Introduction

A growing number of organisations around the world are turning to diversity branding to promote their commitment to a diverse and inclusive workplace (De Meulenaere and De Boom 2024). There are two main types of diversity branding: internal and employee branding. Regarding internal branding, management’s intention to introduce diversity plays an important role in internal communication (Mampaey et al. 2020). Diversity improves employee attitudes (McKay et al. 2007) work-life balance, and benefits (Sharma et al. 2024). Regarding employee branding, to attract talented individuals and establish a position as an employer of choice, it should be explained that they have a diverse and inclusive workforce (Jonsen et al. 2021). Employer branding is particularly important for large companies that do business in multinational markets and have multinational employees (Martin et al. 2011). Furthermore, diversity has significant appeal for Generation Z, who respect diversity and seek free work styles (Pandita 2022). Recently, to send a positive signal of diversity to stakeholders inside and outside an organisation, companies frequently adopted a specific measure in the form of a chief diversity officer (CDO) (Mehta et al. 2021). CDOs can potentially foster a culture of diversity in an organisation, attract individuals from both inside and outside the organisation, and bring financial benefits (Bishop-Monroe et al. 2021).

However, diversity is not a panacea (Ishikawa 2024). It is not easy to enjoy the effects of diversity, and it will not work without the right management and environment (Nguyen et al. 2022). Therefore, diversity is a double-edged sword because it can both promote and hinder organisational performance (Kelemen et al. 2020; Triana et al. 2014). Several previous studies emphasised the negative aspects of diversity. First, diversity is likely to create conflict because individuals tend to feel closer to those who are similar to themselves (Perry et al. 2022; Triana et al. 2014). In particular, when individuals with differing demographics come to the workplace, negative attitudes toward them are strengthened (Brief et al. 2005). Consequently, in some cases, the introduction of diversity reduces employees’ sense of belonging, organisational productivity, profitability, and stock market performance (Jung et al. 2023; Tang 2024).

To obtain the positive effects of diversity, the environment that utilises it should be strengthened before introducing it (Holmes et al. 2021). However, despite the widespread recognition of diversity’s dilemma, limited concrete knowledge exists on how to utilise it (Bogilović et al. 2021; Ishikawa 2024). The reason for this is that diversity has only been considered regarding demographic diversity, such as gender and race, and its combination with other aspects of diversity has not been taken into consideration (Chidambaran et al. 2022). Consequently, this study aimed to fill this gap. This study categorised diversity into three categories—demographic diversity, perceived diversity, and skill diversity—and established the following research question: ‘Is it necessary to combine demographic diversity with perceived diversity or skill diversity to achieve the effect of demographic diversity on employee loyalty?’ The company utilised in this study was Panasonic Connect Co., Ltd., which developed, manufactured, and sold devices, software, and solutions for companies and local governments. The company is a global company with approximately 28,300 employees (including approximately 12,400 in Japan and approximately 15,900 overseas) (Panasonic Connect Co., Ltd. 2024-a). The aforementioned research question was examined by applying structural equation modelling to data from an online survey of 3000 full-time employees working at the Japan offices of Panasonic Connect Co., Ltd. To our knowledge, this is the first study to analyse diversity from three perspectives and clarify the specific conditions for achieving the benefits of diversity.

Literature review and hypothesis development

Demographic diversity

Diversity is defined as ‘the degree to which a firm advocates fair human resource policies and socially integrates underrepresented employees’ (McKay et al. 2008: 352). The perspective of underrepresented employees is predominantly based on demographic diversity, such as gender (e.g., women) and race (e.g., black individuals) (Avery et al. 2008). Regarding companies, demographic diversity creates several social categories (Cobb and Stevens 2017). This diversity promotes organisational innovation (Østergaard et al. 2011) and contributes to corporate performance (Liu et al. 2014; Reguera-Alvarado et al. 2017). However, the opposite effect of diversity has recently attracted attention. Individuals tend to unconsciously believe that ‘people who look different also think differently’ (Proudfoot et al. 2024). Therefore, diversity creates conflict (Brief et al. 2005; Triana et al. 2014), and concerns have been raised that it may reduce organisational performance (Jung et al. 2023; Kelemen et al. 2020; Tang 2024). Owing to this dual nature, diversity is called a double-edged sword (Triana et al. 2014), and corporate leaders are urged to recognise the fact that demographic diversity entails risks (Kunze et al. 2021).

