Volume 23
Number 3
2002

[Table of Contents]

Work and life stressors and psychological distress in the Canadian working population: A structural equation modelling approach to analysis of the 1994 National Population Health Survey

Abstract

Work stressors are increasingly recognized as potentially important determinants of mental health status. We examined such relationships using a structural equation modelling approach with data on adult, working Canadians who participated in the first wave of the National Population Health Survey (NPHS). Work stressors formed a composite construct with paths from psychological demands, decision latitude, work social support and job insecurity, each measured through a reduced version of the Job Content Questionnaire. Life stressors also formed a composite construct composed of chronic stressors and recent life events. Psychological distress was the outcome, mediated by the latent effect constructs of mastery and self-esteem. Work stressors had consistently positive total effects on distress (sum of standardized path coefficients from 0.004 to 0.153 across gender-occupation strata), with all of these effects mediated through reduced self-esteem and mastery (work stressors to these mediators: -0.188 to -0.413). Life stressors had larger positive total effects on distress (0.462 to 0.536), with the majority of these effects direct.

Key Words: Canada; cross-sectional; health surveys; mental health; models-statistical; occupational health; psychology-industrial; stress-psychological

Introduction

Problem and research question

“Work stress” has been increasingly recognized both as a health outcome and as an important health determinant by the public health community. Analysis of the 1985 US National Health Interview Survey, for example, produced estimates of 11 million workers experiencing health-endangering levels of mental stress at work.¹ Although it is generally accepted that work stressors can adversely affect mental health,² some argue that life stressors are far more important than work stressors as determinants of psychological well-being.³ Fortunately, the majority of researchers include both of these key domains in their research: work/job/intra-organizational stressors and non-work/family/life/extra-organizational stressors.^4–8 Psychological mediator variables such as self-esteem and mastery between stressors and health outcomes have been recognized as important to the understanding of stressor-health outcome relationships.^4,7 Yet the relative importance of each stressor domain and the extent of mediation by psychological variables is contested among health practitioners, workplace parties and policy makers. Hence, our research question was:

“Among working Canadians, what relative contribution do work stressors and life stressors make to explaining variation in levels of distress, both directly and indirectly through self-esteem and mastery as mediating variables?”

We were further interested in assessing the extent to which stressor-distress relationships varied between genders^7,9 and across occupational strata.¹⁰ We drew on the results of previous analyses of the NPHS 1994/5 data,^11–16 relevant scientific literature and current debates among workplace health practitioners to hypothesize a structural model linking the main constructs of interest. Our analysis was facilitated through the use of structural equation modelling (SEM) techniques that permit simultaneous estimation of both measurement and structural components.¹⁷ Comparisons of model estimates by gender-occupational strata provided results to explore the extent to which gender-occupational contexts modulate the relationships of interest.

Formulation of the structural model

Organizational psychologists have long argued that a variety of factors are likely relevant to psychological well-being.¹⁸ Work stressors, including lack of social support, have been associated with a variety of adverse health outcomes, including depression and distress.¹⁹ Occupational health researchers have often used measures of depressive symptoms or psychological distress as indicators of job strain.²⁰ A wide variety of life stressors are also seen as contributing to adverse mental health in general and depression in particular.^21,22 These two broad domains, work and life stressors, are therefore postulated to be important determinants of distress among a working population, with higher levels of stressors being associated with higher levels of distress (see direct paths in Figure 1.)

A variety of psychological variables may mediate the relationship between stressors and distress i.e., be impacted by work stressors and in turn affect distress.²³ Mediation is in contrast to moderation (“interaction” or “effect modification” in epidemiological terms), where the psychological variable affects the strength of a stressor-strain relationship, but is not itself affected by stressors e.g., self-esteem in Jex and Elacqua, 1999.²⁴ As a mediator, self-esteem can be influenced by work factors, as eloquently argued by Locke and colleagues, 1996:²⁵ “We consider the realm of work to be crucial to self-esteem, because it is through work that people master reality and sustain their existence ... thus work which one chooses as a value and which is done rationally, in full mental focus, is a source of self-esteem”. Closely related is mastery, which declines as role strains reduce the extent to which people feel that they are able to manage their lives.²⁶ Hence, we postulated that paths would go from self-esteem and mastery to distress (indirect paths in Figure 1). The extent of mediation, from minimal, where most stressor effects would be direct to distress, to substantial, where the majority of stressor effects would be indirect through the mediators to distress, was left open.

