Friday, March 12, 1999
3:45 - 5:45 p.m.
An Application of Kanter's Theory of Structural Power to Hospital Change
Lendrum BL, 1,4 Woodward CA,2 Shannon HS, 2,3 Cunningham C, 1,2 McIntosh J, 1,2 Brown JA,2 Rosenbloom D, 1. 1Hamilton Health Sciences Corp.,2Department of Clinical Epidemiology and Biostatistics, McMaster University, 3Institute for Work and Health, 4School of Nursing, McMaster University.
A premise of re-engineering is that re-designed jobs are larger and more complex as decision-making moves closer to the point at which action occurs. Two corollaries of this premise are that decision latitude, and job influence should increase, and that quality of work life is enhanced for workers.
Kanter’s (1993) theory of structure and power in organizations is one way to view the structure, process and outcomes of re-engineering. The theory suggests that formal power and informal power in association with access to job-related empowerment structures will result in positive personal impacts for workers and work effectiveness for the organization.
Earlier reports on the examination of Kanter’s theory to hospital change have described low levels of formal power (with no change in decision latitude, and declines in job influence) and declines in informal power for registered nurses (RNs), registered practical nurses (RPNs) and other health professional (OHPs) holding staff positions in an academic teaching hospital (Lendrum et al, 1998). While selected measures of access to opportunity structures increased, access to power structures decreased.
This presentation describes changes in personal impact and work effectiveness over a two year period for randomly selected RNs (n=186), OHPs (n=105) and RPNs (n=39). Negative changes were noticed in many measures of personal impact. For example, emotional exhaustion and anxiety increased for nurses (medium effect size). Work effectiveness measures also declined. Nurses and other health professionals perceived staff-organization relations as deteriorating. Quality improvement activities were also seen as decreasing. Their overall impression of the hospital as a place to work and receive care declined.
Despite efforts to improve quality of working life through re-engineering, the personal impacts detrimental to workers were seen and work effectiveness measures have declined, possibly to the detriment of the hospital. The next stage of analysis will test the causal relations implied by Kanter’s theory of structural power.
CORRESPONDING AUTHOR: Bonnie L. Lendrum, Hamilton Health Sciences Corporation, McMaster Division, Room 2G4A, Hamilton, Ontario, Canada. L8N 3Z5
Career Action Centre: TheHealth of those who Use it during Rapid Change
*McIntosh JM1 2, Woodward CA2, Cunningham CE1 2, Brown JA2, Shannon H2 3, Lendrum BL1 4, Rosenbloom D1. 1 Hamilton Health Sciences Corporation; 2 Department of Clinical Epidemiology and Biostatistics, McMaster University; 3 Institute for Work and Health; 4 School of Nursing, McMaster University.
Rationale: Re-engineering the hospital workplace is aimed at effecting profound and rapid change. It may have an effect on the quality of the work environment for staff, many of whom may lose their jobs. Can the availability and use of a Career Action Centre (CAC) in a teaching hospital during re-engineering, protect the health of staff, and their employability?
Procedure: A longitudinal study over a period of two years, which reflected pre, during, and post redesign. A CAC provided free services, whereby staff could learn career management, including self assessment, transferable skills, writing resumes, job interviews and understanding the changing nature of work, through consultation seminars and workshops. This was part of a larger longitudinal study. A random sample of 900 hospital employees representing some 21% of the workforce were surveyed yearly. The questions formed constructs measured by scales whose psychometric properties were known and the internal consistency of all scales was checked. The measures included: job characteristics; psychological distress; personal resources; spill over of home and work; the hospital as a place to work; competence of staff; and, the quality of the care provided.
