Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
8 Job Design and Organ~zaffonal Vanables INTRODUCTION Jobs in which VDTs are used are not purely "VDT jobs," even when a VDT dominates everything else about the job. A total job is defined less by the equipment used than by the outcomes to be achieved, the methods and procedures to be followed, the skills and abilities demanded, and the general set of circumstances under which the work is done. Jobs can be carefully designed to make a work experience satisfying and productive, but they usually are not; most jobs develop with little real planning, and any planning that does take place is more likely to consider the equipment than the person who uses it. If equipment is poorly designed, constructed, or maintained, a worker is not likely to be either productive or satisfied; in the extreme example of unsafe equipment, a worker's complaints are likely to anticipate health and safety hazards. Even equipment that is poor but not physi- cally dangerous can negatively affect mental health or the sense of well-being and, perhaps, job performance. Workers' Complaints and Job Structures VDTs, like any other work technology, can be used properly or improperly. The research evidence examined in previous chapters has indicated little likelihood that VDT use involves serious health or safety hazards, but there is evidence that some jobs in which VDTs are central parts of the work equipment are associated with many kinds of worker complaints--including (but not only) com- plaints related to vision. On some jobs, including some where visual contact with a display unit is relatively intensive, com- plaints seem rare. Although it may be that the number of com- plaints is related to the quality of VDT equipment, there has not been any research testing that hypothesis. In the absence of such 173
174 research showing a substantial difference across equipment in performance decrements on a standard task, for example, it seems more useful to consider whether the design of a job as a whole is associated with the variations in complaints. Perhaps some aspects of jobs are associated with VDT technology but are not necessary concomitants of VDT use: an example would be the availability of immediate feedback (from either the work or the supervisor or both) about productivity. Perhaps other aspects of the job, such as low status, close supervision, or absence of oppor- tunity to plan work sequences, lead to dissatisfaction and health problems independently of whether there is a VDT involved in a job. We note that complaints are less likely to be reported by workers in "good" jobs--those in which a worker is well paid, has had substantial training or education, and has varied kinds and levels of responsibility (National Institute for Occupational Safety and Health, 1981; Smith et al., 1981~. Rather, most of the com- plaints are reported by workers in jobs in which a single task (e.g., data entry) dominates the workday, the pay is relatively low, and a worker's responsibility is almost entirely limited to working continually and avoiding errors. Such jobs are not ordinarily described as good; indeed they can be seen as deadly jobs in the sense that they seem to stifle, if not kill, human initiative, creativity, and sense of achievement. The question that demands answering is whether people who complain and develop various symptoms--including visual or musculoskeletal discomfort, physiological changes, or dissatisfaction with the job--are responding to the equipment, the basic nature of the job, or their own perception of the job and its opportunities and limitations. Complaints exist. They must be taken seriously by researchers. Should the response simply be concentrated efforts to find causal relationships between characteristics of VDT equipment and visual complaints or problems? We think not: it seems unlikely that such an approach, by itself, will offer the greatest probability of payoff in terms of human well being. Rather, we believe that research should also consider total job design. No systematic research program has yet been undertaken to study job design specifically for jobs in which VDTs are used. Consequently, there is no adequate body of knowledge concerning the psychosocial aspects of VDT work and the resulting mental and physical strains (see Table 2.1 in Chapter 2 for an evaluation of existing studies). Studies have not examined the wide array of psychosocial stressors that can influence how VDTs are used and how their use may influence workers' wellbeing. However, there has been research that shows the kinds of psychosocial stressors that affect employee well-being in other kinds of situations.