Gender/occupational differentiation

The effects of work stressors on distress may also be shaped by social contexts.^10,27 “Work contexts” and “family contexts” have been described as background factors for psychological distress.⁶ Sociodemographic factors have also been depicted as independent predictors of depression in the vulnerability model elaborated by Phelan and colleagues.⁴

Studies relevant to gender comparisons present contrasting findings. Some reports show similar relationships between stressors and distress for both genders.^28,29 But some demonstrate gender differences. For example, women's distress levels showed reductions by social integration/resources for social support⁷ and work stressors had a stronger effect on distress among men.⁶ Highly routine jobs were more strongly associated with increased distress among women³⁰ and women were more affected by stressful personal events than men.³¹

Fewer studies inform occupational strata comparisons, as many regression analyses control for occupational grade/level and cannot examine differences in relationships within a strata (e.g., higher demands being confounded with higher employment grade in Stansfeld et al., 1997).³² Several studies have shown greater life stressors among those of lower occupational categories and income (e.g., Stephens et al., 1999).³³ Joint examination of gender/occupational strata is rare. Comparisons are made difficult by the differential gender distributions across occupational grades and problems associated with the unmeasured factors varying across such strata.^34,35 Fortunately, the NPHS provides relatively large, representative samples of both working women and men, allowing us to create separate gender/occupational strata.

Secondary data

Population

The Canadian National Population Health Survey recruited a large, nation-wide sample of randomly selected participants.^36,37 The 1994–1995 cycle was the first wave. A complex, stratified, multi-staged design identified approximately 20,000 households, excluding people living on Indian reserves, at military bases, in institutions, and in some remote areas of Ontario and Quebec. In each household, one adult, chosen at random, was asked more detailed questions. Response rates were 88.7% at the household level and 96.1% at the individual level within households.

We selected all participants in the public use data set (Statistics Canada, 1995)³⁸ who answered the detailed questions, were aged 18 to 64 years old inclusive, were currently in paid employment, and responded to an abbreviated version of the Job Content Questionnaire (JCQ)(see below): in total, 4230 adult working men and 4043 adult working women. Responders to the JCQ (88%) were comparable in age and income distributions to the NPHS working population of interest. However, women respondents were less likely to be working part-time (71.5% respondents versus 78.6% of non-respondents, p=0.02) and respondent men were less likely to be blue collar (48.7% respondents versus 54.7% non-respondents, p=0.006), college/university graduates (38.5% versus 45.5%, p=0.003) and married/common law (70.5% versus 78.0%, p=0.0008).

Respondents were asked if they had worked for pay or profit in the past 12 months, with up to six jobs recorded. The main job was classified according to the sixteen-point Pineo occupational prestige classification.³⁹ These were grouped into Lower Pineo (skilled, semi-skilled & unskilled) and Upper Pineo (supervisor, semi/professional & management) for the purposes of occupational stratification.

Measures by domain

Work stressors

Respondents completed a version of the JCQ⁴⁰ with respect to their current main job. The JCQ included questions on work psychological demands, decision latitude (control), physical exertion, job insecurity, and social support at work. Although the larger pool of items from the JCQ has undergone extensive validity testing,⁴¹ only 12 items were included in the abbreviated NPHS version. Responses were based on a five-point scale (0=strongly agree; 4=strongly disagree), modified from four-point standard versions. Psychological demands were assessed by asking how hectic the job is and the degree of freedom from conflicting demands made by others. Decision latitude or “control” was measured through two dimensions: skill discretion (learning new things, level of skill and “doing things over and over”) and decision authority (freedom to decide how to do the job and how much say a worker has about what happens in the job). The assessment of work social support asked about exposure to hostility or conflict from co-workers, supervisor help in getting the job done and co-workers' help in getting the job done. Job security and physical effort were each tapped with a single item. For our purposes, both decision latitude and work support were reverse scaled to connote “lack of” each as a work stressor, with higher scores connoting higher values for that stressor (lack of control and lack of work support, respectively).