Results: Some 81% of the 881 eligible employees responded. Sixty seven percent (67%) in year ’96 found the CAC to be helpful or very helpful compared with 59% in ’97. Initial analysis shows that those attending the CAC differed from the others. They were more anxious, more depressed, and reported greater emotional exhaustion. Attendees also had poorer active problem solving ability, spent more hours on the job, and, experienced family issues spilling over into the job. They were more confident about the transferability of their skills, and more ready for change. They experienced their job as less clear, and had a poorer perception of the quality of patient care. Relatively fewer employees who left the hospital and attended the CAC (41%) found work than those who left without using the Centre (59%). Whether the amount of CAC use has a positive effect on measures of psychological distress, personal resources, perception of quality of patient care, the employer, and employability will be presented. Conclusion: Hospital staff may have less than optimal mental health during major organizational change and use of a CAC has important implications.
CORRESPONDING AUTHOR: John McIntosh, Hamilton Health Sciences Corporation, 1200 Main St. W., Hamilton, Ontario, Canada L8N 3Z5
Welfare to Work Outcome Expectations Vs. Mental Health
Kenneth S. Thompson, MD, Institute for Public Health and Psychiatry, Western Psychiatric Institute and Clinic, Pittsburgh, PA; Russell Schuh, MA, Administrative and Policy Studies, University of Pittsburgh, Pittsburgh, PA
The strong association between poverty and mental health status has been consistently reported in mental health research for several decades. However, the relationship between poverty and psychological problems is complex. Poverty can take many forms, and responses to any form can be highly individualistic. Similarly, mental health problems can also be complex. The poor have been found to be twice as likely to suffer major depression than the more affluent. They have been found to be at 3.5 times greater risk for alcohol abuse and almost 8 times greater risk for schizophrenia. Those living in neighborhoods characterized by chronic high levels of unemployment, sub-standard or overcrowded housing, and high rates of interpersonal violence are most likely to suffer from depressive or stress-related symptoms.
Welfare to work and time limits on eligibility for support are, on the other hand, molar and undifferentiating solutions to complex problems. The interventions offered tend to be of a one-size-fits-all construction and of a bottom-line orientation. Reports of mental health problems that influence successful participation in training preparatory to work are only beginning to emerge.
We present the results of two pilot projects, one completed and one in process. A third project, integrating the experience of these two, is in the planning and negotiation stage. The first project demonstrated (N=24): that screening by program staff could identify participates at high risk for failure to complete training; identified mental health problems as a potential problem for this population; and suggested that increasing intervention intensity for high risk trainees might improve completion rates.
The second project (N=60) emphasized on-site mental health screening by a senior clinician as a method for targeting services to those in need. To date, the results support and extend the experience of the first project: 30% of those screened (18/60) were identified as having mental health problems that could influence training outcomes; 3 of the 18 identified were considered to have chronic and severe problems and were referred for treatment. All were accepted into treatment and were also exempted from sanctions by welfare. This represents 5% of the total screened; a lower rate than the literature led us to expect.
15 trainees were provided additional screening through a short term intervention. The program reported an overall drop out rate of 40%. The drop out rate for those in short-term intervention was only 20%, a finding consistent with the findings of the first project.
The third project is being designed to screen participants, to increase training intensity where indicated, and to provide on-site short-term mental health services and referral for targeted participants.
Relationships among Depression, Paid Work and Unpaid Work
Joan M. Griffin, Ph.D., Barbara Curbow, Ph.D., Jacqueline Agnew, Ph.D., Johns Hopkins School of Public Health, David LeGrande, R.N., M.A., Communications Workers of America, Jeffrey V. Johnson, Ph.D., Johns Hopkins School of Public Health
OVERVIEW: Numerous reports in the literature suggest that rates of depression among women are increasing, especially among working women. One hypothesis for this increase has been that women experience stress from their work on the job and at home, and the combined effect can result in psychological strain and depression. This study examined the relationship among depression and strain from work on the job and at home.
METHODS: Unionized customer service representatives from a large telecommunications company were asked to complete a survey regarding the demands, control, social support, and emotional burdens of their work on the job and at home. Of the 632 questionnaires sent to union members, 43% or 265 were returned. Of those, 218 were used in the final analysis (218/265, 82%). All participants were women, most were married or living with a partner (56%), had more than a high school education (67%), and were parents (76%). They averaged 42 years of age and 18 years on the job.