Perceived diversity

Perceived diversity is defined as ‘the degree to which members are aware of one another’s differences, as reflected in their internal mental representations of the unit’s composition (Shemla et al. 2016)’. Generally, whether a personnel system works depends on employees’ perceptions of the policy and the organisational culture (Alfes et al. 2012). The same is true for diversity. Before introducing demographic diversity, an important challenge is to establish an organisational culture that will make use of it (Holmes et al. 2021). This is because while diversity initiatives are welcomed by the target group, they may be negatively viewed by non-target employees (men, in the case of gender diversity) who make up the majority of the organisation (Kukula et al. 2024). Unless managers start by fostering an inclusive culture, they will not be able to reap the benefits of hiring diverse talent and will not be able to prevent talent loss (Russen and Dawson 2023). Therefore, to maximise attribute diversity’s potential, employees’ positive perceptions of the initiative should be strengthened (Bos-Nehles and Veenendaal 2019; Tajeddini et al. 2022). This point is also supported through the lens of intergroup contact theory, which suggests that contact between groups with different identities can reduce prejudice and xenophobia (Allport 1954; Pettigrew and Tropp 2006). In the context of organisational diversity, contact between employees with diverse values is also expected to have a prejudice-reducing effect (Pettigrew 2021). Thus, visible diversity such as demographic diversity is not enough; invisible employee perceived diversity is important (Bogilović et al. 2021).

Social exchange theory (Jiang 2024) explains the relationship between workplace diversity culture and employees’ positive attitudes (sense of belonging). This theory is one of the most influential conceptual paradigms in organisational behaviour (Cropanzano and Mitchell 2005). Employees build exchange relationships by seeking fair treatment in exchange for their labour contributions to the company (Barakat et al. 2016; Zhou et al. 2018). Therefore, by communicating a commitment to diversity to employees, companies express their intention to provide employees with appropriate rewards and support (Men et al. 2024). Given that a highly diverse workplace reflects an organisation that is free of prejudice, adheres to fair treatment, and cares about the well-being of employees, employees feel reassured that they are part of the same group (Mansoor et al. 2021).

Accordingly, the following hypotheses were constructed as conditions for utilizing an organisation’s demographic diversity.

H1-1: Employees’ perceived diversity positively strengthens the relationship between demographic diversity and organisational identification.

H1-2: Employees’ perceived diversity positively strengthens the relationship between demographic diversity and meaningfulness of work.

H1-3: Employees’ perceived diversity negatively strengthens the relationship between demographic diversity and burnout.

Skill diversity

In addition to perceived diversity, another invisible diversity is skill diversity. Skill diversity refers to differences based on the skills of members of an organisation (Tonoyan and Olson-Buchanan 2023). Regarding skills, discussions of skill levels tend to be the focus; however, skill diversity is an important organisational performance predictor (Almaatouq et al. 2024). In particular, skill diversity is more necessary now than ever because organisations need to perform complex tasks within global teams, create new businesses, and balance environmental considerations with economic value (Álvarez Pereira et al. 2024; Liu and Cheng 2024; Zhang 2020).

The need for skill diversity can be understood from the perspective of resource dependence theory. According to this theory, resources (tangible and intangible) are key to an organisation’s survival, emphasizing the importance of the ability to acquire and maintain them (Pfeffer and Salancik 1978; Pfeffer and Salancik 2003). Previously, the focus was on how companies could reduce their dependence on external constraints; however, recently, resource dependence theory emphasised improving efficiency through internal resources in addition to external resources (Wry et al. 2013). A company’s intangible resources include its employees’ skills, knowledge, and experience (Post et al. 2015), thereby introducing the perspective of skill diversity. In addition to demographic diversity, older employees who accumulated valuable experience in the industry and younger employees who have technical skills complement each other to enable strategic decision-making (Ali et al. 2014).

Accordingly, the following hypotheses were constructed as conditions for utilizing an organisation’s demographic diversity.

H2-1: Organisational skill diversity positively strengthens the relationship between demographic diversity and organisational identification.

