Friday, March 12, 1999
12:30 pm - 2:00 pm
The Effects of Shiftwork on Health in the Older Worker
Carlla S. Smith, Ph.D., Bowling Green State University (USA), Chet Robie, Ph.D., University of Houston (USA), Simon Folkard, Ph.D., University College of Swansea (UK), Jane Barton, Ph.D., University of Sheffield (UK), Evelien Spelten, Ph.D., Academic Medical Center (Holland), Peter Totterdell, Ph.D., Lawrence Smith, Ph.D., University of Sheffield (UK), Giovanni Costa, M.D., University of Verona (IT)
Although the evidence to date strongly suggests a causal relationship between shiftwork and the development of health problems, researchers have not reached consensus about the specific processes by which the shiftwork-illness relation develops. One model that has received considerable empirical support is the process model of shiftwork and health proposed by Smith et al., (1998). This model proposes that the process by which shiftwork influences health is mediated by various organizational and personal factors, coping behaviors, and proximal or short-term outcomes.
The applicability of Smith et al.’s model to older workers is unknown. This is a particularly important omission because the negative effects of shiftwork are thought to be exacerbated with age. Based upon the results of Smith et al., (1998) and prior research on older workers, we developed hypotheses on how older shiftworkers’ individual differences, sleep and family disruptions, coping behaviors, and proximal outcomes affect their health.
Three samples were used to test the study hypotheses. Due to sample size considerations and few guidelines about what age range defines the older worker, we included any worker over the age of 35 in our analyses. These data were drawn from the larger data sets used in Smith et al., (1998).
The hypothesized relationships were tested within a modified path model. The overall model fit was good across the three samples. The individual path coefficients for the three over-35 samples were generally as strong as or stronger than the estimates for the total (full) samples reported in Smith et al., (1998). Specifically, across all three samples, older shiftworkers with inflexible sleeping habits experienced increased sleep disturbances. These sleep disturbances provoked the use of disengagement (avoidant or passive) coping behaviors and feelings of fatigue. Fatigue and somatic anxiety were both independent predictors of strain, either in the form of cardiovascular or digestive symptoms. Several sample-specific effects were also noted. Given the magnitude of these effects and the graying of the workforce, this research represents one initial step towards understanding the nature and extent of shiftwork’s effects on older workers’ health.
CORRESPONDING AUTHOR: Carlla S. Smith, Ph.D., Department of Psychology, Bowling Green State University, Bowling Green, OH 43404, USA
Women, Bathrooms, and Work
Ingrid Nygaard, MD, University of Iowa College of Medicine, Marc Linder, PhD, University of Iowa College of Law
It is commonly accepted that work may negatively impact hand, neck, or back muscles. Little attention has been paid to the muscle responsible for urine storage and subsequent elimination—the bladder. Over three decades ago, Jack Lapides, a prominent urologist, convinced of a causal relationship between voiding infrequently and urinary tract infections, opined, "Our social mores and attitudes have conspired to encourage an infrequent voiding pattern...Employment practices encourage urine retention..." Yet today, bathroom "privileges" are often dictated by employers, such that workers must bear the cost of delaying voiding.
We undertook a descriptive questionnaire study to acquire more information about issues related to voiding in the workplace in a large group of public school teachers. We assessed whether certain behavioral modifications reported by patients, voiding infrequently at work and restricting hydration at work, predispose women to urinary tract infections. The surveys contained questions about voiding habits at work, allotted breaks, bladder complaints, including urinary tract infections and incontinence, and bathroom access issues related to menstruation. Seven hundred ninety-one of 1492 teachers mailed back a completed questionnaire (response rate 53%). The mean age of the group was 43.1 + 10.0 years (range 22-73). Seventy-nine percent of teachers reported a need to urinate at times other than official breaks. Nineteen percent wore a pad while working to accommodate unanticipated urinary leakage. Only 15% reported being able to void whenever they needed to. Half of the teachers (397 of 791) stated that they made a conscious effort to drink less while at work specifically to decrease the anticipated frequency of urination. The number of voids during the work day increased by age group, from 2.3 in women aged 22-40 to 2.9 in those over 50 years. For none of the age groups was there an association between infection and infrequent voiding at work, with about 15% of all the women reporting a urinary tract infection in the preceding year. Teachers who stated that they made a conscious effort at work to drink less had a higher infection rate (21%) than women who did not drink less than they would like (10%), OR 2.21 (95% CI 1.45-3.38).
