International Review of Industrial and Organizational Psychology 2011

By Gerard P. Hodgkinson and J. Kevin Ford

The twenty-sixth volume in the most prestigious series of annual volumes in the field of industrial and organizational psychology, providing authoritative and integrative reviews of the key literature in the field  

• All chapters written by established experts and all topics carefully chosen to reflect the major concerns in both the research literature and in current practice
• Presents developments in such established areas as stress and well-being, consumer behavior, employee trust, deception and applicant faking, the assessment of job performance and work attitudes, and the employment interview
• Newer topics explored include methodological issues in the development and evaluation of multiple regression models, and the psychological impact of the physical office environment
• Each chapter offers a comprehensive and critical survey of the chosen topic, and each is supported by a valuable bibliography

• Language: English
• Publisher: Wiley
• Release date: Apr 6, 2011
• ISBN: 9781119996248

Book preview

Chapter 1

STRESS AND WELL-BEING ARE STILL ISSUES AND SOMETHING STILL NEEDS TO BE DONE: OR WHY AGENCY AND INTERPRETATION ARE IMPORTANT FOR POLICY AND PRACTICE

Kevin Daniels

School of Business and Economics, Loughborough University, Leicestershire, LE11 3TU, UK

Stress and well-being are well-known, well-researched, and well-theorized areas of industrial and organizational (I/O) psychology. There are numerous reviews, meta-analyses, and influential publications on methods going back decades (e.g., Cass, Faragher, & Cooper, 2002; Cooper, Dewe, & O’Driscoll, 2001; Frese & Zapf, 1988; Jackson & Schuler, 1985; Kasl, 1983; Rick, Thomson, Briner, et al., 2002; Spector, Zapf, Chen, et al., 2000; Warr, 1987; Zapf, Dormann, & Frese, 1996). The area is characterized by high levels of methodological sophistication, including intervention studies (e.g., Jackson, 1983; Wall, Kemp, Jackson, et al., 1986), diary and experience sampling methods (e.g., Totterdell, Wood, & Wall, 2006; Xanthopoulou, Bakker, Demerouti, et al., 2009), multi-method designs (e.g., Frese, 1985), large-scale longitudinal studies (e.g., Bosma, Marmot, Hemingway, et al., 1997; Vahtera, Kivimäki, Pentti, et al., 2000), and qualitative investigations (e.g., Dick, 2000; Hepburn & Brown, 2001).

Moreover, all this scientific effort has found its way into policy and practice. For example, the World Health Organization, the International Labour Organization (ILO), and the European Union (EU) have all issued guidance emphasizing the need to assess causes of stress and take preventive action to eliminate these causes at source (ETUC, 2004; ILO, 2001; Leka & Cox, 2008; Leka, Griffiths, & Cox, 2003). Many developed countries have sophisticated national surveillance schemes to monitor work-related stress and well-being in the working population (Dollard, Skinner, Tuckey, et al., 2007). One particularly good example of how the science of stress has influenced policy and practice is the adoption by the UK’s Health and Safety Executive, the UK Government’s agency with responsibility for health and safety at work, of ‘The Management Standards for Work-Related Stress’ (see Cousins, MacKay, Clarke, et al., 2004; MacKay, Cousins, Kelly, et al., 2004). These Management Standards are underpinned by the applied psychological and epidemiological sciences related to stress and well-being.

The science has found its way into guidance and policy. Surely then those of us concerned with research in this area ought to find something else to do? In this chapter, I argue much still needs to be done and that recent, and not so recent but largely ignored, perspectives on stress and well-being illuminate ways in which research, policy, and practice can develop further. The rationale for making the statement that much still needs to be done can be broken down into four main areas.

First, this is still a lively area of research. With at least three specialist journals that take a psychological perspective on work stress and well-being (International Journal of Stress Management, Journal of Occupational Health Psychology, Work and Stress), and with relevant work continuing to appear frequently in some of the major I/O psychology, organizational behaviour, and management journals (e.g., Human Relations, Journal of Applied Psychology, Journal of Management, Journal of Occupational and Organizational Psychology, Journal of Organizational Behavior), it is clear there is still much research being published capable of meeting stringent requirements of peer review in international journals.

Secondly, the area is practically important. Clearly, worker health is important in its own right, notwithstanding healthcare costs (Manning, Jackson, & Fusilier, 1996) and potential relationships with work performance (Jex, 1998; Taris & Schreurs, 2009). A survey of 29 766 workers in 31 European countries found 22.3% claiming to be affected by stress at work (Parent-Thirion, Fernández Macías, Hurley, et al., 2007). In the UK, on the basis of various regular surveys (UK Labour Force Survey, The Health and Occupation Reporting network), the Health and Safety Executive claims in 2008–2009 that: 415 000 individuals in Britain suffered illnesses because of work stress; 16.7% of working Britons found their work stressful; there is an annual incidence rate of work-related stress, depression, or anxiety of 760 cases per 100 000 workers; and that 11.4 million working days were lost to stress. The interesting point about the UK picture is that the Health and Safety Executive claim that these figures have remained relatively stable over most of the past decade, in spite of increased awareness and the introduction of the Management Standards for Work-Related Stress in 2005–2006.¹

Thirdly, there remain questions over the ability of the dominant approach to research and policy, based on linking psychosocial job characteristics to health and well-being, to account for the experience of well-being at work (Briner, Harris, & Daniels, 2004; Hart & Cooper, 2001). If policy and practice are based on research that provides incomplete or inaccurate theoretical explanations, then it is worth asking whether policy and practice are sub-optimal and how they can be improved.