175 Work on psychosocial stressors in jobs, and particularly work that emphasizes the desirability of a good fit of the character- istics of the job and the characteristics of the worker, seems ver y much worth considering in evaluating complaints about VDTs. Accordingly, this chapter identifies the kinds of psychosocial characteristics of jobs that have been associated with employee well-being in other kinds of work, offers what might be called a pretheoretical point of view about some of the relationships among these characteristics, and indicates how such characte r- istics can be considered in designing jobs and research in order to eliminate or reduce workers' complaints and symptoms of physical and mental problems. Defining Psychosocial Stress and Strain Strains refer to complaints and include negative affective states or changes in attitude, performance decrements, and poor bio- logical functioning. Strains often result from stressors, including stressors on the job. Most of the literature on the psychogenic aspects of work that effect strains focuses on emotional states and physiological responses (e.g., heart rate, cholesterol level, and blood pressure). Strains are usually not diseases themselves, but may be risk factors for mental or physical illnesses. In many cases, a link between a condition of the work environment and those risk factors may have health consequences that are poorly understood. For example, is one's life really shortened if one has a heated argument with a coworker that raises one's blood pressure for a few minutes? Is an elevated heart rate due to anger unhealthy but an elevated heart rate due to working on a complex arithmetic problem not (or less) unhealthy? As for the health consequences of emotional reactions to work, although one does not expect work life to be euphoric, workers probably prefer to experience satisfaction rather than dissatisfaction with their job and a sense of psychological well-being rather than feelings of anxiety, depression, and anger. The reader should keep this discussion in mind when evaluating the review of literature that follows. Objective job stress (or stressors) refers to conditions of the work environment that represent some demand (such as, workload on a person's abilities or represent some deficit of supplies or resources (such as salary, praise, opportunities to participate) to meet some need of the employee. Elective job stressors refers to an employee's perceptions and reports of these environmental conditions. This chapter focuses on psychosocial rather than physical stressors. Physical stressors refers to the direct physical effects
176 of physical forces (mechanical, heat, light, and chemical, for example) on psychological and physiological well-being; psycho- social stressors refers to the symbolic meaning of actions and of conditions of the environment. Psychosocial stressors may come from several sources, including the design of the job (such as pacing, control over tasks, sequencing, quantity, and nature of tasks), the psychosocial environment of work (such as the degree of supportive relations with superiors, peers, and subordinates), and the broader organizational system (such as the reward and influence structures and the isolation of the person from fiscal and political turbulence in the larger environment of which the organization is a part). FRAMEWORK FOR STUDYING PSYCHOSOCIAL STRESSORS IN VDT WORK Person-Environment Fit We find it useful to think of the misfit between the character- istics of a worker and of the work environment as a potential precursor of psychological and physiological strain. This concept of misfit is used here as a heuristic device for evaluating the extent to which VDT technology, the social environments in which VDTs are used, and the nature of the human operator contribute to healt~related outcomes. The basic hypothesis is that the goodness of fit between characteristics of a person (P) and the environment (E) can be thought of as a predictor of well-being (see, e.g., French and Kahn, 1962; Levi, 1972; French et al., 1974) and performance (McGrath, 1976~. By considering both a VDT operator and the work environment, one can start to explain situations in which one worker is overwhelmed by a task that does not overwhelm another person, as well as situations that are likely to overwhelm or not to overwhelm most workers. It is noteworthy that with one exception (Johansson and Aronsson, 1980), none of the studies of psychosocial stressors in VDT work (summarized in Table 2.1) even mentions the possible contribution of personal characteristics to health outcomes. We can think about two kinds of goodness of P-E fit in work situations (French, et al., 1974~. The first is the fit between the person's needs (or preferences, desires, values, etc.) and the related supplies for these needs in the job environment. A variety of needs might be thought about in terms of person-environment fit, such as needs for cognitive stimulation, for social support or social interaction, for structure and clarity in task definition and task demands, for esteem, and for achievement. Many of these needs are discussed in more detail below with regard to VDT work.
177 A second type of fit is concerned with the match between demands in the job environment and the relevant person's abilities to meet those demands. For example, one can think of the fit between the demands of a VDT job for accuracy and speed and a person's ability to do such work with accuracy and speed. These two kinds of fit, need~supplies and demands-abilities, may not always represent two mutually exclusive classifications. Excessive job demands, for example, may threaten a worker who has strong security needs as much as one who has inadequate abilities. Nevertheless, the distinction is used as a reminder that dimensions of fit must be considered both from the viewpoint of an employee's characteristics and motives and from the viewpoint of a job's demands. Objective and Subjective Fit The concept of P-E fit also distinguishes between objective and subjective fit. Objective P and objective E are operationally defined as P and E measured free of any bias in self-reporting. Subjective P and E refer to the perceptions of these parameters by the person (employee); consequently, they are measurable only by self-reporting. Subjective fit is expected to be related to objective fit, but the relationship can be less than perfect. The relationship can be weakened by a personts subjective distortions (e.g., defense mechanisms such as denial and repression) or by lack of information about the objective state of P or E because such information may be hard to provide or intentionally withheld. Although objective misfit may have to be changed in order to improve the objective conditions of work, the pathway by which the reduction of such misfit reduces psychogenic ill-being is hypothesized to be partly and significantly through a worker's perceptions of the misfit. This hypothesis is supported by the work of Kraut (1965), French and Caplan (1972), and Frankem haeuser (1980) and is an important point in both theory and application. If a well-intended change in work conditions is not perceived as such, an objective improvement in work conditions will unintentionally generate strain. For example, the introduction of VDTs to make a work environment quieter and to make text editing more pleasant may be negatively perceived by employees because it generates a misfit with regard to their technological word-processing skills. Rather than feeling more satisfied, the employees may feel just the opposite, at least until their objective skills change. To take another example, a VDT operator may perceive good fit with the job ("I key in 300 characters per minute; the job only requires that I key in 250"), when the objective fit is poor (the real typing speed of the person is only 250; and the job
178 actually requires a speed of 400, but this has not been pointed out to the VDT operator). In terms of person~nvironment fit, the two characteristics of a person, needs and abilities, along with the two characteristics of the environment, supplies and demands, combine to determine how the ways in which VDTs are used in the workplace affect workers' wellbeing. ~ STRESSORS FOR STUDY IN VDT WORK Given the paucity and inconclusiveness of the psychosocial literature on VDTs, one can only suggest possible stressors for study. The stressors discussed below were selected primarily because of circumstantial evidence or because other research on psychosocial stressors has suggested that they may have important and general effects on workers' well-being. Control The amount of control a worker has may have an important influence on stress in VDT-related work. Control can be exercised over the onset of a specific stressor, its intensity, its termination, or any combination of these (Cohen, 1980~. In studying VDT use, researchers might examine the extent to which an operator can control (a) the introduction into the workplace of the specific hardware or software itself and the alteration of his or her role in using the new technology; (b) the amount of incoming work and associated deadlines; (c) the variety of the work content (such as data entry, data acquisition, or some mix; complex as well as simple interactive tasks; etch; (d) the amount of time spent continuously at the VDT and the scheduling of such time (such as massed versus spaced periods of time), and interactions with other persons (such as telephone and in-person interruptions). It should be clear from this list, which is not exhaustive, that "control" is not a unitary concept; there are many aspects of VDT worl<~to be examined from the perspective of control. lit is not our intent to advocate the concept of P-E fit as the particular device for studying VDTs and well-being; rather, it is meant as an example and is meant to provide a framework for discussing job design. Readers interested in empirical research involving the theory may consult a number of references (e.g., Harrison, 1976; Kulka, 1979; Caplan, 1983~.