Life stressors

Measures in this domain covered both particular events and chronic strains, as advocated by other stress researchers.⁴² The chronic stress index⁴³ included 18 questions about ongoing concerns with children, spouse, housework and family – found in developmental work to have lasted on average more than five years. The index was adjusted for marital status and children in the home. The measure of recent life events^44–46 consisted of 10 questions requiring yes/no responses about the respondent or someone in his or her family in the last 12 months. Recent life events included major financial crisis, change of job, demotion, cut in pay and increase in arguments. The recent life events score was adjusted to take into account social roles (e.g., marital status, presence of children).

Psychological mediators

Self-esteem was tapped with six questions from the classic Rosenberg (1965)⁴⁷ scale. Respondents answered on a five-point scale from “strongly disagree” to “strongly agree” and responses were summed to a maximum of 24, higher scores indicating stronger self-esteem. Mastery, reflecting the extent to which individuals believe that their life chances are under their control,⁴⁸ were measured with seven questions on a five-point scale, yielding scores ranging from 0 to 28, with higher scores reflecting superior mastery. Cronbach's a for self-esteem and mastery were both > 0.65 and similar across gender/occupational strata.

Mental health outcome

From the relatively wide range of mental health indicators available in the NPHS,³³ we chose the distress measure based on the short form of the Composite International Diagnostic Interview (CIDI) with six questions. Cronbach's for distress was also > 0.65 and similar across gender/occupational strata.

Analytical approach

Formulation of measurement models

A variety of approaches have been taken to modelling with JCQ scales: as independent continuous variables⁴⁹ or di/trichotomized scores,³² job strain as demand/control differences,⁵⁰ ratio¹⁵ and median-based dichotomized interactions,^31,51 and iso-strain as interactions with work social support,^9,52 among others. Creating a composite latent construct⁵³ of work stressors⁵⁴ leaves open the nature of the relationships between the various scales. Conceiving of individual work stressor scales as formative, rather than reflective in the classical measurement sense, reduces concerns about the distinct nature of the work stressors e.g., job insecurity compared to other scales in the JCQ,⁴¹ and low consistency associated with limited numbers of component items e.g., psychological demands in the NPHS version.¹⁴ The different timing and nature of the chronic stressor and recent life events indices⁵⁵ also supported construction of a composite construct, where each index is formative of life stressors. In contrast, where items reflect one underlying construct and demonstrate relatively high measures of internal consistency as a group of items,⁵⁶ such as items making up self-esteem, mastery and distress, formulation as latent effect constructs was deemed most appropriate.¹⁷

Model estimation and testing

Structural equation modelling was performed using EQS version 5.7b.⁵⁷ Since EQS does not accommodate survey weights, potentially biasing standard errors,⁵⁸ we used SAS version 6.12⁵⁹ to calculate covariance matrices of the variables using survey weights provided by Statistics Canada.³⁸ The co-variance matrices for each stratum were used as input to EQS.

Each of the four gender-occupational stratum models was estimated separately, and paths within strata were compared using standardized estimates.^6,60 Maximum likelihood estimation was used. Covariances among residuals of the endogenous variables were initially fixed at zero.

As the chi-square is highly sensitive to sample size and distributional assumptions,⁶¹ five other measures of the overall goodness of fit were used.⁶² The goodness of fit index (GFI) and the adjusted goodness of fit index (AGFI) were chosen for their low sensitivity to methods of estimation.⁶³ The root mean square error of approximation (RMSEA) and the comparative fit index (CFI) are least sensitive to sample size⁶³ and Bollen's IFI is the least biased due to non-normality of variables.⁶⁴ The GFI, AGFI, CFI, and IFI range in value from 0 to 1, with a value of greater than .9 indicating a good fit. RMSEA values range upwards from 0, a perfect fit, through to .05, a good fit, up to .08, a fair fit, and > .1, a not acceptable fit.⁶⁵

The Lagrange Multiplier (LM) test⁵⁷ was used to suggest the addition of potentially significant paths and the deletion of insignificant ones (p>0.05). Correlations of some errors between sub-scales of the items of latent variables were suggested by the LM test and so item errors were allowed to co-vary freely. The strength of the associations represented by the standardized estimates of the paths were judged according to Cohen's (1992)⁶⁶ criteria for multiple analysis of variance i.e., small=0.02, medium=0.15 and large=0.35. Proportions of effects and ratios of standardized path coefficients were calculated using a spreadsheet to facilitate relevant comparative statements.