Further study is warranted to determine whether modifying behavioral factors at work can reduce the incidence of urinary tract infections. However, few would deny the humanitarian appeal of allowing workers autonomy over their bladder and bowel function.
CORRESPONDING AUTHOR: Ingrid Nygaard, MD, Department of Obstetrics and Gynecology, University of Iowa College of Medicine, Iowa City, IA 52242 USA
Gastrointestinal Problems in Auto Factory Shift Workers
Claire C. Caruso*, M.S.N, R.N., Sally L. Lusk, Ph.D., R.N., F.A.A.N., and Brenda Gillespie, Ph.D., University of Michigan
Research question was: what is the relationship between the shift work and self reported 1) GI symptoms; 2) GI medications; and 3) GI diagnoses. The study design was a secondary analysis from a cross-sectional survey: "Noise effects on cardiovascular and stress related diseases" by Sally L. Lusk, Ph.D. Clock-in and clock-out times were collected post hoc for the 28 day period before the questionnaires. The sample was 343 factory workers, employed 5 years or more at the plant, in a skilled or unskilled job and either white or black race. Dependent Variables. A GI symptom scale summed responses concerning nausea, heartburn, abdominal pain, loss of appetite, and diarrhea or constipation. GI diagnosis and GI medication were dichotomous yes/no variables indicating any self report of same. Independent Variable was shift work. Subjects worked fixed day or evening shift. Since their work schedules showed considerable variability in work times and overtime, four shift work variables were used: assigned shift, number of hours worked, number of night hours between 12mn and 5am, and schedule variability. 48 subjects worked a regular day shift from 6am to 3pm. Control Variables were: demographic variables; lifestyle variables - smoking, alcohol, aspirin or NSAID, noise exposure, hearing protection use; and stress related variable - trait anxiety. Data Analysis Multiple regression and logistic regression tested the effect of the shift variables while controlling for the demographic, lifestyle and anxiety variables.
Results 225 subjects were on day shift, 118 on evenings. Evening shift had more: women, black race, unmarried, slightly younger, employed slightly less years, slightly more work hours, 84% more night hours and more variability in clock-in and out times. GI Symptoms were associated with age, gender, smoking, noise exposure and anxiety. After controlling these, evening shift was associated with more symptoms (p = .02) and night hours showed a non-significant trend for more symptoms as night hours increased. GI diagnoses were reported by 23 subjects so only bivariate tests were used. Evening shift had more persons with a GI diagnosis (p = .02). A GI diagnosis was associated with less schedule variability (p = .02). GI Medication Use was associated with anxiety, noise and hearing protection use. After controlling these, hours worked approached significance (p = .09) and a schedule variability by noise interaction was significant (p = .008). In low noise as schedule variability increased, GI medication use probability increased. In high noise, this relationship reversed. Hearing loss in the high noise group might account for at least some of this effect.
Conclusions Evening shift, increased schedule variability, hours worked and night hours showed some modest adverse relationship to GI function.
CORRESPONDING AUTHOR: Claire C. Caruso, M.S.N., R.N., School of Nursing, University of Michigan, 1045 Western Dr., Ann Arbor, MI 48103, USA
Role of Shift Change Parameters on Health and Cognitive Function in Automotive Workers
Susan P. Proctor, D.Sc.*, Roberta F. White, Ph.D., Boston Environmental Hazards Center and Depts of Neurology and Environmental Health, Boston University Schools of Medicine and Public Health; Thomas G. Robins, M.D., Dept. of Environmental and Industrial Health, University of Michigan School of Public Health.
A two-year prospective cohort study of automotive workers was conducted to examine the neurobehavioral effects of petroleum naphtha exposure (White et al, 1994) and the effects of overtime and machine-paced work on behavior and cognitive function (Proctor et al., 1996). This paper focuses on the acute shiftwork effects on cognitive function, mood, self-reported health symptoms, and blood pressure (BP) measures in this cohort. It was predicted that shift workers (defined as persons working non-day shifts, such as second or third shifts) would differ from day-workers (defined as those working first shift) on these outcomes and that shift change parameters (i.e., length of time worked as a shift worker, counter-clockwise change in shift) may modify the acute effects of shiftwork.
Overtime and shiftwork parameters were calculated from payroll records. Persons with a history of head injury or loss of consciousness, native language other than English, and use of medications that affect concentration on test day were excluded from the analyses.
