Influence and Inequality: Worker Identities and Assessments of Influence over Workplace Decisions

: Though research shows that diversity and equity in the workplace lead to more innovation and other positive outcomes for organizations, businesses often struggle to accomplish their diversity, equity, and inclusion goals. Promoting employee voice is one strategy to support equity; however, employee perceptions of who has a voice at work may be increasingly unbalanced in the post-2020 workplace. Thus, drawing from an original survey dataset of workers across 20 countries and regions ( n = 9197), we use logistic regression to explore how sociodemographic characteristics and perceptions of inclusion at work predict whether participants believe they help inﬂuence important decisions at work (our measure of employee voice). Across our global sample, we found that although feelings of inclusion predict the perception that one has more voice, workers who belong to groups historically marginalized in the workplace due to gender, education level, compensation type, leadership status, and self-identiﬁed “minority” status report lower levels of agreement with the statement of voice. We conclude that while promoting feelings of inclusion is one strategy for achieving a greater diversity of employee voices at work, organizations should also take concrete steps (such as diversifying leadership) to reach equity more fully regarding voice.


Introduction
Recruiting and retaining a diverse workforce is top of mind at many organizations going into 2023, and with good reason.For instance, research finds that diverse work teams are often more productive and innovative than homogenous teams (Sturt 2014;Yeager and Nafukho 2012).However, organizations have sometimes struggled to realize their diversity goals, and this problem is increasingly magnified three years after the onset of the COVID-19 pandemic.Past research has shown that supporting employee "voice" (say or influence over workplace policies and decisions) is related to more inclusion, equity, and positive workplace experiences for employees (Bailey and Madden 2016;Bell et al. 2011;Holland et al. 2011;Prouska and Kapsali 2021;Spencer 1986;Wilkinson et al. 2020;Wilkinson and Fay 2011).However, research suggests that employee voice may be increasingly unequal among worker groups due to larger workplace trends and upsets over the last few years (Henly et al. 2021).Thus, in the current economic climate, it is helpful to investigate whether there are disparities in perceptions of voice among worker groups based on identity statuses.It would also be beneficial to investigate this issue on a global scale.Doing so can shed light on why organizations worldwide continue to have trouble fully realizing their diversity, equity, and inclusion goals and provide insight into what might be done in the future to increase their success in this regard.
In this paper, we use data from an original survey of employees in 20 countries and regions (n = 9197) to explore employees' perceptions of their "voice" at work.Controlling for feelings of inclusion at work, we use logistic regression to explore how demographic characteristics relate to employee agreement with the sentiment that they can influence important decisions in the workplace (our measure of employee voice).Our interdisciplinary approach bridges insights from sociology, human resources, and management literature to explore connections between broader patterns of social inequality and employee perceptions of "influence" as a measure of employee voice at work.This approach has sometimes received less attention in the literature on voice, which tends to focus on contextual and managerial issues (Donaghey et al. 2011;Dundon and Gollan 2007;Wilkinson et al. 2018).
Our results demonstrate significant differences in perceptions of influence or voice among surveyed employees that fall along identity lines, and we argue that these disparities challenge organizational diversity, equity, and inclusion efforts.These findings have important implications.If organizations can cultivate voice among all-rather than selectemployees, they can move closer to reaching their diversity and inclusion goals.