Results

The sociodemographic characteristics of the population are set out in Table 1. Greater proportions of workers with lower education and income and proportionately more of the youngest workers can be seen among lower Pineo groups.

Variations in distributions of model variables across the four gender/occupation strata are apparent in Table 2. Covariance matrices of the analysis variables were too large for reproduction here (25 25 for each stratum) but are available from the authors upon request.

Adjustments to the hypothesized model were required to improve the fit of the stratum-specific models (Table 3). In addition to item correlations, LM tests suggested the addition of a path from self-esteem to mastery. All final models had a fair fit according to all the measures used, with less than 1% of the standardized residual having a value of greater than 0.2 in absolute value.

In the composite measurement models, work stressors most consistently reflected large contributions from lack of control (all paths above 0.5) and job insecurity (paths from 0.264 to 0.545). Intriguingly, psychological demands actually had negative relationships with work stressors except for lower Pineo men, though all were small in magnitude. Strengths of paths varied across gender/occupation strata, particularly for lack of social support: 0.039 among upper Pineo women to 0.630 among lower Pineo women, with intermediate values for men. Chronic stress was more important than recent life events for life stressors but both were of similar magnitudes across strata.

With respect to structural paths, both work and life stressors had significant positive total path coefficients to distress in all four stratum-specific models (see table 4). Work stressor total effects were consistently smaller (minimal to medium using Cohen's criterion) than life stressor total effects (all large). Ratios of work to life stressor total effects ranged from 0.01 for upper Pineo women to 0.33 for upper Pineo men, with lower Pineo women (0.20) and men (0.23) intermediate.

With respect to mediators, work and life stressor effects were consistently negative or absent, with effects on mastery consistently of greater magnitude than those on self-esteem. All of the effects of work stressors on distress were indirect i.e., mediated through mastery and self-esteem. In contrast, the majority of the effects of life stressors on distress were direct: indirect/total ratios from 0.54 among lower Pineo women to 0.93 among upper Pineo men, with lower Pineo men (0.60) and upper Pineo women (0.70) intermediate.

Discussion

Our analyses show associations of work stressors with levels of distress among the broad Canadian working population. Our results are consistent with literature showing relationships between work stressors in general, and job demands and lack of control in particular, and mental health outcomes.^31,51 Work stressor-distress relationships were almost entirely mediated by mastery and self-esteem. From an analytical perspective, these results support the inclusion of such mediator variables on the paths between work stressors as predictors and distress as an outcome rather than examining multiple independent associations between potential exposures, mediators and outcomes in a less-than-clear structure, resulting in the observation of “trivial” relationships.⁶⁷ From an applications perspective, our findings suggest that human resource interventions that either reduce work stressors, bolster self-esteem or increase mastery at work may have similar effects on distress.

The overall similarities in valence of structural paths across genders support the idea that the characteristics and dynamics of work stressors and mental health are similar for women and men.^28,29 The majority of structural paths were similar at different levels of occupational prestige, in keeping with other work stress literature.³¹ Our construction of work stressors as a composite variable demonstrated variation in contribution by different measures across the different strata. Paths from lack of control were the greatest in magnitude, consistent with the important role of control or decision latitude in predicting health impacts associated with hierarchies at work.⁶⁸ Yet they varied in magnitude particularly between women's occupational strata, indicative of variations in meaning or conditions noted by the scale developers.⁴¹ Job insecurity was the next most consistently important measure, in keeping with the strong health impact of labour market vulnerability.⁶⁹ Intriguingly, lack of work social support was the most important for lower Pineo women and of larger magnitude for this stratum than any other. The incongruous negative valence for psychological demands in three strata, may indicate the extent to which the reduced set of JCQ items inadequately represent the contribution of specific work stressors to the overall composite latent construct.

Unfortunately, imprecise measurement in the measures contributing to work stressors may have reduced structural path coefficients from work stressors to mediators and distress as well. The tradeoffs inherent in bargaining for inclusion of work stressor content in national surveys (fewer items versus nothing) are only too apparent. Without some items, analyses such as those presented here would not be possible, but with too few items the impact of work stressors may be underestimated relative to life stressors. Such underestimation was exacerbated by other measurement challenges. Financial stressors associated with changes at work and job promotion were both potential recent life events contributing to life stressors. Further, the life stressor domain represented more of the life course while comparative cumulative work stressor information over a working life was not available. Work histories with imputed scores from job-exposure matrices based on aggregated data, as have been used in cohort studies, would make this possible.⁷⁰

