has several significant implications for the field of EDI and the creative industries. It provides a more nuanced understanding of how different firm attributes may shape attitudes toward EDI, enabling policymakers to design more targeted and effective interventions. It also offers insights into the theoretical foundations for explaining firm attitudes toward EDI as a function of their measurable attributes, borrowed from innovation theory, social identity theory, ambidexterity theory, etc. These insights can inform future research on the mechanisms underlying the relationship between firm attributes and EDI attitudes. The paper also contributes to theory development by highlighting the importance of firm attributes in shaping attitudes toward EDI, expanding upon existing knowledge in the field.

2. Materials and Methods

Our research focuses primarily on South-East Wales and most prominently on Cardiff and the Cardiff Capital Region, even though we have received answers to our survey from other regions of Wales as well. We have focused primarily on Wales, as we had encountered a large number of creative businesses from this region during our recent research and policy work. We do not have any reason to assume that the geographical scope of our study impedes the generalization of our results to the creative industries at large. We have gathered data on 328 creative firms between 2019 and 2021. The dataset that we work with is effectively a cross-section, even though some respondents to our survey answered at different points in time. We have chosen the firms to survey essentially by their subsector of activity. Consequently, the respondents in our dataset operate in one of the creative industries identified by the DCMS (See, for instance https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/203296/Classifying_and_Measuring_the_Creative_Industries_Consultation_Paper_April_2013-final.pdf (accessed on 15 January 2023)).

We have identified the potential respondents by gathering all of their email contacts from the FAME dataset (provided by Bureau van Dijk). We have updated contact information for emails that bounced back through thorough desk research. We have ensured that each firm that we contacted was active in one of the creative sectors as identified by the DCMS. Given that we blindly solicited answers, we acknowledge that there might be an incidence of voluntary response bias in our sample. Nevertheless, the questionnaire was unusually broad, and the EDI-related questions were not primary focus. Instead, most questions focused on innovation, collaborations, sources of funding for R and D, etc. Therefore, it is highly unlikely that the few EDI-related questions (amounting to less than 3% of all questions asked) were a crucial point in shaping our respondents’ willingness to participate in our survey. A total of 75% of the actual responding individuals identified themselves as the founders/directors/managers/vice presidents of their companies, while 25% of respondents were professionals (designers, programmers, artists, etc.).

The sectoral breakdown of the respondents to our survey is in

Table 1. Sectoral breakdown of respondents to our survey.

Sector	Percentage Frequency in Our Dataset
Advertising and marketing	6.1%
Architecture	1%
Crafts	1.8%
Design and designer fashion	2.8%
Film, TV, video, radio and photography	33.4%
IT, software and computer services	10.1%
Museums, galleries and libraries	1.8%
Music, performing and visual arts	20.3%
Other creative sectors	17.2%
Publishing	5.6%

below. Note that the proportions in the table are broadly in line with the composition of creative industries in South-East Wales (Fodor et al., 2021 [50]). An overview of the general EDI patterns of the creative firms in the region can be found in Komorowski et al. (2021) [51].

The data that we have collected are very wide and is applicable to a large number of topics and research questions. In particular, we have collected data on innovation (attitudes, results and spending), skills, corporate culture, management, etc. The data that we use for this article include hard quantitative information on the following firm-level characteristics: turnover (revenue by company in GBP), profit ratio (as a percentage of turnover), the number of full-time employees, the age of the company, R and D spending and the number of ongoing R and D projects. We summarize the values of the responses for these variables in

Table 2. Summary statistics of the most important firm-level attributes used in our study.

Variable	Average	Standard Deviation	The Number of Respondents Providing Information about the Variable
Turnover	£ 800,937	5,947,491	202
Profit margin	26.6%	29.9	132
FTE employees	9.9	31	186
Age of the firm	12.7 yrs	15.8	244
# of ongoing R&D projects	2.3	2.8	188
R&D expenditure per year	£ 23,508	127,121	98

below.

The “soft” information in our dataset is about attitudes towards various discrimination-related issues. The questions track 5 aspects of discrimination and representation, i.e., those related to (i) women, (ii) people of different races, (iii) the LGBTQ2+ community, (iv) people with disabilities in the creative sector and (v) the discrimination of minority groups. Each respondent stated whether they agree or disagree with these 5 statements, e.g., “People with disabilities are underrepresented in my business or business network”—and the five possible answers were (i) strongly disagree, (ii) somewhat disagree, (iii) neither agree nor disagree, (iv) somewhat agree and (v) strongly agree. The data is encoded from 1 to 5 on an ordered scale with 1 corresponding to “strongly disagree” and 5 corresponding to “strongly agree”. The percentage breakdowns of the answers to these questions are given in the

Table 3. Percentage breakdown of answers to the questions related to attitudes towards questions of representation and discrimination.

Question: […] Are Underrepresented in My Business/Business Network	Strongly Agree	Agree	Neither Agree nor Disagree	Disagree	Strongly Disagree
Women	17.74%	21.37%	20.16%	14.11%	26.61%
People with disabilities	29.84%	33.47%	25.81%	4.44%	6.45%
The LGBTQ2+ community	17.74%	20.16%	37.5%	10.89%	13.71%
Different races	27.02%	27.02%	25.81%	11.69%	8.47%
Minority groups face discrimination in my business network	17.81%	17%	33.6%	10.53%	21.53%

.

