Volume 3, Issue 3: Autumn 2022. Special Collection: Technology, Work, and Inequality. DOI: 10.1037/tmb0000090
Electronic performance monitoring (EPM) technologies are widely used by organizations, but research has shown that workers feel EPM is unacceptable when it violates their needs for autonomy and transparency. These negative consequences of EPM have the potential to contribute to broader issues of workplace inequality, such that more vulnerable workers are disproportionately affected. Reasoning about how worker status may affect perceived autonomy and transparency with regard to EPM led to the prediction that workers with lower socioeconomic status [SES] (education and income) and job status (job zone, job security, tenure) experience more EPM that they perceive to be unacceptable (perceived unacceptable EPM [EPM-U]). Additionally, we explored whether workers endorse justifications for EPM based on consent, transparency, and autonomy differentially based on SES and job status. Results from a sample of 267 diverse workers showed that tenure predicted EPM-U, but other indicators of SES and job status did not. Supplementary analyses examining different operationalizations of perceived EPM acceptability as outcomes also supported the effect of tenure. On average, participants endorsed the consent and transparency EPM justifications at a high rate but not the autonomy justification, and status differences did not affect these rates. Results showed that workers in general feel that EPM is only sometimes acceptable, depending on how the organization justifies it. We discuss theoretical implications for technology in the workplace, practical implications regarding how EPM should be implemented in ways that benefit all workers, and suggestions for future research.
Keywords: electronic performance monitoring, organizational justice, inequality, socioeconomic status, job status
Supplemental materials: https://doi.org/10.1037/tmb0000090.supp
Acknowledgments: The authors thank current members of the WAVE Lab for help in developing this article.
Disclosures: The authors have no conflicts of interest to disclose.Data Availability: The authors are unable to share the raw data set for this study publicly. In the institutional review board (IRB) protocol for this research, the authors specifically noted that only members of the research team would have access to the raw data and did not stipulate an exception for the sharing of de-identified data.
Study materials and code for primary and supplementary data analyses and visualizations are available at https://osf.io/vam7j/.
Correspondence concerning this article should be addressed to Tara S. Behrend, Department of Psychological Sciences, Purdue University, 703 Third Street, West Lafayette, IN 47907, United States [email protected]
Concerns about data privacy and lack of control over personal information are pervasive in public discourse about technology. A 2019 Pew Research survey reported that most people are concerned about how their personal data are being collected and used by organizations (Auxier et al., 2019). Responses indicate that these concerns are largely driven by individuals feeling that they lack autonomy over their data and that organizations are insufficiently transparent in how they collect and use personal data. These apprehensions are likely worsened by publicized instances of organizations misusing the data of employees and patrons (Lapowsky, 2019; Shook et al., 2019). As a result, policymakers and others have begun to debate laws and norms for the collection and use of personal data by organizations.
Relevant to the conversation about data privacy and technology is an organizational practice known as electronic performance monitoring (EPM). Employers use a variety of EPM technologies to digitally record and analyze data related to employee job performance (Ravid et al., 2020; West, 2021). Prior work has attempted to categorize the many purposes and targets of EPM, but often the most salient information provided to monitored employees is about the device or tool itself. These technologies may measure a variety of work- and non-work-related behaviors including task performance (e.g., tracking company-provided email), organizational citizenship behaviors (e.g., energy use, waste disposal), and counterproductive work behaviors (CWBs; e.g., bag searches, drug tests; see Tables 1 and 2, and the online Supplemental Materials, for more EPM examples). Though organizations can use data collected through EPM to their employees’ benefit, EPM can also lack transparency and violate workers’ autonomy, leading workers to consider these practices unacceptable. Indeed, data practices of employers are among the top sources of privacy concerns for Americans (Auxier et al., 2019; Rainie & Duggin, 2016).
Some evidence exists regarding the reasons that employees may feel that EPM practices are unacceptable; however, research has yet to address issues of inequality that may be exacerbated by the negative effects of EPM. In other words, do EPM technologies contribute to workplace inequality by resulting in certain workers experiencing more forms of EPM that they perceive to be unacceptable? This question is related to the larger concern that technological progress in the workplace may benefit some and burdens others, instead of benefitting all (Coovert & Thompson, 2013). Studies of reactions to EPM have largely focused on characteristics of the monitoring itself, and reviews of the literature have concluded that they significantly influence reactions (Ravid et al., 2020; Stanton, 2000). For example, perceived levels of invasiveness and transparency of EPM can influence its acceptability to workers, with high invasiveness and low transparency consistently evoking negative reactions (Alder & Tompkins, 1997; Bhave et al., 2020). Theoretical frameworks including self-determination (Gagne & Bhave, 2011) and organizational justice (McNall & Stanton, 2011) are useful in explaining why workers react negatively to EPM technologies and can help explain why certain workers are particularly likely to experience EPM practices that they perceive as unacceptable.