Cross-sectionally for years 1 and 2, shift workers (n=90 in year 1, n=89 in year 2) were significantly more likely to be younger, current smokers, have worked fewer years in the plant, have lower scores on basic academic knowledge, have a history of emotional illness, and have worked less time on current shift compared to day-workers (n=129, n=71 in years 1 and 2 respectively). And, a shorter length of time worked on shift and a counter-clockwise change in shift were significantly correlated with shiftwork. In multiple linear regression analyses, few specific shiftwork effects on cognitive function, increased health symptomatology, or increased BP were noted in this study, especially after controlling for other covariates. In year 1, increased mood was largely affected by overtime work. In year 2 (when less overtime was worked), mood was significantly influenced by the number of weeks worked on current shift with those who had worked longer on current shift reporting less tension, depression, anger, and fatigue. In year 2, a counter-clockwise change in shift and working a machine-paced job were significantly associated with increased BP measures.
CORRESPONDING AUTHOR: Susan P. Proctor, D.Sc., Boston Environmental Hazards Center (116B-4), 150 So. Huntington Ave., Boston, MA 02130.
Development of Strains in Shiftworkers: Effects of Locus of Control
Jody R. Hoffman*, M.A., Carlla S. Smith, Ph.D., Bowling Green State University, Lawrence Smith, Ph.D., University of Leeds, Paul Norman, Ph.D., University of Sheffield, and Simon Folkard, PhD., University of Wales-Swansea
The purpose of the present study was to investigate the process by which a variety of domain-specific measures of locus of control were related to the development of proximal and distal strains in a sample of rotating shiftworkers. Locus of control has received substantial attention in the stress literature, but little in the shiftwork literature. It was predicted, for example, that external social locus of control would contribute to the development of disruptions in the shiftworkers’ social life. Such disruptions would influence the development of short-term strains (i.e., fatigue and job dissatisfaction), which, in turn, would affect distal outcomes (i.e., general well-being). Similar predictions were made for work, health, and sleep locus of control.
Data were collected from 99 nurses employed at a number if hospitals in Great Britain. All worked rotating shifts. These nurses completed a questionnaire that included scales from the Standard Shiftwork Index (SSI; Barton et al., 1995).
Using the item covariance matrix, the hypothesized model was tested using LISREL VII. The initial proposed model did not provide an acceptable fit to the data, according to a number of fit indices. Thus, the model was refined, using theoretical considerations and modification indices. The final model did provide an acceptable level of fit according to most of the indices used, c 2 (26) = 36.6, p = .12, Goodness of Fit = .93.
The hypothesized process by which locus of control affects the development of strains in shiftworkers was generally supported by this study. Examining the significant paths in the model, it can be seen that external locus of control influenced heightened perceptions of social disturbances, as hypothesized. External work locus of control influenced the development of work disturbances. Greater levels of social disturbances and external sleep locus of control provoked fatigue, while work disturbances provoked job dissatisfaction. Job dissatisfaction and fatigue, in turn, affected general well-being. External health locus of control was also related to decreased general well-being.
The results of this study add to our knowledge of individual characteristics that contribute to the development of strains in shiftworkers, and the process by which this occurs. Future research could examine the effects of locus of control on the development of strains, in conjunction with other individual (and situational) characteristics. Such research will greatly aid in the design of programs to aid shiftworkers.
CORRESPONDING AUTHOR: Jody R. Hoffman, M.A., Department of Psychology, Bowling Green State University, Bowling Green, OH, 43403, USA
Effect Of Downsizing On Health Of Employees: What Are The Moderators?
Jussi Vahtera¹, Mika Kivimäki2, Jaana Pentti¹
¹ Finnish Institute of Occupational Health, Finland, 2 University of Helsinki, Finland
Background Reduction of personnel by businesses and other organizations, ie. downsizing, is common in industrial countries, but little is known about its effects on the health of employees and about factors that moderates these effetcs.
Methods We used employer’s records to investigate the relation between downsizing and subsequent absenteeism because of ill health in 981 local-government workers who remained in employment in Raisio, south-western Finland, during a period of economic decline (1991-1995). Data were separated into three time periods: 1991, before downsizing; 1993, major downsizing in some workplaces and occupations; and 1993-1995, after downsizing. Two measures of downsizing, by occupation and by place of work, were used. Data on sick-leave was obtained from records kept by the health care unit in Raisio. We also investigated, whether the effects of downsizing were dependent on other predictors of sick leave.