Fourthly, recent theoretical developments in the allied field of job design have indicated that I/O psychologists might be changing the way they think about work. Mirrored by changes to more knowledge intensive, interdependent, and information technology mediated work, the cognitive aspects of job design are theoretically more prominent (Grant, Fried, Parker, et al., 2010; Hodgkinson & Healey, 2008; Morgeson & Humphrey, 2006; Parker & Ohly, 2008). Moreover, there is growing research interest in the ways workers shape the content of their own jobs, also known as job crafting (Berg, Wrzesniewski, & Dutton, 2010; Clegg & Spencer, 2007; Wrzesniewski & Dutton, 2001). This interest in the cognitive aspects of job design and job crafting points to the importance of workers’ interpretation of their work and agency in shaping their work for understanding work-related health and well-being.

The purpose of this chapter is to examine in greater depth whether we can develop effective policy and practice based on the historically dominant approaches to job characteristics, well-being, and stress. I identify several weaknesses with this approach. I then describe alternative approaches to understanding job characteristics and their impact on well-being that place much greater emphasis on individual agency and interpretation. This also considers potentially adaptive forms of coping. I then examine the practical and methodological implications of these alternative approaches. I will start by outlining what stress and well-being are.

STRESS AND WELL-BEING AS AN AREA OF STUDY

Stress and psychological or subjective well-being are best conceived as areas or an area of study rather than as one or a set of dependent variables (Diener, Suh, Lucas, et al., 1999; Lazarus & Folkman, 1984). In I/O psychology, research generally relates to the impact of the work environment, in conjunction or in isolation of other processes, on psychological and psychosomatic health (see e.g., Cooper, Dewe, & O’Driscoll, 2001; Warr, 2007). Although research into stress has dominated I/O psychology, research on the positive influence of work on health is also important (Bakker, Schaufeli, Leiter, et al., 2008; Quick, Cooper, Gibbs, et al., 2010; Macik-Frey, Quick, & Nelson, 2007). Therefore, work in the areas of stress and well-being can be seen as complementary or even synonymous at a broad level.

One reason for treating stress and subjective well-being as areas of study rather than definable physiological or psychological states concerns specifying the exact nature of the concepts (Diener, Suh, Lucas, et al., 1999; Mason, 1975). For example, stress symptoms have been conceptualized as having physiological, behavioural, cognitive, and affective aspects (Cox, 1978). Subjective well-being is more focused on psychological states, but can include affective states, judgments of life or domain specific (e.g., job) satisfaction, cognitive states such as aspiration (Andrews & McKennell, 1980; Diener, 1984; Ryff & Keyes, 1995; Warr, 1990, 1994), and psychosomatic health (van Horn, Taris, Schaufeli, et al., 2004).

However, there is growing evidence that some aspects of health typically researched under the rubrics of stress and well-being share common features, some common influences, and the potential for common policy approaches to reduce their incidence and improve well-being (Lunt, Fox, Bowen, et al., 2007). These areas of health are known as ‘common health problems’ because they have a high incidence in the adult population in developed economies but do not have an easily identifiable physical basis or a well-established dose–response relationship between an identified pathogen and the disease outcome (Lunt, Fox, Bowen, et al., 2007). Examples include anxiety, depression, muscular–skeletal problems, and cardiorespiratory illnesses. Importantly, for work stress and well-being research, there is evidence that common health problems might all be linked to psychosocial work characteristics such as job autonomy and social support (Lunt, Fox, Bowen, et al., 2007). Moreover, there is evidence that specific aspects of work-related well-being, including cognitive, affective, and psychosomatic components, are related to a single underlying latent dimension, with the affective component having the closest relationship to this single dimension (van Horn, Taris, Schaufeli, et al., 2004).

There are other reasons for thinking that affect is a centrally important factor in research on stress and well-being. First, a common thread in much research in this area is that unpleasant affective reactions to work can be deleterious to psychological health, physical health, and behaviours related to performance (Danna & Griffin, 1999; Schwartz, Pickering, & Landsbergis, 1996; Spector & Goh, 2001). Secondly, affective reactions are thought to be the central components of well-being (Diener & Larsen, 1993; Warr, 1994) that mediate the influence of the environment on more cognitive and summative aspects of well-being, such as competence, aspiration, autonomy, integrative functioning, and satisfaction (Andrews & McKennell, 1980; Diener, 1984; Ryff & Keyes, 1995; Warr, 1990, 1994).

Affective well-being can be defined as the relative experience of pleasant affects to unpleasant affect (Diener & Larsen, 1993). This locates well-being as a phenomenological experience. The dominant view of affect represents affective experience as located in a two-dimensional space, where the dimensions are positive and negative affect (Watson & Tellegen, 1985). Specific affects cluster within this space (Haslam, 1995), so that more specific dimensions of affective well-being can be identified that are related to affective states such as anxiety, anger, tiredness, pleasure, and enthusiasm (Daniels, 2000). Affective well-being is domain specific (Warr, 1990). At least in relation to the work domain, the two major dimensions of affect may have different weightings in relation to affective well-being or be represented by a hierarchical structure (Cropanzano, Weiss, Hale, et al., 2003), with the more specific elements (e.g., anxiety, anger, enthusiasm) being most closely related to a superordinate dimension related closely to displeasure–pleasure (Daniels, 2000).