National Population Health Survey data were limited in other ways. Measures of work-family interaction or conflict⁷¹ were absent, despite demonstration of their effects on mental health⁷² and recognition of their growing importance.⁷³ Alternative formulations of relationships between stressors and mental health may also challenge our analyses. Some models interpret stressor valuation as consequent to levels of psychological symptoms⁷⁴ or to core evaluations, such as self-esteem, as determinants of life and job satisfaction.³ Without reference to corroborative data on stressor evaluations by others than the worker her/himself,^75,76 such alternative formulations are difficult to counter directly. Conceptually helpful are theoretical approaches that reinforce the extent to which self-esteem or other core evaluations can be assaulted by stressors and are thus outcomes rather than predictors.²⁶ Hence, contributions of our analyses to the understanding of cause must be tempered,⁶⁷ as the data set does not provide the precise measures, independently assessed, with clear temporal relationships (i.e., not cross-sectional) necessary to construct causal diagrams based on classic epidemiological contrasts.⁷⁷

Future research on work stressors and mental health must continue to struggle with measurement constructs and to work with modelling techniques that permit incorporation of measurement as well as structural relationships, as advocated by Hurrell and colleagues (1998).²⁰ In the meantime, promotion of “good” work organization⁷⁸ or “healthy workplaces”⁷⁹ as ways of reducing levels of work stressors is one of the applications of health research. Our analyses here provide support to mental health practitioners, workplace parties and policy makers taking actions on reducing work stressors, drawing on work stress intervention literature^80–82 and to existing health promotion campaigns to change determinants of health.⁸³

Acknowledgements

This project was sponsored by the Institute for Work & Health. The Institute, an independent, not-for-profit research organization, receives support from the Ontario Workplace Safety & Insurance Board. Financial support was provided by Health Canada NHRDP grant # 6606-6406.

References

1. Shilling S, Brackbill RM. Occupational health and safety risks and potential health consequences perceived by U.S. workers. Public Health Rep 1987;102:36–46.

2. Kalimo R, El-Batawi MA, Cooper GL Psychosocial factors at work and their relation to health. World Health Organization: Geneva, 1987.

3. Judge TA, Locke EA, Durham CC, Kluger AN. Dispositional effects on job and life satisfaction: the role of core evaluations. J Appl Psychol 1998;83(1):17–34.

4. Phelan J, Schwartz JE, Bromet EJ, Dew MA, Parkinson DK, Schulberg et al. Work stress, family stress and depression in professional and managerial employees. Psychol Med 1991;21:999–1012.

5. Frone MR, Russell M, Cooper ML. Antecedents and outcomes of work-family conflict: testing a model of the work-family interface. J Appl Psychol 1992;77(1):63–78.

6. Lai G. Work and family roles and psychological well-being in urban China. J Health Soc Behav 1995;36:11–37.

7. Pugliesi K. Work and well-being: gender differences in the psychological consequences of employment. J Health Soc Behav 1995;36:57–71.

8. Hendrix WH, Summers TP, Leap TL, Steel RP. Antecedents and organizational effectiveness outcomes of employee stress and health. In: Crandal R, Perrewe PL, editors. Occupational stress: a handbook. Washington, DC: Taylor & Francis, 1995:73–92.

9. Vermeulen M, Mustard C. Gender differences in job strain, social support at work and psychological distress. J Occup Health Psychol 2000;5(4):428–440.

10. Muntaner C, O'Campo PJ. A critical appraisal of the demand/control model of the psychosocial work environment: epistemological, social, behavioral and class considerations. Soc Sci Med 1993;36(11): 1509–1517.

11. Beaudet MP. Depression. Health Reports 1996;4:11–24.

12. Wade TJ, Cairney J. Age and depression in a nationally representative sample of Canadians: a preliminary look at the National Population Health Survey. Can J Public Health 1997;88(5):297–302.

13. Patten SB. Performance of the composite international diagnostic interview short form for major depression in community and clinical samples. Chronic Dis Can 1997;18(3):109–112.

14. Wilkins K, Beaudet MP. Work stress and health. Health Reports 1998;10(3):47–62.

15. Cole DC, Ibrahim SA, Shannon HS, Scott F, Eyles J, Goel V. Job strain, job satisfaction and emotional distress among Canadian workers: a gender analysis of the 1994 National Population Health Survey. PREMUS-ISEOH'98, Helsinki, Finland, 1998a.