Results There was a significant association between downsizing and medically certified sick leave. The rate of sickness absence was 2.3 times greater (95% CI 2.0 - 2.7) after major downsizing than after minor downsizing (adjusted for health status before downsizing and demographic data). The corresponding rate ratios for musculoskeletal disorders and trauma were 5.7 (4.1 to 8.0) and 2.7 (1.7 to 4.2), respectively. The effects of downsizing by workplace depended on the age structure of the staff. When the proportion of employees who were older than 50 years was high, downsizing increased individual risk of absence because of ill health 3.2 to 14.0 times, depending on diagnostic category. When the proportion of employees over 50 years was low, downsizing had only slight effects on health. Other risks factors that increased rates of sick leave after downsizing were age over 44 years, a large workplace, poor health before downsizing, and high income.
Conclusion Downsizing as a risk to the health of employees. But this varies according to individual factors, such as age, socioeconomic status, and health, as well as factors related to place of work, for example, size and age structure of the staff.
CORRESPONDING AUTHOR: Jussi Vahtera, Finnish Institute of Occupational Health, Hämeenkatu 10, FIN-20500 Turku, Finland
Mediators of the Relationship Between Downsizing And Health
Mika Kivimäki, Ph.D., Finnish Institute of Occupational Health (FIOH) and University of Helsinki, Finland, Jussi Vahtera, M.D., and Jaana Pentti, B.Sc., FIOH
It has been demonstrated that downsizing, i.e. reduction in number of personnel in an organization, prejudices the health of employees. However, little is known about what mediates the effects downsizing on health.
We used employers’ records and questionnaire survey information to investigate associations between downsizing, changes related to work, social relationships, health habits, and sickness absence in 764 local-government employees who remained in employment after downsizing in Raisio, a town in south-western Finland, from 1991 to 1995, a period of economic decline. Data was assembled for three periods: before downsizing (1990 to 1991); during major downsizing affecting some places of work and some job categories; and after downsizing (1993 to 1995). We obtained data on downsizing and sickness absence from records kept by Raisio, including records kept by its occupational health-care unit. We obtained information on changes related to work, social relationships and health habits by means of two questionnaire surveys.
Major downsizing was associated with high levels of perceived job insecurity; increased psychological and physical demands of work; decreases in aspects of control over work, such as discretion relating to use of skills, authority to take decisions, participation in decision-making, and predictability relating to work; impaired relationships with colleagues; high levels of rumour and decrease in alcohol consumption. Of these changes, reduction in control over work, especially discretion relating to use of skills, and increased physical demands of work, appeared to act as the mediating mechanisms of through which downsizing affects health. It is therefore possible that risks to health associated with downsizing could be reduced. Our findings suggest that measures need to be focused on work contents. Our findings do not indicate that traditional health-promotion measures such attempting to limit alcohol consumption or smoking help reduce risks to health of those who remain in work after major downsizing.
CORRESPONDING AUTHOR: Mika Kivimäki, Ph.D., Department of Psychology, Finnish Institute of Occupational Health, Laajaniityntie 1, FIN-01620 Vantaa, Finland
The Relationship of Layoff Threat and Mastery to Work Performance
Sarah Y. Moore, Ph.D., Leon Grunberg, Ph.D., University of Puget Sound, Edward Greenberg, Ph.D., University of Colorado at Boulder
Research on the long-term effects of layoffs on survivor's work-related behaviors and work performance is sparse. There is some tentative laboratory evidence that suggests that downsizing can increase absenteeism and turnover, and decrease work performance. As one example, Brockner, Davy, and Carter (1985) found that, as compared to a control group with no layoff threat, students who witnessed a confederate losing his/her "job" in a laboratory setting and who had low levels of self-esteem actually had the highest performance increases between pre- and postlayoff conditions. In the present study, we extended this laboratory experiment to a large field sample of workers. Consistent with Brockner et al's. (1985) findings, we hypothesized that: (1) workers who had received a warn notice would show the greatest increases in work performance, and (2) workers with low levels of personal mastery and who had received a warn notice would show the greatest increases in work performance.
A random sample of white and blue collar employees of a very large manufacturing firm that had recently undergone several rounds of highly publicized layoffs responded to a mail survey (62% response rate). Measures included a single item which asked respondents to indicate, a "Yes" (9.1%) or "No" (90.9%) if they had received a warn notice not within the last two years, but within the past five years (i.e., a three year period). Pearlin and Schooler's (1978) seven item scale was used to assess participant's level of attributions of overall personal mastery (alpha = .82). Performance data (i.e., yearly supervisory ratings) were collected from company records and merged with the data collected from the self-reports. For the current study, individual level performance change was reflected by the difference scores created by subtracting a worker's performance ratings made four years ago from his or her performance ratings made one year ago (usable N = 536). Given that the most recent layoff cycle occurred three years prior to the study, this procedure generally enabled us to capture whatever pre to postlayoff performance change had taken place.