The main aim of the analysis is to provide evidence of systematic correlations between the firm attributes that we have collected and answers to the questions regarding discrimination and representation. We regress each set of answers to EDI questions on each measurable firm-level characteristic. We use ordered logit regressions to do so. We report the significant regression coefficients and odds ratios that allow for a straightforward interpretation of the results. Naturally, we do not claim causality in our analysis and, therefore, these magnitudes are merely indicative. We also do not exclude the possibility of reverse causality in our regressions (meaning that some of the attitudes towards EDI may affect firm performance). Despite the establishment of causality, our empirical studies give an idea of how concrete, measurable changes in firm size, turnover, profitability, etc. are associated with the probability of the answers that the companies’ representatives give to our EDI-related questions.

When analyzing survey data with categorical responses, such as “strongly agree”, “agree”, “neutral”, “disagree” and “strongly disagree”, it is important to choose an appropriate statistical method that accurately captures the relationship between the independent variables and the categorical response variable. While Ordinary Least Squares (OLS) regression is commonly used to analyze survey data, it is not the optimal choice for modeling categorical responses with more than two categories. In such cases, using an ordered logit model is a better option (Cameron and Trivedi, 2013 [52]). The model is based on the proportional odds assumption which states that the coefficients of the independent variables are constant across all levels of the response variable (Long, 1997 [53]). This means that the association of an explanatory variable with the likelihood of being in a higher category of the response variable (in our case, “agreeing more” with the possible statements) is the same, regardless of the current category of the response variable.

In contrast, OLS regression assumes that the dependent variable is continuous and normally distributed which is not the case with categorical response variables (Agresti, 2002 [54]). When OLS regression is applied to categorical data, it can produce biased and inefficient estimates of the regression coefficients, leading to incorrect statistical inferences (Long, 1997 [53]). OLS regression also assumes that the residuals are normally distributed and homoscedastic which is often not the case with categorical data.

Furthermore, using an ordered logit model allows for the interpretation of the regression coefficients as the change in the log odds of being in a higher category of the response variable is associated with a one-unit increase in the independent variable (Agresti, 2002 [54]). This provides a more meaningful interpretation of the results compared to OLS regression which only provides the change in the mean value of the dependent variable.

2.1. The Hypotheses to Test Empirically

It has been suggested in the previous literature (based primarily on anecdotal evidence) that larger, older firms with higher turnover and more employees may be less concerned with EDI-related issues. We explore whether this hypothesis has any empirical underpinning and present the results of our inquiries in Section 3 below. They suggest that there, indeed, is a significant correlation between firm size (turnover, employees), age and lax attitudes towards discrimination.

One possible explanation for why larger, older firms may be less concerned with discrimination and underrepresentation is the idea of institutionalization. Institutionalization occurs when an organization becomes embedded in its routines, procedures and structures, leading to resistance to change. In the case of discrimination and underrepresentation, larger, older firms may have been operating with discriminatory practices for a long time, making it difficult for them to recognize or change these practices. Research has shown that institutionalization can lead to resistance to diversity initiatives, making it difficult for larger, older firms to address issues of discrimination and underrepresentation (Greenwood, Hinings and Suddaby, 2002 [37]).

Moreover, larger, older firms may also feel that they have already done enough to address discrimination and underrepresentation in the workplace. For instance, they may argue that they have implemented equal employment opportunities policies, diversity training programs and other initiatives designed to promote diversity and inclusion. However, research suggests that such initiatives may not always be effective in addressing discrimination and underrepresentation (Cox and Blake, 1991 [55]; Kalev, Dobbin and Kelly, 2006 [56]). Instead, organizations need to take a more proactive approach to diversity and inclusion by implementing policies that target the root causes of discrimination and underrepresentation (Kalev et al., 2006 [56]).

Larger, more established firms also tend to be less innovative and more risk averse than smaller firms (Bartlett & Ghoshal, 1990 [42]). This risk aversion may translate into a reluctance to take on initiatives that may be perceived as risky, such as recruiting more diverse employees or implementing policies to address underrepresentation. Furthermore, larger, older firms may not feel the need to address underrepresentation if they believe that it does not affect their bottom line. This is true even though research suggests that diversity and inclusion can have a positive impact on a firm’s financial performance (Catalyst, 2018 [57]; McKinsey & Company, 2015 [58]). Smaller firms may, therefore, be more likely to view EDI as essential for their survival, particularly in the highly competitive creative industries. They may also be more likely to view diversity and inclusiveness as a source of innovation and creativity rather than a burden or a risk. For example, a study by Edmondson and colleagues (2015, [59]) found that smaller firms were more likely to engage in creative problemsolving than larger firms and that this was partly due to the diversity of their workforce.

These insights, as well as those outlined in our theoretical framework, result in the following testable hypotheses.