Background Employee Voice and Power Inequalities in the Workplace
The concept of "employee voice" has historically been a concern of scholars in businessrelated fields, such as management, human resources, and industrial relations.However, theoretical insight from sociology can provide nuance to interpreting and studying this concept.Though specific definitions can differ, "employee voice" is most simply articulated as the extent to which employees believe they have a "say" in work activities and decisions (Wilkinson et al. 2018(Wilkinson et al. , 2020;;Wilkinson and Fay 2011)."Employee voice" as a concept dates back to the 1970s and has arguably most often focused on how employee feedback can lead to positive, constructive changes in the workplace (Bashshur and Oc 2015;Gunawardana 2014;Hirschman 1970;Holland et al. 2016;Maynes and Podsakoff 2014;Morrison 2014;Wilkinson et al. 2018).Scholars have found strong relationships between the proliferation of employee voice and positive outcomes for employees and organizations, including improved performance and professional relationships among employees in a given workplace context (Kim et al. 2022;Li et al. 2020).
Over time, scholars have conceptualized employee voice slightly differently depending on the academic field, with some discussions centering employees as individuals and others focusing on employees as a collectivity (Kaufman 2015;Wilkinson et al. 2020).For instance, industrial and labor relations literature often investigates voice from the perspective of worker groups, asking questions about how employees "assert and protect their interests" in the workplace context (Wilkinson et al. 2020).Literature from the fields of organizational behavior and human resource management differ from industrial and labor relations in that they sometimes take a more individualized or mixed-level approach, investigating issues such as employee participation in decision-making and the conditions under which individuals' voices are encouraged or discouraged (McCabe and Lewin 1992;Morrison 2014).However, scholars have recently pointed out that although literature diverges, insularity is probably a more significant issue than difference among fields (Wilkinson et al. 2020).There is a need for more dialogue and synthesis of ideas among these fields and exploration into how micro/individual, meso/organizational, and macro/societal influences on voice can converge (Kwon 2017).
It is vital to map the micro, meso, and macro influences on employee voice because social systems and individual behaviors work together in any given context (Kwon 2017;Wilkinson et al. 2020).While in general, much work on voice has focused on the meso and micro factors at play (Klaas et al. 2012;Mowbray et al. 2015), macro social forces that exist outside of-but still influence-the workplace environment have received less attention (Kwon 2017).One reason for the lack of focus on macro elements may be a perceived lack of control on the part of organizations.For example, organizations may be able to make some changes to company culture and climate through training leaders on issues of inequity, but they cannot eradicate racism or sexism in society.However, social power inequalities that exist outside of the workplace can relate to voice patterns within the workplace (Brooks 2018;Wilkinson et al. 2018).For instance, sociological research has shown that broader social inequalities based on race, gender, and social class are replicated in the workplace and impact employees' individual experiences at work (Alfrey and Twine 2017;Giazitzoglu and Muzio 2021;Ispa-Landa and Thomas 2019).
Concurrently, research has shown that individuals who belong to groups that have historically been (or are currently) minoritized in the workplace often have less voice and influence than others (Bell et al. 2011;Wilkinson and Fay 2011).This evidence suggests that macro forces such as prevailing racism, classism, and sexism embedded in society's structure make their way into the workplace despite organizational efforts to remove these issues locally.Research has shown that organizations often fail to eradicate inclusion issues in the workplace.Rather than address structural inequalities, firms often try to force individual behavior and attitudinal changes among leadership-a strategy that typically is not successful (Dobbin andKalev 2016, 2022).Extant literature on employee voice does not always link individual experiences at work to broader social inequalities; however, patterns inside and outside the workplace often overlap (Brooks 2018;Wilkinson et al. 2018).
Sociological research has long made connections between the micro and the macro (Armstrong et al. 2006), and certain groups of scholars (such as scholars of intersectionality) have also pointed out that personal experiences are reflective of complicated structural forces connected to individuals' identity statuses (Collins 1990(Collins , 2019;;Crenshaw 1991;Madfis 2014;Nash 2018).In the last few years, some literature on employee voice has tried to speak to this point; however, there is still a greater need for research on voice that considers the importance of divergent experiences among employees-particularly employees who belong to groups that have been historically marginalized in the workplace (Kaufman 2015;Wilkinson et al. 2018).
To date, sociological and other social science research has shown that individuals' experiences in workplace contexts are linked to broader structural inequalities of race, class, gender, ability status, and other identity locations (Alfrey and Twine 2017;Brown and Moloney 2019;Rosette et al. 2018;Tatli and Özbilgin 2012).These social inequalities manifest in employees' lives and have implications for voice.Yet, when it comes to research on voice, the links between the structural and the individual are sometimes forgotten in lieu of individual employee characteristics or specific company contexts.Company context is still important, yet the broader patterns of inequality present at the societal level influence these contexts.Thus, these broader patterns are a significant challenge to change.

