4

Educational and Workplace Contexts

EDUCATION AND WORK AS PLACES THAT AFFECT COGNITIVE RESERVE

Eric Grodsky, Professor of Sociology and Educational Policy Studies at the University of Wisconsin–Madison, started the discussion by talking about cognitive reserve theory. Cognitive reserve can be developed early in the life course and can be built and sustained through cognitive work in later years. Education and work are important in this context, providing ways to preserve old brain pathways and build new ones. Having cognitive reserve helps to delay cognitive decline even when there are biological manifestations associated with cognitive decline.

Grodsky said it is important to look at the right levels of geographic analysis when researching how education influences cognitive function. Much of the research focuses on states (e.g., Avila et al., 2020; Soh et al., 2023b; Walsemann et al., 2022b), and some looks at districts (Moorman et al., 2019) or schools (Mantri et al., 2019; Moorman & Khani, 2024; Sisco et al., 2015; Walsemann et al., 2022a). Nevertheless, 82 percent of the variation in learning associated with teachers, schools, or districts occurs within classrooms (Chingos et al., 2015), and researchers do not have good data on what happens in classrooms. Also, based on Grodsky’s research, about 80 percent of the variation in cognitive skills around age 60 is attributable to educational opportunities and achievements that are observable by the 12th grade. Grodsky et al. (2024) said these statistics suggest where intervention in learning should appear. When asked whether state-level data were important given that policies are often made at the state level, Grodsky

agreed that states are important in setting the context but said that what happens in the classroom will have the most immediate impact.

Some educational measures associated with dementia include the average number of days kids attend at the state level, school segregation, and the demographic composition of the schools, Grodsky said. Findings are mixed regarding whether the high school curriculum is associated with cognitive function in later life. Other aspects that may be important to cognitive function include academic expectations and how much effort the students put in. At the classroom level, important factors include small class size for early grades, social belonging, and feeling supported in the classroom. There is a tremendous amount of work indicating that teachers matter, but what it is about teachers that matters is less clear. There are observational studies about how teachers dialogue with students, and additional work about how the classrooms are set up, but no clear findings.

According to Grodsky, work complexity may also matter, with apparent benefits from placing greater demands on workers’ cognitive skills or greater opportunities to further enhance those skills. Work complexity varies with education but also varies among those with the same level of education. Measuring work complexity requires making further distinctions—that is, among occupations, industries, sectors of the economy, kinds of places where people work, firms, jobs, and where work really happens.

Grodsky said two main tools for looking at work are the Dictionary of Occupational Titles, which started in 1939 and was last updated in 1991, and O*NET, which replaced it in 2000. The Dictionary of Occupational Titles provides information on the complexity of working with data, with people, and with things. O*NET has a different logic but also includes measures of complexity based on worker responses to surveys.

Generally, Grodsky said, jobs rather than occupations provide greater predictive power. Based on a paper by Deming and Kahn (2018), only 2 to 20 percent of the variance required by Burning Glass Institute data on job ads was between occupations, while the vast majority of the variance in the complexity of work is within occupations. A separate paper by Autor and Handel (2013) reported that 46 percent of the variance in managing or supervising people was between occupations, and about 49 percent of the variance with regard to routine work was between occupations. If one decomposes the log of earnings or wages into components across occupations, establishments, and jobs, occupations account for about 30 to 55 percent of the variance (Avent-Holt et al., 2020).

According to Grodsky, occupational complexity as defined by the Dictionary of Occupational Titles is associated with the level of hippocampal volume and whole-brain atrophy (Boots et al., 2015). Complexity with regard to working with people and data predicts cognitive skills at both age 77 (Andel et al., 2007) and at 70, net of IQ at age 11 and of years of

education (Smart et al., 2014). Greater complexity with regard to working with people and things reduces the risk of dementia but not of Alzheimer’s Disease, net of years of education (Kröger et al., 2008). Grodsky believes these statistics understate the relationship between work complexity and cognitive decline because they are based on occupations, not jobs. He said that within individual jobs, over time daily activities may change (such as when a professor is promoted or starts running a study), becoming more complex. Performing the same task over 25 years probably has less benefit than having varied cognitive demands.