16. Cole DC, Ibrahim SA, Shannon HS, Scott F, Eyles J, Goel V. Job demand/control, work factors and depressive episodes among Canadian workers: a gender analysis of the 1994 National Population Health Survey (NPHS). 1^st International ICOH Conference on Psychosocial Factors at Work. Copenhagen, Denmark, 1998b.

17. Hoyle RH, Smith GT. Formulating clinical research hypotheses as structural equation models: a conceptual overview. J Consult Clinl Psychol 1994;62(3):429–440.

18. Warr P. Study of psychological well-being. Br J Psychol 1978; 69:111–121.

19. Karasek R, Theorell T. Healthy Work: stress, productivity and the reconstruction of working life. New York, NY:Basic Books, 1990

20. Hurrel JJ, Nelson DL, Simmons BL. Measuring job stressors and strains: where we have been, where we are, and where we need to go. J Occup Health Psychol 1998; 3(4):368–389.

21. Turner RJ, Wheaton B, Lloyd D. The epidemiology of social stress. American Sociological Review 1995;60:104–125.

22. Kaplan GW, Roberts RE, Camacho TC, Coyne JC. Psychosocial predictors of depression. Prospective evidence from the Human Population Laboratory Studies. Am J Epidemiol 1987;125:206–220.

23. Kelloway EK, Barling J. Job characteristics, role stress and mental health. J Occup Psychol 1991;64:291–304.

24. Jex SM, Elacqua TC. Self esteem as a moderator: a comparison of global and organization-based measures. J Occup Organiz Psychol 1999;72:71–81.

25. Locke EA, McClear K, Knight D. Self-esteem and work. Int Rev Industrial and Organizational Psychology 1996;11:1–32.

26. Pearlin LI. Roles strains and personal stress. In: Kaplin HB, editor. Psychosocial Stress: Trends in Theory and Research. Orlando et al., Academic Press, Harcourt Brace Jovanovich, Pubs.,1983, 3–32.

27. Fenwick R, Tausig M. The macroeconomic context of job stress. J Health Soc Behav 1994;5:266–282.

28. Barnett RC, Marshall NL, Raudenbush SW, Brennan RT. Gender and the relationship between job experiences and psychological distress: a study of dual-earner couples. J Pers Soc Psychol 1993;64:794–806.

29. Schwartzberg NS, Dytell RS. Dual-earner families: the importance of work stress and family stress for psychological well-being. J Occup Health Psychol 1996;1(2):211–223.

30. Roxborough S. Gender differences in work and well-being: effects of exposure and vulnerability. J Health Soc Behav 1996;37: 6265–277.

31. Niedhammer I, Goldberg M, Leclerc A, Bugel I, David S. Psychosocial factors at work and subsequent depressive symptoms in the Gazel cohort. Scand J Work Environ Health 1998;24(3):197–205.

32. Stansfeld SA, Fuhrer R, Head J, Ferrie J, Shipley M. Work and psychiatric disorder in the Whitehall II Study. J Psychosom Res 1997;43(1):73–81.

33. Stephens T, Dulberg C, Joubert N. Mental health of the Canadian population: A comprehensive analysis. Chronic Dis Can 1999; 20(3):118–126.

34. Mergler D. Adjusting for gender differences in occupational health status. In: Messing K, Neis B, Dumais L, editors. Invisible: Issues in women's occupational health. Charlottetown, PEI: Gynergy Books, 1995:236–251.

35. Kilbom Å, Messing K, Thorbjörnsson CB. Women's health at work. In: Kilbom Å, Messing K, Thorbjörnsson CB, editors. Solna, Sweden: National Institute of Working Life, 1998.

36. Tambay JL, Catlin G. Sample design of the National Population Health Survey. Health Reports 1995;7(1):29–38.

37. Hood SC, Beaudet MP, Catlin G. A healthy outlook. Health Reports 1996;7:25–32.

38. Statistics Canada, Health Statistics Division. National population health survey, 1994-95, Public use microdata files. Ottawa, Ontario, 1995.