Poor affective experience is often a diagnostic criterion for poor psychological health (Lindsay & Powell, 1994) whilst good affective experience is an indicator of good subjective well-being. Physical health and well-being are not usually defined as central aspects of subjective well-being; rather, links between affect and physical health are explained by other processes. In tripartite models, factors related to the psychosocial work environment (e.g., job characteristics such as the level of job demands) and personality factors combine to produce an emotional response, which in turn might influence the course of a specific disease, if the person has a physiological susceptibility to develop that disease (Schwartz, Pickering, & Landsergis, 1996) or a vulnerability to engage in behaviours that might influence susceptibility (Lunt, Fox, Bowen, et al., 2007). This susceptibility need not be genetic, but can be transient (e.g., salt in diet) or a combination of both (e.g., cholesterol levels). Moreover, affective reactions to the work environment might also influence how symptoms are experienced, presented to medical practitioners, and diagnosed (Daniels, Jones, Perryman, et al., 2004; Lunt, Fox, Bowen, et al., 2007).

In summary, although stress and subjective well-being are complex areas of study that subsume a range of specific behavioural, physiological, cognitive, and affective elements, the centrality of affect to well-being may underpin a range of processes that suggest certain features of the work environment influence, however distally, a range of aspects of psychological and psychosomatic health and well-being.

MANAGING PSYCHOSOCIAL HAZARDS

Adverse job characteristics are also know as psychosocial hazards because of the amount of research indicating they are risk factors for poor well-being and ill-health (MacKay, Cousins, Kelly, et al., 2004; Rick et al., 2002). Examples include high job demands, low job control, low support from co-workers, lack of role clarity (Cousins et al., 2004; Rick, Thomson, Briner, et al., 2002) amongst many others embedded in typologies of job characteristics (e.g., Cooper & Marshall, 1976; Karasek & Theorell, 1990; Warr, 1987). Such psychosocial hazards are often treated as if they are objective features of the work environment (MacKay, Cousins, Kelly, et al., 2004; Schaubroeck, 1999), from which it is possible to determine workers’ exposure to the hazards, and then primary prevention strategies can be developed to limit the level of exposure to the hazard (Cox, Griffiths, Barlowe, et al., 2000; Schaubroeck, 1999). As psychosocial hazards have been identified as risk factors, a sensible approach to policy and practice might focus on controlling these hazards.²

This approach is embedded in policy and guidance on the implementation of policy from many supra-national organizations (ETUC, 2004; ILO, 2001; Leka & Cox, 2008; Leka, Griffiths, & Cox, 2003). In the EU, there is no distinction in law between risks to workers’ physical health and risks to psychological health, and employers are required to conduct risk assessments and take preventive action to reduce exposure to health and safety risks (ETUC, 2004). One of the most sophisticated approaches to policy on work-related stress has been developed and implemented in recent years in the UK. This is the UK Health and Safety Executive’s Management Standards for Work-Related Stress.

This approach provides a validated questionnaire to enable employers to assess psychosocial hazards in their workforce (an indicator tool), a process to take action when hazards are identified, and target states to be achieved in the workforce as a whole as indexed by the indicator tool (Cousins, MacKay, Clarke, et al., 2004; MacKay, Cousins, Kelly, et al., 2004). Because the Management Standards approach is targeted at controlling psychosocial hazards, rather than their consequences, and regulating their level against pre-determined targets, the approach is very strongly predicated upon changes in job and organizational processes in the effective management of stress and well-being (MacKay, Cousins, Kelly, et al., 2004). The Standards approach is concerned with regulating six psychosocial hazards: demands, (low) job control, (low) support, (poor) relationships at work, (poor) role clarity, and management of change. The approach has many positive features, not least of which include: the development of the standards from existing scientific evidence (Rick, Thomson, Briner, et al., 2002); the care, attention to detail, and involvement of multiple stakeholder group inherent in operationalizing the Standards (Cousins, MacKay, Clarke, et al., 2004); and the relative ease with which the Standards allow UK companies to comply with UK and EU Health and Safety Legislation in the area of psychological health and well-being.

Examining the assumptions underpinning the Standards provides a useful test case as to whether similar approaches could be effective in other countries, as well as whether the UK’s approach could be developed to enhance policy and guidance for managers.³ The Standards approach is based on several assumptions. These assumptions are not all unique to the Standards and many elements are present in World Health Organization, ILO, and EU guidance (ETUC, 2004; ILO, 2001; Leka & Cox, 2008; Leka, Griffiths, & Cox, 2003). Because the Standards are based on prevailing scientific practice, the Standards also reflect assumptions that are commonly held by I/O psychologists working in the work stress and health area, or at least assumptions concerning how the results of research can be turned into practice (see e.g., Hurrell, Nelson, & Simmons, 1998; Schaubroeck, 1999).

Some of the assumptions underpinning the Standards approach can be stated as follows. First, job characteristics are objective properties of jobs and relatively stable. Secondly, job characteristics have probabilistic relationships with well-being and health, which are usually assumed to be monotonic. Thirdly, organizations can take action so that relatively stable job characteristics can be changed to new, stable levels, which then will lead to improvements in health and well-being because stress-related problems will be prevented. Fourthly, because of probabilistic relationships between job characteristics and well-being and health, differences between individuals are largely irrelevant to the prevention of stress-related ill-health.⁴ However, each of these assumptions is problematic.