The present study examines whether EPM technologies contribute to inequality at work based on the socioeconomic status [SES] (e.g., education, income) and job status (e.g., job zone, job security, tenure) of workers. Practically, SES and job status are important variables due to growing economic and educational disparities in society (International Labour Organization, 2020). Further, these factors are pertinent to the study of EPM because, to a large extent, individuals with higher status decide who has certain rights and privileges in organizations (Anderson & Brion, 2014; Keltner et al., 2003). Decisions made about EPM by those with more organizational status and power may result in lower status workers, who have less influence over these decisions, disproportionately feeling that EPM violates their needs for autonomy and transparency. As a result of these status disparities, we expect that workers with lower SES and job status will experience more instances of EPM that they perceive to be unacceptable. We also explore whether SES and job status explain differences in the endorsements of certain justifications for EPM. In doing so, we contribute to the EPM, privacy, and workplace and technological inequality literature, and policy discussions regarding EPM and the rights of individual workers.
Two of the focal concepts in this article, inequality and performance, require further definitional clarity. Inequality can take on different conceptual definitions across studies and fields of study. For example, inequality can be defined in terms of outcomes (e.g., organizational distribution of rewards; Baron & Pfeffer, 1994) or in terms of treatment (e.g., discrimination; Avery et al., 2018). In the current article, we define inequality in the context of EPM as the disproportionate experience of EPM that is perceived to be unacceptable by certain workers. Thus, inequality is the product of the extent to which organizations monitor certain workers (i.e., treatment) and the extent to which workers perceive the EPM that they experience as unacceptable (i.e., outcome).
We define performance using a broad conceptualization that includes both task performance and contextual performance (Borman & Motowidlo, 1993). Importantly, this conceptualization includes CWBs (Sackett & DeVore, 2001) such as employee theft, which are often the target of workplace surveillance practices (e.g., metal detectors, bag scanners). This definition also includes organizational citizenship behaviors, which are voluntary behaviors directed at supporting coworkers or general organizational functioning (Borman & Motowidlo, 1993). These behaviors are less frequently targeted by monitoring but may be captured by some forms of EPM (e.g., energy usage and waste).
The use of EPM technologies in the workplace creates a challenge for organizations of balancing their rights and interests with those of their workers (Culnan et al., 1994; Selmi, 2006). Workers have the right and need to self-regulate, which includes having control over their environment and personal information (i.e., autonomy) and transparency regarding the use of technology to gather and analyze their personal information, while organizations have the right and need to monitor and regulate their workforce (Bhave et al., 2020). There is, however, often a power imbalance in this relationship. Workers are likely to feel more compelled to surrender their rights and interests for fear of the consequences of noncompliance (e.g., termination) and because they are operating on the organization’s time and property. This tension has risen to the forefront of the regulatory debate over how to balance the rights of individuals and organizations. The protection of individuals is often the focus of both scholarship (Levinson, 2009) and policy (e.g., General Data Protection Regulation [GDPR]).
Even when EPM is used in the pursuit of legitimate organizational interests, the extent to which workers perceive EPM as unacceptable is determined, in part, by how workers perceive its characteristics. Recently, a typology of EPM characteristics was developed as an organizing framework for studying how they affect reactions to monitoring (Ravid et al., 2020). Among these characteristics are EPM invasiveness and transparency, which are defined as, “the amount, target, and systematic constraints placed on EPM use” and, “the extent to which employees are provided information about the characteristics of monitoring” (p. 104). Perceptions of EPM characteristics, including invasiveness and transparency, are a function of the technology itself (e.g., whether the technology targets behavior or physiology) and how the organization implements the technology (e.g., whether workers have information about and control over how they are monitored). These factors affect workers’ sense of autonomy and transparency in relation to EPM. In at least some circumstances, workers may feel that the EPM they experience violates autonomy and transparency to an unacceptable degree, which we define as perceived unacceptable EPM (EPM-U). In the next section, we discuss two theoretical frameworks and empirical findings in the existing literature to explain why workers consider EPM unacceptable when they are invasive and lack transparency. Based on this theoretical and empirical grounding, we then discuss why low-status workers may disproportionately experience EPM-U.
Organizational justice theory is an approach that has been used to explain EPM reactions. Procedural justice, one aspect of organizational justice, is particularly relevant to perceptions of EPM acceptability. This concept is understood as the perceived fairness of organizational policies and procedures that are used to make decisions (Colquitt et al., 2013). Components of procedurally fair organizational practices include ones that apply equally to all employees and ones that are open to some level of input (Rupp et al., 2014). Workers tend to consider practices that violate fairness norms to be unjust.
Research has examined several EPM characteristics related to invasiveness and transparency that explain whether workers see them as procedurally just. For instance, if workers are informed that EPM is job relevant and they are given greater voice in the implementation of monitoring, procedural justice perceptions are more positive (Ambrose & Alder, 2000). Workers also consider physiological monitoring to be less justified than other forms of monitoring due to its perceived invasiveness (Kaupins & Coco, 2017). Further, the relationship between EPM and perceptions of fairness is mediated by the extent to which individuals perceive monitoring as an invasion of privacy (Alge, 2001). Thus, EPM tends to be considered unacceptable when it is seen as invasive and is implemented without input from workers because it violates feelings of procedural justice.
