Perceived Job Insecurity and Health Across the Life Course · Perceived Job Insecurity and Health Across the Life Course . ... Tausig. I am very grateful ... Perceived Job Insecurity
Post on 03-May-2018
221 Views
Preview:
Transcript
Perceived Job Insecurity and Health Across the Life Course
by
Paul Glavin
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy
Sociology University of Toronto
© Copyright by Paul Glavin 2012
ii
Perceived Job Insecurity and Health Across the Life Course
Paul Glavin
Doctor of Philosophy
Sociology University of Toronto
2012
Abstract
Job loss and unemployment have been consistently shown to have deleterious consequences for
health. However, less is known about how insecure employment experiences and the threat of
job loss influence well-being. Given the high levels of uncertainty associated with threatened and
insecure employment, perceptions of job insecurity are thought to constitute a potent form of
work stress because the ambiguity over a future undesirable event—job loss—undermines
coping strategies and attempts at stress reduction. It has been suggested, then, that health
penalties should be greatest with prolonged exposure to this threat. Further, since the meaning of
job loss likely varies across working life, individual reactions may be contingent on life course
position. Drawing from the Stress Process Model and the life course perspective, this dissertation
explores whether the two factors of timing and duration influence the health penalties associated
with perceived job insecurity, along with its impact on personality traits that are fundamental to
well-being.
iii
Findings reveal the detrimental social-psychological and health implications of perceived job
insecurity based on a national panel study of American workers surveyed in 2005 and 2007.
Health penalties associated with perceived job insecurity are greatest for middle age workers
reporting prolonged exposure to the threat of job loss. In addition, a personality trait—a high
sense of personal control over one’s life—is demonstrated to alleviate the stress of perceived job
insecurity; but this trait is itself prone to erosion with prolonged exposure to insecure
employment. Collectively, this dissertation contributes to knowledge about the social-
psychological processes through which insecure employment impacts individual well-being, and
how these processes are shaped by age as a key social status and life course marker.
iv
Acknowledgements
This dissertation would not have been possible without the support and guidance from a long list
of people. I am very grateful for their help and encouragement throughout the project, and offer
them my wholehearted thanks.
I thank Scott Schieman as my mentor and chair of the dissertation. His guidance throughout the
course of my academic training has been critical; he has influenced every aspect of my scholarly
development and I cannot express my gratitude to him enough. To my dissertation committee
members, Sandy Welsh and Ronit Dinovitzer, I thank you for your time and valuable insights on
the dissertation. Thanks also go to my internal and external examiners, Brent Berry and Mark
Tausig. I am very grateful for their support and their willingness to participate in the dissertation.
I would like to thank all my friends for their support and understanding throughout my graduate
training. I could not have gotten through the process without them. To my family, I thank them
for their love, guidance and belief in me. Without the support from my parents and my brothers
and sisters, the idea of pursuing a PhD would have been nothing but a dream. I thank them for
giving me the confidence and desire to seek out and achieve my life goals.
Finally, I thank Marisa for her unyielding support. The journey here was not always an easy one.
But you have always been at my side, through good and bad, and I know that without you, this
dissertation would not have been completed. I thank you for your love, patience and above all,
your willingness to be there for me when I needed you most.