Exposures to stress from both paid and unpaid work were conceptualized and evaluated using the constructs in the Job Strain Mode (demands, control, and social support) and emotional labor associated with work. Exposures to stress from paid were measured using questions from the Job Content Instrument (JCI) as well as job specific measures developed for this population. Exposures to stress from unpaid work were measured using a new set of instruments developed by the authors. The Center for Epidemiologic Studies Depression Scale (CES-D) was used to evaluate depressive symptoms.
RESULTS: Using the CES-D, over 32% of the participants reported moderate symptoms of depression. Over 22% had been diagnosed with depression by a physician, while 16% had been diagnosed with an anxiety disorder. In bivariate analyses, depression was significantly associated with high job demands and emotional exhaustion from the job, but not low control or low social support on the job as expected. For exposures related to unpaid work, depression was associated with high home demands, low social support at home, and high emotional management. In the hierarchical logistic regression models used for the multivariate analysis, however, high job demands, high emotional labor at work, high emotional management at home, and, unexpectedly, high rewards at home were significantly associated with depression, when controlling for age, marital status, having school-aged children, body mass index, smoking status, and trait anxiety.
CONCLUSIONS: In this study, participants with high job demands and who have high emotional requirements for their work on the job and at home were more likely to suffer from depressive symptoms.
CORRESPONDING AUTHOR: Joan M. Griffin, Ph.D., International Centre for Health and Society, University College London, 1-19 Torrington Place, London, WC1E 6BT, UK
Temporary Work, Psychological Contract, and Stress
Olympia M. Kyriakidou*, MPhil, University of Surrey
Today’s business environment is changing and with it the very nature of the employment relationship. A new type of employment relationship is evolving in terms of more part-time work, more job-sharing, and greater degrees of flexibility based on short-term contracts, in which the idea of a ‘job’, and especially the concept of a ‘job for life’ is no longer relevant or appropriate (Pemberton, Herriot &Bates, 1994).
These recent changes in the nature of the workplace have created various stresses and strains. Previous research into employment contracts has considered the impact of contractualization and the qualitative changes people have felt-both in terms of life chances and in terms of self-concept. For the majority the move has been largely viewed negatively. Security of tenure-with its concomitant predictability of payment for services rendered, life/workplace routine and sense of belonging, has largely been replaced by insecurity over employment prospects, concern over promotional prospects and hesitation in taking on long term large financial burdens such as mortgages, loans, etc.
However, little work has been conducted on the social and psychological implications of such large-scale restructuring and its impact on the employees’ well-being. Large scale redundancies and changes in the nature of the employment contract for those remaining within the workforce could have considerable psycho-social impact. This paper looks at some of the latest psychological changes in features of the employment contract for two types of employees : those on a temporary contract and those on a permanent contract. Issues such as job insecurity, role ambiguity, corporate identification and commitment, self-esteem, job importance and satisfaction, and organizational culture are investigated and their relationship with employee stress is explored. It is of particular interest to examine how employees relate to and manage their involvement in their organizations. The important psychological features of the employment contract are identified and their perceived influence on stress are examined. The Psychological Contract Model is used to frame the findings obtained and the need for new forms of psychological contracts is proposed.
Preliminary analysis of the results shows that job insecurity and role ambiguity are associated with increased work stress , and the stress of organisational culture is the strongest predictor of job dissatisfaction and work stress.
CORRESPONDING AUTHOR: Olympia m. Kyriakidou, MPhil, University of Surrey, Department of Psychology, Guildford, Surrey, GU2 2RZ, UK.
Understanding the Fatigue Index Score
Vincent Cantwell, Chair, The Human Factors Group
In its efforts to understand the scope and depth of the issues surrounding human performance in the maritime industry, the U.S. Coast Guard has sought to improve the manner in which data is being collected and reported by U.S. Coast Guard Marine Investigating Officers. Specifically, a new and scientific approach to estimating the probability of fatigue that is present or causal to an incident has been proposed. The purpose of this computation is to give Investigating Officers a general tool by which to evaluate the presence of fatigue observed in persons associated with the incident. This approach uses a formula that is based on mathematical analysis of accident and incident data already contained in the Coast Guard database. Elements of the formula include hours worked and slept in the previous 24, as well as (seven) subjective symptoms (forgetful, distracted, less motivated, sore muscles, difficulty keeping eyes open, desire to sit or lay down).