First Assumption: Job Characteristics are Objective Properties of Jobs and Relatively Stable

This assumption has two parts. First, job characteristics are objective. Put another way, job characteristics are properties of jobs and independent of the person doing the job – two people doing exactly the same job are assumed to be exposed to the same level of hazards. Secondly, job characteristics are relatively stable – over time, people more or less experience the same level of job characteristics. Without either part of this assumption, the idea that organizational action can lead to lasting improvements in jobs across people is not strictly tenable.

The assumption of objectivity is easy to question (Daniels, 2006). An overview of some of the main journals in this area indicates that the most prevalent means of assessing job characteristics is by questionnaire, and the use of a questionnaire to assess psychosocial hazards is embedded in the Management Standards approach (Cousins, MacKay, Clarke, et al., 2004; MacKay, Cousins, Kelly, et al., 2004). Many critiques of self-report measures have been published that highlight limitations such as self-reports are subject to numerous biases, including those linked to trait affect (Brief, Burke, George, et al., 1988; Spector, 1992) and social interaction (Salancik & Pfeffer, 1978).

However, other methods used to assess job characteristics do not return objective assessments either (see Daniels, 2006). Using reports from line managers or researchers’ observations of work merely reflects someone else’s perception of a job and cannot access cognitive activities that typify many jobs in advanced knowledge-based economies. Aggregating reports from all job holders with the same job title reflects shared perceptions rather than necessarily any objective reality. Even where self-reports from geographically separated jobs are aggregated, common institutional practices, such as socialization, training, or industrial relations, may condition perceptions (cf. Scott, 1995). Coding and analysing jobs based on nationally representative databases is also problematic, as such codes represent occupations rather than the local conditions of specific jobs in specific organizations.

A commonly suggested solution to the problem of imperfect measures of objective job characteristics is triangulation of results across several methods, on the assumption that common patterns of relationships indicate something about the objective nature of certain job characteristics. However, triangulation does not always work: job characteristics assessed by different methods do not always correlate sufficiently well to indicate the different methods are assessing the same thing, and different methods can produce associations with different indices of health or make independent contributions to predicting the same indicators of health (Daniels, 2006). These problems with attempting to assess ‘objective’ job characteristics led to the conclusion that a better strategy might be to abandon the idea that job characteristics somehow reflect unitary concepts, but that different methods reflect different facets of job characteristics (Daniels, 2006). For example, a job characteristic like job control does not exist per se; rather, different methods for assessing job control assess different facets of job control that may be interdependent but are nonetheless different from each other. In this scheme, self-reports of job characteristics assess perceptions of work, nothing more.

The assumption of stability is also questionable. Treating job characteristics as stable does not reflect the dynamics of organizational life (Peterson, 1998; Weiss & Cropanzano, 1996), that proximal events can have a stronger influence on well-being than more distal events (Pillow, Zautra, & Sandler, 1996), or empirical evidence that levels of job characteristics can change from week to week (e.g., Totterdell, Wood, & Wall, 2006), from day to day (e.g., Daniels & Harris, 2005), and even within the same day (e.g., Daniels, Boocock, Glover, et al., 2009). Moreover, changes in the nature of work, with a shift to more complex and interdependent jobs, have also questioned the assumption of treating jobs characteristics as stable (Grant & Parker, 2009). In the 27 European countries in 2005, including former Warsaw Pact transition economies, 81% of workers reported solving unforeseen problems of work, 69% reported learning new things at work, and 72% reported assessing the quality of their own work (Parent-Thirion, Fernández Macías, Hurley, et al., 2007). Such is the nature of modern working environments that Grant and Parker stated that changes in the nature of work ‘challenge fundamental assumption that employees passively carry out static jobs and tasks assigned by managers’ (2009: 342).

Second Assumption: Job Characteristics Have Probabilistic Relationships with Well-Being and Health

The number of reviews and meta-analyses synthesizing evidence indicates that there are reliable linear associations between indicators of job characteristics and indicators of well-being and health (e.g., Cass, Faragher, & Cooper, 2002; de Lange, Taris, Kompier, et al., 2003; Jackson & Schuler, 1985; Rick, Thomson, Briner, et al., 2002). The problem here is not so much whether the assumption is tenable, rather whether the probabilities are high enough and there is enough specificity in the nature of the relationships to be able to develop organizational practices that will improve health and well-being (Rick & Briner, 2000). The problem of poor-specificity and poorly specified dose–response relationships is well known in the area (de Jonge & Dormann, 2006; Lunt, Fox, Bowen, et al., 2007). Most of the policy guidance in the area is typified by a tacit assumption of linear relationships between job characteristics and health, although more complex relationships are entirely feasible (Karanika-Murray, Antoniou, Michaelides, et al., 2009).

One problem with poor specificity is that in any one instance of organizational change, it will not be possible to specify a priori what will be the best health and well-being outcomes to measure in order to determine whether the changes actually had any impact on well-being. To some extent, this can be overcome in assessment studies by taking a broad range of measures and including affective well-being amongst this range, because of its centrality to well-being. A more important problem is that even if the idea is tenable that changing job characteristics can improve well-being at a population level, because of poorly specified and stochastic relationships, managers might be required by health and safety law to implement costly organizational change to redesign jobs to protect health with no guarantee of any local health benefits, yet the changes may be sub-optimal for operational and financial performance.