Regression analysis revealed that receiving a warn notice significantly predicted performance change in the anticipated direction (Beta = .478, p = .03) such that those who had received a warn notice significantly increased their performance over time more than those who had not received such a notice. The second hypothesis was also supported (mastery * warn interaction; Beta = -.470; p = .03). For workers receiving a warn notice, mastery and performance change were negatively related. Mastery and performance change, however, were positively related for workers who had not been warned. Although these data do not permit causal inferences, the results do point to the possibility of unanticipated relationships between mastery and work performance under conditions of possible job loss.
CORRESPONDING AUTHOR: Sarah Y. Moore, Ph.D., Department of Psychology, University of Puget Sound, 1500 N. Warner, Tacoma, WA 98416, USA
Work Role Importance: Moderating the Job Insecurity-Outcome Relationship
Monica A. Hemingway, Ph.D.
With the recent trend towards organizational downsizing, increasing job insecurity has become a pervasive and potent stressor. Lack of job security is thought to have as severe an impact on stress-related outcomes as actual job loss itself, and to result in many of the same psychological and physiological outcomes.
Identity theory states that strong involvement in a social role (e.g., worker, parent, or spouse) will be associated with more negative reactions to stressors that impede role performance. Job security may be central to perceptions of effective work role performance because insecurity is associated with increased ambiguity regarding expectations, job-related goals, and opportunities for advancement, all of which can undermine an individual’s ability to perform effectively.
The current study examined the moderating effects of work involvement, prioritization of work versus nonwork (i.e., family, community, leisure, and religion), and financial need to work, on the relationship between job insecurity and outcomes (general health, job satisfaction, organizational commitment, and intent to quit). The 311 participants were primarily female (56%), married (53%), worked full-time in the service industry (41%), and had a mean age of 34. Overall, participants rated the importance of nonwork, and of family in particular, more highly than the importance of work (73.95 vs. 26.05 out of a possible 100 points). Over half (56%) of the participants were the primary breadwinner in the household, indicating a fairly strong financial need to work.
Moderated regression analysis demonstrated that a financial need to work moderated the relationship between job insecurity and outcomes. Specifically, those individuals who were the primary breadwinner in their households demonstrated lower organizational commitment, lower satisfaction, and a greater intent to leave their job. The degree of work involvement did not moderate the job insecurity-outcome relationship. However, those for whom work was important showed an increased likelihood of leaving their job and were less committed to the organization when faced with increased job insecurity as compared to those for whom work was not important. The importance of leisure time in one’s life was also a significant moderator. When faced with job insecurity, individuals for whom leisure was important demonstrated considerably less satisfaction with work.
These results are consistent with identity theory and indicate the importance of personal characteristics and circumstances in determining an individual’s reaction to job insecurity. Those with a low financial need to work, and those who placed less importance on the work role, exhibited less negative reactions to the possibility of job loss.
CORRESPONDING AUTHOR: Monica A. Hemingway, Ph.D., 4234 Town Court North, Lawrenceville, NJ 08648, USA
Mental Health Before, During, and After a Close-Down
Lennart Hallsten, Ph. D., National Institute for Working Life, Solna, Sweden
Although a large number of studies have examined the relations between job loss and mental health, no study has described mental health development over all the phases (before-during-after) of a lay-off process. In the following study, however, mental health data have been collected at four occasions, covering an entire lay-off process: Before, during and two times after a close-down of a public agency. The aim of the study is to delineate the mental health development for the laid-off employees, and to relate their development pattern to three models: The constancy, shock and cumulative stress models. The constancy model corresponds to no change at all in mental health, the shock model to a short-term decrease during the anticipation phase followed by stabilisation or adaptation, while the cumulative stress model corresponds to successive decreases in mental health.
Work environment and health were generally surveyed among employees in central authorities in Sweden during the 1980s, and some these employees lost their jobs during the economic recession in the early 1990s. A prospective study of the civil servants within the first public agency to closed, the Swedish Board of Education, has been carried out. The close-down occurred unexpectedly in 1991, and data on mental health and labour market connection for the laid-off employees were collected in 1985-89 (before), 1991 (during), 1992 and 1994 (after). The same questionnaire and mental health scale was adopted as in the surveys from the 1980s. Archival mental health data from 1985-89 were available for about half of the laid-off group, and complete data from all the measurements were obtained from 139 individuals, who were representative of the study group. Due to the new work contracts and the differentiated labour market measures in Sweden, the laid-off employees were categorized into those with, and without, new permanent jobs. Eight intraindividual health developmental patterns related to the three models were discerned.