The formula results in a value known as the "Fatigue Index Score,"or FIS. If the FIS is > 50, there is an 80% likelihood that fatigue was one of the causes to the incident. Accordingly, Investigating Officers are instructed to conclude that fatigue is a probable cause of the accident, or at least one of the probable causes. For a FIS score of <50, there is similarly an 80% likelihood that fatigue was NOT a cause. It is unknown if Investigating Officers have been instructed to make specific notation to this effect.
Application of this formula has been tested and found reliable in correctly identifying fatigue as causal to the incident in about 80% of the trials. Accordingly, there is roughly a 20% chance that the computed score will be in error, (Type I, or Type II errors). Certain factors are not considered in the formula though. Such as, the timing of work or sleep as compared to the circadian cycle, quality of sleep previous, and other environmental or individual factors that might promote or be responsible for the subjective sensations listed. Given the liability and exposure that mariners are subject to, the probability that the such an approach will discourage honest and accurate reporting seems high.
Nevertheless, the Fatigue Index Score methodology may have some predictive value. It is possible that the FIS may serve to highlight areas or work-rest patterns of concern. Organizations and personnel are therefore encouraged to use the Fatigue Index Score methodology to review their routine and high demand operational workloads for the potential of fatigue, provided conclusions are appropriately constrained.
CORRESPONDING AUTHOR: Vincent Cantwell, Chair, The Human Factors Group, Postal Drawer 498, Linthicum, MD 21090, USA
Preventing Self-report bias in Occupational Health Research
Stewart I. Donaldson, Ph.D.*, Claremont Graduate University, and Elisa J. Grant-Vallone, Ph.D., California State University, San Marcos
Theoretical advances in occupational health are highly dependent on empirical confirmation and disconfirmation. Theoretical perspectives supported by numerous empirical studies become dominant in the field. In contrast, theories followed by a collection of null or mixed empirical results become controversial and often fade away. Unfortunately, these trends can occur despite limitations that exist in the methodologies used to gather data about occupational health behavior.
Accurate measurement of behavior in organizations is essential for advancing the field of occupational health. Despite its importance, measurement in organizational settings is often referred to as one of the main shortcomings of this area. The problem is that researchers must rely to a large extent on self- reports. Such measures are common because they are relatively easy to obtain and are often the only feasible way to assess constructs of interest. However, studies which rely on self reports are suspect for two primary reasons: 1) self- reports are prone to many kinds of response bias, and 2) inferences about causal relationships may be inflated by the problem of common method variance. The purpose of this presentation is to critically evaluate methodological and statistically approaches for preventing self-report bias in occupational health research. Using a longitudinal, multitrait-multimethod (MTMM) dataset, we will illustrate when self-report bias is most likely to distort conclusions in occupational health research. A framework will be presented that argues motivational self-report bias is the function of at least the following four factors: 1) the nature of the construct of interest, 2) the "true state of affairs" (i.e., the true score on the construct versus random or systematic error), 3) dispositional characteristics of the respondent, and 4) situational characteristics of the measurement situation or environment. The audience will be shown how to use this framework to understand and control for self-report bias in occupational health research.
CORRESPONDING AUTHOR: Stewart I. Donaldson, Ph.D., Department of Psychology, Claremont Graduate University, 123 E. 8th Street, Claremont, CA 91711, USA
An Investigation of "Challenge" and "Hindrance" Related Stress among Managers
Marcie A. Cavanaugh*, Ph.D., Cornell University, Mark V. Roehling, Ph.D., Cornell University, Wendy R. Boswell, M.S., Cornell University, & John W. Boudreau, Ph.D., Cornell University
Job-related stress among managers has been described as reaching epidemic proportions (Marino, 1997). While there is converging evidence that most managers report feeling job-related stress, there is less agreement regarding the nature of the relationship between reported stress and various organizational outcomes. In the past, the relationship between job-related stress and negative outcomes has been emphasized. More recently, there has been increasing recognition of the potential positive outcomes associated with job-related stress.
