New Technology, Older Workers: How Workplace Technology is Associated with Indicators of Job Retention

Abstract

Middle-aged and older adults who are employed in precarious, high-strain jobs may face challenges to continued work, risking economic insecurity and poor wellbeing in retirement. Technology in the workplace, an under-studied aspect of work environments, could accommodate aging workers or could add stress to their jobs. This study examines how technology in sales and surveillance at work are related to job satisfaction and planned job exits among approximately 6,000 workers ages 50–69 employed in the low-wage service sector (e.g., retail, pharmacy, grocery, hardware, fast food, casual dining, delivery, and hotel). On-the-job surveillance was related to lower job satisfaction and higher reports of looking for a new job, especially when combined with sanctioning for slow speed of work. However, rewards for speed, and to a lesser extent the use of leaderboards, were associated with higher job satisfaction, demonstrating the potential of technology to enhance the work experience for older employees. The use of sales technologies was not associated with job satisfaction or intentions to look for a new job. These results provide a uniquely detailed portrait of prevailing labor market conditions for aging workers in the service sector and demonstrate how certain kinds of technology matter for older workers’ employment.

Introduction

The aging of the US population will increasingly strain the finances of the Social Security and Medicare programs, while limiting US labor force growth and productivity (Maestas et al., 2023; Yellen et al., 2021). Several policies, such as gradually increasing the age at which beneficiaries receive full retirement benefits, are designed to incentivize longer working lives to increase the number of years in which workers pay into these programs and to reduce the number of years in which they receive benefits (Quinn et al., 2011). However, working into old age may not be attainable for large segments of the aging workforce in America (Abrams et al., 2020; Berger et al., 2021; Martin et al., 2009).

The divide between high- and low-quality jobs in America is growing, and for many, work has become more precarious (Kalleberg, 2011; Raymo et al., 2011; Wallace & Kwak, 2017). Workers of lower socioeconomic status, women, and racial/ethnic minorities are concentrated in jobs with high psychosocial demands, low control and autonomy, and low rewards, jobs associated with poorer worker wellbeing (Growth, 2017; Huffman & Cohen, 2004; Karasek, 1979; Kristal et al., 2018; Storer et al., 2020). In the service sector, workers contend with such challenging conditions as low wages, unpredictable schedules, and little paid time off (Golden & Dickson, 2021; Schneider & Harknett, 2019; Zundl et al., 2021). Workers in this sector are often burdened and exhausted by customer-related social stressors (Dollard et al., 2003; Kim et al., 2012), which can in turn reduce job satisfaction and increase intentions to quit (Walsh, 2011). There is some evidence that older age is associated with being better able to employ strategies to reduce emotional exhaustion and cynicism in the face of customer stressors compared to younger employees (Johnson et al., 2016).

Even so, it is possible that some of the inhospitable aspects of service work may be especially detrimental to the job satisfaction and job retention of older workers if they need flexible accommodations, are more sensitive to a changing workplace, or are more loosely attached to the labor force. Previous research has shown that unpredictable schedules for older workers (ages 50–80) in the service sector are associated with psychological distress, poor sleep quality, work-family conflict, economic insecurity, job dissatisfaction, and intentions to look for a new job (Abrams et al., 2022). Low job satisfaction and intentions to quit are important outcomes to examine in low-wage older workers, because low wealth makes maintaining employment to maximize social security payments especially critical.

Alongside such established stressors as low wages, unstable schedules, and limited time off, workers in the sector are likely to be increasingly exposed to technological surveillance, sales technology, and automation. However, little is known about how technology in the workplace is experienced by the aging workforce. New sales technologies, such as self-checkout, online ordering for in store pick up, and handheld devices to place orders, might replace workers’ jobs, reshuffle customer and worker duties, or assist workers. In doing so, they often result in fewer customer interactions, which could reduce some of the emotional stress of working in the service sector. However, for workers who are enriched by customer interactions or who find learning new technology to be challenging, these changes may increase stress on the job.

A recent review of the literature on how digital technology in the workplace impacts technology-related stress in older workers (ages 50 and over) found mixed results with some studies indicating high techo-stress, some showing both positive and negative effects of technology, and others finding null results for specific aspects of working life like techno-overload and techno-insecurity (Nedeljko et al., 2023). One study compared how age modifies the way workers in the service sector experience stress and burnout from customer interactions, and they found that higher age was positively associated with employees using stress management strategies to cope with customers stressors to ward off emotional exhaustion and cynicism (Johnson et al., 2016). Therefore, the way that sales technology reduces customer interactions may present more of an advantage to young compared to older workers. At the same time, sales technology reduces workers’ control over sales interactions (Karasek, 1979). Therefore, we hypothesize that exposure to sales technology, which bring a reduction in control and an uncertain effect on job demands, will be negatively associated with job satisfaction for older workers.