Another psychological framework that helps explain EPM acceptability is self-determination theory (SDT; Deci et al., 2017). Most broadly, SDT is a metatheory of motivation based on psychological need fulfillment. It posits three fundamental needs: autonomy, relatedness, and competence. A lack of the fulfillment of any of these needs by one’s environment (e.g., workplace) is thought to result in negative psychological outcomes (Ryan & Deci, 2000). The extent to which EPM infringes on workers’ need for autonomy, or the ability to act independently of external forces, is an important aspect of perceived EPM acceptability.
According to SDT, the more that EPM reduces workers’ autonomy, the more negative their reactions should be. As with procedural justice perceptions, EPM that is more invasive and less transparent also degrades feelings of autonomy. Supporting this notion is evidence that workers consider EPM technologies to be less invasive and more acceptable when organizations provide more control over how and when monitoring occurs (McNall & Stanton, 2011; Zweig & Webster, 2002) and how data are used (Alge, 2001). Additionally, higher levels of transparency provided by the organization regarding the purpose of EPM provide greater autonomy by giving workers the ability to judge for themselves whether EPM practices are justified and acceptable (Alder et al., 2006). Thus, the extent to which the implementation of EPM technologies violates workers’ sense of autonomy will lead to lower perceptions of EPM acceptability.
Importantly, perceptions of EPM characteristics may not be the same across all workers. Rather, workers’ perceptions of the invasiveness, transparency, and ultimately acceptability of the EPM they experience may be related to personal and/or job characteristics such as SES and job status. Although we know of no empirical studies that explicitly examine worker status variables, research has explored other worker and job characteristics and has shown that they are related to the characteristics of the EPM that managers choose to implement (Alge et al., 2004) and workers’ reactions to EPM (Holland et al., 2015). In the next section, we explain the theoretical relationship between the status of workers and their experience of EPM-U.
The present study examines SES and job status as factors that influence EPM-U. The SES variables that we focus on are income and level of education. Although definitions of SES can differ across studies, both income and level of education are usually incorporated as indicators in the operationalization of SES (e.g., Mueller & Parcel, 1981) and social class (e.g., Morgan, 2017). Reviews of the literature have found that these variables substantially influence individuals’ thoughts, feelings, and behaviors (Manstead, 2018). Regarding job status indicators, we focus on job zone (i.e., level of preparation or job-specific training necessary for a job) and job security. Using these predictors as indicators of job status is consistent with existing scholarship on workplace inequality (Klandermans et al., 2010; van Dijk et al., 2020). Overall, these SES and job status variables are indicators of personal and job-related resources that confer status in organizational hierarchies and influence workplace outcomes. As we argue below, these status indicators are theoretically linked to perceptions of EPM acceptability because they affect workers’ ability to have a voice and access to information regarding EPM practices.
Although some studies have focused on how individual differences like personality (Zweig & Webster, 2003) and psychological reactance (Yost et al., 2019) influence reactions to monitoring, to our knowledge, still no research has directly studied SES or job status factors as they relate to the experience or perceptions of EPM. However, SES and job status factors are implicated in conceptual understandings of EPM that place it in the broader context of organizational power hierarchies (Ball, 2010; Ball & Wilson, 2000). According to this approach, EPM is a mechanism that may be implemented by those with higher status in an organization with little say or control given to workers with less status. This may result in lower status workers experiencing more EPM. Moreover, even if the amount of EPM experienced by workers of all status levels is similar, this perspective suggests that EPM is likely to be perceived more negatively for workers of lower status.
We propose two explanatory links between worker status and their perceptions of EPM acceptability: employee voice and access to information. In organizations, the extent to which individuals have a voice to be able to influence organizational policies and procedures such as EPM is partly a function of their status (Morrison, 2011). Specifically, workers are often reluctant to voice opinions and concerns due to the belief that they have limited resources with which to control their personal job outcomes and broader organizational outcomes. Being of lower status (e.g., lower income, education, job security, and/or job zone) contributes to this belief, and thus contributes to workers feeling that they have little power with which to voice opinions and concerns (Morrison & Rothman, 2009). For instance, factors such as an individual’s position in the organization (e.g., job zone, tenure) and expertise (e.g., education) will affect their perceived ability to voice input in policy decisions. Applying this idea to EPM, it can be inferred that workers with lower status will have less power to influence what technologies are used to monitor their behavior, what behavior is monitored, and how their data is used. In other words, EPM should result in less autonomy for lower status workers compared to higher status workers. Based on this reasoning, we expect that workers with lower status tend to perceive EPM as less acceptable and thus report more EPM-U.