179 Opportunity to Control: E In the studies on VDT use reported in Table 2.1 in Chapter 2, some employees have complained about lack of control and others have not. Rigid pacing of a job task by a VDT and automatic monitor- ing of the rate and accuracy of task performance are among the concerns expressed by VDT workers about control (Johansson and Aronsson, 1980~. Equipment-paced work, in general, has usually been found to be more likely to produce undesirable emotional and somatic symptoms when compared with self-paced work (Caplan et al., 1975; French et al., 1982~. The research to date does not permit firm conclusions about how variations in control influence employee well-being when work involves VDT use, but research on control per se suggests that there are measurable, undesirable effects on employee well-being. Literature reviewed by Cohen (1980) suggests that lack of control may produce emotional and physiological strain as well as performance deficits and insen- sitivity to the needs of others. Many of these outcomes may, in turn, adversely affect the well-being of other employees, starting a chain reaction of strains and other stressors. For example, highly controlled jobs, such as machine-paced assembly work, can produce elevated adrenalin and noradrenalin levels (Franken- haeuser and Gardell, 1976~. Such jobs may also lead to anxiety, depression, boredom, and somatic complaints (Caplan et al., 1980; French et al., 1982~. Need for Control: P Individual differences in need for control are important. Experi- ments by Lundberg and Frankenhaeuser (1978) suggest that withdrawal of control, or low levels of control, are particularly upsetting for people with a high need for control (Burger and Cooper, 1979~. Among those with high needs for control are Type A coronary-prone people. Experiments show that in uncontrol- lable situations, Type A people have particularly high autonomic reactivity (Glass, 1977~. It is hypothesized that this reactivity may increase the risk of atherosclerosis. If so, VDT work that threatens a person's need for control could increase the risk of coronary heart disease. Reactivity to uncontrollable aspects of VDT work may vary by gender as well as by personality type. In a sample of Swedish workers, Frankenhaeuser (1980~- found that females had lower catecholamine reactivity than males to uncon- trollable stressors, suggesting that females are better able to withstand exposure to jobs involving such tasks. These data suggest that lack of control may produce strain and that there may be individual differences in strain responsivity
180 depending on the person's need for control. (The extent to which too much control can be a stressor is discussed below.) As noted, there are a variety of points at which control can be exercised in VDT work. Distinguishing among those points when assessing the role of control in VDT work may allow specific enough diagnoses of the stressors so that preventive or corrective action can be designed. Participation Participation, a concept closely related to control, refers to having a say in decisions that affect the nature of one's work. In VDT work, participation could extend to decisions regarding the introduction of VDT work into the job, the flow and nature of the work to be processed, and the nature of the VDT hardware and software. Opportunity to Participate: E Cross-cultural surveys in the United States, Italy, Yugoslavia, and Israel indicate that greater perceived control is associated with greater job satisfaction (Tannenbaum et al., 1974~. Multivariate cros~sectional analyses suggest that participation may reduce job dissatisfaction and boredom, allowing employees to influence person~nvironment fit on a number of facets of work, including task complexity, responsibility, and quantitative workload (Caplan et al., 1980; French et al., 1982~. Consequently, participation may be a useful social mechanism for reducing such strain in VDT work when misfit is present in the job. Need to Participate: P People's need to participate has not been well studied; see the discussion above regarding need for control. Predictability and Controllability Unpredictable events are by definition uncontrollable in terms of when their onset will occur, but not all uncontrollable events are unpredictable. For example, a VDT user may know in advance that a computer system serving the VDT will be down for a specified period and that the user will be unable to do anything about it. One of the special aspects of VDT work is unpredictable
181 system breakdown. In time-sharing systems, delays in access, processing, and output may be both unpredictable and uncontrol- lable in that one cannot influence when these delays will occur. Unpredictable Events: E Overall, little research has been done on the effects of unpre- dictable versus predictable stressors. Johansson and Aronsson (1980) had data on six VDT operators who, fortunately (for the study), encountered an unexpected system breakdown and showed elevated levels of adrenalin, systolic and diastolic blood pressure, heart rate, irritation, feelings of fatigue, and boredom. Labora- tory research examining the role of unpredictable noise as a stressor found that it had no effect on performance (Glass and Singer, 1972; Gardner, 1978~. Unfortunately, these studies confound unpredictability and uncontrollability. However, studies of warnings before disasters suggest that people cope better when they have warning even though they cannot control the onset and effects of the disaster (see, e.g., Janis and Mann, 1977~. Tolerance of Unpredictability: P Experiments giving people realistic job previews indicate that such previews reduce job turnover as well as increase job satisfaction (see, e.g., Ilgen and Seeley, 1974; Wanous, 1978~. Whether such previews have their effects because they make the job predictable or because of other mechanisms (such as creating feelings of trust or opportunities for better selection in and out of the job by pro- spective job applicants) is not known. Research could be done to determine the extent to which strains might be reduced by giving VDT operators and those who make demands on them realistic expectations about system reliability. By increasing people's expectations of unpredictability, one might increase their toler- ance of it. No research on this topic in VDT work has been done. As with control, predictability can be influenced at several points in the flow of VDT work. Some strategies for experimen- tally reducing strain might focus on changing P (such as increasing tolerance for unpredictability); other strategies might just as easily focus on making the design of the work environment better fit a person. For example, the flow of work to a VDT operator could be made more predictable by improving the reliability of VDT systems, carefully training those who provide work for the VDT operator, and studying workflow patterns. Even an impend- ing peak in workload might be better tolerated if a VDT operator were given plenty of early warning to allow adjustments to be