Third Assumption: Organizations can Take Action which then Will Lead to Improvements in Health and Well-Being

Perhaps reflecting problems with the appropriate level of specificity to take organizational action to effect beneficial changes in job characteristics, there are questions over the extent to which organizational changes can influence well-being. In a recent review of 90 intervention studies, LaMontagne, Keegel, Louie, et al. (2007) concluded both organizationally focused and individually focused interventions (e.g. coping skills training) had positive effects on employee health. In a systematic review of interventions focused on task restructuring, Bambra, Egan, Thomas, et al. (2007) concluded interventions focused on autonomy had the strongest effects on health. However, in their review of 19 interventions, the strongest evidence came from task restructuring that reduced job autonomy and was accompanied by subsequent deterioration in health rather than from interventions designed to improve job autonomy. Moreover, across all studies, less than 30% of health outcome measures changed in the predicted direction. Two studies found health outcomes changed in the direction opposite to that predicted. Egan, Bambra, Petticrew, et al. (2007) reviewed 18 intervention studies and concluded interventions with a focus on improving workers’ opportunities to exercise control and participate in decisions improve health. Two studies found evidence of deterioration in health subsequent to the intervention, which Egan et al. attributed to confounding due to organizational downsizing. In the other studies, less than 45% of the health outcomes assessed changed in the predicted direction. All of these reviews included studies with either no control groups or non-equivalent control groups.

Randomized control trials reflect the most rigorous designs to make causal inferences concerning the effectiveness of interventions. In a meta-analysis of randomized controlled trials of interventions with working adults from healthy populations, Richardson and Rothstein (2008) found that organizational-level interventions did not have a statistically reliable association with outcome measures (d = 0.14, ns), whereas individually focused interventions tended to have statistically reliable associations with outcome measures. Cognitive–behavioural interventions demonstrated the strongest impact (d = 1.16, p <0.01). Although Richardson and Rothstein found only five studies of organizational interventions that met the criteria for inclusion in the study, their results do raise questions concerning the effectiveness of organizational level interventions given null results under the most rigorous conditions of random allocation to treatment and control groups. Given the findings of the other recent reviews (Bambra, Egan, Thomas, et al., 2007; Egan, Bambra, Petticrew, et al., 2007; LaMontagne, Keegel, Louie, et al., 2007), the best that can be said is that interventions aimed at changing work and organizational structures, practices, and processes might have a subsequent beneficial effect on some health outcomes, but these effects might be specific to interventions that target only certain job characteristics rather than a broad range, and there is a small chance of interventions producing harmful effects.

Fourth Assumption: Differences between Individuals are Largely Irrelevant to the Prevention of Stress-Related Ill-Health

A major issue in developing policy on work-related stress is the extent to which self-reports of health and job characteristics reflect real differences in health and jobs rather than differences between individuals in their personality, attitudes, self-concept, etc. (Daniels, Jones, Perryman, et al., 2004). One reason for this is because self-reports are one of the most efficient means of collecting data from the population in order to monitor the impact of policies developed to reduce stress and increase well-being. Indeed, one major debate has centred on negatively oriented personality traits, and whether failure to control for trait negative affect leads to inflated correlations between self-reports of job characteristics and health and well-being (e.g., Spector, Zapf, Chen, et al., 2000). However, it is also likely that differences between individuals could have a substantive role in the relationship between work and well-being (Spector, Zapf, Chen, et al., 2000). The question for policy then, is not so much whether individual differences influence the development of stress-related ill-health, but whether work has a sufficiently large influence on stress-related ill-health, after taking into account differences between individuals, to warrant interventions targeted at jobs and organizations.

A recent meta-analysis examined 19 longitudinal studies that assessed reports of psychological and physical symptoms, adverse job characteristics and negative oriented personality traits (Ferguson, Daniels, & Jones, 2006). Negatively oriented personality refers to individual differences such as neuroticism, negative affectivity, and low self-esteem. In meta-regression analyses controlling for baseline measures of symptom reports, baseline reports of adverse job characteristics did not have a significant relationship with subsequent assessment of physical symptoms (β = 0.01, ns) but did have a relationship with subsequent reports of psychological symptoms (β = 0.05, p <0.05). In contrast, baseline measures of negatively oriented personality had relationships with subsequent reports of both physical symptoms (β = 0.05, p <0.01) and psychological symptoms (β = 0.09, p <0.01). In all cases, there were relationships between baseline assessments of symptoms and adverse job characteristics, after controlling for negative oriented personality, but the relationships between baseline symptoms and negatively oriented personality were higher. The results from this meta-analysis indicate that the strongest relationships were between negatively oriented personality and symptom reports and that adverse job characteristics did not have a predictive relationship with physical symptoms.

Another issue tied up with the Management Standards and other approaches based on preventing ill-health by changing job characteristics is that they do not address the issue of people who may be in work but are suffering from some form of illness. The issue is significant. For example, over 30% of workers were estimated to have a health-related work limitation in one study in the USA (Lerner, Amick, Malspeis, et al., 2000). Further, in any one workplace, the levels of people suffering from chronic illnesses may be underestimated as not everyone discloses their illness to their line manager (Munir, Leka, & Griffiths, 2005). Moreover, it cannot be assumed that interventions designed to protect health will have a curative effect for those that have already developed health problems.