Hypothesis 1:

The age of the firm is a clear determinant of EDI-related attitudes. The older and more established a company is, the less it is concerned with inclusivity-related issues for any type of protected characteristic. Analogously, the younger the company is, the more it is reliant on including and fostering a diverse workforce.

Hypothesis 2:

The more innovative a firm is, the more preoccupied it is with EDI-related challenges. This insight stems from ambidexterity and innovation theories.

Hypothesis 3:

Firms that are more efficient (and thereby generate higher profit margins) are more likely to exploit individual skillsets regardless of social composition and institutional rigidities. Therefore, they will attribute more weight to EDI-related challenges than less efficient firms.

2.2. The Empirical Strategy

Our empirical strategy relies on a two-step iterative method that involves the following.

• We test each hypothesis above by regressing the attitude scores regarding each protected characteristic (ethnic minorities, disabilities, LGBTQ2+, etc.) on each relevant firm-level attribute (turnover, firm age, the number of employees, R and D investment, the number of R and D projects, profit ratio). We report and discuss only the statistically significant correlations, sorted by attitudes regarding the inclusion of each protected characteristic. If we do not report a particular attitude-firm characteristic pair, it entails that the correlation between them is statistically insignificant in our univariate logit regressions.
• We take the significant correlations that we find in step 1 and we test if they are robust to the inclusion of all other firm-level attributes. This ensures that the firm characteristic actually has autonomous explanatory power as opposed to just being a false proxy for some other firm-level attribute. To the extent that the correlations that we find in step 1 remain statistically significant, we are able to conclude that varying the attribute independently of all other characteristics, indeed, changes EDI attitudes.

Note that the correlations we report below are robust, but they do not establish causality. We have attempted to find instrumental variables in our dataset to establish quasi-natural experiments that could pin down causality. We have strived to find variables that exogenously change firm-level attributes without changing attitudes towards EDI. However, our dataset lacks adequate instruments. One noteworthy candidate variable that we had considered was the amount of funding that the respondents received from public bodies. Upon thorough inspection, we have had to conclude that this variable was not perfectly adequate, as funding schemes are often tied to implementing EDI initiatives. All other candidate variables exhibited similar issues. Consequently, at this iteration of our research, we are constrained to examining robust correlations only.

3. Results

As stated above, we document the statistically significant correlations stemming from our logistic regression analysis. All the pairs between firm-level attributes and EDI questions that are not in the text below came back as statistically insignificant. This implies that on at least a 90% confidence level, we were unable to reject the null hypothesis that there exists no effect of the firm attribute on any given EDI-related attitude.

The structure of the section below is uniform for each EDI-related aspect that we study. We first show the logistic regression coefficient(s) and the odds ratio(s) of all variables that exhibit a significant statistical association with EDI attitudes. Then, we provide a brief intuitive explanation of what our results suggest. We repeat this process for the underrepresentation of women, people with disabilities and different ethnicities, the LGBTQ2+ community and for the discrimination of minorities in general.

3.1. Underrepresentation of Women in the Creative Industries

The only measurable firm-level characteristic that shows a significant statistical relationship with the perception that women are underrepresented in the creative industries is firm age. This correlation is negative, meaning that the older the firm is, the less it believes that the underrepresentation of women in the creative industries is prevalent. We present the ordered logit regression coefficient as well as the odds ratios in

Table 4. Results of an ordered logit regression of the answers to the statement “Women are underrepresented in my business or business network” (on a scale from 1 to 5, 1 being “strongly disagree”) on the age of the firm. **: Significant on a 95% confidence level.

Dependent Variable: Answer to the Statement “Women Are Underrepresented in My Business or Business Network”	Coefficient	Standard Error
Age of the firm (regression coeff.)	−0.0147 **	0.007
Age of the firm (odds ratio)	0.985 **	0.007
Pseudo R-squared	0.01	N/A
Number of observations	240	N/A

.

The odds ratio is the exponential of the regression coefficient and has a relatively straightforward interpretation. It shows how the likelihood of a firm “strongly agreeing” with the statement above changes with a unit increase in firm age. Conversely, it also shows how an increase in firm age switches the firm’s answer from “strongly disagreeing” with women being underrepresented to any other “higher” category (“somewhat disagreeing” to “strongly agreeing” with it).

In this regression, this interpretation above implies the following. If we take two firms, A and B, with firm B being one year older than firm A, then firm B is 1.5% less likely (one minus the odds ratio of 0.985, i.e., 0.015 or 1.5%) to choose the “strongly agree” answer to the statement above than any other possible answer. Conversely, firm B is also 1.5% less likely to choose any other answer than “strongly disagree” than firm A. To put this into an even simpler context, if firm B was 10 years older than firm A, it would be 10.84% less likely to choose the “strongly agree” answer to women’s underrepresentation than firm B (The odds ratio for a 10-year change in firm age is the exponential of 10 times the regression coefficient, i.e., 0.8916. The complementary of that is 0.1084, i.e., 10.84%). Conversely, younger firms are more likely to “strongly agree” with the statement that women are underrepresented in the creative industries.