Workers’ status may also affect how informed they are about organizational policies and procedures including EPM. Organizations are political in nature; members seek to gain advantages in organizational outcomes for themselves or for groups to which they belong. Access to resources and information regarding decision-making processes is a significant part of these political dynamics (Drory & Romm, 1990). Higher status employees have greater personal and organizational resources, such as income, education, and job security, which render political advantages. Thus, workers’ SES and job status may determine, in part, how informed they are and whether they are provided with justifications regarding policies and procedures. Applying this reasoning to the EPM context, workers’ SES and job status may relate to the extent to which they have access to information and justifications regarding EPM practices. This expectation aligns with the understanding of EPM as a means of organizational control implemented with little input from lower status workers (Ball & Wilson, 2000). In this case, EPM can act as a form of informational control resulting in lower status workers having worse perceptions of EPM transparency. In turn, they are likely to perceive EPM as less acceptable and thus report more experience of EPM-U.
We offer two hypotheses based on the arguments in the preceding section:
Hypothesis 1: Workers with lower SES (income and education) will experience more EPM-U (i.e., perceive the EPM that they experience to be less acceptable).
Hypothesis 2: Workers with lower job status (job zone and job security) will experience more EPM-U (i.e., perceive the EPM that they experience to be less acceptable).
Additionally, to gain a more complete understanding of how workers’ status influences their experience of EPM, we explore whether status influences the reasons why workers consider EPM to be acceptable or unacceptable. In other words, do workers endorse different justifications for monitoring based on their status? Although there are a multitude of justifications that organizations provide for monitoring workers, we focus on three in particular: (a) If the monitoring is done with my consent; (b) If the organization transparently justifies the monitoring; (c) If the monitoring increases my autonomy (e.g., I can telework more). We chose to focus on these justifications because of their relevance to EPM consent, transparency, and autonomy, respectively. Based on our earlier arguments, we believe justifications focused on these issues should be relevant to worker status. Rather than offering predictions regarding endorsement of these consents, transparency, and autonomy EPM justifications, we explore potential SES and job status differences in each as research questions.Research Question 1: Does SES (income and education) influence workers’ endorsement of EPM justifications based on (a) consent, (b) transparency, and (c) autonomy? Research Question 2: Does job status (job zone and job security) influence workers’ endorsement of EPM justifications based on (a) consent, (b) transparency, and (c) autonomy?
The data were collected from 270 MTurk workers (56.1% male) in February 2021. The survey was managed by CloudResearch who contacted the participants and managed the payments. MTurk workers tend to be older, more ethnically diverse, and characterized by more work experience than undergraduate student samples. These factors make MTurk a more desirable sample when studying workers (Woo et al., 2015). No a priori power analysis was conducted. The sample size was determined based on practical consideration and an examination of existing studies of the relationship between EPM and worker attitudes.
Participants completed an online Qualtrics survey with a completion time of about 20 min. After providing informed consent, participants were presented with a list of 43 forms of EPM (see online Supplemental Materials, for the full list) and were asked two or three questions about each. First, they were asked to what extent they would accept being monitored in that way (1 = never acceptable, 2 = sometimes acceptable, or 3 = always acceptable). Participants responding “sometimes” received a follow-up question which presented a list of 10 EPM justifications that organizations might provide and asked them to indicate which ones, if provided by their organization, would make that form of EPM acceptable. The same list of justifications was provided for all forms of EPM. Third, all participants were asked to indicate whether they had experienced each form of EPM in their job (yes or no). Last, respondents searched their occupation on O*NET (https://onetonline.org) to report their O*NET occupation (Standard Occupational Classification [SOC]) code and their job title, and then completed a demographics section. All respondents received a monetary payment for participation.
We are unable to share the raw data set for this study publicly. In the institutional review board (IRB) protocol for this research, we specifically noted that only members of the research team would have access to the raw data and did not stipulate an exception for the sharing of de-identified data. Study materials and code for primary and supplementary data analyses and visualization are available at https://osf.io/vam7j/.
Our primary outcome variable, EPM-U, was derived from the two EPM acceptability and experienced EPM survey questions described above (see the Procedure section). If participants indicated that they had experienced a form of monitoring and that it was “never acceptable,” that was tallied as one instance of EPM-U. The final score was the total EPM-U ranging from 0 to 43.
We also derived a variable indicating the amount of EPM experienced by participants that they perceived to be acceptable (perceived acceptable EPM [EPM-A]) for use in supplementary analyses. This variable was calculated with the same survey questions used to derive EPM-U. If participants indicated that they had experienced a form of monitoring and that it was “always acceptable,” that was tallied as one instance of EPM-A. The final score was the total EPM-A ranging from 0 to 43.