Thus, Grodsky said, both measures of education and measures of work understate the importance of cognitive reserve by being subject to aggregation bias, while more precise measures should provide stronger statistical associations. Surveys such as the National Longitudinal Survey of Youth (NLSY), the National Longitudinal Study of Adolescent to Adult Health (Add Health), and HRS could address this by asking people about their current work (rather than their occupations) and asking what people do at work. Other surveys such as High School and Beyond and the National Longitudinal Survey of 1972 (NLS-72) can ask about work retrospectively. NLS-72 is going to do that.

Grodsky said a newly developed resource is state longitudinal data systems, something that the U.S. Department of Education incentivized in the 1990s and early 2000s.¹ These data systems could be used to develop a data infrastructure for future studies of cognitive change over the life course.

Grodsky was asked about the implications of changes in education, such as whether lessons based on educational conditions in the 1950s are relevant in the 2020s. Grodsky replied that many educational reforms do not last (citing Tyack & Cuban’s 1997 Tinkering Toward Utopia); the tools may change, but the basic technology of how one teaches does not change a lot.

THE WORKPLACE AS A SOURCE OF STRESS OR MITIGATOR AGAINST STRESS

Deborah Carr, Director of the Center of Innovation in Social Science and A&S Distinguished Professor of Sociology at Boston University, focused on the workplace as a source of stress that could affect cognitive health, while also noting that positive aspects of work such as a supportive work culture can mitigate stress. She began with an overview of cumulative stress exposure models, discussing how stress is socially patterned and perpetuates social inequalities (Figure 4-1). Social and economic disadvantage in early life places some people on a course toward further stress accumulation.

___________________

¹ https://nces.ed.gov/Programs/SLDS/

Disadvantageous early life exposures include exposure to poverty, having parents with low education or low socioeconomic status, experiencing bereavement prematurely, and exposure to systemic racism. Chronic and prolonged stress exposure has direct effects on one’s body through physiological stress exposures, which could undermine cognitive outcomes. Persons with early socioeconomic adversity also are at an elevated risk of midlife disadvantages, such as poor-quality jobs or physically stressful work.

Carr said stress proliferation may occur at the family level, since there is assortative mating in couples on the basis of socioeconomic status. One partner’s stress, such as exposure to workplace discrimination or dangerous work conditions, can spill over and cause stress for other family members. Stress crossover also can lead to physiological stress responses, directly affecting cognitive impairment. Stressful encounters and stress spillover also may trigger unhealthy coping behaviors such as smoking, unhealthy eating, or substance use, which can affect cardiovascular health and consequently cognitive health.

Carr said one particular source of workplace stress that may affect cognitive health is shift work, or schedules other than the more standard timing between 7 a.m. and 6 p.m. Such work schedules apply to roughly 20 percent of U.S. workers. Some of the behavioral factors associated with these schedules are disrupted or short sleep, higher rates of substance use, poor diet, anxiety and depression, disruptive circadian rhythm, cardiovascular risk and metabolic disorders, obesity, strained relationships at home, heightened work and family conflicts, and families that are less well-functioning because they spend less time together (Abrams et al., 2022; Boivin et al.,

2022; Rivera et al., 2020; Torquati et al., 2019). Using data from the UK Biobank, Chang et al. (2024) found that three different measures of shift work (shift work vs. non-shift work, frequency of shift work, and type of shift work) all were related to increased rates of dementia.

Workplace discrimination is another dimension of employment that may affect cognitive health, Carr said. Perceived interpersonal discrimination is typically measured in surveys using questions such as whether or how often the survey respondent was treated as not as smart or not as good as others. Surveys also capture institutional discrimination, such as not being hired or promoted on the basis of one’s age, race, gender, and so on. A robust literature links perceived discrimination to physical and mental health, and a small yet emerging literature links it to cognitive functioning (Carr & Namkung, 2021). Seldom does discrimination show up as producing a direct effect; rather, it operates through either physical health or mental health. The effects also vary across different measures of cognition. For example, a study among 3,304 older adults showed the greatest effects on executive function and processing speed but less on episodic memory (Zahodne et al., 2020).