The correlation between the age of the firm and attitudes towards women’s underrepresentation remains strongly significant after the inclusion of all other firm attributes as shown in

Table 5. Results (regression coefficients only) of an ordered logit regression of the answers to the statement “Women are underrepresented in my business or business network” (on a scale from 1 to 5, 1 being “strongly disagree”) on the age of the firm. ***: Significant on a 99% confidence level.

Dependent Variable: Answer to the Statement “Women Are Underrepresented in My Business or Business Network”	Coefficient	Standard Error
Age of the firm	−0.14 ***	0.006
Number of R&D projects	−0.07	0.23
Total R&D investment (in pounds)	0.00	0.00
Turnover (in pounds)	0.00	0.00
Profit (profit to turnover ratio)	0.0001	0.01
Number of full-time employees	0.099	0.074
Pseudo R-squared	0.07	N/A

. This allows us to conclude that independent of R and D activities, the size of the workforce, turnover and profits, the age of the firm decreases the perception that women are underrepresented in the industry. This entails that even amongst the most innovative firms, an older one is significantly more likely to take a lax attitude towards women’s underrepresentation. Analogously, younger firms, independent of all other firm-level attributes consider that women’s underrepresentation is a serious issue to tackle.

3.2. Underrepresentation of People with Disabilities in the Creative Industries

The only measurable firm-level characteristic that shows a significant statistical relationship with the perception that people with disabilities are underrepresented in the creative industries is firm innovativeness, measured as the number of ongoing R and D projects (R and D in the context of our survey was defined as costly activity undertaken to develop a new product, service, experience or procedure). This correlation is positive, as shown in

Table 6. Results of an ordered logit regression of the answers to the statement “People with disabilities are underrepresented in my business or business network” (on a scale from 1 to 5, 1 being “strongly disagree”) on the age of the firm. **: Significant on a 95% confidence level.

Dependent Variable: Answer to the Statement “People with Disabilities Are Underrepresented in My Business or Business Network”	Coefficient	Standard Error
Number of R&D projects (regression coeff.)	0.111 **	0.056
Number of R&D projects (odds ratio)	1.12 **	0.06
Pseudo R-squared	0.01	N/A
Number of observations	188	N/A

. This entails that the more R and D projects a firm is leading, the more it believes that the underrepresentation of people with disabilities in creative industries is prevalent. We present the ordered logit regression coefficient as well as the odds ratios below.

The interpretation of the regression is the following. If we take two firms, A and B, with firm B leading one more R and D project than firm A, then firm B is 12% more likely to choose the “strongly agree” answer to the statement above than any other possible answer. Conversely, firm B is also 12% more likely to choose any other answer than “strongly disagree” than firm A.

This correlation, while interesting, is not robust to the inclusion of all other observable firm-level attributes. Its coefficient turns negative and insignificant once total R and D investment, the age of the firm, turnover, profits and the number of employees are included in the specification. Consequently, it appears as though there are no significant drivers of the attitudes toward the underrepresentation of people with disabilities in our dataset.

3.3. Underrepresentation of the LGBTQ2+ Community

There are two measurable firm-level characteristics that show significant statistical relationships with the perception that the LGBTQ2+ community is underrepresented in the creative industries. These are firm age and innovativeness, measured by the number of ongoing R and D projects. The correlation with firm age is negative (as

Table 7. Results of ordered logit regressions of the answers to the statement “The LGBTQ2+ community is underrepresented in my business or business network” (on a scale from 1 to 5, 1 being “strongly disagree”) on the age of the firm and on the number of R and D projects. *: Significant on a 90% confidence level.

Dependent Variable: Answer to the Statement “The LGBTQ2+ Community Is Underrepresented in My Business or Business Network”	Coefficient	Standard Error
Age of the firm (regression coeff.)	−0.011 *	0.006
Age of the firm (odds ratio)	0.988 *	0.006
Pseudo R-squared (for the reg. on firm age)	0.004	N/A
Number of observations (for the reg. on firm age)	240	N/A
Number of R&D projects (regression coeff.)	0.092 *	0.053
Number of R&D projects (odds ratio)	1.097 *	0.058
Pseudo R-squared (for the reg. on R&D projects)	0.01	N/A
Number of observations (for the reg. on R&D projects)	188	N/A

shows), meaning that the older the firm is, the less it believes that the underrepresentation of the LGBTQ2+ community in the creative industries is prevalent or problematic. The correlation with innovativeness on the other hand is positive, meaning that the more R and D projects a firm lead, the more likely it is to consider the underrepresentation of the LGBTQ2+ community a problem.

The interpretation of the regression is the following. If we take two firms, A and B, with firm B leading one more R and D project than firm A, then firm B is 9.7% more likely to choose the “strongly agree” answer to the statement above than any other possible answer. Conversely, firm B is also 9.7% more likely to choose any other answer than “strongly disagree” than firm A.