For the laid-off group as a whole, mental health decreased sharply during the anticipation period, but afterwards it gradually increased, and a full recovery had occurred three years after the closure, both for those with (70%) and without (30%) new permanent jobs. Thus, this developmental trend was most congruent with the shock model. The only difference between the two groups was that those re-employed in new permanent jobs had better mental health at all phases of the close-down. Analyses of individual mental health patterns showed that the shock and constancy patterns were most common (50 and 30% respectively), both for those with, and without, new permanent job positions. Labour market position just had a slight influence on the mental health development pattern, as had demographic variables. Cumulative stress patterns were only seen in exceptional cases, suggesting that, in general, the lay-off did not have any critical long-term health effects.
Sexual Harassment and Workplace Violence
Chair: Julian Barling, Ph.D.; Presenters: Patricia Biles, MA, C. Gail Hepburn, Ph.D., E. Kevin Kelloway, Ph.D., Katherine Lippel, LL.M., Sharon Parker, Ph.D.; Discussant: Mark Braverman
Both workplace sexual harassment and workplace aggression and violence are widespread. The purpose of this symposium is to explore a variety of issues concerning. In the first paper, Gail Hepburn and her colleagues report on a study of the predictors of workplace violence for senior executives in the federal civil service. Based on a sample of 1,000 respondents, they showed that role conflict, skill underutilization, intragroup conflict, quantitative workload and workload variance predicted workplace aggression and violence. Sharon Parker and her colleagues then discuss their study, which examined the performance-based consequences of workplace sexual harassment for police personnel. Using two separate samples, they show that performance-based anxiety was significantly greater in male-dominated groups. Julian Barling and his colleagues report on their study of workplace aggression and sexual harassment, among social workers and nurses who perform all their work-related duties in their clients' own homes. They show how fear of recurrent violence mediates any negative effects of the harassment and aggression on organizational loyalty, turnover intentions, neglect of duties and interpersonal job performance. Patricia Biles will then present OSHA's prevention response to workplace violence in the healthcare, social service and night retail industries. These four aspects of OSHA's prevention plan emphasize management commitment and employee involvement, worksite analysis, hazard prevention and control, and training and education. Finally, Katherine Lippel will analyze compensation and prevention issues within the Canadian environment. The presentation will highlight jurisdictional differences, and reflect on the relevance of existing approaches to compensation and prevention.
Mark Braverman, the discussant for the symposium, will focus on both scientific and practical issues raised.
CORRESPONDING AUTHOR: Julian Barling, Ph.D., School of Business, Queen's University, Kingston, Ontario K7L 3N6, Canada.
OSHA Workplace Violence Prevention Guidelines
Patricia D. Biles, M.A., Occupational Safety and Health Administration (OSHA)
Violence in the workplace is a serious problem in this country. Nearly 1,000 workers are murdered, and over a million are assaulted in the workplace each year. According to the BLS Census of Fatal Occupational Injuries (CFOI) for 1997, homicide continues to be the second leading cause of death for all American workers. There were 856 workplace homicides in 1997, accounting for 14% of the total 6,218 fatal work injuries in the United States.
According to Department of Justice statistics, assaults and threats of violence against Americans at work number 1.7 million a year. Retail sales workers were the most numerous victims, with 330,000 being attacked each year. They were followed by police, with an average of 234,200 officers victimized. The risks of homicide are much greater in certain industries and occupations. The taxicab industry has the highest risk at 41.4 per 100,000 persons. Job-related homicides in retail trade account for almost half of all workplace homicides.
Some of the factors which may increase a worker’s risk for workplace assault, as identified by the National Institute for Occupational Safety and Health (NIOSH), are:
- Contact with the public
- Exchange of money
- Delivery of passengers, goods, or services
- Having a mobile workplace such as a taxicab or police cruiser
- Working with unstable or volatile persons in health care, social services, or criminal justice settings
- Working alone or in small numbers
Under the Occupational Safety and Health Act of 1970 (the OSH Act, or the Act), the extent of an employer’s obligation to address workplace violence is governed by the General Duty Clause. Section 5(a)(1) of the OSH Act, or P.L. 91-596 (the "General Duty Clause") provides that:
"Each employer shall furnish to each of his employees employment and a place of employment which are free from recognized hazards that are causing or are likely to cause death or serious physical harm to his employees." 29 U.S.C. 654(a)(1)
It is, therefore, OSHA’s commitment to encourage employers to develop workplace violence prevention programs.