The identities that render one vulnerable to workplace discrimination (e.g., having a disability, belonging to an ethnic or racial minority group) also are documented risk factors for cognition, consistent with cumulative risk perspectives. Perceived weight discrimination also is related to cognitive function (Sutin et al., 2020). Carr argued that workplace discrimination also includes situations such as when blue-collar workers and those from lower SES backgrounds are in jobs where aging may affect their ability to perform the physical labor required, whereas white-collar workers have largely sedentary jobs that are not physically demanding and thus age-related functional limitations do not affect their capacity to carry out their job duties.

Carr said ageism can be another source of stress and is associated with increased risk of multiple physical, mental, and cognitive health conditions (Levy et al., 2020). It also is experienced unevenly, with some studies suggesting that persons with fewer socioeconomic resources are most vulnerable to ageist treatment. Persons with lower levels of education and more physically demanding jobs experience stressors that can take a toll on their physical appearance, while also having fewer economic resources to alter their appearance to meet a youthful ideal (e.g., through teeth whitening, hair coloring). As a result, ageism is both a consequence and a perpetuator of inequality, which may bear directly and indirectly on cognitive health.

Carr said work-family overload is another type of stress. Midlife workers are grappling with elder care, childcare, and in some communities custodial grandparenthood. An AARP study estimated that nearly 70 percent of family caregivers report difficulty in balancing career and caregiving

responsibilities, with long-term impacts for the economy.² Carr commented that often the level of work-family overload is sharper for those who have greater early life adversity. Family members tend to rely on one another for instrumental and emotional support, yet family members also tend to share socioeconomic positions and the stressors that may accompany low SES. As a result, the members of one’s social network may lack the physical, economic, or emotional wherewithal to provide support. For example, Mayeda et al. (2020) found that rates of memory loss varied depending on family arrangements, with different patterns depending on whether women were mothers or nonmothers, working or nonworking, and single or married. When queried why some of Mayeda et al.’s results appeared paradoxical (e.g., that being a single working mother showed less stress than some other combinations), Carr observed that ultimately more data are needed; for example, if a married working mother does not wish to work but is forced to work due to economic need or due to a spouse’s work-limiting health condition, then that may create stress.

Carr said a related and important data issue is that data on stress are often collected at the individual level, but it is important to look at dyadic or even full family relationships to properly measure stress in one’s social networks. Extensive research documents spousal crossover when it comes to stress and mental health, yet far fewer studies consider cognitive health as an outcome. Results on spousal stress contagion tend to be asymmetrical, with women and men often not equally affected by the other partner’s stress level (Liu et al., 2024).

Another data issue described by Carr is what types of work-related stressors to consider, as well as which stressors are modifiable (Omura et al., 2022). Studies based on the National Survey of Daily Experiences and other daily diary-based data resources have found that small-scale stressors are important, such as car troubles or traffic jams driving to work. If small-scale stressors persist frequently and over a long time period, they have implications for cardiovascular health and consequently cognitive health.

Carr said the length of time on a stressful job is also important. Ideally, longitudinal studies would obtain complete job histories, including not only changes in employers but also changes within an employer. The effects of some stressors are duration dependent in complex ways. Some people experience acute short-term physical or mental health symptoms immediately after the onset of a stressor but then adapt in the longer term. Alternatively, long-term and persistent exposure to particular work stressors may severely undermine one’s health and well-being.

___________________

² https://press.aarp.org/2024-5-16-US-Workforce-Report-70-Caregivers-Difficulty-Balancing-Career-Caregiving-Responsibilities

Carr said work environments may also be changing with the growth in telework and home-based work; we do not yet have answers on how to measure stressful contexts and interactions in such situations.

This page intentionally left blank.