Technology is also changing work in the service sector by enabling new forms of monitoring and surveillance of workers (Bernhardt et al., 2022). Workers have long been observed by managers and video surveillance, but now can also be tracked via wearable badges or bands, cash register data, handheld devices, and other means (Bernhardt et al., 2022; Van Oort, 2019). The data generated by such tracking are sometimes displayed publicly on leaderboards or used to sanction or reward workers for their speed and performance (Bernhardt et al., 2022). The term Digital Taylorism has been used to apply Fredrick Taylor’s principles of scientific management to the way that technology is used to control and standardize workers to optimize efficiency (Konuk et al., 2023). Levy and Barocas’ theory of “refractive surveillance” describes how electronic data collected about customers, such as how busy a store is at a given time, are used to control workers, for example with just-in-time scheduling practices (Levy & Barocas, 2018). These practices appear to benefit the employers’ interest in managerial control, at some cost to workers.

The resulting pressure from surveillance to increase the pace of work can directly impact workers’ physical health and safety via injury risk, as well as impact stress-related health effects (Bernhardt et al., 2022). Qualitative research has revealed how surveillance can intensify workers’ anxiety and sense of insecurity (Van Oort, 2019), but it remains unknown whether experiences differ by age. Older workers may not mind such monitoring and gamification if it gives them an opportunity to demonstrate, and be rewarded for, their high level of experience and performance. On the other hand, the loss of autonomy and dignity associated with surveillance may feel infantilizing to older workers given their experience and tenure on the job (Bernhardt et al., 2022). It is possible that the impact of surveillance depends on how the data collected is used to reward or sanction employees. Because surveillance has the potential to increase job demands by incentivizing faster work while inherently reducing worker autonomy (Karasek, 1979), we hypothesize that exposure to surveillance technology and the use of surveillance data to reward or punish employees will be negatively associated with job satisfaction for older workers.

There is limited research on what specific aspects of job quality shape job retention among older American workers in low-wage front-line occupations, despite the fact that there are over five million workers ages 55 and older employed in the service sector in the United States, representing 15% of workers in this age range (U.S. Bureau of Labor Statistics, 2021). The dearth of evidence is at least in part due to limited data on dimensions of job quality relevant to this subsector. Other national datasets that cover older adults and collect information about workplaces such as the Census, the Health and Retirement Study, and the Panel Study of Income Dynamics do not specifically sample workers in low-wage service sector occupations and do not ask respondents detailed questions about those specific work environments, such as schedule unpredictability and workplace technology. In this study, we use a uniquely targeted and rich data source to examine exposure to sales technology and surveillance in the workplace by age. Second, we examine how these technologies are associated with job satisfaction and planned job exits among older workers to determine how technologies that are becoming more common in the service sector shape the prospect of working longer among low-wage workers.

Methods

Data

We draw on survey data collected by The Shift Project, which contains detailed measures of job quality and job retention for a large sample of older workers in the United States. The Shift Project’s ongoing data collection efforts have surveyed 117,000 respondents either currently working at or recently separated from one of approximately 150 of the largest service-sector (retail, pharmacy, grocery, hardware, electronics, general merchandise, fast food, casual dining, delivery and fulfillment, hotel) employers in the United States across 13 bi-annual cross-sections from Fall 2016 through Fall 2022. Example firms represented in The Shift Project data include Walmart, Home Depot, Walgreens, Safeway, Costco, Subway, Dollar General, Cracker Barrell, Amazon, UPS, Marriott, and Hobby Lobby. We focus on the sub-sample of approximately 6,000 workers aged 50–69 who responded to new survey items regarding technology in the workplace that were fielded in the 2020 – 2022 waves of the survey and who were not missing data on model covariates.