H2-2: Organisational skill diversity positively strengthens the relationship between demographic diversity and the meaningfulness of work.

H2-3: Organisational skill diversity negatively strengthens the relationship between demographic diversity and burnout.

Hypothetical model

To properly evaluate diversity, it is necessary to comprehensively analyse the main factors of employee attitudes. Thus, the extant literature on employee attitudes was reviewed and the following six factors were extracted: purpose (Bhattacharya et al. 2023; Porfirio et al. 2024; Qin et al. 2022), product/service in charge (Kato and Koizumi 2024), psychological safety (Al-Refaie 2015; Carmeli et al. 2010), office (Danielsson and Bodin 2008; Mitchell 1992; Rothe et al. 2012), skill acquisition (Ko 2012; Kong et al. 2015; Peng et al. 2020), and wage (Al-Refaie 2015; Dodanwala and San Santoso 2021; Ge et al. 2021). Based on these factors, a hypothetical model (Model 1) was constructed, as shown in Fig. 1. Based on Model 1, the hypotheses were tested using Model 2 that added the interaction between demographic diversity and perceived/skill diversity (Fig. 2).

Fig. 1
figure 1

Hypothetical model (Model 1).

Full size image
Fig. 2
figure 2

Hypothetical model (Model 2-1 & Model 2-2).

Full size image

Method

An online survey was conducted on 3000 full-time employees of Panasonic Connect Co., Ltd. in Japan from September 9 to September 27, 2024. All study participants provided informed consent via the online survey, and the study design was approved by the Meiji University ethics review board. The inclusion criteria for the participants were (a) full-time employees, (b) aged 20–60s, and (c) the highest level of education obtained was high school or higher. To effectively perform structural equation modelling analysis, the sample size should be 400 or more (Hoyle and Gottfredson 2015). As this study deals with gender, more than 400 men and 400 women were required to avoid the conclusions being biased towards one gender. The gender proportion at Panasonic Connect is 82.4% male and 17.6% female (Panasonic Connect 2024-b). Therefore, using SurveyMonkey’s sample size calculator (SurveyMonkey, 2025), the required sample size was calculated to be 3063 for the 11,600 employees at Panasonic Connect Co., Ltd. in Japan (Panasonic Connect, 2024-a), with a confidence level of 99% and a margin of error of 2%. Thus, the overall sample size was set to 3000 people to ensure that it contains more than 400 women. As shown in Table 1, the distribution of respondent attributes meets the above conditions. In addition, ages were collected evenly to avoid generation-specific dependent results.

Table 1 Distribution of respondent attributes.
Full size table

As presented in Table 2, the questions were based on extant literature and included three items for each of the following factors: loyalty, organisational identification, meaningful work, burnout, purpose, product/service in charge, psychological safety, office, skill acquisition, wage, and demographic diversity. In addition, participants were asked about perceived and skill diversity, which corresponded with the hypothesised moderation effects.

Table 2 Survey item list.
Full size table

In the survey design of this study, a critical issue is the trade-off between addressing social desirability bias and the quality of a single-item survey. Contamination of survey data by social desirability response bias is a major threat (King and Bruner 2000). To improve the quality of a survey, questions should be designed to cover all factors related to the objective variable, while concealing the purpose from the subjects (Bullock et al. 2017; Kato et al. 2024). Specifically, as the subject of interest in this study is diversity, an area of strong social significance, it is susceptible to social desirability bias (Boring and Delfgaauw 2024). In surveys about issues where ideal social norms, such as diversity, are widespread, respondents tend to overreport answers that are in line with those ideals and underreport answers that are socially undesirable (Krumpal 2024; Schell et al. 2021). Even if responses are anonymous, people place importance on conforming to ideals in surveys, which are a forum for expressing their identities and beliefs (Brenner and DeLamater 2016). Social desirability bias is a serious problem that can alter the conclusions reached in marketing experiments (Larson 2019). Therefore, the survey had to be designed so that the participants were not made aware of the true purpose of diversity. As this study focused on the three types of diversity, the survey would have included a total of nine questions on diversity if three questions were asked about each type. This would make the subjects aware of the purpose of the survey, that is, to investigate diversity, and hence, the answers would be biased. Thus, studies on diversity, that is, a topic with clear social norms, are bound to overestimate the effects of diversity unless researchers pay closer attention to the fact that respondents are not always revealing the truth about what they think (Boring and Delfgaauw 2024; Lüke and Grosche 2018).