Analogously, if firm B is a year older than firm A, it is 1.2% less likely than firm A to choose the “strongly agree” answer to the statement above than any other possible answer. Conversely, firm B is also 1.2% less likely to choose any other answer than “strongly disagree” than firm A. Put into a more straightforward context, if firm B is 10 years older than firm A, it is 10.4% less likely than firm A to choose the “strongly agree” answer to the statement above rather than any other answer (The odds ratio for a 10-year change in firm age is the exponential of 10 times the regression coefficient, i.e., 0.896. The complementary of that is 0.104, i.e., 10.4%).

The correlation between the age of the firm and attitudes towards the LGBTQ2+ community’s underrepresentation remains statistically significant after the inclusion of all other firm attributes. This allows us to conclude that independent of R and D activities, the size of the workforce, turnover and profits, the age of the firm decreases the perception that LGBTQ2+ are underrepresented in the industry. On the other hand, the positive and significant correlation between the number of R and D projects and this perception of underrepresentation disappears after the inclusion of all available firm-level attributes. This entails that the number of R and D projects was—on its own—absorbing some of the partial explanatory of other variables which are now included in the specification shown in

Table 8. Results (regression coefficients only) of an ordered logit regression of the answers to the statement “The LGBTQ2+ community is underrepresented in my business or business network” (on a scale from 1 to 5, 1 being “strongly disagree”) on the age of the firm. *: Significant on a 90% confidence level.

Dependent Variable: Answer to the Statement “The LGBTQ2+ Community Is Underrepresented in My Business or Business Network”	Coefficient	Standard Error
Age of the firm	−0.06 *	0.03
Number of R and D projects	−0.07	0.21
Total R and D investment (in pounds)	0.00	0.00
Turnover (in pounds)	0.00	0.00
Profit (profit to turnover ratio)	−0.007	0.01
Number of full-time employees	−0.009	0.04
Pseudo R-squared	0.04	N/A

below. Consequently, innovativeness is not a robust driver of the attitudes towards the underrepresentation of the LGBTQ2+ community while the age of the firm is.

3.4. Underrepresentation of Various Ethnicities in the Creative Industries

The only measurable firm-level characteristic that shows a significant statistical relationship with the perception that people with different ethnicities are underrepresented in the creative industries is firm innovativeness, measured as the number of ongoing R and D projects. This correlation is positive, meaning that the more R and D projects a firm is leading, the more it believes that the underrepresentation of people with different ethnicities in the creative industries is prevalent. We present the ordered logit regression coefficient as well as the odds ratios in

Table 9. Results of an ordered logit regression of the answers to the statement “People with different ethnicities are underrepresented in my business or business network” (on a scale from 1 to 5, 1 being “strongly disagree”) on the number of R and D projects a firm leads. ***: Significant on a 99% confidence level.

Dependent Variable: Answer to the Statement “People with Different Ethnicities Are Underrepresented in My Business or Business Network”	Coefficient	Standard Error
Number of R and D projects (regression coeff.)	0.179 ***	0.063
Number of R and D projects (odds ratio)	1.19 ***	0.07
Pseudo R-squared	0.02	N/A
Number of observations	188	N/A

.

The interpretation of the regression is the following. If we take two firms, A and B, with firm B leading one more R and D project than firm A, then firm B is 19% more likely to choose the “strongly agree” answer to the statement above than any other possible answer. Conversely, firm B is also 19% more likely to choose any other answer than “strongly disagree” than firm A.

This correlation, while interesting, is not robust to the inclusion of all other observable firm-level attributes. Its coefficient turns negative and insignificant once total R and D investment, the age of the firm, turnover, profits and the number of employees are included in the specification. Consequently, it appears as though there are no significant drivers of the attitudes toward the underrepresentation of people with disabilities in our dataset.

3.5. Discrimination against Minorities

There are three measurable firm-level characteristics that show significant statistical relationships with the perception that minorities are discriminated against in the creative industries (The level of R and D expenditure also shows a statistically significant correlation with this EDI-related question. Nevertheless, the regression coefficient is so small (−0.00001) that it has no tangible economic interpretation and is, therefore, omitted). These are turnover (revenue), the number of employees and profit rates. The correlations with turnover and the number of employees is negative (see

Table 10. Results of ordered logit regressions of the answers to the statement “Minorities are discriminated against in my business or business network” (on a scale from 1 to 5, 1 being “strongly disagree”) on revenue, the number of employees and on the profit ratio. ***: Significant on a 99% confidence level. **: Significant on a 95% confidence level.

Dependent Variable: Answer to the Statement “Minorities Are Discriminated against in My Business or Business Network”	Coefficient	Standard Error
Revenue (in £ 100 K) (regression coeff.)	−0.05 ***	0.01
Revenue (in £ 100 K) (odds ratio)	0.95 ***	0.02
Pseudo R-squared (for the reg. on revenue)	0.02	N/A
Number of observations (for the reg. on revenue)	200	N/A
Number of employees (FTE) (regression coeff.)	−0.011 **	0.005
Number of employees (FTE) (odds ratio)	0.988 **	0.005
Pseudo R-squared (for the reg. on employees)	0.01	N/A
Number of observations (for the reg. on employees)	182	N/A
Profit ratio (in %s) (regression coeff.)	0.01 **	0.005
Profit ratio (in %s) (odds ratio)	1.01 **	0.005
Pseudo R-squares (for the reg. on profit ratios)	0.01	N/A
Number of observations (for the reg. on profit ratios)	132	N/A

), meaning that more revenue the firm generates and the more people it employs, the less it believes that discrimination against minorities is prevalent in the creative industries. The correlation with profit rates on the other hand is positive, meaning that the more efficient a firm is, the more likely it is to consider discrimination against minorities as a problem.