The Shift Project data used in this study were collected using a novel sampling method that relies on targeted advertising on Facebook and Instagram to both construct the quasi-sampling frame and to field recruitment. This convenience sampling approach is described in detail and the validity assessed in a recent publication (Schneider & Harknett, 2022) and these data have been widely used to study the effects of unstable scheduling on worker health and wellbeing (Abrams et al., 2022; Harknett & Schneider, 2020; Schneider & Harknett, 2019), the effects of labor standards on job quality (Harknett et al., 2021; Harknett & Schneider, 2020; Schneider, 2020), as well as the sources of inequality in exposure to precarious work (Brick et al., 2023; Storer et al., 2020). The Shift Project data capture a large and predominantly underserved sector of the workforce. The majority (72%) of Shift Project respondents are either members of racial/ethnic minority groups or low-income (< $35,000/year). These data provide a unique opportunity to understand detailed aspects of working conditions in a specific sector of the workforce that is not often the focus of retirement-related research.

Measures

A key advantage of the Shift Project data is the detailed measurement of multiple dimensions of job quality and the measurement of emerging technology, which to our knowledge has not been captured in other large surveys on work characteristics. We measure exposure to any technology in sales (customers order online or with an app, customers use device to order in store, employee uses device to place order, self-checkout, employees use handheld device at checkout, and other sales technology). In addition, our survey measures workers’ exposure to any surveillance (speed tracked via video recordings, badge or band, manager observation, cash register data, handheld device carried, or other means). We ask workers whether this surveillance is posted publicly on leaderboards and whether it is used to sanction them for slowness (with assignment to less desirable tasks, less desirable schedules, or to being fired) or to reward them for fast work (with bonus pay, gift cards, paid time off, or other rewards). These measures were informed by prior qualitative research (Ball, 2010; Van Oort, 2019).

We make use of two key outcome variables in this cross-sectional analysis. First, we model the relationship between multiple kinds of technology in the workplace and older workers’ current level of job satisfaction, given that satisfaction has shown to be a key precursor to job transitions (Griffeth et al., 2000). Those who are very satisfied with their job (nearly half of respondents) are compared to those only somewhat satisfied, not too satisfied, or not at all satisfied. We also examine respondents’ reports that they were somewhat or very likely to look for a new job in the next three months, compared to the majority of respondents who said they were not at all likely to look for a new job.

Analytic Strategy

After characterizing exposure to these technologies by workers’ age, we test the associations between these technologies and our two outcomes of job satisfaction and likelihood of looking for a new job in older workers (ages 50–69). As depicted in Equation 1, we pool across our repeated cross-sections and estimate a series of linear probability models that include each kind of technology and surveillance (TS) together and cycle through our two dichotomous outcomes of interest (Y). In robustness checks, conclusions were consistent in logit models and in ordinal models using the full number of categories. The model controls for individual-level (X) demographic characteristics (age in years, gender, race/ethnicity, highest grade of education, marital/cohabitation status, parental status, and English-as-second language) and work characteristics (holding a second job, household income, hourly wage, usual hours, job tenure, being a manager, union coverage, workplace type, and a scale of five aspects of unpredictable schedules as previously defined (Abrams et al., 2022)). Finally, we include month (MO), year (Y), and state (ST) fixed-effects. Values for each of these covariates can be found in Table 2. Survey weights adjust sample demographics according to their distribution among hourly service sector workers in the American Community Survey.

$${\text{Y}}_i=\alpha +\beta T{S}_i+{\mathit{\Lambda }}_{\mathit{\text{k}}}{\mathit{\text{X}}}_{\mathit{\text{i}}}+ST+MO+\text{Y}+\in$$ (1)

Results

We begin with a description of “Variation in Exposure Across Age Groups” in our sample – an examination of age difference in exposure to technology in sales, surveillance, and related sanctioning/rewards. In the next section, “Technology, Surveillance, and Older Workers’ Outcomes,” we report results from models testing how exposure to technology and surveillance were associated with job satisfaction and intention to look for a new job among workers ages 50–69.

Variation in Exposure Across Age Groups

Table 1 compares exposure to technology in sales, surveillance, and related sanctioning/rewards among older workers (ages 50–69) and younger workers (ages 18–49). For most types of technology, younger workers reported statistically significantly more exposure. For example, technology in sales was reported by 91% of younger workers and 87% of older workers, and surveillance was reported by 74% of younger workers and 64% of older workers. Even so, substantial portions of older workers report technology in their workplace. For example, almost two-thirds of older workers report speed tracking at work, and 60% of older workers report any sanctioning for slowness such as being fired, given a worse task, or given a worse schedule.

Table 1.