One factor that may influence whether job-related stress is associated with positive outcomes or negative outcomes is the source of the stress. Researchers examining the relationship between stress source and positive and negative outcomes have identified several sources of stress that may lead to negative outcomes (categorized as hassles or negative stressors in the literature) and sources of stress that may lead to positive outcomes (categorized as uplifts, job challenges, or positive stressors) (Matteson & Ivancevich, 1987). However, there appears to be some debate in the literature regarding specifically what work demands or circumstances are related to positive or negative outcomes.
A better understanding of the relationship between various sources of stress and organizational outcomes for managers is needed. First, we make the distinction between two sources of stress that may differentially relate to organizational outcomes. These two sources of stress are challenges and hindrances. Second, we investigate the extent to which reported challenge and hindrance related stress among a sample of 1,886 U.S. managers is associated with three organizational outcomes.
The results indicate that challenge related stress was significantly positively related to job satisfaction and significantly negatively related to job search. Hindrance related stress was significantly negatively related to job satisfaction and significantly positively related to job search and turnover.
These results suggest that examining stress as two separate constructs (challenges and hindrances) will increase our understanding of the relationship between stress and organizational outcomes for managers.
CORRESPONDING AUTHOR: Marcie A. Cavanaugh, Ph.D. School of Industrial and Labor Relations, Cornell University, 393 Ives Hall, Ithaca, NY 14853, USA
Perceived Stressors of 386 Working Individuals Referred for Stress Management
John E. Garrison, Ph.D., M.P.H., Lahey Clinic
This Poster Presentation describes the use of the Brief Computerized Stress Inventory (Brief CSI) to assess the perceived stressors of working individuals referred to Stress Management Groups offered in a large medical center. The investigation was conducted to answer the questions "Who are these stressed patient's and what's bothering them?," as part of a larger study to evaluate and refine the Medical Center's Stress Management Program. The subjects of this study were 386 patients referred for participation in Stress Management Groups in the Dept. of Psychiatry at the Lahey Medical Center in Burlington, MA. Mean age of the patients was 45.8 years, 50% were male, 67% were married, with a mean educational level of 15.3 years. Referrals were initiated by physicians and other health care providers, mental health clinicians, and patients themselves. This is a unique research population in that all participants were defined as "stressed" by themselves and/or their clinicians and were experiencing enough mental or physical distress to present themselves to a medical center for assistance. The Brief CSI was chosen to assess the stressors of these individuals. The Brief CSI is an 115 item assessment device which systematically identifies the respondents' perceived stressors (among other variables). For the present study, the Brief CSI was manually completed by 386 consecutive, employed outpatients referred for stress management during calendar years 1995 through 1998. Patients rated their "Overall Stress Level" as well as severity of ten specific "Sources of Stress". The Brief CSI was computer-scored and cumulative data for the 386 patients was analyzed for perceived stressors. The top stressors for the group were "Time Pressures", "Inability to Make Life Changes", and "Physical Appearance". Women experienced significantly more stress in most areas of their lives compared to men. Only in the area of "Friends/Social Life" did men report more stress than women. "Work" and "Physical Health" tied for fourth out of the ten "Sources of Stress" in the respondents' lives. Statistical comparison of this patient group with a (non-clinical) normative sample of 3954 east coast workers indicated that the patient group experienced a significantly greater amount of "Overall Stress" as well as greater stress from all of the "Sources". Working individuals presenting for stress management services in this medical center appear to be very suitable candidates for a Stress Program.
CORRESPONDING AUTHOR: John E. Garrison, Ph.D., Dept. of Psychiatry and Behavioral Medicine, Lahey Clinic, 41 Mall Rd., Burlington, MA 01805.