The interpretation of the regressions is the following. If we take two firms, A and B, with firm B generating £ 100 K more turnover than firm A, then firm B is 5% less likely to choose the “strongly agree” answer to the statement above than any other possible answer. Conversely, firm B is also 5% less likely to choose any other answer than “strongly disagree” than firm A.

Analogously, if firm B employs one more full-time worker than firm A, it is 1.2% less likely than firm A to choose the “strongly agree” answer to the statement above than any other possible answer. Conversely, firm B is also 1.2% less likely to choose any other answer than “strongly disagree” than firm A. Put into a more straightforward context, if firm B has 10 more employees than firm A, it is 10.4% less likely than firm A to choose the “strongly agree” answer to the statement above rather than any other answer (The odds ratio for a 10 employee change is the exponential of 10 times the regression coefficient, i.e., 0.896. The complementary of that is 0.104, i.e., 10.4%).

Finally, if firm B’s profit-to-turnover ratio is one percentage point higher than that of firm A, it is 1% more likely than firm A to choose the “strongly agree” answer to the statement above than any other possible answer. Conversely, firm B is also 1% more likely to choose any other answer than “strongly disagree” than firm A. Put into a more straightforward context, if firm B has a profit-to-turnover ratio that is 10 percentage points higher than that of firm A, it is 10.5% more likely than firm A to choose the “strongly agree” answer to the statement above rather than any other answer (The odds ratio for a 10 percentage point profitability increase is the exponential of 10 times the regression coefficient, i.e., 1.105. This is the numerical equivalent of 10.5%).

The correlation between the profitability of the firm and its attitudes toward the discrimination of minorities remains statistically significant after the inclusion of all other firm attributes, as

Table 11. Results (regression coefficients only) of an ordered logit regression of the answers to the statement “Minorities are discriminated against in my business or business network” (on a scale from 1 to 5, 1 being “strongly disagree”) on the age of the firm. **: Significant on a 95% confidence level.

Dependent Variable: Answer to the Statement “Minorities Are Discriminated against in My Business or Business Network”	Coefficient	Standard Error
Age of the firm	−0.05	0.035
Number of R and D projects	−0.07	0.23
Total R and D investment (in pounds)	0.00	0.00
Turnover (in pounds)	0.00	0.00
Profit (profit to turnover ratio)	0.03 **	0.01
Number of full-time employees	0.014	0.04
Pseudo R-squared	0.09	N/A

shows. This allows us to conclude that independent of R and D activities, the size of the workforce, turnover and profits, the profitability of a firm increases the perception that minorities are discriminated against in the industry. On the other hand, turnover and the number of employees cease to be a significant determinant of the perceptions towards discrimination in this specification.

Note that, while profitability in and of itself is a significant positive driver of awareness regarding EDI issues, it does not correlate to the age of firms in our sample at all. The correlation coefficient between the two variables has a negligible value of 0.011. We highlight the age of firms again at this point as that is the most robust negative driver of concern for EDI-related issues. The lack of correlation in our sample between profitability and the age of firms suggests, while contradicting the previous literature somewhat, that more established firms do not necessarily face a tradeoff between shortterm profits and social responsibility. Instead, it would appear that a potential association between profitability and firm age (possible in other samples) could make it appear that there is a tradeoff between pursuing profits and EDI initiatives. This then creates simultaneity bias, which the results stemming from our sample may resolve. As it is the age of firms that makes companies’ attitudes more lax towards EDI-related issues at any level of profitability, the entrenched organizational practices at these firms seem to show a lack of awareness towards EDI without any particular regard for the impact that such attitudes have on profitability. This suggests that firms may have lexicographic preferences for a reluctance to change fundamentally. Such strategic preferences may point to the need for top-down policy initiatives that instill EDI-related performance indicators instead of the companies themselves. If more established firms consciously ignore the link between profitability and diversity, then other policy interventions (such as education and policy briefs) might be insufficient on their own.

3.6. A Summary of Our Results

Our rigorous correlation analysis confirms Hypothesis 1 that we set out in our study from the get go. Indeed, as Hypothesis 1 suggests, older and more established firms tend to view numerous issues of underrepresentation and of discrimination as less problematic compared to younger ones. This result holds even when we control for other firm-level attributes, implying that the age of the firm shares an autonomous relationship with EDI-related attitudes.

We also find some robust empirical support for Hypothesis 3, predicting that more profitable and efficient firms will consider EDI-related issues to be of significant importance. While this relationship does not hold across all protected characteristics, it is highly robust regarding discrimination against minorities. More profitable companies, independent of their age, size or innovativeness, consider that minorities are indeed discriminated against.