39. Pineo PC. Revisions of the Pineo-Porter-McRoberts socioeconomic classification of occupations for the 1981 census. Program for Quantitative Studies in Economics and Population (QSEP), Faculty of Social Sciences, McMaster University. QSEP Research Report No. 125. Hamilton, Ontario, 1985:17 pp.

40. Karasek R. Job content instrument questionnaire and user's guide. (unpublished work, 1985).

41. Karasek R, Brisson C, Kawakami N, Houtman I, Bongers P, Amick B. The Job Content Questionnaire (JCQ): an instrument for internationally comparative assessments of psychosocial job characteristics. J Occup Health Psychol 1998;3(4): 322–355.

42. Pearlin LI. The sociological study of stress. J Health Soc Behav 1989;30:241–256.

43. Wheaton B. Chronic Stress: models and measurement. Paper presented at meeting of the Society for the Study of Social Problems, Cincinnati, OH, 1991. Cited in Wheaton B (1994).

44. Dohrenwend BS, Krasnoff L, Askenasy AR, Dohrenwend BP. Exemplification of a method for scaling life events. The PERI Life Events Scale. J Health Soc Behav 1976;19:205–229.

45. Avison WR, Turner RJ. Stressful life events and depressive symptoms: Desegregating the effects of chronic stress and eventful stressors. J Health Soc Behav 1988;29: 253–264.

46. Turner RJ, Avison WR. Innovations in the measurement of life stress: Crisis theory and the significance of event resolution. J Health Soc Behav 1992;33:36–51.

47. Rosenberg M. Society and the adolescent self-image. Princeton, N.J. USA: Princeton University Press, 1965.

48. Pearlin LI, Lieberman MA, Menaghan EG, Mullan JT (1981). The stress process. J Health Soc Behav 1981;2:237–356.

49. Fletcher BC, Jones F. A refutation of Karasek's demand-discretion model of occupational stress with a range of dependent measures. J Organiz Behav 1993;14: 319–330.

50. Karasek RA, Triantis KP, Chaudrhy SS. Co-worker and supervisor support as moderators of associations between task characteristics and mental strain. J Occup Behav 1982;3:181–200.

51. Bourbonnais R, Brisson C, Moisan J, Vézina M. Job strain and psychological distress in white-collar workers. Scand J Work Environ Health 1996;22:139–145.

52. Amick BC, Kawachi I, Coakley EH, Lerner D, Levine S, Colditz GA Relationship of job strain and iso-strain to health status in a cohort of women in the United States. Scand J Work Environ Health 1998;24(1): 54–61.

53. Bollen K, Lennox R. Conventional wisdom on measurement: a structural equation perspective. Psychol Bull 1991;1110(2): 305–314.

54. MacCallum RC, Browne MW. The use of causal indicators in covariance structural models: some practical issues. Psychol Bull 1993;114:553–541.

55. Cohen P, Cohen J, Teresi J, Marchi M, Velez CN. Problems in the measurement of latent variables in structural equations causal models. Applied Psychological Measurement 1990;14(2):183–196.

56. Edwards JR, Bagozzi RP. On the nature and direction of relationships between constructs and measures. Psychological Methods 2000;5(2):155–174.

57. Bentler PM. EQS structural equations program manual. Encino, CA: Multivariate Software, Inc., 1995.

58. Kaplan D, Ferguson AJ. On the utilization of sample weights in latent variable models. Structural Equation Modeling 1999; 6:305–321.

59. SAS Institute Inc. SAS/STAT User's Guide. Version 6. SAS Institute Inc., Cary, NC, 1990.

60. Bryne BM. Structural equation modeling with EQS and EQS/Windows: basic concepts, applications, and programming. Thousand Oaks, CA: Sage, 1994

61. Hu LT, Bentler PM. Evaluating model fit. In: Hoyle RH, editor. Structural equation modeling: Concepts issues and applications. Newbury Park, CA: Sage, 1995: 76–99.

62. Chou CP, Bentler PM. Estimates and tests of fit in structural equation modeling. In: Hoyle RH, editor Structural equation modeling: concepts, issues and applications. Newbury Park, CA: Sage, 1995:37–55.

63. Fan X, Thompson B, Wang L. Effects of sample size and model specification on structural equation modeling fit indexes. Structural Equation Modeling 1999;6:56–83.

64. West SG, Finch JF, Curran PJ. Structural equation models with non-normal variables. Problems and Remedies. In: Hoyle RH, editor, Structural equation modeling: Concepts issues and applications. Newbury Park, CA: Sage, 1995:56–75.