182 made in other pending work. Experiments that vary such param- eters, as well as studies that assess the predictability of different aspects of VDT-related work rather than characterize job predict- ability at a global level, are needed. Such studies may lead to knowledge about how to estimate the strain-producing contribu- tions of different parts of the workilow and how to locate points, if any, of intervention. Complexity The complexity of work performed using a VDT can vary dramati- cally, from that of a programmer or scientist attempting to create or discover new algorithms to that of a clerical worker confined to entering meaningless (to the operator) data (such as strings of invoice numbers run together on a code sheet). Studying complexity in VDT-related work is likely to be complicated. It would be ideal to define and measure complexity in terms of standard units of cognitive information load. This ideal can be met in the laboratory (Berlyne, 1958), where an investigator can carefully define '~its of information" and can carefully vary the presentation (e.g., complexity of a visual pattern) of stimuli to a viewer. In the workplace, however, the problem of comparing different types of work in terms of bits of information remains unsolved. As an example, how should a secretary doing word processing have workload defined in terms of bits of information? What constitutes a bit in information theory when comparing the work of entering first~raft text and of making revisions? If the secretary's phone rings while VDT work is being done, is the interruption included as part of the normal definition of complexity as bits of information? And so on. Recognizing this problem, social scientists have tended to define complexity in the work environment in terms of a number of stressors: (a) relatively vague definitions of the occupational role (role ambiguity; (b) changes in tasks from day to day (variety); (c) contact with people (which may involve unpredict- ability, uncontrollability, and variety); (d) time sharing rather than linear processing of multiple role responsibilities; and (e) work with several groups outside one's own immediate work group (Kohn and Schooler, 1969; Caplan et al., 1980~. A VDT job might be defined as complex with only a few of these elements present in almost any combination. For example, a mathematician at a VDT might have little contact with people (simplicity), a relatively vague definition of the job and of work from day to day (complexity), and task variety across time some days routine and some days involving a high level of problem solving (complexity).
183 These elements of complexity have not been examined in any systematic way as components of jobs involving VDTs. Some of the studies discussed in Table 2.1 in Chapter 2 compare VDT operators whose occupations look as though they might vary in level of complexity, requirements for cognitive skills and knowl- edge, and the level of complexity desired by employees. In such studies the extent to which employee (person) versus task (environ- ment) characteristics account for group differences in well-being cannot be determined because no differentiation between these characteristics has been made. The separate elements of complexity also have not been examined in VDT studies, so the effect of complexity in VDT work on well- being is unknown. Although research on complexity and well-being has not been conducted on VDT-related work, it has been done in a wide vari- ety of occupations. The studies indicate that complexity can affect employee wellbeing. In a cros~sectional survey study of over 2,000 employees in 23 different occupations ranging from unskilled work to highly skilled professions (Harrison, 1976; Caplan et al., 1980), poor P-E fit on job complexity was associated with emotional strain and somatic complaints (such as heart beating hard, sweaty palms, and trouble sleeping). It is important to note that too little as well as too much complexity produced these outcomes. When one considered only the absolute level of complexity in the job (E) or only the absolute level of desired complexity (P), the findings were not as striking. Lack of complexity--i.e., repetitive, monotonous work- has been found to be a key predictor of job dissatisfaction in a number of studies in the United States, some of which were of random samples of the U.S. workforce, and in Scandinavia (see, e.g., Walker and Guest, 1 952; Gardell, 1 97 1; Barnowe et al., 1 973~. With regard to physical health, research in Scandinavia suggests that lack of complexity in work can increase the risk of accidents, headaches and nervous disturbances, and abnormal catecholamine excretion, although the results of this study are themselves complex (Johansson et al., 1978~.2 Too much complexity, as aptly characterized by the Peter Principle (Peter and Hull, 1969), is believed to lead to feelings of incompetence, emotional strain, and poor performance. Aside from the study of 23 occupations, discussed above (Harrison, 1976; Caplan et al., 1980), very little is known about excess complexity in tasks. But it is a very important concept when studying ~ 'The interested reader is advised to consider the findings in their entirety. They deal with how different aspects of complex and noncomplex tasks may affect different physiological and emotional responses, a case of response specificity.
