The Blackwell Companion to Sociology (83 page)

Social Inequality, Stress, and Health

357

According to the job strain (demand-control) model of Karasek et al. (1981),

more demanding jobs are assumed to require more effort, but not necessarily to

generate distress. Those with decision latitude in their job have greater flexibility to decide how best to meet the demands of their job and are therefore predicted

to experience little or no distress. The jobs hypothesized to cause the most

distress are those that combine high demands (workload) and low decision

latitude, and this combination of job characteristics is labeled ``high strain.''

Two Swedish studies were the first to report an association in men between job

strain and the risk of a heart attack (Karasek et al., 1981; Alfredsson et al.,

1982). A US study reported evidence from two separate national health surveys

that men in occupations that other incumbents rated as high in demands and low

in decision latitude were more likely to have had a myocardial infarction

(Karasek et al., 1988). There have now been more than twenty studies, with

the vast majority showing that those in high strain jobs have a 20±300 percent

greater risk of CV disease (reviewed in Belkic et al., 2000).

One plausible mechanism by which job strain may increase CV risk is by

gradually increasing employees' resting blood pressures (BPs), eventually leading to the development of hypertension. Using 24±hour ambulatory blood pressure

monitoring and controlling for age, body mass, and race/ethnicity, the Work Site Blood Pressure Study found that men in high strain jobs had systolic BPs that

were 6±7 mmHg higher, both at work and at home, than those in jobs with high

decision latitude or low demands (Schnall et al., 1992). This same pattern was

observed when participants were re-evaluated three years later. Particularly

interesting was the finding that those classified as having high strain jobs at

both evaluations had systolic BPs that were 11 mmHg higher than those in non-

high strain jobs on both occasions (Schnall et al., 1998).

Like the demand-control model, the effort±reward imbalance model of Sieg-

rist (Matschinger et al., 1986; Siegrist et al., 1990; Siegrist, 1996) has two

dimensions. It is hypothesized that the combination of high effort and low

reward is pathogenic. Occupational rewards include aspects of control such as

job stability and promotion prospects, as well as wages, fringe benefits, and

status consistency. The combination of high effort and low reward has been

found to predict heart attacks (Siegrist et al., 1990).

A common feature of these and several other models of stress is the import-

ance of control. Whether a demanding situation is experienced as a challenge or

a stressor may well depend on the degree of control and flexibility the individual has in determining how best to fulfill the demands. It is important to realize that control is both a personality characteristic (as in locus of control, sense of

mastery) and a situational factor. Jobs vary enormously in the degree of control incumbents have over how to perform their tasks. In the Whitehall II Study, a

second longitudinal study of a large cohort of British civil servants, low control at work was able to account for more of the SES gradient in new cases of

coronary heart disease (over five years) than any other factor, including the

traditional risk factors (Marmot et al., 1997). In this same data set, the effort±

reward model predicted coronary heart disease better than the job strain model

(Bosma et al., 1998).

358

Joseph E. Schwartz

Are the Effects of Stressors Psychologically Mediated?

Most models of stress and health tacitly assume that (a) stressors cause an

emotional response (stress), (b) stress causes physiological changes in the body, and (c) these physiological changes contribute to morbidity and mortality (pathway ACD in figure 24.4). A major implication of this model is that if one can

intervene at the second stage ± for example, by teaching individuals stress

management skills ± one should be able to reduce the health consequences of a

stressful environment. A less obvious implication is that if, for some reason, an individual does not become stressed by a situation that others find stressful, then he or she should not exhibit a physiological response. However, it is worth

considering the possibility that environmental stressors can directly impact

physiology (pathway BD in figure 24.4; see LeDoux, 1996), and that part of

the association between emotions and physiology is spurious. This is consistent

with the finding of Feldman et al. (2000) that emotional responses to acute

stressors were not substantially correlated with cardiovascular (blood pressure

and heart rate) responses. Perhaps job strain/stress is bad even for those who

do not experience their jobs as stressful. If so, it would probably be more

effective to alter work environments than to alter how individuals respond to

these environments.

Animal Models of Stress

Sociologists interested in the role of stress on health should not ignore the

findings from well controlled experimental studies of animals. Many animal

researchers have used electric shock as a stressor. In one experiment, paired

rats were exposed to shocks (Weiss, 1972). One member of each pair had no

control over the situation, whereas the other could avoid the shock by pressing

its nose to a panel. The rats with no control had higher levels of circulating

cortisol and developed more gastric lesions than those rats who could avoid

the shocks. Using this same stress paradigm, exposure to unavoidable shock

Figure 24.4 Two models of the linkage from psychosocial stressors to morbidity and mortality.

Social Inequality, Stress, and Health

359

has been shown to suppress several parameters of the immune system and

promote tumor growth, leading investigators to conclude that lack of control,

or helplessness, alters the physiological response to a stressor (see review by

Shavit, 1991).

In an extensive program of research on cynomolgus monkeys, Kaplan, Man-

uck, and collaborators have investigated the impact of social stability, experi-

mentally manipulated, on dominant and non-dominant monkeys (Kaplan et al.,

1982, 1991). In the socially stable condition, three sets of five male monkeys

lived together for 22 months. In the unstable condition, 15 males were rotated

among three cages 14 times during the 22 months. Much more overt conflict was

observed among monkeys in the unstable condition, because the dominance

hierarchy had to be re-established every 4±12 weeks. At the end of the 22

months, the coronary arteries of the dominant monkeys in the unstable condi-

tion were found to have substantially greater atherosclerosis than those of the

dominant animals in the socially stable condition and the subordinate animals in either stability condition.