v
Table of Contents Chapter 1. Introduction................................................................................................................ 1
Theoretical and Methodological Contributions ......................................................................... 2
The Stress Process Model ................................................................................................... 3
The Life Course Perspective ............................................................................................... 4
Age Contingencies in the Stress of Job Insecurity .............................................................. 5
Overview of the Dissertation ..................................................................................................... 6
Goals of the Dissertation ............................................................................................................ 7
References ....................................................................................................................................... 8
Chapter 2. Perceived Job Insecurity and Health Across the Life Course ............................. 10
Introduction .............................................................................................................................. 10
Literature Review ..................................................................................................................... 12
Perceived Job Insecurity and Health ................................................................................. 12
Job Insecurity Across the Life Course .............................................................................. 14
Job Insecurity and Absenteeism ........................................................................................ 18
Data and Methods .................................................................................................................... 20
Measures ........................................................................................................................... 20
Plan of Analyses ............................................................................................................... 24
Results ...................................................................................................................................... 25
Descriptives ....................................................................................................................... 25
Multivariate Analyses ....................................................................................................... 25
Discussion ................................................................................................................................ 27
Conclusion................................................................................................................................ 31
References ..................................................................................................................................... 32
vi
Chapter 3. Control in the Face of Uncertainty: The Sense of Personal Control and the Link between Job Insecurity and Distress ........................................................................... 44
Introduction .............................................................................................................................. 44
Literature Review ..................................................................................................................... 45
Job Insecurity and Health .................................................................................................. 45
Does the Sense of Control Buffer the Stress of Job insecurity? ....................................... 46
Sense of Control as a Mediator ......................................................................................... 48
Age Contingencies ............................................................................................................ 49
Data and Methods .................................................................................................................... 50
Measures ........................................................................................................................... 51
Plan of Analyses ............................................................................................................... 54
Results ...................................................................................................................................... 55
Descriptives ....................................................................................................................... 55
Multivariate Analyses ....................................................................................................... 55
Discussion ................................................................................................................................ 57
Age Contingencies ............................................................................................................ 59
American-Canadian Comparisons .................................................................................... 60
Conclusion................................................................................................................................ 62
Notes ........................................................................................................................................ 63
References ..................................................................................................................................... 64
Chapter 4. The Consequences of Job Insecurity and Job Degradation for the Sense of Personal Control..................................................................................................................... 74
Introduction .............................................................................................................................. 74
Literature Review ..................................................................................................................... 75
The Sense of Personal Control .......................................................................................... 75
Job Insecurity and Feelings of Powerlessness .................................................................. 76
Job Degradation as a Mediator ......................................................................................... 78
vii
Contingencies: Duration and Life Course Timing ............................................................ 80
Data and Methods .................................................................................................................... 81
Measures ........................................................................................................................... 82
Plan of Analyses ............................................................................................................... 84
Results ...................................................................................................................................... 85
Descriptives ....................................................................................................................... 85
Multivariate Analyses ....................................................................................................... 86
Discussion ................................................................................................................................ 88
Duration and Timing ......................................................................................................... 88
The Contribution of Job Degradation ............................................................................... 90
Conclusion................................................................................................................................ 92
References ..................................................................................................................................... 93
Chapter 5. Conclusions and Discussion .................................................................................. 101
Overview of Contribution and Themes .................................................................................. 101
Declining Job Security .................................................................................................... 101
The Duration and Life Course Timing of Job Insecurity ................................................ 102
The Sense of Control and the Link between Job Insecurity and Distress ....................... 102
Examining the Sense of Control as an Outcome of Job Insecurity ................................. 103
Summary ................................................................................................................................ 104
Limitations ............................................................................................................................. 105
Future Research ...................................................................................................................... 107
References ................................................................................................................................... 110
Appendix I .................................................................................................................................. 112
viii
List of Tables Table 2.1 Means and Proportions of all Study Variables ............................................................. 37
Table 2.2 Regression of Psychological Distress, Anger, and Self-rated Poor Health at wave 2 on
Perceived Job Insecurity ............................................................................................................... 38
Table 2.3 Regression of Days Absent from Work at wave 2 (logged) on Perceived Job Insecurity
....................................................................................................................................................... 40
Table 3.1 Means and Proportions of all CAN-WSH Study Variables .......................................... 68
Table 3.2 Means and Proportions of all WSH Study Variables ................................................... 69
Table 3.3 Regression of Psychological Distress on Focal Independent Variables (CAN-WSH) . 70
Table 3.4 Regression of Psychological Distress on Focal Independent Variables (WSH)........... 72
Table 4.1 Means and Proportions of all Study Variables ............................................................. 97
Table 4.2 Regression of Excessive Job Pressures and Decision-making Latitude at Wave 2 on
Perceived Job Insecurity ............................................................................................................... 98
Table 4.3 Regression of Sense of Personal Control at wave 2 on Perceived Job Insecurity ........ 99
ix
List of Figures
Figure 2.1 Age Differences in the Effect of Job Insecurity on Psychological Distress ................ 41
Figure 2.2 Age Differences in the Effect of Job Insecurity on Anger .......................................... 42
Figure 2.3 Age Differences in the Effect of Job Insecurity on Self-rated Poor Health ................ 43
Figure 4.1 Age Contingencies in the Association between Insecurity and Sense of Control ..... 100
x
List of Appendices
Appendix I. Predicting the Hazard of Attrition .......................................................................... 112
1
Chapter 1 Introduction
Social, economic and political forces over the last three decades have changed the structure of
employment relations in the United States (Kalleberg 2011). One key aspect of the changing
landscape of work in recent decades has been the decline in secure and stable employment. As
American firms have responded to global competition and economic uncertainty by adopting
mean and lean organizational strategies that include downsizing and the favoring of fixed term
contracts over permanent positions, job insecurity has become an increasingly widespread
experience for workers (Fullerton and Wallace 2007; Vallas 1999). Industrial restructuring and
the decline of the manufacturing-base in the United States have also led to the disappearance of
many once-secure blue-collar jobs (Kalleberg 2009). These economic and labour market trends
have led to a widespread erosion of the post-World War II social contract between capital and
labour—a contract that specified the exchange of worker commitment for employer-guaranteed
secure employment (Farber 2010).