184 changes in work technology. The move per se from non-VDT to VDT technology may appear to future VDT operators as a significant increase in complexity even if, from the point of view of those who designed the hardware and software, the VDT will "greatly simplify things." Once an employee becomes familiar with the new technology, however, the perception of excess complexity may disappear. Anecdotal evidence (e.g., Leo, 1980) indicates that fear of computers and allied technology may occur because the machines are viewed as "too complicated." These fears seem to know no boundaries of occupation and can be found in executive as well as in line positions. We do not know how widespread such fear is and what effect it has on physical and mental health when VDTs are introduced. It would seem useful to conduct systematic research on the magnitude of such effects and the educational conditions that might overcome such fears. It is likely that these effects are secular. Younger people, increasingly exposed to VDT technology in school and recreation, will probably have fewer problems accepting VDT technology in work than will older people. The rate at which elements of a job become more complex has been referred to as complexification. Large-scale, ecological analyses suggest that as complexification increases in societies, does mental illness, suicide, and bankruptcy (Terreberry, 1968~. Toffler (1970) argues that "future shock" consists of a high rate of complexification without the resources to adapt to it. If this is the case in VDT work, then it would be valuable to monitor the rate of change in VDT software and hardware technology as a predictor of the well-being of employees. To our knowledge no such research has been done. Role Ambiguity Role ambiguity refers to the extent to which a given work role clearly defined for an employee. Cros~sectional correlational analyses indicate that too much ambiguity relative to a person's preferred level of clarity~mbiguity is associated with job dim satisfaction, depression, anxiety, and anger-irritation (Harrison, 1 978~. Role ambiguity may be pervasive in the workforce (Kahn et al., 1964), but of 33 facets of work, role ambiguity ranked in the lower one-third in its effects on job satisfaction in multivariate analyses of data from a random sample of the U.S. workforce (Barnowe et al., 1973~. Lack of challenge and lack of complexity were the - most important causes of occupational strain. Is role ambiguity an important stressor in VDT-related work? The variable has not been studied in VDT-related work. Any effects of too little
185 ambiguity in highly structured, input~riented clerical VDT work may be highly correlated with lack of task complexity and variety. There is no evidence of any association between role ambiguity, per se, and physical health outcomes (see Caplan et al., 1980~. Threat of Unemployment Some VDT research (e.g., Johansson and Aronsson, 1980) has found that VDT operators, particularly those in clerical occupations, worry about the possibility that their jobs might be eliminated in the future by a computer. How might such worry affect the well-being of these operators? We know of one study of the effect of the threat of job loss and the actual job loss on physical and mental well-being (Cobb and Kasl, 1977~; it followed employees from the time of the announced job loss to two years after termination. As evidenced by higher levels of anxiety and serum cholesterol, the anticipation period was more stressful than the time of actual termination. Lazarus (1966) reviews other literature suggesting that the threat of an unpleasant event is more strain-producing than the actual event. No research has assessed the prevalence and strength of threatened job loss perceived in various VDT jobs in relationship to physiological and psychological manifestations of strain. Conse- quently, we cannot draw conclusions about its effects on well- being in VDT work. Quantitative Workload Up to this point we have considered qualitative aspects of the work. Is it controllable? How complex is it? And so on. There is also the factor of quantitative workload, the sheer amount of work to be done in a given unit of time regardless of its qualitative nature. The pace and flow of work are important aspects of quantitative workload. A study in Sweden (Johansson and Aronsson, 1980) presents evidence suggesting that the nature of the human use of VDT technology influences the level of workload over the course of the day. This study showed that the heaviest use of computer systems occurs in the afternoon and that with increased use there are greater delays in response time by the VDT in interactive and batch modes and greater probabilities of unscheduled breakdowns in the system. Johansson and Aronsson note that VDT workers appear to cope with this pattern by increasing the pace at which they work in the morning: the increased pace in the morning is insurance against system delays in the afternoon. This self-pacing