Workers’ illnesses need not have their generation in work. However, work may affect levels of well-being for workers with illnesses. Munir, Yarker, Haslam, et al. (2007) found that after taking into account illness type and severity, distress in those at work with chronic illness was related to restrictions in ability to perform work due to illness, presenteeism (i.e., attending work when feeling unwell), poor management of illness at work, and low workplace support. Reasons for higher levels of presenteeism amongst those with a chronic illness could be attributed, at least partly, to people trying to avoid disciplinary action for absences (Munir, Yarker, & Haslam, 2008). Workers may be more likely to disclose illness if they feel they may receive support (Munir, Leka, & Griffiths, 2005), although those suffering from anxiety and depression are less likely to receive support in the form of adjustments to their work tasks to take into account their illness (Munir, Jones, Leka, & Griffiths, 2005).

Summary

Approaches to enhancing well-being and reducing stress-related ill-health based on the management of psychosocial hazards or job characteristics may be sub-optimal because some of the underpinning assumptions are questionable, at least to some extent. First, job characteristics may not necessarily be objective and stable properties of jobs that are independent of the person performing the job; secondly, the low specificity and stochastic nature of dose–response relationships between psychosocial hazards and specific outcomes means that attempts to alter stable and objective properties of jobs are not guaranteed to be successful; thirdly, the evidence from intervention studies is that changing psychosocial job characteristics may have limited success in protecting health and well-being; fourthly, differences between individuals, including in their health status, are important factors in preventing stress-related ill-health and enhancing well-being.

In the following sections, perspectives will be introduced that go someway to addressing these problems. These perspectives emphasize the agency of workers in shaping their jobs, the interpretation of work events and adaptive adjustment to adverse work conditions, and the role of workers, individually and collectively, in helping or hindering the effectiveness of interventions designed to protect health and improve well-being through altering work characteristics.

REVISIONING JOB CHARACTERISTICS

If job characteristics are neither necessarily stable nor objective, then there are questions as to what measures of job characteristics actually assess. To address this issue, Daniels (2006) suggested abandoning the idea of a unitary concept of job characteristics, instead breaking them down into different, interdependent facets that correspond to what is actually being measured by different methods. For example, self-reports of job autonomy assessed by questionnaire provide information on one facet of job autonomy, but managers’ reports of subordinates’ autonomy provide information on another facet of job autonomy. Daniels suggested three facets of job characteristics: latent, which reflects facets of job characteristics embedded in organizational and technological processes; perceived, which can be disaggregated into a person’s perception of their own job and a person’s perception of someone’s else job; and enacted job characteristics. Enacted job characteristics are jobs as they come to be shaped by job incumbents and others, and in Daniels’ (2006) scheme, reflect how interdependencies with others and proactive behaviour come to shape jobs (cf. Grant & Parker, 2009). In the following sections, each facet of job characteristics identified by Daniels (2006) is described in more detail.

Latent Job Characteristics

Latent job characteristics are the institutional and technological processes that influence work tasks (Daniels, 2006). Examples might be contractual hours as indicators of quantitative workload or temporary versus permanent employment contracts as an indicator of job security. In some ways, these might be considered to be the most ‘objective’ indicators of job characteristics, as they can be assessed with techniques that do not rely on perceptual data. However, it is unlikely that assessing latent job characteristics provides a sound basis for finding close relationships between work and well-being or health. This is because they neither reflect what people do, how they perceive what they do, nor what happens to them, all of which could have closer relationships with well-being and health. For example, taking contractual work hours as an index of quantitative workload does not reflect how many hours many people work and whether they volunteer or feel coerced into working longer hours. The choice to work longer hours itself may alter the perceptions of working longer hours, thus possibly influencing well-being and health (Lazarus, 1999). However, latent job characteristics may provide constraints within which people act, but may themselves be subject to change as individuals, alone or collectively, formally renegotiate contractual terms, organizational processes, and how technologies are used (Grant & Parker, 2009). Latent job characteristics may also provide information that influences perceptions of job characteristics (Daniels, 2006).

Perceived Job Characteristics

Perceived job characteristics reflect generalized perceptions of a person’s job, and are assessed by reports of job characteristics often collected by structured questionnaire measures that ask observers to rate jobs as they are perceived to be over an extended or unspecified period (Daniels, 2006). Potential perceivers include the job incumbent, co-workers, managers, customers, and researchers. Although researchers’ perceptions are often treated as if they reflect an objective assessment of a job, they are still perceptual data and are likely to provide only a partial picture of what it is people do in their day-to-day work.

While researchers’ perceptions may not influence incumbents’, managers’, co-workers’, or customers’ perceptions, they are likely to be mutual influences between the perceptions of those people that interact with the job incumbent in the course of performing their job. Partly this might because they are all observing the same phenomena; but also because there may be a shared awareness of latent job characteristics; because of social information exchange within an organization (Salancik & Pfeffer, 1978); because of shared socialization, training, and other processes that may be common to a job within an organization, sector, or profession (DiMaggio & Powell, 1983); and/or partly because powerful groups of workers are able to claim and then protect the right to certain job characteristics, such as skill use, regardless of job descriptions, technology, and other factors related to latent job characteristics (Noon & Blyton, 1997).