The growth of job insecurity has drawn attention from scholars interested in its consequences for
worker health (Burgard, Brand, and House 2009; Tausig and Fenwick 1999). While some have
examined this issue using macroeconomic data (Fenwick and Tausig 1994; Tausig and Fenwick
1999), job insecurity is often examined with respect to workers’ perceptions of the security of
their current job—referred to as ‘perceived job insecurity’ (Greenhalgh and Rosenblatt 1984).
Research indicates that individuals who perceive their job to be insecure are more likely to report
anxiety, depressive symptoms, and poor physical health (Barling and Kelloway 1996; Burgard et
al. 2009; De Witte 1999). As a result, the perception of job insecurity is considered an important
emerging source of work stress (Sverke and Hellgren 2002).
Despite the growing body of evidence linking perceived job insecurity to poor health, the
specific mechanisms and conditions describing this process remain under-identified. Central to
this is the extent that the timing and duration of job insecurity represent contextual factors that
influence its health penalties. Comparatively little is known about whether the health penalties of
perceived job insecurity are universally experienced, or if they depend on the length of the
experience, and the stage in working life that it is encountered. Additionally, the role of
2
personality in these processes remains unclear; particularly with regard to whether personality
traits mediate or moderate the connections between job insecurity and health. In sum, while the
existing literature is suggestive of these factors and social-psychological linkages, it is limited
by: (1) the absence of a solid theoretical framework upon which to collectively integrate
predictions about timing, duration and personality contingencies, and (2) its general reliance on
small-convenience based and cross-sectional samples that make generalizations to wider
populations difficult, and that limit propositions regarding causality.
This dissertation addresses these limitations by using the Stress Process Model and insights from
the life course perspective to consider the social-psychological processes through which
perceived job insecurity impacts individual health, and how these processes are shaped by age as
a key social status and life course marker. To this end, I principally utilize survey data from a
nationally representative and longitudinal survey of American workers conducted in 2005 and
2007; this is an ideal data source for examining whether the health penalties of perceived job
insecurity vary according to the duration of the experience and its timing in the life course. The
survey also includes a state of the art measure of the sense of control, an important personal
resource that has been shown to buffer the effects of stressful experiences on health (Mirowsky
and Ross 2003), but one that has received inadequate attention with regards to workers facing the
threat of job loss. Integrating insights from the Stress Process Model and the life course
perspective to examine these unique data, I advance understanding on the nature of perceived job
insecurity as a stressor, and the processes through which it connects to mental and physical
health.
Theoretical and Methodological Contributions I extend existing theoretical and empirical research on perceived job insecurity by combining
insights from the Stress Process Model and the life course perspective (Elder 1998; Pearlin et al.
1981). Here, I conceptualize the experience of job insecurity as one that operates uniquely over
time, and one that is personally negotiated within temporal and contextual constraints. Drawing
from the Stress Process Model, I argue that the duration of job insecurity and the role of
personal resources represent key factors that influence its health penalties. Integrating insights
3
from the life course perspective with this theoretical framework enables a consideration of the
role of life course timing—and therefore age contingencies—in the process. I outline these
theoretical perspectives in turn, and their integration.
The Stress Process Model
The Stress Process Model represents a central theoretical framework in the sociological study of
stress (Pearlin et al. 1981). The model identifies the ways that stressors influence health, and the
mediating or moderating influence of psychosocial resources (Pearlin 1999; Thoits 2006). The
model describes the ways that social statuses and roles influence exposure to stressors, along
with the available resources for dealing with stress, and the ways that stressors lead to particular
physical and mental health problems. At its core, the model can be split into three domains that
link together to describe the stress process. First, different sources of stress are distinguished,
such as whether a stressor is an acute, event-based strain, or a durable, chronic strain. Second, the
model outlines how personal resources, including feelings of mastery and the sense of control,
moderate the effects of stressors on health. Third, it specifies the various outcomes of stress (e.g.
mental and physical health, behavioral problems) and how they pertain to different stressors.