Does Stress Affect Absenteeism? A Meta-analytic Response
Yitzhak Fried, Ph.D., Wayne State University, Arie Shirom, Ph.D., Tel Aviv University, Cary Cooper, Ph.D., University of Manchester, Joel Ager, Ph.D., Wayne State University, Kim Stepanski, M.A., Wayne State University
The present meta-analysis investigates the relationship between three major work stressors (role ambiguity, role conflict, role overload) and work absence. The relationship between each of the stressors and absenteeism was significant, although relatively weak. Using the Schmidt-Hunter 75% rule, the potential existence of moderators was realized. Report style (self-report vs. organizational record) was hypothesized to moderate the relation between work stressors and absenteeism. As hypothesized we found that, on average, the relation between role ambiguity and absenteeism was higher when the data was collected from the organization's record. No significant differences were found for role conflict or role overload, although in both cases the trend of the results was in the hypothesized direction. Furthermore, satisfaction was hypothesized to mediate the relationship between overall stress and absenteeism. Satisfaction was found to fully mediate this relationship.
Development and Validation of a Work-Family Orientation Scale
Robert P. Delprino, Ph.D., Karen O'Quin, Ph.D., Buffalo State College
Several theories of an individual's work-family orientation have been identified to explain the relationship between work and family. Three prominent theories in the field of work and family include spillover theory, compensation theory, and segmentation theory. Despite the fact that such models of work-family orientation have existed for more than a decade, few attempts have been made to identify whether employees' orientation can be identified and measured. The purpose of this two-part study was to develop and validate a work-family orientation scale.
The first study focused on the development of the work-family orientation scale and work-family issues scale. Participants included 597 law enforcement officers from 21 agencies in Western New York. Participants completed a 21-item work-family orientation scale designed to measure the three theoretical orientations identified in the literature (spillover, compensation, segmentation). Also a 15-item work-family issues scale designed to measure the impact of the officer's job on their family was completed. Data reduction (Factor analysis and Reliability analysis) was used to examine which items formed reliable subscales and which would be likely candidates to retain in the future.
In the second study, modified versions of the work-family orientation scale and work-family issues scale were completed by 1,632 police officers from three states. As in the first study, data reduction was used to examine which items formed reliable subscales.
The results of both Studies 1 and 2 supported two of three theoretical models. Spillover and segmentation, which the literature suggests might be different, were not differentiated by officers in either study.
The results suggest that these two theoretical orientations are simply endpoints of the same continuum. The compensation model was well supported by the present data. Results showed encouraging evidence for the validity of the work-family orientation subscales. Work-family spillover was strongly related to reports of negative effects on the family while compensation was not significantly related to that variable. Work-family spillover was significantly and highly correlated with several outcome variables: occupational effects, personal stress, overall stress and number of reported health problems. Compensation was significantly correlated with only two of these outcomes: personal stress and overall stress and the magnitude of the correlations were low. Suggestions are offered as to the value of these findings and how such knowledge of work-family orientation may be used by organizations to assist their employees with work-family issues.
CORRESPONDING AUTHOR: Robert P. Delprino, Ph.D., Department of Psychology, Buffalo State College, 1300 Elmwood Avenue, Buffalo, NY 14222, USA
Organizational Commitment as a Predictor of Turnover
Raymond N. Wolfe, Ph.D., and Thomas H. Feeley, Ph.D., State University of New York, College at Geneseo
Workers stay in their jobs for two main reasons: affective ties to the organization, and costs associated with quitting. Allen and Meyer’s (1990) commitment scales purport to measure these two motivational variables, along with a third, normative commitment. We pitted these scales against Mowday, Steers, & Porter’s (1979) Organizational Commitment Scale as predictors of turnover among employees of two supermarkets in Buffalo, NY (N = 213 at Store A and 104 at Store B). In data from Store A, scores on Allen and Meyer’s scales measuring affective ties and costs associated with quitting correlated significantly and in the expected direction with turnover after a 4-month interval. These rs did not withstand cross-validation, however; in data from Store B, none of the correlations with turnover approached significance. In both data sets, scores on Mowday et al.’s scale were unrelated to turnover.