CORRESPONDING AUTHOR: Patricia D. Biles, M.A., Department of Labor, Occupational Safety and Health Administration (OSHA), 200 Constitution Ave., N.W., N-3107, Washington, D.C., 20210
Outcomes of Workplace Aggression and Violence for Executives in the Canadian Federal Public Service
C. Gail Hepburn, Ph.D., University of Guelph, Wayne Corneil, Sc.D., Occupational Health and Safety Agency, Health Canada, and Julian Barling, Ph.D., Queen's University.
Aggression and violence in the workplace are on the rise. Typically, discussions of these behaviors have focused on their prevalence or attempted to develop profiles of the perpetrators, the "disgruntled" workers. Research on workplace aggression's impact on the well-being of the victims is limited. We examined this impact for Executives in the Canadian Federal Public Service.
We surveyed all executive level employees in the Canadian Federal Public Service. Executives completed self-report questionnaires and indicated if they had been victims of verbal harassment or torment, threats of physical harm, and/or actual physical assault. Executives also indicated the perpetrator of the violence (e.g., coworker, supervisor, stranger). Executives further reported on their well-being and several workplace factors (e.g., anxiety, job satisfaction).
A representative subset of the executive population, based on age, gender, province of work, and executive level, returned the questionnaire (N=1,143). The impact of being victims of aggression on the executives' health and well-being is examined, with a focus on the severity of the effect as it relates to the source or perpetrators of the aggression. These findings and their implications will be discussed.
CORRESPONDING AUTHOR: Julian Barling, School of Business, Queen's University, Kingston, Ontario, Canada, K7L 3R9.
Behind Closed Doors: Organizational and Personal Consequences of Sexual Harassment and Workplace Violence for In-Home Workers
Julian Barling, Queen’s University, A. Gail Rogers, Queen’s University and E. Kevin Kelloway, University of Guelph
We develop and test a structural model predicting personal and organizational consequences of workplace violence and sexual harassment for health care professionals who work inside their client’s home. The model suggests that workplace violence and sexual harassment result in fear of their recurrence in the workplace, which in turn predicts negative mood (anxiety and anger) and perceptions of injustice. In turn, fear, negative mood and perceived injustice predict lower affective commitment and enhanced withdrawal intentions, poor interpersonal job performance, greater neglect and cognitive difficulties. Using Lisrel 8, the results supported the model and showed that the effects of workplace violence and sexual harassment on organizational and personal outcomes are indirect, being mediated by fear and negative mood. Conceptual implications for an understanding of sexual harassment and workplace violence, and research directions for future research, are suggested.
CORRESPONDING AUTHOR: Julian Barling, Ph.D., School of Business, Queen’s University, Kingston, ON, K7L 3R9, Canada
Aggression in the Workplace: Compensation and Prevention in Canada
Katherine Lippel, L.L.M., Université du Québec à Montréal
Acts of aggression in the workplace are a major cause of work-related disability and death in all North American jurisdictions. While violence in the workplace has been clearly identified as a priority issue in the field of occupational health and safety in the United States, most Canadian administrative agencies have put little emphasis on the prevention of violence as an occupational health priority, and no published Canadian studies have specifically examined workers' compensation issues.
This paper examines workplace violence as an occupational health and safety issue, in Canada. More specifically, we will examine how state institutions (administrative tribunals and regulators) view occupational health and safety issues arising from workplace violence. Three aspects will be examined: 1. legislative mechanisms that have been introduced to prevent some forms of workplace violence; 2. access to compensation for disability arising out of incidents of physical or verbal aggression in the workplace, and finally 3. issues regarding the impact of violent incidents on experience rating of employers.
Controversy surrounds compensability of injury caused by aggression in the workplace. Recent decisions have questioned the compensability of consequences of incidents of workplace violence, including murder, ascribing the violent events to personal or societal factors and implicitly absolving the employer of any obligation to prevent violent incidents or to compensate for their repercussions. Others have concluded that such claims are compensable, and have awarded compensation to workers disabled following incidents of physical or verbal aggression, including incidents of sexual or racial harassment. Decisions regarding access to compensation indirectly dictate the necessity of better prevention mechanisms, as increased compensation costs often lead to changes in prevention priorities.