Within the stress process model, perceived job insecurity is situated as a role stressor that
involves the anticipation of an undesirable event that is both uncontrollable and unpredictable
(Pearlin et al. 1981). Given that the threat of job loss may be experienced for extended periods of
time, and may not have a clear start or end point, it is typically considered a chronic stressor
rather being than experienced as an acute event like actual job loss (Wheaton 1999). It is argued
that its uncertain, uncontrollable nature represents part of the mechanism that makes job
insecurity stressful, given that high levels uncertainty can have paralyzing effects on the coping
process (Lazarus and Folkman 1984). The proximal health penalties of perceived job insecurity
are expected to manifest themselves through increased anxiety and negative emotional states
such as feelings of anger. However, if the experience is prolonged it may also lead to more
serious mental and physical health problems (Burgard et al. 2009; Sverke and Hellgren 2002). It
is expected that these penalties may be less for those with high levels of perceived personal
control, also referred to as the sense of control—a personal resource that may help to mitigate the
stress associated with uncertain, unpredictable, and tenuous employment experiences (Pearlin et
al. 2005).
4
The sense of control is the generalized belief that one can control meaningful events in their life
(Wheaton 1983). As a psychological disposition that promotes active coping strategies for
dealing with the daily challenges of life, the sense of control is linked to healthy mental and
physical functioning (Mirowsky and Ross 2003; Wheaton 1983). Within the framework of the
Stress Process Model, a high sense of control is expected to reduce the effects of stressors on
health. However, in addition to this buffering effect, the sense of control may play a further role
in the stress of perceived job insecurity. Since the sense of control is not a static aspect of
personality but rather one that is shaped by objective, social conditions (Pearlin and Skaff 1996),
the powerlessness to control one’s economic future may indirectly influence health by eroding
the sense that one controls important aspects of their life. The sense of control may therefore
constitute not only a buffering process in the links between perceived job insecurity and health,
but also a mediating mechanism through which the threat of job loss leads to health problems.
While the Stress Process Model offers an interpretive and predictive framework for
conceptualizing perceived job insecurity as a stressor, inadequacies exist with respect to its
treatment of age and life course contingencies. Despite the fact that it acknowledges the potential
role of social statuses in shaping the experience and outcomes of stress, until recently the model
had remained relatively silent on how the stress process is shaped by life course position. As
Miech and Shanahan (2000:163) observe, much of the stress literature continues to rest on the
“implicit conception of the “ageless adult” who experiences the same stressors and reacts to them
the same way from age 18 until the end of life.” Some, however, have highlighted the benefits of
an allegiance between the Stress Process Model and the life course perspective (Pearlin and
Skaff 1996). I respond to this call by integrating insights from the life course perspective to more
fully flesh-out predictions regarding life course contingencies in the stress of perceived job
insecurity.
The Life Course Perspective
Core to the life course perspective is the observation that temporality and historicity matter for
understanding life experiences and events (Elder 1998). As part of this, the perspective outlines a
set of principles that emphasize the importance of considering the timing and sequencing of life
experiences and how they influence outcomes later in one’s life—principles that are useful to the
study of stress (George 2007). Life course researchers argue, for example, that while individuals
5
are active agents who exert some control over their responses to life experiences and future life
course trajectories, these reactions and choices are nevertheless constrained by a predefined
social structure and historical context not of their choosing. Additionally, the life course principle
of linked lives observes the interdependence of individuals within social networks and the family;
thus outlining how life events and experiences are shaped by broader social contexts in which
individuals are embedded (Elder 1998).
With respect to the stress of perceived job insecurity, I argue that the life course timing of job
insecurity should shape the meanings and consequences that individuals attach to the threat of
job loss, and subsequent health penalties. Specifically, I suggest that labour market experiences
and family obligations that vary over working life, along with expectations about acceptable
work conditions, may be relevant factors that influence both the level of threat that an individual
attaches to job loss, and their perception of their available resources to deal with the threat.