65. Browne MW, Cudeck R. Testing structural equation models. In: Bollen KA, Long JS, editors. Alternative ways of assessing model fit. Newbury Park, CA: Sage, 1993:136–162.

66. Cohen J. A power primer. Psychol Bull 1992;112:155–159.

67. Kasl SV. Measuring job stressors and studying the health impact of the work environment: an epidemiologic commentary. J Occup Health Psychol 1998;3(4):390–401.

68. Bosma H, Marmot MG, Hemingway H, Nicholson AC, Brunner E, Stansfeld SA. Low job control and risk of coronary heart disease in Whitehall II (prospective cohort) study. Br Med J 1997;314(7080):558–565.

69. Lavis J, Amick B. Labour markets and health: a framework and set of applications. In: Tarlov A, editor. Health and its determinants. New York, NY: Free Press, 1999.

70. Johnson J, Stewart W, Fredlund P, Hall EM, Theorell T. Stress Research Reports: psychosocial job exposure matrix: an occupational aggregated attribution system for work environment exposure characteristics. National Institute for WHO Psychosocial Centre. Stockholm, Sweden, 1990.

71. Frone MR, Russell M, Barnes GM. Work-family conflict, gender and health-related outcomes: a study of employed parents in two community samples. J Occup Health Psychol 1996;1(1):57–69.

72. Bolger N, Delongis A, Kessler RC, Wetherington E. The contagion of stress across multiple roles. J Marriage and the Family 1989;51:175–183.

73. Westman M, Piotrkowski CS. Introduction to the special issue: work-family research in occupational health psychology. J Occup Health Psychol 1999;4(4):301–306.

74. Daniels K, Guppy A. Stressors, locus of control, and social support as consequences of affective psychological well-being. J Occup Health Psychol 1997;2(2): 156–174.

75. Spector PE, Fox S, Van Katwyk PT. The role of negative affectivity in employee reactions to job characteristics: bias effect or substantive effect. J Occup Organiz Psychol 1999;72:205–218.

76. Ostry AS, Marion S, Green L, Demers P, Hertzman C. Reliability and validity of two “expert” methods for measuring psychosocial job strain. Institute for Work & Health Working Paper # 107. Toronto, Institute for Work & Health. 2000;17 p.

77. Greenland S, Pearl J, Robins JM. Causal diagrams in epidemiological research. Epidemiology 1999;10(1):37–48.

78. Lindstöm K. Psychosocial criteria for good work organization. Scand J Work Environment Health 1994;20:123–133.

79. Robson LS, Polanyi MF, Kerr MS, Shannon HS, Eakin J, Brooker A-S, et al. What is a ‘healthy workplace?' In: Vink P, Koningsveld EAP, Dhondt S, editors. Proceedings from Human Factors in Organizational Design and Management. North-Holland: Elsevier Science, 1998:539–544.

80. Johnson JV, Johansson G, editors. The psychosocial work environment: work organization, democratization and health. Amityville, Baywood Publishing Company, Inc., 1991.

81. Karasek R. Stress prevention through work reorganization: a summary of 19 international case studies. ILO Conditions of Work Digest 1992;11(2),23–42.

82. Hepburn CG, Loughlin CA, Barling J. Coping with chronic work stress. In: Gottlieb BH, editor. Coping with chronic stress. New York, NY: Plenum Press, 1997:343–366.

83. Health Canada: Health Determinants Partnership. Making connections: social determinants of health posters/campaign. Canadian Government Publication, 1999.

Donald C Cole, Selahadin Ibrahim, Institute for Work & Health and Department of Public Health Sciences, Univeristy of Toronto, Toronto, Ontario

Harry S Shannon, Institute for Work & Health and Program in Occupational Health and Environmental Medicine, McMaster University, Hamilton, Ontario

Fran E Scott, John Eyles, Department of Clinical Epidemiology and Biostatistics and Institute of Environment and Health, McMaster University, Hamilton, Ontario

Correspondence: Dr Donald C Cole, Senior Scientist, Institute for Work & Health, 481 University Ave, Ste 800, Toronto, Ontario, Canada M5G 2E9;
Fax: (416) 927-4167; E-mail: dcole@iwh.on.ca

[Previous] [Table of Contents] [Next]