186 appears to lead to elevated adrenalin levels in the morning. The data are not precise enough to indicate the extent to which the elevation of adrenalin precedes the anticipation of high workload or results from the high workload itself. Furthermore, the health consequences of these elevations are unknown. In a longitudinal study of a 21-day planned (i.e., anticipated, predictable) computer shutdown at a large university, there was evidence that the workload generated in anticipation of the shutdown was more likely to lead to anxiety and to elevated heart rates among people who had traits like those of the Type A coronary-prone person (Caplan and Jones, 1975~. These people have an inner sense of time urgency, a tendency to express pref- erences for competitive situations, and a preference for deadline pressures. Later work by Glass (1977) suggested that the elevated levels of anxiety and heart rate may have been part of a reaction to threatened, strong needs for control that characterize Type A persons. Such research suggests that there may be individual differences in response to workload pressures created by VDTs. It may be important to assess the extent to which excessive quantitative workload (E) signals a threat to some operators' strong needs for control but poses no threat to others for whom this need is not as salient (P). Laboratory research indicates that quantitative overload (E > P) can increase serum cholesterol levels and maintain high pulse rates (Sales, 1969) and that both too little (E < P) and too much work (E ~ P) can increase noradrenalin and adrenalin (Frankenhaeuser et al., 1971; Prankenhaeuser and Gardell, 1976~. Field studies have found similar effects of deadline pressures on the cholesterol levels of medical students facing academic exams (e.g., Horwitz and Bronte-Stewart, 1962), on tax accountants approaching the April federal income tax deadline (Friedman et al., 1958), and among professional staff at the National Aero- nautics and Space Administration, whose telephone calls and meetings were observed and tallied (French and Caplan, 1972~. There is also evidence that heavy quantitative workload may disrupt the circadian rhythm of adrenal cortisol (Caplan et al., 1979~. Such effects may occur in VDT-related work that involves poor P-E fit on workload. Multivariate analyses of stressors in data from a national sample survey of the workforce (Barnowe et al., 1973) indicate that workload, per se, ranks near the bottom of 33 facets predic- tive of job dissatisfaction. The analyses suggest that it is not hard work itself that is the key stressor on effects; rather, the culprit may be excessive workload that is also uncontrollable or carries with it misfit with regard to complexity. This analysis is con- firmed by results of a study of 23 occupations that showed the effects on psychological strain (such as depression and anxiety) of
187 P-E misfit on quantitative workload were greater for persons who also had P-E misfit on job complexity, either too little or too much complexity (Caplan et al., 1980; French et al., 1982~. Persons with the right amount of complexity in their work rela- tive to what they wanted were the least affected by large amounts of work, even when the workload exceeded their preferred levels. Overall, these findings suggest that the study of VDT operations should consider both the quantitative and qualitative nature of the workload. Prescription for OverloadDeadline Plus Delay We have discussed both delays and deadlines as potential stressors in VDT work. Together they can greatly increase quantitative overload. Deadlines create quantitative workloads by setting a time limit during which a given amount of work must be accom- plished; delays exacerbate the pressure of a deadline, reducing the amount of time for completing the job. There is some evidence of individual differences in tolerance f or delays (P). Type A persons tend to be impatient (Friedman and Rosenman, 1959~. Furthermore, Type A coronary~rone persons are described as having a sense of time urgency sense that no matter how much time there is, it is never quite enough. In studies of VDT workers, Johansson and Aronsson (1980) found that operators preferred having delays of less than 5 seconds when awaiting a system reply to a VDT input. If Type A persons are faced with delays they feel are excessive, they might be more upset than Type B persons by the delay. Research by Glass (1977) indicates that Type A people have a strong need for control. Long delays on a VDT are periods during which an operator gives up control to the machine; for Type A people, the longer the delay, the more the need for control may be threatened. It may be the loss of control, rather than the quantitative workload, that is especially strain-producing about VDT delays, but stress research on the health consequences of these delays, short of a computer breakdown, has not been reported. In general, the meaning of poor P-E fit on workload to the VDT operator may be important to assess before attempting to determine the effects of such misfit on strain. Delays induced by computer system overloading may not be as threatening to a scientist faced with only vague and perhaps self-imposed dead- lines as they are to a newspaper reporter attempting to meet a press deadline. And overload in VDT-related work may signal to some VDT users that their work is important and, consequently, that their VDT-related jobs are secure; in contrast, underload may signal a possible reduction in workforce.