Exposure to technology and surveillance in the workplace by age group (N= 19,129, weighted)

	Younger workers (18–49)		Older workers (50–69)		Age group differences
	Mean	SE	Mean	SE	P
Any sales technology	91.4%	0.003	87.1%	0.006	<0.001
Customer use device to order online	79.1%	0.005	77.7%	0.007	0.112
Customer use device to order in store	15.7%	0.004	12.2%	0.005	<0.001
Employee use device to place order	43.9%	0.006	34.1%	0.008	<0.001
Self check out	27.9%	0.005	39.8%	0.008	<0.001
Employees use handheld device at check out	20.4%	0.004	14.6%	0.006	<0.001
Other sales technology used	2.9%	0.002	2.9%	0.003	0.823
Any surveillance of speed	73.8%	0.005	63.9%	0.008	<0.001
Via video recordings	16.4%	0.004	11.9%	0.005	<0.001
Via badge or band	2.4%	0.002	0.5%	0.001	<0.001
Via observation by manager	46.4%	0.006	39.2%	0.008	<0.001
Via cash register data	32.4%	0.005	23.8%	0.007	<0.001
Via handheld device I carry	6.8%	0.003	5.0%	0.003	<0.001
Other	2.1%	0.002	2.9%	0.003	0.010
Leaderboard at work	42.6%	0.006	29.5%	0.008	<0.001
Any reward for speed	36.0%	0.006	28.8%	0.008	<0.001
Gift card	14.1%	0.004	8.1%	0.005	<0.001
Paid time off	6.6%	0.003	2.1%	0.003	<0.001
Bonus pay	10.2%	0.004	4.1%	0.003	<0.001
Other	5.2%	0.003	5.1%	0.004	0.764
Any sanction for slowness	79.3%	0.004	60.2%	0.008	<0.001
Fired if slow	52.1%	0.006	36.7%	0.008	<0.001
Worse task if slow	64.7%	0.005	43.8%	0.008	<0.001
Worse schedule if slow	57.1%	0.006	38.4%	0.008	<0.001

Note. Age group differences were tested with Pearson’s chi-square test.

Table 2 shows descriptive statistics among our main sample of older workers for the outcomes of job satisfaction and interest in new jobs, as well as the demographic and work-related covariates. When weighting the sample, 57% of respondents were female, average age was about 58 years old, 68% of respondents were non-Hispanic white, and 85% had lower education than a bachelor’s degree. The average hourly wage was about $15.50, 11% worked a second job, 15% were union members, and half had been in their job for six years or longer. The most common workplaces were retail stores (25%), grocery stores (24%), restaurants (14%) and big-box stores (11%). Nearly half of respondents reported being satisfied with their job, and about one third reported they would likely look for a new job in the next three months.

Table 2.

Distribution of demographic and work covariates in sample of older workers ages 50–69 (N=5,934, weighted)

Demographic covariates	Mean	SE
Gender
Male	42.8%	0.009
Female	57.2%	0.009
Age	57.97	0.08
Race/ethnicity
White, NH	67.5%	0.009
Black, NH	9.5%	0.007
Hispanic	13.1%	0.007
Other or two or more, NH	9.8%	0.006
Educational attainment
No degree or diploma earned	2.9%	0.003
High school diploma/GED	36.3%	0.008
Some college	32.3%	0.008
Associate’s degree	13.1%	0.005
Bachelor’s degree	11.5%	0.005
Master’s degree/Advanced degree	3.8%	0.003
Marital/cohabitation status
Married, living with spouse	43.7%	0.008
Living with a partner	11.1%	0.005
Not living with a spouse or partner	45.2%	0.008
Has children	75.1%	0.007
Non-English language at home	14.9%	0.007
Work covariates
Has a second job	11.4%	0.006
Hourly wage	15.49	0.11
Usual hours worked/week	34.51	0.18
Manager	17.1%	0.006
Union member	15.0%	0.006
Unpredictable schedule scale (0–5)	1.55	0.019
Tenure in job
Less than 1 year	13.0%	0.006
1 year	8.0%	0.005
2 years	8.9%	0.005
3 years	8.3%	0.005
4 years	5.8%	0.004
5 years	5.8%	0.004
6 or more years	50.2%	0.008
Household income
Less than $15,000 per year	8.4%	0.005
$15,000 to $24,999 per year	17.8%	0.006
$25,000 to $34,999 per year	19.2%	0.006
$35,000 to $49,999 per year	19.4%	0.006
$50,000 to $74,999 per year	17.2%	0.006
$75,000 to $99,999 per year	9.6%	0.005
$100,000 to $149,999 per year	6.0%	0.004
$150,000 or more per year	2.4%	0.003
Workplace
Big-box superstore	10.8%	0.005
Department store	4.5%	0.004
Retail store	24.8%	0.007
Grocery store or food market	23.9%	0.007
Restaurant	13.7%	0.006
Fast food place	8.5%	0.005
Coffee shop or cafe	1.4%	0.002
Hotel or motel	0.1%	0.001
Warehouse	1.3%	0.002
Fulfillment center	0.4%	0.001
Delivery vehicle	1.0%	0.002
Convenience store or gas station	1.7%	0.002
Drugstore or pharmacy	3.1%	0.002
Other workplace	4.7%	0.004
Outcomes
Satisfied with job	47.3%	0.008
Look for new job within 3 months	30.8%	0.008