The present study focused on three EPM justifications due to their relevance to inequality in EPM. The first justification, consent, was, “If it was done with my consent.” The second justification, transparency, was, “If the organization transparently justified it.” The third justification, autonomy, was, “If it was done to increase my autonomy (e.g., I could spend more time working from home).” The outcome of interest was the extent that a participant endorsed each justification across the forms of EPM that they said were sometimes acceptable. Participants differed in the total number of justifications provided since they were only asked about justifications when they indicated that a form of monitoring was sometimes acceptable. Thus, the outcome variable for each justification was the proportion of the number of times a participant endorsed a justification over the total forms of EPM that they said were “sometimes acceptable.” This resulted in three outcome variables representing the extent that participants endorsed the consent, transparency, and autonomy justifications as making EPM acceptable.
Participants self-reported the closest estimate of their personal (not household) annual gross income to the nearest thousand (e.g., $43,100 would be rounded to $43,000).
Participants self-selected their highest degree attained on a scale of 1–5 from the options of some high school, high school diploma, associate degree, bachelor’s degree, or graduate degree.
Since participants provided their O*NET job codes as a part of the survey, we were able to gather the job zone associated with that occupation. This variable was created by O*NET as an indicator of the occupational preparation and expertise needed to enter a job (e.g., prior job experience, on-the-job training). The values of this variable are on a scale of 1–5 for little or no preparation needed, some preparation needed, medium preparation needed, considerable preparation needed, and extensive preparation needed.
Participants self-reported the degree to which they perceived their job to be secure over the next several months on a single item. Responses were recorded on a Likert-type scale with values ranging from 1 (very insecure) to 5 (very secure).
Participants self-reported the amount of time in years (rounded to the closest year) that they occupied their current job. This variable was not initially hypothesized as a predictor of EPM-U along with the other independent variables; however, it was collected within the same survey.
Participants were asked to provide their age and sex. No other demographic information was requested.
Three participants provided jobs for which no associated O*NET code could be determined. All three were removed from the data for the primary analyses resulting in a final sample size of 267. We began by examining the demographic variables of interest for our sample (see Table 3). Participants had a mean annual gross income of $60,224 (SD = $42,928), which is above the most recently reported U.S. national average of $56,310 (Bureau of Labor Statistics, 2020), and the majority were well educated (graduate degree = 15.7%, bachelor’s degree = 50.1%, some college = 23.3%, no college = 29.9%, all had at least a high school diploma). Regarding job status, respondents reported moderate levels of job zone (M = 3.38, SD = 1.03) and high levels of job security (M = 4.17, SD = 0.91). Average tenure was 9.94 years (SD = 8.18). Regarding other demographic variables, participants were on average middle aged (M = 39.77, SD = 10.86) and evenly distributed in terms of sex (male = 56.2%). Finally, based on self-reported O*NET job codes, our sample included a wide range of occupations and fields.
Next, we examined the composition of our sample in terms of EPM-U. We found that most participants reported experiencing seemingly few EPM practices that they considered unacceptable in their jobs (see Figure 1). On average, the total EPM-U was .58 (SD = 1.09), the maximum number reported was six, and 184 participants experienced zero EPM-U. Since we computed EPM-U as a function of both experienced EPM and acceptance of EPM, we also examined these variables individually. On average, participants thought that EPM was “sometimes” acceptable (M = 1.83, SD = 0.35) and they experienced 14.6% (M = 6.31, SD = 5.05) of the 43 EPM practices (see Table 1, for most common forms of EPM experienced and Table 2 for least acceptable forms of EPM). Further, only five participants did not experience any EPM. Thus, participants were not universally accepting of EPM and for the most part experienced several forms of EPM in their jobs.
Given this understanding of our sample, we then tested our hypotheses regarding the effects of SES and job status on EPM-U. Because there was so little variance on the EPM-U variable, we recoded it as binary (0 = zero instances of EPM-U, 1 = at least one instance of EPM-U) and tested our hypotheses with logistic regression. All predictors were centered for this analysis. We regressed this outcome variable on income, education, job zone, and job security. Results showed that none of the SES or job status variables significantly predicted the odds of EPM-U (see Table 4). Although job tenure was not one of the originally hypothesized predictors, we included it in an additional analysis as an indicator of job status. We note that this analysis was exploratory since job tenure was not initially hypothesized as a predictor at the outset of this study; however, the decision to include job tenure as a predictor was made on substantive grounds at the initial stages of analysis. We found that tenure did significantly predict EPM-U (see Table 4) such that workers with less tenure had higher odds of reporting at least one instance of EPM-U. Thus, we did not find evidence that workers with lower SES have a higher likelihood of experiencing EPM-U (Hypothesis 1) but did find some evidence that workers with lower job status, namely in the form of lower job tenure, have a higher likelihood of experiencing EPM-U (Hypothesis 2).