Materials and Methods
The data in this paper are derived from an annual study conducted by the authors' research institution.The data profiled here were gathered in 2021 and comprised a convenience sample of employees across 20 countries and regions on six continents (excluding Antarctica) and a wide range of industries.Potential participants had to be employed at companies with at least 500 employees to qualify for the survey.The authors' research team built the survey using Alchemer software and used Lucid marketplace to screen and administer the survey.Since Lucid is a sample aggregator that works with contracted panel providers, we cannot say for certain how many potential respondents received a survey invitation.However, approximately 13,259 participants were screened for the survey, and 9289 returned completed surveys eligible for analysis.
According to Coppock and McClellan (2019), Lucid is the largest sample aggregator for online respondents.Lucid enables direct-to-respondent sampling through the marketplace platform, aggregating respondents from various contracted panel providers (Coppock and McClellan 2019).Panel providers are responsible for respondents' compensation; compensation varies from cash, gift cards, or reward points for various merchandise.Ultimately, our sample is one of convenience and is subject to the statistical errors associated with convenience studies.However, our sample is sufficiently large enough to generate meaningful exploration of employee voice and inequality.

Survey and Instrumentation
We posted a survey on the Lucid recruitment platform.Respondents were able to choose if they would like to participate, with no harm occurring if the respondent chose not to participate.Further, respondents could exit the survey at any time with no penalty to their quality score with the panel provider.After informed consent, the survey asked respondents a series of questions about inclusion sentiment within their workplace.We detail these questions in Table 1.This battery of questions includes five statements covering: the perceptions of equity of access to opportunities; comfort with discussing diversity and inclusion with leaders; perceptions of leader appreciation for their identities; the belief that the organization is interested in understanding, rather than categorizing, them; and the belief that their opinions are fairly represented.Respondents were presented with a Likert scale (1-5) and prompted to answer with their level of agreement.These questions are appropriate for factor analysis, with more details appearing in the findings below.These questions capture respondents' feelings and perceptions around the general inclusivity of their workplaces, rather than specific actions their workplaces have taken to be more inclusive.In addition to these inclusion sentiment questions, we asked respondents a series of demographic questions.Upon completion, the survey directed respondents to a debriefing page and credited them for participating.The survey results profiled here are the latest in a series of surveys administered annually by the authors' research team since 2018.Each year, these surveys ask questions about individuals' experiences at work.For example, employees answer questions about workplace engagement, burnout, inclusion, experiences with their leaders, and other aspects of their employee experience.In addition, the survey asks respondents a series of questions on the perceived opportunities available to them at work.These statements are also measured with a Likert scale (1-5), gauging the level of agreement.One of these statements, the extent to which participants believe they help influence important decisions at work, serves as the response variable for our analyses.