188 Responsibility for Persons Some jobs involving VDTs--for example, air traffic control work also involve direct responsibility for the lives and well-being of other people. Given that fact, is it possible that the responsibility and not the VDT exposure produces strain? Might the two factors have joint effects? Although no studies of VDT work have con- sidered responsibility as a competing stressor, there is enough literature linking responsibility to ill-being to justify serious consideration. Responsibility: E Research suggests that responsibility for persons plays a role in the development of coronary heart disease (see, e.g., Russek, 1965~. Studies of foremen and people in jobs involving both responsibility for persons and role conflicts, have shown that they are particularly prone to peptic ulcer (see, e.g., Doll and Jones, 1951; Gosling, 1958; Pflanz et al., 1966~. Of particular interest is the study by Cobb and Rose (1973) of the universe of U. S. medical records of air traffic controllers and commercial pilots, all of whom entered their occupations passing the similar standardized medical examinations. The data (corrected for age effects) showed that of the two occupational groups, air traffic controllers had developed the highest incidence of high blood pressure and peptic ulcer. Furthermore, the disease rates were highest for controllers at airports with high-density, compared with low~ensity, air traffic. The data do not allow one to isolate the extent to which these occupational differences in illness represent effects caused by quantitative workload, complexity, or responsibility for persons. Responsibility: P With regard to coronary heart disease, Kasl (1978) notes that some large sample epidemiological surveys have shown that employees at higher levels of management have lower levels of coronary heart disease (e.g., Lee and Schneider, 1958; Pell and d'Alonzo, 1963~. This finding runs contrary to the responsibility hypothesis. Perhaps responsibility (E) is not the key stressor, but rather whether the amount of responsibility exceeds one's resources (P). Individual differences in preference for responsibility and in need to discharge responsibility (P) versus supplies of authority and resources to meet that need (E) have rarely been examined. One study found that too little as well as too much responsibility (P-E
189 misfit) was associated with high levels of cholesterol. Neither the amount of responsibility desired (P) nor the amount demanded (E) predicted high levels of cholesterol (Caplan, 1971; French et al., 1974~. This finding has not been replicated. Given the possibility that responsibility for persons may affect employee well-being, this aspect of work should probably be measured in studies of VDT operations in which such responsibility varies as part of the task. In this way, one could statistically control for psychosocial effects of responsibility when exploring any link between exposure to the physical aspects of VDTs and health. Role Conflict Conflicting demands from different persons in one's work does not seem like an obvious concomitant of VDT work. Should it be suspected in some VDT operations, there is literature on it from other work situations (e.g., Kahn et al., 1964; Kahn and Quinn, 1970; Miles, 1976~. Its main effects appear to be job-related tension (Kahn et al., 1964) and anger-irritation (Ca plan et al., 1980~. Social Support Social support can be defined as affirmation of someone's attitudes and beliefs, liking, trust and respect, and certain kinds of direct assistance (Katz and Kahn, 1978~. There are a variety of other definitions (see Caplan, 1979), but they generally share the above properties. Some VDT work may, by its organization, reduce social interaction and thereby reduce opportunities for social support among workers. We have already noted that, at least in theory, certain types of stressors can occur as part of the way in which \lDT-related work is organized (heavy or excessive workload, boring work that lacks adequate complexity, and so on). If the design of a job reduces chances for social support, the undesirable effects of VDT work on well-being may be magnified. Evidence suggests that loss of social support reduces well- being both directly and via its effect as a buffer between stressors and workers. Having a nonsupportive supervisor in the workplace has been associated with increased risk of mortality from coronary heart disease (Medalie, et al., 1973~. Lack of social support at work has also been associated with dissatisfaction with work and with depression (see, e.g., Caplan et al., 1980~. When stressors such as those that have been described above are present in the work environment (heavy workload, excessive
190 responsibility, too little control, etc.), social support may buffer their effects on well-being. Buffering refers to interaction: social support reduces the relationship between the stressor and health consequence. Buffering has been found with regard to a number of healt~related outcomes (see reviews by Cobb, 1976; House, 1981~. These outcomes include serum cholesterol elevations among employees losing their jobs (Cobb and Kasl, 1977; Gore, 1978), escapist drinking among a national random sample of employees experiencing various levels of occupational stressors,3 emotional strain in a sample of 23 occupations ranging from blue to white collar (LaRocco et al., 1980), and ulcer symptoms among factory workers (House and Wells, 1978~. In other studies (such as Pinneau, 1975; Frydman, 1981), however, social support has not buffered the effects of stressors on strain. Although the conditions under which buffering does and does not occur are not understood, the evidence of both the direct and buffering effects of social support on well-being suggests that it is an important health-related predictors In VDT-related work, social support takes the form of supervision as well as relation- ships with peers and with any clients who interact with the VDT user. Accordingly, studies of VDT-related work may encounter healt~related outcomes that are partly affected by the social support available to the VDT user. By studying that support, it would be possible to determine the extent to which DOT work generates certain types of interpersonal frictions and generates needs for social support as a buffer. It would be possible to determine the extent to which any lack of social support is not unique to the VDT work, but is part of the larger organizational climate of the worksite. When studying the role of social support in VDT-related work as a buffer of strain or, when low, as a stressor, it may be useful to view supervisory support in terms of two major factors that have been repeatedly identified in leadership studies. These factors can be generally described as task and process orientation (see, e.g., Bales, 1950; Bowers and Seashore, 1966; House and Mitchell, 1975~. Task orientation, sometimes called initiating structure, deals with supervision that provides information to a subordinate about the definition and execution of the work role. Process orientation, sometimes called consideration, refers to the provision to the subordinate of mutual trust, respect, considera- tion for feelings, and general rapport (Stodgill, 1974~. A review of the literature (House and Mitchell, 1975) indicates that employees Robert P. Quinn, associate research scientist, Institute for Social Research, University of Michigan, personal communication, 1973.