Next, we explored our research questions regarding the different EPM justifications endorsed by participants. First, we examined the makeup of the sample in terms of the mean rate at which participants endorsed each justification (expressed as decimals). Participants endorsed the consent (M = .75, SD = .30) and transparency (M = .64, SD = .35) justifications at high rates. For the autonomy justification, participants endorsed it at a very low rate (M = .10, SD = .20). Next, we used multiple regression to test whether participants’ SES and job status predicted the rates at which each of the three justifications was endorsed. We could not calculate the rate of justification endorsement for four participants because they did not respond “sometimes acceptable” for any form of EPM and thus were not presented with any justification prompts. All four were excluded from the following analyses (n = 263). All predictor variables were centered. Findings across the three multiple regression models did not demonstrate significant effects (see the online Supplemental Materials, for regression results). Therefore, we did not find evidence that SES and job status influence the reasons why workers consider EPM to be justified.
Several supplementary analyses were conducted to address the insufficient variation with the EPM-U variable and to further explore the relationship between worker status and EPM. First, we examined the relationships between the SES and job status variables (including tenure) and EPM-A. This count variable was substantially less zero inflated and showed much more variation (M = 2.80, SD = 3.36; see Figure 2); thus, we used Poisson regression to regress the SES and job status variables on EPM-A. Predictor variables were centered. Results showed that tenure and income were significant predictors such that workers with higher incomes and tenure experienced more EPM-A (see Table 5). Although it was significant, the effect size for income was quite small (b = 0.002).
Next, we examined both EPM-U and EPM-A relative to the total amount of EPM experienced. Both EPM-U and EPM-A were transformed by dividing each by total EPM experienced resulting in continuous variables ranging from 0 to 1. This transformation changed the interpretation of each outcome from a count of EPM experienced that was perceived as unacceptable or acceptable to the proportion of EPM experienced that was perceived as unacceptable or acceptable. We could not calculate these outcome variables for five participants because they did not experience any EPM. All five were excluded from the following analyses (n = 262). Both outcomes were regressed onto the SES and job status variables (including tenure) in multiple regression models with predictor variables centered. Only tenure predicted proportion of EPM-U (see Table 6). For workers with less tenure, a higher proportion of the EPM they experienced was considered unacceptable, which matched the results of the primary analysis of EPM-U. No SES or job status variables predicted proportion of EPM-A (see Table 7).
In this study, we explored how the effects of EPM technologies may exacerbate inequality in the workplace along the dimensions of SES and job status. Our primary research question was whether lower status workers disproportionately perceive the EPM that they experience to be unacceptable. Because EPM affects workers’ privacy, autonomy over their work environment, and informational transparency, it is important to develop a thorough understanding of who is most negatively affected by these increasingly prevalent technologies.
We hypothesized that workers with lower SES (Hypothesis 1) and job status (Hypothesis 2) would report more EPM-U. Due to limited variability in our dependent variable, we first tested these hypotheses with logistic regression (i.e., whether the independent variables predicted the odds of experiencing at least one instance of EPM-U) using data from a sample of workers in a diverse set of jobs. Our analyses showed that tenure has a small but significant effect on the odds of EPM-U, but the other status indicators did not. Supplementary analyses were conducted to explore other operationalizations of perceived EPM acceptability. These analyses first showed that income and tenure significantly predicted the amount of EPM-A. The size of both effects was small. Additionally, tenure was found to predict the proportion of EPM experienced that was unacceptable but none of the SES or job status variables predicted the proportion of EPM experienced that was acceptable. Overall, we found evidence that tenure has a small but consistent effect on differences in the perceived acceptability of experienced EPM and that income has a small but significant effect specifically on EPM-A (but not proportion of EPM-A).
These findings have both theoretical and practical implications for technology and inequality at work and the implementation of EPM. First, it is heartening that for most of the SES and job status indicators that we examined, we did not find evidence that lower status workers disproportionately experience EPM-U. Organizations undoubtedly have a lot to gain by collecting data on their employees via electronic monitoring; however, it is clear that invasive and nontransparent EPM can have deleterious effects on workers (Ravid et al., 2020). The consequence of the tension between these realities is that workers may feel that they are being treated unjustly by their organization as a result of the organization pursuing its interests (Ball & Wilson, 2000). As a result of lower status workers having less voice and access to information regarding EPM practices, we speculated that these workers may disproportionately perceive the EPM that they experience as unacceptable, thus magnifying workplace inequality. The fact that our findings only consistently support this assertion for one (tenure) out of the five SES and job status indicators that we examined is potentially encouraging. That is, one possible explanation for this pattern of results is that organizations are being mindful about the negative consequences of technology-related decisions for the privacy and autonomy of all workers. Of course, there are alternative explanations for the lack of significant effects for most of the SES and job status variables, which we discuss in the Future Research and Limitations section.
Although tenure was not initially hypothesized as a predictor of EPM-U, we see it as an important indicator of job status. Being a longer tenured employee can confer status and influence in several ways including having more years of experience, obtaining seniority, and gaining social centrality within the organization. Indeed, job demands–resources theory considers tenure to be a valuable job-related resource that allows workers to effectively cope with job stress and negative events (Grandey & Cropanzano, 1999; Hobfoll et al., 2018). Longer tenured employees can potentially use these resources to gain greater voice and information transparency regarding how they are electronically monitored and, as a result, EPM is less likely to be seen as unacceptable. For workers without the advantages of tenure, autonomy and transparency regarding EPM may be harder to attain and, thus, EPM is more likely to be seen as unacceptable.