Nonetheless, single-item surveys have a negative reputation in the scientific community (Allen et al. 2022). Using a single-item measure was even considered one of the surest ways to receive a rejection letter from a journal editor (Wanous et al. 1997). In contrast, multi-item scales are considered more reliable for measuring complex social objects (Churchill 1979; Diamantopoulos et al. 2012; Jacoby 1978). However, some studies claim that the same survey quality can be expected with single-item measures because multiple-item scales are redundant, and the respondent’s perception can be sufficiently confirmed with a single item (Bergkvist and Rossiter 2007; Jordan and Turner 2008; O’Brien et al. 2022; Trail et al. 2023). Furthermore, single-item questionnaires have the advantage of shorter survey times, thereby positively impacting the respondent’s concentration levels, and hence, improving the quality of their responses (Castro et al. 2023). Owing to these advantages, research is underway into the optimal conditions for using single items. For example, single-item measures work effectively for questions that subjects can clearly understand and judge, such as satisfaction with their own weight (Allen et al. 2022). Thus, both survey methods have advantages and disadvantages (Castro et al. 2023; Sloan et al. 2002). Moreover, knowledge regarding single-item scales is still evolving. Therefore, it has been mentioned that viewing all single-item scales as representing weak research designs is counterproductive and hinders the progress of organisational science (Matthews et al. 2022).

To summarise, while the use of single-item scales has been well researched, the problem of data contamination by social desirability bias has been overlooked for many years. Both issues cannot be addressed simultaneously and completely. Therefore, this study placed emphasis on social desirability bias and avoided including many questions regarding diversity, compared to the number of questions regarding other factors. Specifically, only the base demographic diversity was allocated three questions, the same as the other factors, and the other two diversity types were made into single items, thereby limiting the number of diversity-related questions to five. All responses were rated on a five-point scale (1 = not at all applicable, 5 = very applicable).

Subsequently, structural equation modelling was used for verification. First, the exploratory factor analysis for observed variables No. 1–33 in Table 2 confirmed the concerns about common method bias. As the same respondents were asked about the objective and explanatory variables, a bias may be observed, in which causal relationships are overemphasised. Next, confirmatory factor analysis was performed. The validity of the factor analysis was determined using Cronbach’s α, average variance extracted (AVE), and composite reliability (CR). Structural equation modelling was applied based on the model shown in Fig. 1. Model 2-1 was constructed by adding the moderation effects of perceived diversity, which corresponded to hypothesis 1, and Model 2-2 was constructed by adding the moderation effects of skill diversity, which corresponded to Hypothesis 2. Data analyses were performed using R version 4.2.2.

Results

Table 3 shows the correlation matrix of the observed variables. When the values for correlation variables are 0.9 or higher, concerns about multi-collinearity arise (Hair et al. 1998; Kattan and Abduljawad 2019). This level has not been reached for the three diversity types that are the subject of interest in this study. Moreover, common method bias is a concern when the first factor explains more than half of the variance (Fuller et al. 2016; Podsakoff and Organ 1986). As shown in Table 4, the proportion of variance explained by the first factor is 7%, which does not meet the above criteria, as a result of the exploratory factor analysis. In this study, the main factors related to employee attitudes were comprehensively evaluated so as not to overestimate the diversity—the subject of interest—which enabled precise verification. As presented in Table 5, the confirmatory factor analysis results indicated that each construct met the Cronbach’s alpha standard of 0.7 (Hao et al. 2022). The standard for AVE is 0.5 (Bagozzi and Yi 1988); however, if the AVE is less than 0.50 and the CR is higher than 0.6, the convergent validity of the construct is still appropriate (Na-Nan and Saribut 2020). Therefore, the confirmatory factor analysis of this study was valid.