Analytic Strategy
Based on our background research, we have two hypotheses.They are as follows: H1: Employees who feel a sense of inclusion will feel a greater sense of employee voice.
H2: Controlling for a sense of inclusion, employees who belong to historically marginalized groups (due to gender, compensation type, education, leadership status, sexual orientation, and self-identified "minority" status) will have lower perceived levels of employee voice.
For our analyses, the response variable we used was employee responses to the question, "I help influence important decisions at work."Although the original response was measured ordinally, we created a dichotomous variable for our analyses, comparing those who agreed or strongly agreed with all other responses (neutral or did not agree).Inde-pendent variables included all the demographic variables listed in H2 and a dichotomous variable based on the inclusion battery.After assessing the battery (discussed more in the findings below), we then created a dichotomous variable from the battery results to represent those who felt a general sense of inclusion at work versus those who did not.
We used StataMP 17 to generate chi-square tests and estimate a logistic regression.Our response variable for the logistic regression is the "influence" statement, with the model also including gender (cisgender men, cisgender women, transgender, or gender not identified); identification as a minority, broadly defined (either yes or no); compensation type (either salary or hourly); identification as a people leader at work (yes or no), highest level of education (less than high school to postgraduate school), sexual identity (straight or LGBQ+), and inclusion sentiment (yes or no).In addition to the inclusion battery, we chose these variables because previous research has shown that individuals who identify membership with these demographic groups have historically faced discrimination, unfairness, and/or underrepresentation in the workplace (Bobbitt-Zeher 2011; Newman 2002; Ozeren 2014; Schilt 2010; Van Laer and Janssens 2011).Although race/ethnicity is another category that could be included in analyses like ours, we did not include it due to the complexity of measuring race and ethnicity across an international sample.Instead, we used the "minority" variable which includes, but is not limited to, employees who identified as racial/ethnic minorities in their own workplace contexts.The question from which this variable is derived asked respondents whether they self-identified as a "minority".While in a U.S. context other terms are often preferred, we elected to use the term "minority" in our survey question for two reasons: first, due to its translatability for our international sample, and second, in an attempt to best capture employees who self-identify as belonging to any minority group (whether that be due to race/ethnicity, religion, disability, and so on), as "minority" status often relates to historical (and/or current) marginalization and exclusion in the workplace.
In addition to the variables described above, we also included respondent age and country, as it is reasonable to conclude that the dependent variable may change based on generational dynamics and influences by geographic area.The countries and regions sampled include: Argentina, Australia, Brazil, Canada, China, France, Germany, Hong Kong (analyzed separately from China), India, Japan, Mexico, Philippines, Singapore, Saudi Arabia, United Arab Emirates, Russia, South Africa, and South Korea, the United Kingdom, and the United States.In addition to fielding the survey in English, we also contracted a translation service provider to translate the survey into 10 additional languages for some of the locations under study: Spanish (Argentina and Mexico), Portuguese (Brazil), Chinese (China and Hong Kong), French (France), German (Germany), Hindi (India), Japanese (Japan), Russian (Russia), Arabic (Saudi Arabia and United Arab Emirates), and Korean (South Korea).The specific number of respondents from each geographic location (and from every demographic group) can be found in Table 2.A total of 9289 participants returned completed surveys.However, we dropped participants who did not describe their compensation type as either "hourly" or "salary" (n = 92), which left 9197 participants for chi-square analyses.Some participants also declined to give information on their compensation type and were removed from the regression analysis (n = 2609).This brought our final sample down to 6497 for the regression.

Results
Below, we first detail our results in assessing the workplace inclusion sentiment scale described above.Second, we discuss our regression model that explores relationships between inclusion and voice as well as between demographics and voice.

Inclusion Sentiment Scale, Factor Structure, and Internal Consistency
To assess the suitability of the inclusion sentiment data for factor analysis, we estimated Cronbach's Alpha, Bartlett's Test of Sphericity, and the Kaiser-Meyer-Olkin measure of sampling adequacy (KMO).We found that α = 0.806, indicating strong internal consistency across the five items.The Bartlett test indicated that we can reject the null hypothesis that variables are not intercorrelated (χ 2 = 15171.403,df = 10, p < 0.001), and the KMO test indicated factor analysis suitability (0.782).We conducted a principal component analysis to explore factor structure.The model specified converges into one factor with all loadings greater than 0.60.We present the factor loadings in Table 1.We conclude that the factor structure of the inclusion battery is sound and appropriate for further statistical analysis.