191 in some jobs want task leadership more than process leadership, while in other jobs the reverse may be true, and in still other jobs some more equal mix may be desired. Following House's theory (House and Mitchell, 1975), in particularly noncomplex work, such as that involving highly repetitive VDT input of data, employees may experience more satisfaction with the job if the supervisor is process oriented, because there is little in the way of task infor- mation and structure that is required (or perhaps that can be changed). The best a supervisor can do under such circumstances is to provide a pleasant and agreeable environment as compensa- tion for the potential stress of an uninteresting job. In highly professional VDT-related work, such as the work of scientists and engineers, however, VDT operators may want task orientation-- that is, technical information. In such jobs, satisfaction derives primarily from structural and informational guidance and from the intrinsic nature of the job, not from leadership that is aimed at creating a pleasant social environment. Thus, the extent to which a style of supervision leads to VDT operator satisfaction may depend on whether that style meets the needs of the operator, which probably vary by type of VDT work. Person-environment fit with regard to social support, rather than either amount of support (E) or amount desired (P) may be an important predictor of strain. Is there anything about the physical parameters of VDT work that might influence social support? In highly computerized environments, it is possible (although it need not be) for VDT employees to be denied the social support that can derive from interaction with coworkers and others. For example, in the interest of reducing glare, cubicles may be erected that block of f vision in all directions except to the screen. In highly computer- ized operations, even some forms of supervisory feedback, particularly negative feedback, may come to the employee in the form of computer-generated messages, sometimes called error messages (such communication, even if containing words of praise, does not constitute true social support because it lacks the element of human response). Such situations and the extent to which a task is highly confined to a terminal itself when deadlines and other pressures arise may reduce opportunities for social support and may deny people the opportunity to satisfy basic needs for social interaction (Murray, 1938~. There has been no systematic study of the extent to which different types of VDT work have an effect on social support. Given the key role of social support in the literature on psychosocial stressors and well-being, it may be worthwhile to investigate whether VDT work in some conditions creates a new form of sociotechnical system (Emery and Trist, 1960) with important effects on employee wellbeing.
192 DISCUSSION AND CONCLUSIONS The VDT, like the typewriter, is a technology that can be used in ways that abuse or augment the human resources of society. In this chapter we have attempted to indicate the types of stressors that might be associated with VDT work when the work is not organized with the well-being of the user in mind. In our opinion, most, if not all, of the stressors we have reviewed (unlike the physical properties of the VDT itself) are not inherent to VDT technology and software but depend on how the VDT work is structured. Therefore, there is a great deal of freedom to make work at a VDT as pleasurable or as painful as work at a desk with a typewriter. If there are health risks inherent in VDT work that derive from psychosocial stressors, there is no compelling evidence in the literature. As reviewed earlier, the VDT literature on psychosocial stressors is, with a few exceptions, inconclusive because the designs of the studies have not allowed conclusiveness. VDT use occurs in jobs that vary greatly in their complexity, responsibility for others, quantitative workload, control over the work pace, and so on. Consequently, it would be accurate to conclude that there is no such thing as VDT work if it is defined as a particular occupational title or condition. The concept of VDT work is relatively meaningless for use in studies of occupational health because the concept refers to too diverse a set of con- ditions. Systematic studies of the relative contributions of each psychosocial stressor as a component of the context in which the VDT is used would provide a much greater increase in knowledge about the psychogenic health effects of VDT use than would studies that compare a VDT group and a non-VDT group, as has been done in some designs. If more detailed evaluations of psychogenic factors in VDT work and well-being are carried out, P-E fit theory may provide a useful conceptual device for systematically examining the relationship of the VDT user and the work environment to user well-being. The concept of P-E fit can help generate the options for improving fit between a person and a job and can involve both changes in the person (selection, training, and so on) and in the environment (managerial and coworker increase in resources, workflow, variety, complexity, increased opportunity for par- ticipation, and so on). Once areas of misfit are identified, one can turn to choices between how much of P and how much of E should be changed. Some of the strategies for reducing stressors have been discussed elsewhere (e.g., French and Caplan, 1972; Caplan et al., 1980; French et al., 1982) as well as the strategies for introducing change (e.g., Bowers and Hauser, 1977; Katz and Kahn, 1978~. The selection of P or E (or a mix) as the target for
193 lmprov~g P-E fit in VO1 work in which misfit exists should neces=dly consider both the acceptance of the strategy by me VOl user as well as the costs of the strategy in terms of human welshing. Acceptance is relatively easy to assess assessing costs may be quite another matter. But it is that assessment that is at the heart of research on psychosocial stressors and how they affect employee wellbeing. ~ is that assessment that largely remains to be done with regard to psychosocial aspects of VDT use.