Animal research can also be used to identify factors that protect against the

effects of stress. In a fascinating program of research at McGill University,

Meaney and his colleagues have demonstrated that the amount of licking and

grooming behavior of a mother mouse toward her newborn pups permanently

alters brain physiology. Those pups who experience more licking and grooming

(LG‡) during the first three weeks of life develop a faster-acting negative feedback mechanism (relative to pups receiving less licking and grooming), causing

plasma levels of the stress hormone cortisol to return to normal more quickly

following the termination of a stressor. As adults, the LG‡ mice are less fearful in new environments, are less reactive to a loud noise, and exhibit less hippocampal neuron loss and fewer spatial memory deficits at later ages (Meaney et

al., 1991; Liu et al., 1997). Subsequent research has shown that this effect varies across genetic strains of mice, such that maternal licking and grooming (LG)

matters more in genetically vulnerable strains, and matters very little in genetically hardy strains (Anisman et al., 1998).

Most recently, Meaney's group has conducted an elegant series of studies

documenting a high degree of consistency in the LG behavior of mothers from

one litter to the next and a high degree of intergenerational transmission of LG

behavior (i.e. a high correlation between the LG behavior of mothers and that of their grown daughters). They then demonstrated that by having the daughter of

a low-LG mother be raised by a high-LG mother, the daughter would grow up to

become a high-LG mother herself and her daughters would also become high-LG

mothers (Francis and Meaney, 1999). Similarly, through a manipulation involv-

ing handling of the pups, the researchers were able to make the naturally low-LG

mothers engage in high levels of LG, which was also transmitted intergenera-

tionally. These studies document that early social environment has profound and

permanent effects on a mouse's stress physiology, that effects that might well be assumed to be genetically transmitted are in fact socially/behaviorally transmitted across generations, and that a psychosocial intervention can foster

health-promoting behavior in mothers which is then transmitted to future

360

Joseph E. Schwartz

generations. At this point, we can only speculate about possible parallels with

parenting behavior towards newborn human infants.

Collectively, these and other studies have begun to document a variety of

physiological pathways by which stress(ors) may promote the development of,

or alter one's vulnerability to, disease. Current research is investigating the

contribution of these pathways to observed associations between chronic stress

exposure and illness in humans (for example, the association of work stress with blood pressure and cardiovascular disease). At a more general level, there is

increased interest in trying to estimate the extent to which stress and other

psychosocial factors can explain the association of socioeconomic status with

morbidity and mortality.

25

Two Research Traditions in the

Sociology of Education

Maureen T. Hallinan

The field of sociology of education is primarily a product of twentieth-century

scholarship. The roots of the discipline may be traced to the late nineteenth- and early twentieth-century writings of Weber, Durkheim, Waller, and Parsons.

Building on the theoretical foundation laid by these sociologists, a rich and

comprehensive body of knowledge has accumulated over the past century

about the institution of education and its role in society.

In this chapter, I illustrate research in the sociology of education by examining two research traditions in which systematic scholarship has accumulated over

the past several decades. For each area I discuss theoretical and empirical studies, including both basic and applied work. I highlight these two traditions because

they include conceptually rich and analytically rigorous research and because

they encompass studies that are particularly relevant for contemporary Ameri-

can education. The areas are the organizational analysis of schools and the

transition from school to work.

The Organizational Analysis

Analysis of Schools

Conceptual Models of School Organization

Bidwell's (1965) seminal work on the school as a formal organization laid the

foundation for a large body of research by sociologists of education on

the organizational analysis of schools. Relying on Weber's (1946) description

of a bureaucracy, Bidwell depicted the school as possessing characteristics of a bureaucracy, including a functional division of labor, staff roles defined as

offices, a hierarchical arrangement of positions, and a set of rules of procedure.

At the same time, Bidwell emphasized that schools, in contrast to other

362

Maureen T. Hallinan

organizations, possessed à`structural looseness'' in the relationships among

their various components.

Early studies of school effects were based on an organizational perspective on

schools. Coleman et al.'s (1966) landmark study on equality of educational

opportunity estimated an input±output model of schools in which variation in

school resources and student characteristics predicted student achievement. This input±output model failed to specify the mechanisms that linked organizational

resources to student outputs and hence became known as thè`black box''

model. While research in this genre accomplished little in terms of explaining

the processes that affect learning, it did contribute important insights into

schooling. Not the least of these insights was demonstrated in the Coleman

Report, namely that family background plays a predominant role in student

achievement and must be considered along with the resources of schools when

studying school effects.

In an effort to specify more precisely how schools affect student achievement,

researchers in the late 1970s and 1980s focused on the processes that produce

learning. A recognition of the importance of the context of learning led to the

formulation of à`nested layers'' model of school organization. From this per-

spective, the outputs at one level of the organization become the inputs to

another level. For example, Gamoran and Dreeben (1986) discussed how

resource allocation affected teaching. They described the way school officials