We find limited support for Hypothesis 2, suggesting that innovativeness is positively correlated with the concern for EDI-related issues. While univariate correlations across a number of protected characteristics seem to suggest that this is the case, the relationships break down after the inclusion of further control variables. This entails that left on their own, innovativeness measures absorb the effects of other explanatory variables, such as firm age or profitability.

4. Discussion

4.1. Research Limitations Due to Identification Issues

One potential alternative explanation for our findings above could be that larger firms are more likely to have inherent prejudices and, thus, view discrimination as acceptable. However, this explanation is unlikely, as research has shown that most firms strive to present themselves as committed to equality and diversity even if they do not necessarily practice it (Cox, Lobel, and McLeod, 1991 [60]). Additionally, larger firms are often subject to more scrutiny and pressure to appear inclusive which would make it more difficult for them to openly express discriminatory views (Bendl and Schmidt, 2010 [61]). Therefore, it is unlikely that our findings are due to inherent prejudices in larger firms.

Another possible alternative explanation for our results (that we also address above) is that larger firms may have already implemented measures against discrimination and underrepresentation. However, this explanation is also unlikely. Previous research has shown that larger firms tend to have less diverse workforces and leadership teams than smaller firms (Richardson and Sawyer, 2001 [62]). Additionally, larger firms tend to have more bureaucratic structures that make it difficult to implement and enforce diversity and inclusion policies (Jackson et al., 2003 [63]; Ely and Thomas, 2001 [64]). Additionally, larger firms often struggle with issues, such as “diversity fatigue”, where initiatives become routine and lose their impact over time (Kulik, 2004 [65]). Furthermore, larger firms are often more focused on short-term financial gains which can lead them to prioritize profits over diversity initiatives (Bartel and Borjas, 1981 [66]). Therefore, it is unlikely that our findings are due to larger firms already having implemented measures against discrimination and underrepresentation.

4.2. Correlation versus Causation

While our study does not address causality between firm attributes and attitudes toward underrepresentation and discrimination, it still contributes significantly to the literature by highlighting important patterns and correlations. The documentation of these correlations can inform future research and policy interventions. It can also serve as a starting point for identifying potential causal mechanisms.

Our study utilizes a cross-sectional dataset which limits our ability to establish causal relationships between firm attributes and EDI attitudes. While our analysis provides evidence of systematic correlations between the variables, we cannot rule out the possibility of reverse causality or the impact of unobserved factors influencing both firm attributes and EDI attitudes. Longitudinal studies tracking the same firms over time could offer more robust evidence of causal relationships and shed light on the dynamics of EDI attitudes and their determinants.

Several studies have emphasized the importance of documenting such correlations in the context of discrimination and underrepresentation. For example, Smith et al. (2019, [67]) argue that understanding the relationship between gender and representation in organizations can help identify potential barriers to representation and inform interventions to address them. Similarly, Robinson et al. (2020, [68]) note that the identification of correlations between underrepresentation and factors, such as organizational culture and leadership, can help organizations develop more effective diversity and inclusion strategies.

Therefore, while establishing causality is important, documenting correlations between firm attributes and attitudes toward underrepresentation and discrimination is still a valuable contribution to the literature. This documentation can inform future research and guide policy interventions.

4.3. Further Research Limitations

First, our research focuses primarily on South-East Wales, specifically Cardiff and the Cardiff Capital Region. Although we believe that the findings are generalizable to creative industries at large, the geographical scope of our study may limit the applicability of our results in other regions and countries. Future research could expand the scope of the analysis to include creative firms from different regions and countries to further validate our findings and ensure their broader applicability.

Second, our analysis relies on self-reported data from firms which may introduce biases due to social desirability or the potential for misreporting. Respondents might provide answers that they perceive to be more socially acceptable or favorable rather than accurately reflecting their true attitudes toward EDI. Future research could employ alternative data collection methods, such as experimental designs or objective measures of EDI policies and practices, to mitigate these biases and obtain more accurate assessments of firms’ attitudes toward EDI.

Third, other factors besides the ones that we have measured may influence firms’ attitudes toward EDI, such as organizational culture, leadership styles, or industry-specific factors. Including a broader range of firm attributes and potential determinants of EDI attitudes in future research could provide a more comprehensive understanding of the factors shaping firms’ attitudes towards EDI and inform the design of more targeted interventions.

Fourth, our study examines creative industries as a whole which may obscure important differences between subsectors. Creative industries are diverse and encompass a wide range of activities from advertising and architecture to film production and software development. The factors influencing EDI attitudes may vary across subsectors due to differences in business models, workforce characteristics or market dynamics. Future research could explore the role of subsector-specific factors in shaping firms’ attitudes toward EDI and identify potential barriers and opportunities for promoting EDI within specific creative subsectors.

Last, our study does not directly examine the effectiveness of different EDI policies and practices in addressing issues of discrimination and underrepresentation. While our findings suggest that certain firm attributes are associated with more positive attitudes towards EDI, further research is needed to determine the specific policies and practices that are most effective in promoting EDI in the creative industries. This could involve conducting case studies of successful firms or conducting experimental evaluations of different EDI interventions to identify best practices and inform policy recommendations.