Technology, Surveillance, and Older Workers’ Outcomes

Tables 3 contains the key results from regression models testing how exposure to technology and surveillance were associated with each outcome among older workers. We did not find strong evidence of an association between exposure to any of the sales technologies measured in the survey (such as online ordering, self-checkout, or using a device to order in the store) and job satisfaction or looking for a new job in unadjusted models or in models adjusting for other technology at work, demographic factors, work characteristics, and fixed effects. Confidence intervals for these coefficients crossed the null, indicating that these are not statistically significant associations, but also suggesting that the associations of sales technology with job satisfaction and intentions to look for a new job could be as large as −0.058 and 0.069 respectively. Therefore, we conclude that there is only minimal suggestive evidence that exposure to sales technology shapes these experiences of work for older adults.

Table 3.

Results from two multivariate linear regression models predicting each outcome by exposure to technology and surveillance among workers ages 50–69 (N=5,934, weighted)

	Coefficient	SE	p-value	95% CI
Job satisfaction
Any sales technology	−0.012	0.024	0.615	(−0.058, 0.034)
Any surveillance of speed	−0.046	0.017	0.006	(−0.079, −0.013)
Any sanction for slowness	−0.188	0.016	0.000	(−0.219, −0.156)
Any reward for speed	0.129	0.018	0.000	(0.095, 0.164)
Leaderboard at work	0.050	0.018	0.006	(0.014, 0.085)
Look for new job
Any sales technology	0.025	0.022	0.260	(−0.019, 0.069)
Any surveillance of speed	0.034	0.015	0.028	(0.004, 0.064)
Any sanction for slowness	0.094	0.015	0.000	(0.065, 0.123)
Any reward for speed	−0.071	0.016	0.000	(−0.102, −0.040)
Leaderboard at work	−0.017	0.017	0.328	(−0.051, 0.017)

Note. Model adjusted for demographic factors (gender, age, race/ethnicity, education, cohabitation, children, English at home) and work characteristics (has second job, hourly wage, usual hours, manager, union member, unpredictable schedule scale, tenure in job, household income, workplace), as well as the month and year of survey and state of residence. Full models results are shown in supplement.

In contrast, we found that exposure to surveillance was significantly associated with these job outcomes for older workers. In fully adjusted models, surveillance was negatively associated with job satisfaction (B=−0.046, 95% CI=−0.079, −0.013) and positively associated with intentions to look for a new job (B=0.034, 95% CI=0.004, 0.064). Associations were even larger for sanctions and rewards. Being sanctioned for being slow was negatively associated with job satisfaction (B=−0.188, 95%CI=−0.219, −0.156) and positively associated with intentions to look for a new job (B=0.094, 95% CI=0.065, 0.123). For example, those who reported being sanctioned for slowness at work had a predicted probability of job satisfaction of 0.40 compared to 0.59 for those who were not sanctioned for slowness (with all other factors held constant at their means). Looking at the 95% confidence interval, this difference may be as large as22 percentage points. Being rewarded for speed was positively associated with job satisfaction (B=0.129, 95% CI=0.095, 0.164) and negatively associated with intention to look for a new job (B=−0.071, 95% CI=−0.102, −0.040). Leaderboards were only significantly associated with job satisfaction, indicating that older workers had 0.05 greater probability of reporting job satisfaction when their performance was shared on a leaderboard (p=0.006). When job satisfaction was included in the model predicting intents to look for a new job, it was significantly associated and its inclusion resulted in technology and surveillance variables becoming null associations, although longitudinal data would be needed to formally test mediation (results not shown).