Table 3 Correlation matrix.
Full size table
Table 4 Results of exploratory factor analysis.
Full size table
Table 5 Results of confirmatory factor analysis.
Full size table

The structural equation modelling results are presented in Fig. 3. Model 1 indicated a high degree of fit according to the following indices: CFI (comparative fit index) = 0.957, GFI (goodness of fit index) = 0.947, AGFI (adjusted goodness of fit index) = 0.933, SRMR (standardised root mean square residual) = 0.040, RMSEA (root mean square error of approximation) = 0.040. Regarding loyalty, organisational identification (β = 0.763, p < 0.000), meaningful work (β = 0.121, p < 0.000), and burnout (β = −0.067, p < 0.000) all showed significant effects. The most important contributing factor for organisational identification was purpose (β = 0.776, p < 0.000), that for meaningful work was product/service in charge (β = 0.387, p < 0.000), and that for burnout was psychological safety (β = −0.451, p < 0.000). Overall, the most important contributing factor was skill acquisition. When considering all contributing factors of employee attitudes, demographic diversity indicated no significant effect on organisational identification (β = −0.003, p = 0.911), meaningful work (β = −0.008, p = 0.716), or burnout (β = −0.026, p = 0.466).

Fig. 3
figure 3

Results of structural equation modelling (Model 1).

Full size image

Table 6 presents the Model 2-1 and Model 2-2 results. Demographic diversity × perceived diversity indicated no significant effect on organisational identification (β = −0.003, p = 0.838), a positive effect on meaningful work (β = 0.033, p = 0.033), and no significant effect on burnout (β = −0.004, p = 0.844). Therefore, H1-2 was supported, and H1-1 and H1-3 were not supported. Demographic diversity × skill diversity positively affected organisational identification (β = 0.036, p = 0.028), meaningful work (β = 0.037, p = 0.016), and burnout (β = 0.043, p = 0.031). Dissimilar to this study’s assumptions, the opposite effect was found for burnout. Therefore, H2-1 and H2-2 were supported, and H2-3 was not supported.

Table 6 Structural equation modelling results (Models 2-1 and 2-2).
Full size table

Discussion

Although the social significance of diversity is recognised, effectively applying it to employee attitudes is difficult, and concrete knowledge is limited (Bogilović et al. 2021; Ishikawa 2024). This is because diversity is considered only in terms of demographics, such as gender and race, and not in combination with other aspects of diversity (Chidambaran et al. 2022). To enjoy the benefits of diversity, understanding in detail the conditions under which positive effects and negative effects are obtained is necessary. Therefore, this study introduced three categories—demographic diversity, perceived diversity, and skill diversity—and conducted a detailed analysis of the conditions under which diversity can be effective. The results showed that perceived diversity and skill diversity are effective in enhancing the effects of demographic diversity. Nonetheless, the impact of diversity is limited compared to that of major employee attitude factors, such as purpose, products/services in charge, and psychological safety. This is because the purpose of the survey was concealed from the subjects, and the employee attitude factors were evaluated comprehensively so as not to overestimate the effect of diversity. Studies on issues vulnerable to social desirability bias, such as diversity, tend to overestimate its impact (Boring and Delfgaauw 2024; Lüke and Grosche 2018; Krumpal 2024; Schell et al. 2021). Social desirability bias is a serious problem that can distort research results (Larson 2019). Thus, we carefully designed this study to overcome this issue, and hence, the results do not overestimate diversity. In structural equation modelling, effect size indicators have rarely been studied (Gomer et al. 2019). However, as described in the method section, the sample size of this study was appropriately designed; therefore, concerns about an excessive sample size are dispelled.

As shown in this study, demographic diversity alone has no effect on employee attitudes. However, combining demographic diversity with perceived diversity can increase employees’ organisational identification and meaningful work. The result suggests that to make the most of demographic diversity, it is important for employees themselves to empathise with the need for diversity. Moreover, this means that organisations where various human resources can utilise their diverse skills are characterised by a strong sense of belonging and meaningful work. This is because employees want to work for organisations that make the most of their talent (Hughes and Rog 2008). In this way, companies can reap the benefits of diversity by utilizing multiple aspects of it simultaneously. Contrastingly, organisations will not be able to reap the benefits if they only consider visible diversity.