4.4. Research and Policy Implications

4.4.1. Implications for Industry and Practice

Our findings suggest that a one-size-fits-all EDI policy may not be effective for all types of creative firms, given the differences in attitudes towards discrimination and underrepresentation between larger, more established firms and innovative, profitable younger ones. This implies that EDI policies need to be tailored to the specific characteristics of each creative firm and to the challenges that they face.

Diversity management initiatives can fail when they are not tailored to the specific needs of an organization (Cox and Blake, 1991 [55]). Moreover, initiatives that are not aligned with the organizational culture and strategy can create unintended consequences (Jackson et al., 2003 [69]). In the creative industries, research has highlighted the importance of understanding the unique dynamics of each sector, as well as the differences in the experiences of different groups within those sectors (Deuze, 2007 [70]; Banks, 2017 [71]).

Larger firms with more employees and higher turnover may have established structures and processes that are more resistant to change. Overall, entrenched management practices deteriorate firm performance (Salehi and Moghadam, 2019 [72]), despite the fact that they may exert positive effects on innovation to a certain extent (Salehi et al., 2018 [73]), while in other contexts, they may stifle it (Salehi et al., 2018/2 [74]). In any case, larger firms may have more resources to invest in superficial diversity initiatives that do not address the root causes of discrimination and underrepresentation (Schwartz and Carroll, 2003 [75]). In contrast, efficient and profitable firms may be nimbler and may take more risks, thereby being able to adapt to changing circumstances (Hansen et al., 2021 [76]), including those related to EDI. Such firms may be more likely to value diversity and inclusion as drivers of creativity and innovation8. Therefore, policymakers and practitioners should consider the role of institutionalization in shaping firms’ attitudes toward EDI. Strategies that focus on disrupting entrenched routines and structures may be more effective in promoting change in larger, older firms. On the other hand, smaller, more innovative firms may benefit from policies that encourage risk taking and bold projects related to diversity and inclusion.

This underlines the need for a tailored approach to EDI policies is necessary to address the specific challenges faced by each type of creative firm. For larger, more established firms, policies that focus on systemic change and address the root causes of discrimination and underrepresentation may be more effective. This could involve investing in long-term initiatives, such as mentoring programs, leadership training, and policies, that promote transparency and accountability. For profitable firms, policies that focus on promoting diversity and inclusion as drivers of creativity and innovation may be more effective. This could involve initiatives that encourage collaboration across diverse teams as well as policies that recognize and reward diversity and inclusion efforts.

4.4.2. Implications for Theory

This study contributes to theory development by highlighting the role of firm attributes in shaping attitudes toward EDI beyond what has been previously discussed in the literature. The following insights expand upon the existing understanding of these relationships.

The notion of “organizational inertia” could be explored as a possible explanation for larger, older firms’ resistance to EDI. Organizational inertia refers to the difficulty firms face in adapting to new environments and changing internal structures. Future research could examine how organizational inertia affects firms’ ability to recognize and address discriminatory practices and underrepresentation in the workplace.

The impact of “psychological safety” on attitudes toward EDI should be further investigated. Psychological safety refers to the extent to which individuals feel comfortable taking interpersonal risks within a group or organization. In more innovative and efficient firms, employees might feel more psychologically safe which could contribute to the firm’s willingness to address EDI-related issues.

4.4.3. Implications for Further Research

Building on the novel insights from our study, future research could focus on investigating the role of human resource management (HRM) practices in mediating the relationship between firm attributes and attitudes toward EDI. For example, how do HRM practices in larger, older firms differ from those in smaller, more efficient firms, and how do these differences affect EDI implementation?

Further research could also explore how external factors, such as market competition, industry norms, and legal regulations, influence the relationship between firm attributes and attitudes toward EDI. Understanding these external factors may help to identify additional levers for promoting greater diversity and inclusiveness in the creative industries.

5. Conclusions

Overall, our study adds to the growing body of the literature on EDI policies and highlights the importance of considering firm-level characteristics when promoting diversity, inclusiveness and equality in the workplace. Our findings suggest that firm size, turnover and profitability significantly impact attitudes toward discrimination and underrepresentation. In particular, we found a significant correlation between firm size and age and more lax attitudes toward discrimination and underrepresentation. Conversely, more profitable firms that value risk taking and bold projects consider discrimination in creative industries as more of an issue.

Our work is novel in several ways. First, we focus specifically on the creative industries which have received relatively little attention in the literature on EDI policies. Second, we collect data on a sizeable sample of firms, enabling us to establish robust correlations across many aspects related to EDI. Third, we examine a wide range of firm attributes, allowing us to identify those that are most strongly associated with EDI attitudes.

Our findings have important implications for policymakers and industry leaders seeking to promote EDI in the creative industries. Specifically, our results suggest that smaller, younger and more innovative firms may be more receptive to EDI policies than larger, older ones given their differing attitudes toward inclusion. Policymakers and industry leaders should therefore focus their efforts on engaging differently with different types of firms to promote diversity, inclusiveness and equality in the workplace. Our study provides guidance on how these methods of engagement depend on firm attributes that policymakers can observe.