In Canada, occupational health and safety legislation falls under provincial jurisdiction. While legislation regarding prevention of workplace violence has been enacted in some provinces, no workers' compensation legislation specifically addresses the issue of violence, so that information must be gleaned from analysis of caselaw. Our literature review and our legislative overview will look at the situation in all of Canada, while our reflections on workers' compensation and experience rating of employers will rely solely on Québec data. We will present preliminary results of a qualitative and quantitative study of Québec tribunal decisions relating to compensation claims for incidents of physical or verbal aggression in the workplace.
CORRESPONDING AUTHOR: Prof. Katherine Lippel, Département des sciences juridiques, Université du Québec à Montréal, c.p. 8888, Succursale Centre Ville, Montréal H3C 3P8, Québec, Canada
What Is So Distressing About a Little Name-Calling and Teasing? Work Performance Anxiety as an Explanation of the Negative Effects of Harassment
Sharon K. Parker, Mark A. Griffin, and Simon Holdaway
The prevalence of harassment in the work place, especially for women in male-dominated contexts, is now widely acknowledged. However, whilst many studies have documented significant negative mental health effects of such harassment (e.g. Fitzgerald et al., 1997), little research has considered the psychological processes by which harassment leads to these negative outcomes. That is, we know little about why harassment causes distress.
In this paper, we propose that harassment can lead to women having a heightened sense of anxiety about work performance (i.e. work performance anxiety; WPA) which, in turn, causes distress and a greater likelihood of job withdrawal. For example, sexual harassment often makes a woman’s gender more salient than their employment role (i.e. by rendering them as ‘woman’ rather than co-worker or employee), and can lead to women worrying more about their performance, especially in contexts where women’s performance capability is already under scrutiny (as is often the case when women carry out jobs traditionally performed by men). Harassment can also focus on performance issues (such as public berating of mistakes), and it can actually lead to impaired performance; both outcomes that are likely to raise women’s concerns about performance.
We explored the research proposition in two studies involving men and women. Both studies were conducted within police constabularies, but in contrasting occupational groups. In each study, employees completed confidential questionnaires that assessed: the frequency of exposure to harassing behaviors (e.g. name-calling, requests for dates), the level of WPA (e.g. how often they worried about making mistakes), job strain, general psychological distress (assessed by the GHQ-12), and intention to turnover. Survey response rates were over 60% in all cases.
The first study involved a sample police officers in which women made up less than 20% of the workforce (Parker, Holdaway & Griffin, 1998). We surveyed all women (N = 260) and a random sample of men (N = 320). As expected, given both the dominance of men and the fact that police work is traditionally carried out by men, women police officers reported high levels of harassment. They also had very high levels of WPA; levels that were significantly higher than for male officers. This result was consistent with focus groups in which women reported that, because many of their male colleagues felt women were less capable of police work, they felt considerable performance pressure, such as a sense that they ‘constantly had to prove themselves’.
We used LISREL analyses to explore the relationships amongst the variables for both the male and female groups. For both men and women, WPA was significantly associated with job strain, general distress, and intention to leave, supporting the premise that excess levels of WPA are potentially damaging. Most importantly, however, harassment was strongly related to WPA for women; a relationship that was significantly stronger than that for men. Results therefore suggest that, for women in particular, one reason for the negative psychological effects of harassment is that such adverse behaviors serve to heighten performance concerns.
The second study (N = 319) was also conducted within a police constabulary, but it focused on the support staff who provide support to the police. In this case, because there are more women than men (although there are still fewer women in management positions), and because more women carried out traditionally ‘female’ duties, we anticipated that WPA would be less salient for women. This was the case. In contrast to the police officer sample, there were no significant differences between men and women in their reported extent of harassment or their levels of WPA. Moreover, comparing across samples, women support staff had much lower levels of harassment and WPA than women police officers. These results are consistent with literature which shows that harassment is more common in male-dominated contexts, and are consistent with the idea that a masculinized context leads to heightened performance anxiety for women.
Nevertheless, despite the fact that there were few differences in the levels of WPA and harassment between men and women support staff, results of LISREL analyses showed that the patterns of relationships were very similar in this sample as for the police officers. First, showing its potentially damaging effects, WPA was strongly positively associated with job strain, distress, and a intention to leave for both men and women. Second, harassment was strongly associated with WPA for women, and this relationship was significantly greater than that obtained for men. Thus, results suggest that, even within less masculinized contexts, harassment can lead to heightened anxieties about performance for women which, in turn, causes distress.
Together, these studies help us to understand why harassment at work can have such devastating effects for women. In two samples with completely contrasting occupational cultures, harassment was associated with escalated sense of anxiety about work performance for women, which was in turn linked to negative mental health and work outcomes. Exposure to harassing behaviors did not have the same meaning for men.