It is possible that the negative outcomes of exposure to surveillance that we identified above are dependent on the consequences of surveillance in terms of leading to sanctions or to rewards. To explore this possibility, we estimated interactions between surveillance and sanctioning and between surveillance and rewards for each outcome to examine if the consequences of surveillance are heightened or depressed depending on how the monitoring is used by managers. Adjusting for the same set of covariates and fixed effects, we found null interactions for job satisfaction. In the case of looking for a new job, sanctioning for slow speed modified the main effect of surveillance. Figure 1 shows that, in the absence of sanctioning, respondents who do and do not report surveillance both have a 0.25 probability of reporting intent to look for a new job. However, when that surveillance is combined with sanctioning for slowness, the probability of reporting intent to look for a new job rose to 0.37.

Figure 1.

Predicted probability of reporting intention to look for a new job in the next three months according to an interaction between surveillance and sanctioning for slowness for older workers (N=5,934, weighted)

Note. Model adjusted for demographic and work characteristics, as well as month and year of survey and state of residence. All covariates held at their mean when calculating predicted probability.

Discussion

This study used a uniquely targeted and detailed dataset of working conditions in the service sector to provide insights on an often-overlooked aspect of the aging workforce. Overall, our results reveal a nuanced story about how older workers experience technology in their workplaces, not a total resistance to these advances. In line with the mixed literature on techno-stress (Nedeljko et al., 2023), we find that the way that technology is employed determines whether it enhances or reduces job satisfaction and job retention for the aging work force in the service sector. These results have important implications for informing employer and public policies regulating the use of technology for millions of Americans working in retail, pharmacy, grocery, hardware, electronics, general merchandise, fast food, casual dining, delivery and fulfillment, and hotels.

Our first finding is that older workers (ages 50–69) reported slightly lower exposure to most types of technology in sales and all types of surveillance in the workplace compared to younger workers (ages 18–49). They were also less likely to report the presence of leaderboards and the use of sanctioning slowness or rewards for speed. These differences may be due to the industries or employers where older adults tend to work. For example, the majority of older workers in our sample worked in retail stores, grocery stores, and restaurants, whereas few older respondents worked in fast food establishments, delivery vehicles, and fulfillment centers, where certain kinds of technology and surveillance may be more common (Holland et al., 2017; Litwin et al., 2022). It is also possible that older workers select into specific employers or establishments with lower technology use, if that is indeed their preference. While rates of exposure were comparatively lower for older workers, we still observed that most older workers were exposed to some kind of technology in sales, some kind of surveillance, and some kind of sanctioning for slowness.

Technology in sales, such as ordering online, ordering on a device in the store, and self check out, was not significantly associated with lower job satisfaction or increased likelihood of looking for a new job, contradicting our expectations and undermining any assumption that older workers are resistant to or afraid of new technology in the workplace. It could be that technology in sales is not considered new or disruptive to workers. It is also possible that the null finding results from mixed experiences with sales technology, which may reduce job strain in some workers, for example by reducing interactions with difficult customers, but could also increase job strain in others who lose positive customer interactions or face difficulty with the learning curve of new technologies (Kim et al., 2012). The implication of this null finding is that there is not strong evidence that more technology in sales will relate negatively to older workers’ interest in remaining in their roles in the service sector.

As hypothesized, surveillance, which was reported by over 60% of older workers in our sample, was associated with reduced job satisfaction and increased likelihood of looking for a new job, indicating that older workers do not appreciate being monitored by their manager or technology. Considering the job demand and control model (Karasek, 1979), surveillance might intensify job demand by increasing the pressure to work quickly while also threatening the autonomy and control that older workers have earned through long tenure. We found that sanctioning for slowness, again reported by the majority of older workers in this study, reduced older workers’ job satisfaction and interest in staying in their role. In addition, our interaction model revealed that the main effect of surveillance on looking for a new job appeared only when combined with sanctioning for slowness. This finding indicates that surveillance alone may be a more neutral exposure in the case of looking for a new job, unless surveillance data is used to punish workers for their performance.