In contrast to the expectations of this study, demographic diversity × skill diversity had the opposite effect on burnout. From the perspective of demographic diversity, it creates a conflict structure within the organisation (Brief et al. 2005; Proudfoot et al. 2024; Triana et al. 2014). From the perspective of skill diversity, it creates frustration among employees who lack specialised skills. In terms of employee skills, recent changes in personnel evaluation systems are having a negative impact on psychological safety. Traditionally, Japanese companies have a unique culture that differs from that of the West; it places great importance on collaboration within a team (Kameda 2014). Based on the lifetime employment system, salaries are determined by seniority; therefore, age rather than ability is generally the determining factor for salaries. However, as globalisation has progressed, the importance of this practice has declined dramatically, and each employee is now expected to have specialised skills (Pudelko 2006). For example, in 2019, Toyoda Akio, former president of Toyota Motor Corporation, said that it had become difficult to maintain the lifetime employment (Takezawa 2019). Moreover, Panasonic Group, which was the subject of this study, announced the abolition of the seniority-based pay system in 2015 and has since reformed their traditional employment style in Japan (Hirasawa 2015). Specifically, they have shifted from age-based compensation with stable employment to competition-based compensation (Panasonic 2022; Stevenson and Furusawa 2024). Therefore, it is assumed that the combination of conflicting attribute perceptions and personnel evaluation based on skills is causing competition within organisations. Intense competition within an organisation threatens employees’ interpersonal trust (Bunjak et al. 2023) and psychological safety (Brouwer 2016). Based on the above mechanism, it is speculated that the more diverse an organisation is, the more likely it is that a competitive spirit will take root when employees are surrounded by colleagues with diverse specialised skills, increasing the risk of burnout.

Theoretical implications

In response to the current demands for ethical behaviour, companies implement diversity to strengthen their internal branding (Mampaey et al. 2020; McKay et al. 2007; Sharma et al. 2024) and employee branding (Jonsen et al. 2021; Martin et al. 2011; Pandita 2022). However, diversity is not easy to achieve (Nguyen et al. 2022; Triana et al. 2014), and diversity management has become a major concern for organisations globally (Kadam et al. 2020). Several previous studies emphasised the negative aspects of diversity (Brief et al. 2005; Jung et al. 2023; Kunze et al. 2021; Perry et al. 2022; Tang 2024). Although strengthening the environment that utilises diversity is a priority to obtain the positive effects of diversity (Holmes et al. 2021), limited knowledge has been accumulated on this topic (Bogilović et al. 2021). This is because diversity has been examined only in terms of demographic diversity, and its combination with other aspects of diversity has not been considered (Chidambaran et al. 2022). Therefore, this study classified diversity into three categories—demographic, perception, and skill diversity—and analysed their moderation effects. The results indicated that demographic diversity × perceived diversity has a positive and significant effect on the meaningfulness of work. Additionally, demographic diversity × skill diversity has a positive effect on organisational identification and meaningfulness of work. However, demographic diversity × skill diversity has a side effect of increasing the risk of burnout owing to concerns about intensified competition among employees. These results provide new knowledge on specific organisational strategies to utilise diversity, which has been a concern owing to its negative side effects.

Practical implications

This study offers three main practical implications. First, corporate managers should not have excessive expectations of diversity. Recently, in response to social demands, the number of companies appointing Chief Diversity Officers has increased dramatically, and these organisations are adopting diversity indicators to evaluate their business performance; however, such activities have been criticised as ‘woke’ and have come under scrutiny because they have little impact on the performance of their main business (Foss and Klein 2023; Warren 2022; Winston 2023). There are even companies that hypocritically claim to be diversity-conscious, and in many cases, they have deviated from their essential purpose (Crowley and Eccles 2023). Although managers in an organisation should be sensitive to the trends and demands of the times, overemphasis in this regard may lead to other factors being neglected. Therefore, as this study shows, organisations should take an overview of their resources and clearly prioritise the factors that will have the greatest impact. If one wants to increase organisational identification, corporate purpose will have a much greater effect, if one wants to increase the meaningfulness of work, one should be in charge of an attractive product/service, and if one wants to suppress burnout, psychological safety will have a much greater effect. Therefore, designing a workplace that can achieve skill acquisition, which simultaneously achieves all three of the aforementioned, is a promising personnel policy. To not overestimate the object of interest, it is important to extract and evaluate factors from a bird’s-eye perspective.