Alternatively, person–environment fit and organizational socialization mechanisms may explain why longer tenured workers report experiencing less unacceptable and more acceptable EPM. The attraction–selection–attrition model posits that the mutual attraction between an organization and a potential worker, an organization’s choice to select a worker, and a worker’s choice to remain in or leave an organization are based in part on the fit between a worker’s and an organization’s attributes (Schneider, 1987). This model supports the assertion that workers who feel that they have experienced unacceptable EPM, which in this context is a form of misfit, have two options: choose to acclimate to the organization’s EPM practices and become desensitized to them over time, or exit the organization. In either case, tenure becomes a direct proxy for the choice to remain in the organization, and presumably the choice to become more accepting of the organization’s EPM practices.
The present study also has implications for theorizing about the effects of technology in the workplace more broadly (Landers & Behrend, 2017). Our findings lend support to the theoretical position that workplace technology is neither inherently positive nor negative; rather, it depends on how the technology is applied (Cascio & Montealegre, 2016; Coovert & Thompson, 2013; Landers & Marin, 2021). Empirical support for this argument has been found in studies of EPM that document both the upsides (e.g., performance benefits) and perils (e.g., feelings of injustice) of these technologies (Alge & Hansen, 2013). The current data further supports this theoretical position by suggesting that EPM is not seen as being inherently oppressive even for those who, as we argued, seem most likely in theory to perceive it as such.
Although participants did not find EPM to be inherently oppressive, they did not indicate that it was entirely acceptable either. From a practical standpoint, then, it is advised that organizations be cautious and thoughtful when implementing EPM. This forethought should include a consideration of how monitoring workers will affect workers psychologically, especially their sense of autonomy, privacy, and transparency. Additionally, it is imperative for policymakers to develop legal protections from oppressive EPM practices when employers fail to do so (Sprague, 2017). These policies should specify the forms of EPM and general data practices that violate workers’ rights based on empirical data regarding when and for whom EPM is violative, and they should introduce legal safeguards against these practices. Such regulatory steps have been taken in Europe with the GDPR but, in the USA, there are still few regulations that protect individual rights from EPM and organizational use of personal data (Sprague, 2017).
A second aim of this study was to better understand the relationship between status and how workers feel about justifications for EPM. Neither the SES nor job status variables predicted differences in the extent to which participants endorsed the EPM justifications of consent, transparency, or autonomy. Across varying levels of status, participants near-universally endorsed consent and transparency as justifications for EPM. Consistent with previous findings (Alder et al., 2006; McNall & Roch, 2009; Wells et al., 2007; Zweig & Webster, 2002), this suggests that regardless of the power they hold in society or within an organization, people want to have some level of control over EPM by first being provided a transparent explanation of its purpose, and second by being able to use this information to consent (or not) to monitoring. This aligns with our suggestion that indicators of organizational justice, particularly procedural (Colquitt et al., 2013), are explanatory mechanisms for perceptions of EPM acceptability.
Across all status levels, most participants endorsed the autonomy justification at a low rate. It seems workers may be hesitant to surrender privacy and autonomy over their personal information and work environment in exchange for increased autonomy in other areas (e.g., schedule flexibility). It was somewhat surprising that this justification was not endorsed more given the research showing that individuals tend to weigh benefits and tradeoffs of surrendering their privacy in general, known as “privacy calculus” (Acquisti, 2009; Bhave et al., 2020). The nonsignificant relationships between endorsement of this justification and the SES and job status variables suggest that this sentiment tends to be held regardless of status. This finding further demonstrates the salience of justice in the mental calculus that all individuals engage in when choosing whether to surrender their privacy. In this case, transactional autonomy is not sufficient for workers to calculate a net positive when it comes to EPM.
A noteworthy finding of our supplementary analyses is the distinction between the EPM-U and EPM-A variables. The distributions of these variables indicate that workers in our sample were much more likely to say that the forms of EPM they experienced were acceptable rather than unacceptable. This difference suggests an asymmetry between the perception that something experienced in one’s organization is acceptable versus unacceptable. While our data indicate that workers on average feel that EPM is only sometimes acceptable, it may be that given the choice, workers are more likely to acquiesce to their organization’s EPM practices rather than to entirely reject them as unacceptable. Consistent with this notion, research on acquiescence biases shows that lower status individuals are more likely to acquiesce to providing responses that they believe are more favorable (Winkler et al., 1982), particularly when they are responding in reference to someone who they perceive to be of a higher status than themselves (Krosnick, 1999). Further, workers’ seeming reticence to declare their organization’s EPM practices unacceptable is consistent with the power disadvantage of workers relative to the organizations for which they work, as discussed in the introduction to this article. We see this idea as one that future research should examine more directly to better understand how organizational status dynamics influence workers’ experience of EPM and other organizational data practices.