Research indicates that self-reports of job characteristics may have a predictive relationship with indicators of psychological health at least (Ferguson, Daniels, & Jones, 2006). However, how job incumbents generally perceive their work need not be the process through which work influences health and well-being. It is possible that incumbents’ and others’ perceptions influence what happens on a moment-by-moment basis, within weak or strong constraints of latent job characteristics, and these dynamic processes as jobs are enacted are the appropriate units of analysis for studying the psychological and physiological processes linking work to well-being and health.

Enacted Job Characteristics

Enacted job characteristics are events that reflect jobs as they happen (Daniels, 2006). The idea of enacted job characteristics is strongly influenced by the notion of job crafting (Wrzesniewski & Dutton, 2001) and the idea that dynamic events rather than static conditions offer more potential for explaining the experience of work (Peterson, 1998; Weiss & Cropanzano, 1996). Enacted job characteristics are viewed as dynamic and emergent from how an individual does a job, rather than as static and determined by organizational and technological processes. They can reflect behavioural, discursive, or cognitive processes. Daniels (2006) considers that enacted job characteristics are influenced by the job incumbent’s own perceptions of what the job is or should be, what others think the incumbent’s job is or should be, and latent job characteristics.

The incumbent’s own perception of his or her job is an important influence on enacted job characteristics because individuals are active in interpreting and shaping their jobs through their behaviour (Wrzesniewski & Dutton, 2001). For example, a person may decide to leave work early because he or she believes that he or she has autonomy over his or her work schedules. Co-workers can behave in ways that influence enacted job characteristics (Grant & Parker, 2009), for example by providing ambiguous information where the incumbent’s tasks are dependent upon clear communication from co-workers. Moreover, as well as acting as individuals, workers may collectively enact certain job characteristics (Grant & Parker, 2009). Managers are likely to be particularly influential in respect of enacted job characteristics, because of normative perceptions concerning influence over others (cf. Clegg & Spencer, 2007).

Latent job characteristics may act as constraints or facilitators on what can be enacted. Some latent job features, relating to jobs that require high levels of training, skill use, and discretion, may stimulate individuals to enact specific job characteristics to shape what they do and the physical environment in which they work (Grant & Parker, 2009; Pierce, Jussila, & Cummings, 2009). In turn, enacting beneficial job characteristics may lead to circular and dynamic changes in work, so that the potential to enact more beneficial job characteristics is increased (Clegg & Spencer, 2007).

As enacted job characteristics are behaviours as they occur, they may be seen as discrete events or episodes that influence affective experience at work (Weiss & Cropanzano, 1996) and that emerge and change over the course of a working day (Daniels, 2006). Support for giving greater causal credence to recent, on-going events comes from studies of well-being that show more recent events have a stronger effect than events in the more distant past (Suh, Diener, & Fujita, 1996) and that show the impact of major life events on well-being is mediated by daily events (Pillow, Zautra, & Sandler, 1996).

Enacted job characteristics as events or episodes can be thought of in two ways (Daniels, 2006; Daniels, Harris, & Briner, 2004): first, as the locus of cognitive processes that shape affective reactions to events and episodes (Lazarus, 1999); secondly, individuals might enact job characteristics for specific purposes. The first view reflects the many cognitive theories of affect in which the appraisal of how events alter progress towards personal goals is a key determinant of affective reactions to those events (Power & Dalgleish, 2008). These cognitive processes are one way in which enacted job characteristics might influence health through an impact on the cognitive experience of affect and attendant changes in physiological processes and behaviours. The second view posits that individuals enact job characteristics to achieve specific goals or meet particular needs, such as to exercise control, find meaning in work, develop a positive self-image, affiliate with others, or to make the tasks of work consonant with the incumbent’s self-identity (Pierce, Jussila, & Cummings, 2009; Wrzesniewski & Dutton, 2001). Enacting job characteristics for the purposes of achieving specific goals or meeting particular needs echoes goal based theories of affect. However, job characteristics might be enacted by job incumbents specifically to cope with the other enacted job characteristics (Daniels, Harris & Briner, 2004). For example, a person many enact job control to take a break from working on a complex and demanding task that a manager has requested be completed.

One important differentiation between the notion of enacted job characteristics and more traditional approaches that underpin much science and policy is that enacted job characteristics relate to a person doing a job at a specific point in time. In contrast, the dominant approach treats the person and the job as separate entities. This dualist notion of job characteristics and workers inherent in the dominant view is perhaps both artificial and too far from reality to be practically tenable.

The notion of enacted job characteristics, as they apply to understanding stress and well-being, gives central importance to workers as active sense makers of their work as it happens and as agents that can shape their work. This need not imply deliberate and conscious choices on the part of workers; interpretation and agency can also reflect unconscious and habitual processes. In the following sections, cognitive processes, both conscious and deliberate and unconscious and automatic, will be outlined. These processes may help explain both how enacted job characteristics are interpreted and the choices that lead to enacting job characteristics to cope with potentially harmful events and episodes as they occur at work.