Second, it is difficult to improve organisational performance simply by hiring diverse personnel of different ages, genders, and nationalities. When considering diversity, it is easy to focus only on visible characteristics; however, diversity also includes invisible characteristics such as employees’ values and skills. To enhance perceived diversity, organisational managers need to provide employees with training and foster their values. Alternatively, management should ensure that skill diversity is maintained in each department. To effectively utilise diversity within an organisation, the organisation needs to manage diversity from three aspects: demographics, perceptions, and skills. Until now, many organisations that have not been able to enjoy the benefits of diversity have likely only managed visible diversity, such as race and gender. By changing this situation, diversity can be transformed into an asset that contributes to improving corporate performance.

Third, it is important to balance diversity-driven innovation with the risk of burnout. As shown in this study, the combination of demographic diversity and skill diversity simultaneously produces positive and negative effects on employee attitudes. Hence, to mitigate the risk of burnout, it is important to have a strategy that promotes collaboration among experts without intensifying competition, and to design a system that ensures psychological safety.

Conclusion and future directions

Extant literature on organisational diversity emphasised its negative aspects; however, limited knowledge is available on how to solve them. Therefore, this study categorises diversity into three categories: visible demographic, invisible employee perceived and skill diversity. Based on the structural equation modelling results of data from an online survey of 3000 full-time employees at Panasonic Connect Co., Ltd. in Japan, demographic diversity × perceived diversity had a significant positive effect on the meaningfulness of work. This suggests that employees’ increased awareness of diversity may make them more likely to find job meaning in a demographically diverse workplace. Demographic diversity × skill diversity had a positive effect on organisational identification and meaningfulness of work. This suggests that employees want to work for an organisation that makes the most of their talents, that is, one in which they feel a greater sense of belonging owing to the diverse workforce utilizing their diverse skills. However, regarding demographic diversity × skill diversity, a concern exists that it may increase the risk of burnout owing to employees’ competitiveness. Japanese companies have a long history of protecting employees through a stable seniority-based system, as well as of placing importance on cooperation owing to the Japanese culture that values harmony. When this system collapses and skill-based personnel evaluation is introduced, and conflicts arise owing to racial and gender diversity, internal competition intensifies. This is particularly true for employees with weaker specialised skills, who are more likely to experience a loss of psychological safety and experience these disadvantages. Hence, managers who want to reap the benefits of diversity should design a system that ensures psychological safety while introducing diversity in three dimensions (demographics, perceptions, and skills) to avoid any negative effects.

Despite the significant contribution of this study, it has some limitations. First, given that the study was conducted on employees of Panasonic Connect Co., Ltd. in Japan, the findings have limited generalizability. Second, a detailed analysis of cultural distance was not conducted on demographic diversity. For example, from the perspective of Japanese people, it is believed that having Western colleagues is more beneficial than having Korean colleagues in the workplace (Orsini and Magnier-Watanabe 2023). Accordingly, deeper insights would be gained by establishing specific classifications for demographic diversity, as well as for perceived diversity and skill diversity. Third, the classification of diversity was limited to demographic, perceived, and skill diversity. In particular, invisible diversity includes career experience and religious diversity. Fourth, the degree of diversity was not considered. The effect of each aspect of diversity may change depending on the degree. Fifth, in this study, perceived diversity and skill diversity were measured using a single item to conceal the purpose of the study from the subjects and to take into consideration social desirability bias. This is because the subject of this study is diversity, an area of great social significance, and is therefore susceptible to the influence of social desirability bias (Boring and Delfgaauw 2024). Some studies have also argued that multi-item scales are redundant and that respondents’ perceptions are better represented by single-item scales (Matthews et al. 2022; O’Brien et al. 2022; Trail et al. 2023). In other words, unless a survey measures multiple items, it cannot be concluded that it is unreliable. However, to obtain more robust conclusions, it is also important to extend the study of diversity with multiple items while considering social desirability bias.

Sixth, although we assume that diversity affects employee attitudes, the opposite may also be true. To address this limitation, experimental verification using randomised controlled trials is preferable. This approach allows us to evaluate pure causal effects. Thus, accumulating knowledge on this topic from both observational and experimental studies can increase the reliability of our conclusions. These are directions for future studies.