Of the 24 items in Allen and Meyer’s scales, 9 are reverse-scored. Other investigators (Magazine, Williams, & Williams, 1996; McGee & Ford, 1987) have identified these 9 items as psychometric trouble spots. We factor analyzed responses from Store A and Store B separately. The reverse-scored items turned out to be problematic, and for the same reasons as those specified by Magazine et al. and McGee and Ford. In each data set, these items clearly compromised interpretability of the factor structure. In data from Store A, rs with turnover increased (though not significantly) in the expected direction when reverse-scored items were removed. These findings are congruent with an array of results from other instruments used by industrial and organizational psychologists; taken together, the cumulative evidence leads us to question the wisdom of using item reversals in self-report inventories.
The literature on Allen and Meyer’s commitment scales displays a pattern frequently seen in the social sciences: The instrument’s authors keep turning up favorable evidence, whereas other users tend to find serious defects in it.
CORRESPONDING AUTHOR: Raymond Wolfe, Ph.D., Department of Psychology, SUNY College, Geneseo, NY 14454-1401
Creating Notifications Tailored for Individual Workers by Using a Database
*Norio Mishima, M.D., University of Occupational and Environmental Health, Japan, Noboru Iwata, Ph.D., University of South Florida
NIOSH held a workshop on the methodology of worker notification in 1991, where the main subject was to notify workers exposed to toxic substances. As concerns regarding mental health at the workplace have been increasing, many studies using psychometric tests have been conducted worldwide. However, worker notification in this type of research has rarely been discussed. One of the reasons may be that because mental health at the workplace is related to every worker, researchers must explain to all workers in detail whatever test results, i.e. normal or abnormal scores, they have, but up until now it has been very difficult to prepare a large number of these documents methodologically. We have developed a new system in which we can create a notification that explains the respective results of individual workers in plain words. We were able to overcome the difficulty by using the relational database management system. After applying our system to different studies, we realized that it has a high potential ability to be used in many activities. We introduce the outline of our system and discuss its potential application.
First, we decided on a criterion for dividing test scores into several categories in each psychometric test. Then we classified the whole data of workers' scores according to the criterion, and saved the results in a computer file. Secondly, we made up sentences that explained the meaning of each category in every test, and saved these sentences in a different computer file. We also made a file that held personal information. Thirdly, we used relational database software (RDB) to read these files into a single database file. All component files were stored in the respective tables of the same database file. Then we integrated all these tables by making up relationships. RDB provides the basic functions to handle this process. Finally, we printed out notifications. In order to handle these procedures, a computer has to be equipped with RDB and a printer that outputs quality documents.
The most remarkable advantage of our system is that one is capable of making notifications tailored for each worker. When a similar study is conducted again, the previous files can be used repeatedly. Every time a new psychometric test is used, a new component file must be prepared. However, this process will give one a good opportunity to create a database on how to explain the results of psychometric tests. When many kinds of information and experience are accumulated, it will allow one to study further the methodology of worker notification.
CORRESPONDING AUTHOR: Norio Mishima, M.D., Dept of Mental Health, IIES, University of Occupational and Environmental Health (UOEH), 1-1, Iseigaoka, Yahatanishi-ku, Kitakyushu 807-8555, Japan
Occupational Trends in Absences Resulting from Job-related Stress
Bruce J. Bergman and Timothy A. Webster, Bureau of Labor Statistics
The Bureau of Labor Statistics (BLS) collects data on neurotic reactions to stress (NRTS) in the workplace, otherwise known as job-related stress. Data from 1992-96 indicate that although reaching a five-year low of 3,453 cases, the estimate of job-related stress incidents has remained a relatively constant proportion of the total injury and illness cases resulting in work absences. In 1996, job-related stress accounted for almost two-thirds of all anxiety disorders resulting in lost workdays.