REVISIONING THE WORKER AS AN ACTIVE INTERPRETER

The idea that stress and well-being are influenced by how we interpret events and episodes and not the events and episodes themselves has a long history in psychology (for reviews see e.g., Lazarus, 1999; Power & Dalgleish, 2008). Central to many interpretive approaches is the idea that how individuals appraise episodes and events as relevant to their goals influences their affective experience, in turn influencing their well-being (for a detailed review see Power & Dalgleish, 2008). Inherent in these models is the idea that affect signals whether things are going well, and therefore cognitive resources can be directed elsewhere, or whether things are going not so well and something needs to be done (e.g., Frijda, 1986; Oatley & Johnson-Laird, 1987). Perceived rate of progress toward a goal might influence the level of positive affect and perceived impediments might influence negative affect (Carver & Scheier, 1990).

There is evidence to indicate that appraisals of work-related goal progress are associated with well-being (e.g., ter Doest, Maes, Gebhardt, et al., 2006), although these relationships may be moderated by a number of factors. Work-related well-being seems to be associated most closely with progress towards goals that are considered to be more important (Harris, Daniels, & Briner, 2003) or consonant with other goals (Kehr, 2003), including higher order goals close to the self-concept (Maier & Brunstein, 2001). Similarly, well-being seems to be more closely associated with goal progress when progress exceeds expectations, because goals are appraised as difficult (Wiese & Freund, 2005) or because initial appraisals of the likelihood of attaining the goals were low (Pomaki, Karoly, & Maes, 2009).

However, appraisal theories have largely been ignored, or downplayed, by I/O psychologists (see, e.g., Cooper, Dewe, & O’Driscoll, 2001; Dewe & Cooper, 2007). There may be two main reasons for this. First, appraisal-based theories give primacy to individual interpretation. For some I/O psychologists and stakeholders, such as trades union officials, focusing on individual processes may seem like ‘blaming the victim’ for exposure to adverse work conditions (cf. Murphy, 1988). Secondly, assessing job characteristics that most people would ‘appraise’ in the same way might lead to the development of interventions based on changing those job characteristics that most people would appraise in the same way. One basis for this assertion is that work occupies a unique niche because of the economic transaction that takes place, rendering it a situation ripe for common interpretation (Brief & George, 1991). The problem with this perspective is that work might mean different things to different people, with the economic motive being one of only several reasons why people apply for different jobs and pursue different careers (cf. Holland, 1985).

This does not mean that common patterns of appraisals within work groups are unimportant, but rather these patterns might be localized to specific work groups within organizations (Harris, 2002). Moreover, examining individual patterns of appraisals may be one way to incorporate differences between individuals, whether these differences are due to personality traits or health status, more fully into research, policy, and practice in work stress and well-being.

In considering appraisal based theories, four questions are pertinent: (i) What is being appraised? (ii) How do appraisals influence well-being? (iii) How do appraisals come to be? (iv) What happens after appraisal? A fifth question, how people decide to cope with what is being appraised, is dealt with in another section.

What Is Being Appraised?

In most appraisal theories, discrete events or episodes can form the basis of appraisals (e.g., Lazarus, 1999; Power & Dalgleish, 2008). This fits well with notions of enacted job characteristics, which are conceptualized as dynamic. The idea that discrete workplace events or episodes form the basis of affective reactions gained a lot of momentum with the publication of Affective Events Theory (AET, Weiss & Cropanzano, 1996) and subsequent support and extensions of AET (e.g., Fisher, 2000, 2002; Fuller, Stanton, Fisher, et al., 2003). AET suggests that work-related attitudes and behaviours are influenced by work-related affect and that work events mediate the impact of more stable on-going work conditions on affect. AET positions affect as the central reactive phenomenon to work events, which is consistent with viewing affective experience as central to well-being (cf. van Horn, Taris, Schaufeli, et al., 2004) and appraisal theories that are concerned with explaining affective reactions (Power & Dalgleish, 2008).

How Do Appraisals Influence Well-being?

Goal-based appraisal theories indicate that events appraised as influencing goals influence affect (e.g., Lazarus, 1999). In recent dual-process approaches, appraisals are conceived of operating through slow, deliberative, and detailed cognitive processes, or on the basis of fast, restricted, unconscious cognitive processing (Power & Dalgleish, 2008; Smith & Kirby, 2001). These two processes of appraisal reflect the distinction between controlled and automatic processing (Schneider & Chein, 2003; Schneider & Shiffrin, 1977). Controlled processing is constrained by attentional limits, but is flexible and under greater intentional control than automatic processing, and often includes conscious thought. Controlled processing can be influenced by information held in long-term memory, but is influenced more by information attended to in the environment than automatic processing. Automatic processing is unintentional, involuntary, occurs outside of awareness, is only minimally influenced by information in the environment, and there is a strong influence of information recalled from long-term memory. Automatic processing is less effortful, faster, and affords parallel processing of multiple cognitive tasks, but is less flexible or adaptable to changes in the environment than controlled processing. Distinguishing controlled from automatic processes does show promise for broadening our understanding of job design and affect in the workplace (Elfenbein, 2007; George, 2009; Parker & Ohly, 2008).

One of the more sophisticated dual-process models of appraisals has been developed by Power and Dalgleish (2008). In their model, Power and Dalgleish proposed that mental models, or networks of beliefs concerning how events influence goals or affect, are the basis upon which work events are appraised and consequently alter the cognitive experience of affect. Daniels, Harris, and Briner (2004) extended Power and Dalgleish’s model in three ways. First, Daniels, Harris, and Briner developed the model to be applicable to work contexts. Secondly, they specified the processes by which work place events and episodes become the focus