Employee Characteristics, Inclusion, and Voice
Chi-square tests of association proved significant in all but the "sexual identity" variable as outlined in Table 2.However, we elected to keep this variable since there is reason to believe that sexual identity may have an impact on perceptions of voice (Bell et al. 2011).Table 2 also outlines the number of responses we received from each country or region of interest (as well as from each demographic group of interest).
Next, we conducted logistic regression to explore the relationship between employee demographics, inclusion, and our dependent variable of interest, perceptions of employee influence at work.The pseudo R 2 for our model was 0.21, indicating acceptable fit.We present the results of the regression in Table 3. Notably, we found that feeling a sense of inclusion had a strong relationship to feeling a sense of influence at work.We also found that belonging to gendered groups historically marginalized in the workplace, as well as being paid hourly (compared to salaried) and self-identifying as a "minority," led to decreased odds of feeling a sense of influence.People leaders as well as respondents with a postgraduate education (compared to college-educated respondents) also had greater odds of a sense of voice.At the same time, compared to cisgender men and college graduates, cisgender women, trans participants, and participants with less education had lower odds of feeling they had a voice.Interestingly, we did find differences between countries as well.Compared to the United States, workers in Brazil, China, and Mexico had greater odds of feeling they had influence at work whereas workers in Japan, Singapore, Canada, and the United Kingdom had lower odds of feeling they had influence at work.

Discussion
Since the beginning of the COVID-19 pandemic, the landscape of work has changed rapidly.Unfortunately, historical advances toward equity in the realm of work seem to be backsliding in some cases.For instance, scholars have documented worsening inequalities at work among employees across the world along race, gender, and social class lines (Dalessandro and Lovell 2022;Hall et al. 2021;Henly et al. 2021;Mooi-Reci and Risman 2021).Employee voice can be one strategy to encourage more equity and inclusion.However, previous research on voice has often focused on individual employees and organizational contexts.Organizational context matters, yet social inequalities influence workplace environments as well and have an important impact on employees' experiences.Thus, our research sought to explore, accounting for individuals' feelings of inclusion at work, whether membership in identity groups historically marginalized at work matters for employees' perceptions of their voices at work.Our work is unique in its use of a large international sample and yields both practical and scholarly insights.
First, using an original inclusion battery capturing employee sentiment about whether diversity, equity, and inclusion is supported in their workplaces, we found that there is a strong relationship between feeling inclusion and feeling a sense of voice across our international sample.Thus, we found support for our first hypothesis (H1).At the same time, controlling for feelings of inclusion, we found that workers belonging to historically marginalized groups due to gender, compensation type, education, leadership status, and identification with a "minority" identity status (broadly defined) felt a diminished sense of voice.We also found support for our second hypothesis (H2), with the exception that sexual identity did not have an impact on perceptions of voice.Our findings suggest that structural inequalities matter for employees' perceptions of voice and indicate that future research on employee voice should do more work to account for ways to address social inequalities that exist outside of, but also influence, the workplace and employees' experiences in the workplace.
Our findings indicate that, using influence over important workplace decisions as an indicator of voice, disparities continue to exist among certain groups of employees controlling for feelings of inclusion.Although employees ostensibly have more control in the workplace now than they did in the recent past (Coughlin 2021), research has shown that the pandemic generally exposed structural inequalities at work rather than challenged them.This is especially the case when comparing hourly and salaried workers, white workers and workers of color, and men and women (Dunatchik et al. 2021;Goldman et al. 2021;Kantamneni 2020).Further, pandemic-related job losses and furloughs have disproportionately impacted historically marginalized groups, such as women and people of color (Dang and Viet Nguyen 2021;Sáenz and Sparks 2020).One side effect of recent negative workplace trends such as these could be unequal perceptions of voice.In this sense, our findings are not surprising.However, our findings do provide evidence that organizations may miss out on fully realizing the benefits of attracting and retaining a diverse workforce if more is not done to address inequalities of voice.
On the organizational side, while inequalities of voice that fall along identity lines are problematic due to the impact on individual employees, they also counteract organizational efforts to diversify the workforce because a lack of voice relates to decreased job satisfaction and less organizational loyalty among workers (Farndale et al. 2011;Holland et al. 2011).As 2020 brought increased remote work options, workers have begun to have more choice when it comes to their work locations and even their job roles (Coughlin 2021).Employees have good reason to switch jobs if they perceive they are being treated poorly.Thus, organizations that fail to make changes in support of equity will more than likely continue to lose out on employees and the innovation that diverse viewpoints and ideas can bring to workplace settings.
Given our findings, we have some suggestions for how researchers and practitioners might move forward.Our inclusion battery questions capture employees' feelings around whether their workplaces are supportive of diversity, equity, and inclusion.However, our findings demonstrate that in addition to cultivating a sense of inclusion, organizations will need to do more to better cultivate an equitable sense of employee voice.For instance, although it is beyond the scope of our own research, other scholars have suggested taking steps such as targeted recruitment of employees from historically marginalized groups and instituting family-friendly policies (Dobbin and Kalev 2022).These strategies can work because they help address long-running issues of systemic inequality, which-while difficult for organizations to address-are important to counteract whenever possible so employees see their organizations as committed to equity.
Our research is not without limitations.For instance, although we were able to survey a large number of employees across the world, our sample is ultimately one of convenience.Thus, it is not representative of all workers in each of the countries and regions we sampled.Further, some of the demographic groups-such as transgender respondents-were very small, and so our study is less representative of the views of respondents with these identities.However, our study does fulfill a need for more research on the topic of employee voice that utilizes international samples.Helping employees feel a sense of inclusion at work is helpful for equalizing voice, which can in turn help the workplace feel like a place that's receptive to diversity, equity, and inclusion for employees.However, to fully equalize voice, organizations should take steps to ensure that they are doing what they can to address broader social inequalities and taking concrete actions in an attempt to make all employees feel included.Exploring this point further is an avenue for future research.