Managerial and professional specialties, and technical, sales, and administrative support occupations had a higher proportion of NRTS incidents than other occupational groups.
Between 1992 and 1996, eight occupations were involved in at least 50 cases of job-related stress resulting in work absences in at least 4 of the 5 years: 1) Supervisors and proprietors, sales occupations, 2) Cashiers, 3) Managers and administrators, n.e.c., 4) Investigators and adjusters, excluding insurance, 5) Secretaries, 6) Assemblers, 7) General office clerks, and 8) Supervisors, production occupations.
Other occupations with 50 or more cases of job-related stress, but in fewer than 4 of the 5 years, include bank tellers; truck drivers; miscellaneous machine operators; health aides, excluding nursing; insurance adjusters, examiners, and investigators; stock and inventory clerks; bookkeepers, accounting and auditing clerks; receptionists; sales workers, other commodities; licensed practical nurses; registered nurses; and other financial officers.
Registered and licensed practical nurses account for about four percent of all private-sector job-related stress cases reported in the 1996 BLS survey.
Job related stress occurs most frequently in finance, insurance, and real estate (FIRE), with an incidence rate three to four times than that in manufacturing. Over the past 5 years, more than 1 of every 6 cases of job-related stress were reported in FIRE, a division that typically accounts for less than 1 of 40 injury and illness cases in the private sector.
Between 1992 and 1996, job-related stress consistently resulted in lengthy absences, with a median duration ranging from 10 to 33 days. The median 1996 absence due to NRTS was double the level for all injuries and illnesses, and one-third of the total reported cases of NRTS resulted in 30 or more days away from work lost. At the same time, the incidence of one-day cases being reported has steadily risen.
The BLS Office of Compensation and Working Conditions publishes annual estimates of nonfatal injuries and illness sustained in the workplace. These estimates are based on a sample of approximately 165,000 private sector businesses. BLS tabulates national estimates by industry, selected case characteristics, and worker demographics.
Methodological Approaches to Sampling and Response Biases in Health Surveys
Clark Johnson, Ph.D., Randal D. Beaton, Ph.D.*, Shirley A. Murphy, Ph.D., and Ken C. Pike, Ph.D., University of Washington
Data from a longitudinal health survey of two urban fire departments provided an opportunity to examine selection biases . At baseline, surveys were distributed at fire houses. This resulted an initial, "early" response rate of about 50%. When the follow-up surveys were scheduled both fire departments agreed to administer the survey as part of in-service training: An approach which increased the rate of participation to over 75%. "Early" responders (n = 267) were compared with the group who first completed the survey as part of the in-service training ("later" responders -- n = 142). Since a complete data record existed for both groups, a thorough examination of the sampling protocol’s impact was possible.
Results. The "later" group had lower Symptoms of Stress ( SOS ) scores on the Total SOS (t = 2.65; p<.01) and on six of the ten SOS subscales. The "later" respondents were also less concerned about their occupational stressors as measured by their total Sources of Occupational Stress scores (t = 3.96; p< .001).
Conclusions/Significance. These comparisons suggest that participants who responded to initial survey requests tended to endorse more and / or more frequent sources of occupation stress as well as stress symptomatology. This suggests the possibility that health surveys with high risk occupational groups may tend to select a subset of "volunteer" respondents for whom health issues may be a more current personal concern. However, these results require careful interpretation. The statistically significant differences that emerged between the two respondent groups must be balanced with an appreciation for their clinical relevance.
Approaches to analyzing the contribution of response biases were also assessed. A between group comparison of negative affectivity and social desirability revealed that these groups are impacted differentially by intra- and inter-personal influences. Investigators are advised to include qualitative methods that provide a signal to the investigator when surveys are being used as a means to give "voice" to employee distress by "sending a message" to management.
CORRESPONDING AUTHOR: Clark Johnson, Ph.D., Univ. of Washington, Dept. of Psychosocial & Community Health, Box 357263, Seattle, WA 98195-7263, USA