Though we maintain that the aims of this study are novel and the results meaningful, interpretation of our findings should include a consideration of power and sample size. As stated in the procedure, no a priori power analysis was conducted to determine the appropriate sample size. To caveat our findings, and to provide useful information for researchers planning future studies in this domain, we conducted a post hoc power analysis (see online Supplemental Materials, for full results). At 80% power, to find a small effect size (typical of the relationship between EPM and work attitudes; Ravid et al., 2022), sample sizes of at least 300, 52, and 1,289 would be required for the logistic, Poisson, and linear multiple regressions, respectively. With this information, we are confident in our findings about the effect of income and tenure on EPM-A using the Poisson regression and encourage future research to replicate these and our other findings with a larger sample.
Another potential limitation of the primary analyses in this study is the low occurrence and variability in the EPM-U outcome variable. As we noted, one explanation for this observed distribution is that it is reflective of the population. This may be due to organizations implementing EPM in a way that is mindful of its negative consequences. An alternative explanation for our findings is that sampling error resulted in range restriction for EPM-U, thus obscuring the relationship between worker status and perceived EPM acceptability. An examination of the rate of experienced EPM in our sample does not suggest an undersampling of workers who experience EPM. Ninety-eight percent of workers in our sample experienced some form of EPM and 66% experienced at least three forms of EPM. These figures are comparable to or greater than recent estimates of the percentage of organizations that use EPM (50%–80%; Kropp, 2019). Still, we took steps to address this potential limitation with our supplementary analyses using different operationalizations of perceived EPM acceptability whose distributions were substantially less zero inflated and had more variation. The general consistency of the supplementary results with those using the EPM-U binary variable reduces concerns over this possible limitation. However, future research examining perceived EPM acceptability should consider different ways of operationalizing EPM-U or different sampling methods to ensure that this finding is not the result of insufficient variability.
Similarly, future research should explore other factors that may reveal inequality related to EPM. For example, the response to COVID-19 has resulted in an increased use of physiological and health-related EPM (e.g., temperature checks; Kilic, 2020). Given the highly personal nature of the information collected by these forms of EPM, they may be perceived as particularly unacceptable. Research should determine which workers are experiencing these forms of monitoring and what the effects are on feelings of autonomy, privacy, and transparency.
Finally, future research should examine potential moderators of the relationships between worker and job characteristics and perceived EPM acceptability. One potential moderator is the job relevance of EPM, which refers to whether the behavior targeted by monitoring is directly pertinent to job performance. The job relevance of EPM is related to perceived justice and feelings of privacy invasion (Alge, 2001) which, as we argue, are important determinants of perceived EPM acceptability. We speculate that the nature or strength of the relationships between worker and job characteristics and perceived EPM acceptability may depend on whether the EPM is seen as job relevant. For example, the negative effect of tenure on perceived EPM acceptability that we found may be stronger for those who see EPM as less job relevant and weaker or null for those who see EPM as high in job relevance.
Last, we encourage future research to explore how the use of EPM in organizations might alleviate rather than contribute to workplace inequality. EPM can result in positive outcomes for workers under certain conditions. For example, employee performance can benefit from EPM when its purpose is developmental (Karim, 2015) and it is under employee control (Douthitt & Aiello, 2001). Thus, when implemented properly, EPM can serve as a tool to facilitate fair and accurate behavioral feedback that is equally accessible to all workers when they desire the information. Such a system could have the effect of ameliorating inequality-related issues such as bias in managerial performance reviews (Levy & Williams, 2004) and unequal access to, as well as control over, information (i.e., informational justice; Thurston & McNall, 2010). Research approaching the study of EPM from this angle would contribute to a theoretical understanding of the positive effects of EPM, and workplace technology generally, and provide recommendations for how organizations can implement EPM to benefit all workers.
EPM technology has become commonplace in many organizations. Despite the popularity of EPM, workers are concerned about how their personal data are being collected and used, and many questions remain regarding the consequences of EPM. We examined whether workplace inequality is exacerbated by EPM technologies resulting in workers with lower SES and job status experiencing more EPM that they perceive to be unacceptable. We found limited evidence that EPM technologies further inequality based on worker status, which may indicate a reason to be optimistic about EPM practices in organizations. However, workers found the acceptability of most forms of EPM to be questionable and contingent on the circumstances. Thus, EPM technologies should not be thought of as inherently oppressive. Instead, organizations and policymakers should focus on implementing and justifying EPM in ways that preserve workers’ fundamental rights to autonomy, privacy, and reasonable information transparency. The responsible use of EPM is one step that can be taken to ensure that technological progress at work benefits all rather than benefiting some at the cost of others.
Copyright © The Authors 2022
Received January 1, 2021
Revision received June 2, 2022
Accepted June 10, 2022