The Social Security Administration has increased the full retirement age for more recent cohorts of older adults, meaning that workers who exit the work force early (e.g. age 62 rather than 67) receive lower social security benefits for life (Quinn et al., 2011). Low-wage workers are unlikely to have pensions or retirement savings, making them especially reliant on social security payments for financial stability in old age (Ghilarducci et al., 2017). If surveillance and sanctioning in the workplace results in early labor force exits, it could translate to economic precarity and/or utilization of other government programs such as Medicaid and Social Security Disability Insurance. Even if these technologies do not result in labor force exits, they may translate to older service sector workers experiencing low job satisfaction for more years as they wait to reach the full retirement age.

At the same time, one unanticipated finding was that surveillance data has the potential to enhance worker satisfaction when used to reward workers’ performances. Receiving a gift card, paid time off, bonus pay, or another reward for speed was associated with increased job satisfaction and decreased intentions to look for a new job. These rewards were not commonly received by older workers according to our data from large retail and food firms, but they appear to hold promise for employers hoping to improve worker retention.

Another potential avenue to use technology to enhance the experience of older workers can be found in leaderboards. Posting metrics from surveillance publicly in a leaderboard was associated with increased job satisfaction. Perhaps older workers appreciate the opportunity to demonstrate their speed and proficiency on the job or find engagement and motivation in gamification, as many people do in non-work realms like tracking sleep or steps (Ilhan & Fietkiewicz, 2019; Koivisto & Malik, 2021).Future research could examine whether the implementation of rewards or leaderboards relate to changes in older workers’ job satisfaction and retention.

Unlike the European Union and other countries, the United States does not have many regulations or guidelines regarding how employers use new data-driven technologies (Bernhardt et al., 2022). The state of California is an exception, where the California Consumer Privacy Act of 2018 provides residents, including employees, with the right to transparent data collection practices (California Consumer Privacy Act (CCPA), 2018). Bernhardt and colleagues recently put forward policy recommendations, starting with a clear disclosure to workers regarding what technology is being employed, the data being collected, and the rights and protections available to workers. In addition, regulations could encourage employers to collect only necessary data about workers, to give workers access to those data, and to protect the data from misuse and privacy violations. Regulatory agencies including the Occupational Safety and Health Administration can evaluate the potential harm that new technologies pose to workers and enforce standards across workplaces. Such federal legislation to limit invasive monitoring and require transparency, like The Employee Privacy Protection Act, remains pending (House of Representatives, 2014). While these recommendations are not targeted at the aging workforce, they hold potential to increase older workers’ job satisfaction and job retention by shaping technology use in the workplace so that it can be used to enhance worker experience and minimize increased strain.

Limitations

The study’s results should be considered in light of its limitations. First, while survey weights help this national sample represent the population of service sector workers, data are drawn from a convenience sample, not a random sample. The outcomes of job dissatisfaction and intent to look for a new job may be indicators that a worker will leave their role, this study does not observe actual job exits or, importantly, labor force exits. The cross-sectional design also introduces the possibility of reverse causality, for example, if those less satisfied with their job are more likely to recall or report certain kinds of sanctioning. Like other studies on older adults, those included in our sample are selected for employment at ages 50–69, meaning we do not observe the non-trivial population who exit the labor force earlier.

Conclusion

This study used unique and high-quality data to capture an emerging area of practice and research – technology in the service sector. Older workers in food and retail establishments like Walmart, Home Depot, Walgreens, and Amazon tend to face strenuous working conditions and a lack of financial preparedness for retirement. Given social security policies incentivizing longer working lives, it is important to examine specific aspects of job quality that can be modified to improve the satisfaction and retention of older workers. We found that older workers were not entirely opposed to technology in their workplace. On-the-job surveillance was related to lower job satisfaction and higher reports of looking for a new job, especially when combined with sanctioning for slow speed of work. However, rewards for fast speed, and to a lesser extent the use of leaderboards, were associated with higher job satisfaction, demonstrating the potential of technology to enhance the work experience for older employees. Increased regulations regarding technology in the workplace, like those seen in the EU and other countries, could benefit older workers in the US going forward.

Key Points.

• Technology in the service sector is an understudied working condition that may accommodate or strain older workers.

• Sales technology (e.g. self checkout) was not associated with job satisfaction or intentions to look for a new job.

• Surveillance and sanctioning for slowness related to lower job satisfaction and higher intent to look for a new job.

• Rewards for speed and leaderboards demonstrated the potential of technology to enhance the experience for older workers.

• Regulations or guidelines could shape technology use in workplaces to enhance worker experience and minimize strain.