Conclusions
Ultimately, we find that gender, compensation type, leadership status, identification with a "minority" identity status (broadly defined), and education all structure perceptions of voice at work.This issue is especially important now.Although many organizations are beginning to take equity and inclusion initiatives seriously, unsatisfactory responses to pandemic-related issues in the workplace-which disproportionately impacted historically marginalized groups (Choi et al. 2021;Dang and Viet Nguyen 2021)-are likely top of mind for employees.Supporting employee voice is one strategy to advance diversity, equity, and inclusion initiatives, although our findings indicate that inequalities in perceptions of voice still persist when accounting for feelings of inclusion among employees across the world.
Along with individual-level and meso/organizational factors, we find that macro structural inequalities help determine individual employee perceptions of voice (or lack thereof) in the workplace.As social scientists have pointed out, broader patterns such as increases in classed and gendered divisions of labor, as well as worsening worker alienation, since the beginning of the pandemic means that scholars will need to closely explore issues such as workplace equity going forward (Mooi-Reci and Risman 2021).While examining voice is one way to capture disparities, employees' perceptions of their voice may have changed over the last few years, especially given rapid changes to the workplace such as the increase in remote work and changes in individuals' perceptions about the role of work in their lives (Coughlin 2021).Thus, going forward, researchers of employee voice should take into account how widespread, macro social changes matter for employees' perceptions of voice.On a broader level, cultural change that challenges broader patterns of inequality existing outside of individual organizations is needed.However, organizations can still take additional actions aimed at addressing systemic inequalities and support the cultivation of voice among employees, particularly for those belonging to historically marginalized groups.

Table 1 .
Inclusion Sentiment Scale Items and Factor Loadings.
Note: Table1is adapted with permission from the O.C. Tanner Institute.

Table 2 .
Demographic Characteristics and Chi-Square Results (Influence at Work Question).

Table 3 .
Logistic Regression Results (Influence at Work Question).