Additionally the sequencing and proximity of job insecurity relative to a key important stage in
the life course—retirement—represents an additional factor that should shape the meaning that
an individual attaches to the experience. Drawing from the life course perspective, then, I extend
on the Stress Process Model’s treatment of perceived job insecurity as a stressor by
acknowledging the role that individuals play in negotiating the meaning of the experience, but
within a set of social-structural parameters not of their choosing; specifically, the timing and
sequencing of the event in working life.
Age Contingencies in the Stress of Job Insecurity
Integrating insights from the Stress Process Model and the life course perspective, I derive a set
of hypotheses that specify how the length and timing of the experience should act as key factors
that influence its health penalties. First, as a chronic stressor, sustained exposure to the
experience may be required for health penalties to be experienced. Second, age should moderate
the extent that perceived job insecurity is stressful, given the different meanings that individuals
attach to job loss across the life course. To this end, I present two competing hypotheses
regarding age contingencies. The stress of aging hypothesis suggests that health penalties should
be greatest in late-working life, given that older workers experience greater economic penalties
due to job loss, in the form of longer job search times and lower wages after rehire (Hirsch,
Macpherson, and Hardy 2000); penalties that should shape the anxiety associated with the threat
6
of job loss. Alternatively, the stress of middle age hypothesis suggests that health penalties
should peak during middle age when the age-related economic penalties associated with job loss
begin to increase, and when family-related obligations tend to be highest (Erikson 1998). These
age hypotheses extend research on perceived job insecurity by moving beyond the “ageless
adult” assumption in which job insecurity is seen as uniform in its meaning and potency as a
stressor across the life course.
To test these hypotheses, I analyze two waves of panel data from a nationally representative
survey of American workers. The design of the study allows for an examination of both the
duration and life course timing of job insecurity as factors that may influence its health penalties,
as well as changes in personality traits (i.e. the sense of control) that may moderate and mediate
these penalties. Additionally, I supplement these analyses with a second cross-sectional survey of
Canadian workers—a study that has the advantage of being based on a much larger sample,
allowing therefore for a more refined investigation of age contingencies.
Overview of the Dissertation This dissertation draws from the Work, Stress and Health study (WSH), a longitudinal survey of
American workers conducted in 2005 and 2007, and the first wave of its Canadian equivalent:
the CAN-WSH, a national survey of Canadian workers conducted in 2011. The piece in its
entirety includes three independently publishable papers and the present common introductory
chapter, along with a concluding chapter. Each of these papers can be considered independent
projects, but together, they represent an overarching modeling of the stress process applied to the
experience of the threat of job loss.
The first paper (chapter 2) examines the extent that duration and life course timing shape the
health penalties associated with perceived job insecurity. Drawing from panel survey data based
on the WSH study, the paper compares the effects of episodic and prolonged exposure to job
insecurity as they pertain to three indicators of mental and physical health. Additionally, two
competing hypotheses are tested with regards to the life course timing of job insecurity and
subsequent health disparities. The second paper (chapter 3) examines the role of an important
personality trait—the sense of control—in both moderating and mediating the stress of perceived
7
job insecurity. Here, survey data from the first wave of the WSH study and the CAN-WSH
survey are examined to assess direct and indirect links between perceived job insecurity and
mental health. Given, the large sample sizes of these cross-sectional surveys, more refined age
contingencies are also tested to consider how the stress of perceived job insecurity may vary
across the life course. This has the advantage of examining specific stages within middle age, for
example, and how they relate to job insecurity and health. The third paper (chapter 4) considers
how personality traits may be shaped over time in response to job insecurity. This involves an
analysis of the American WSH panel survey data to examine: 1) whether perceived job
insecurity erodes the sense of control over time, 2) potential explanations—in terms of other
aspects of job quality—for this effect, and 3) whether the strength of any effect varies across age.
Chapter 5 summarizes the general findings and contributions of this dissertation and also
discusses its limitations, alongside future directions and plans for further research.
Goals of the Dissertation Three goals motivate this research. Job insecurity is an increasingly common labour market
experience in the United States and Canada—one that is not simply a result of periodic economic
and financial crises, but the result of long-term structural economic changes that will not reverse
in the foreseeable future (Fullerton and Wallace 2007). The ‘human’ impact of these economic
and labour market trends is an issue of vital importance and one that guides this dissertation. As
such, perceived job insecurity is an experience that merits further scholarly examination in order
to understand its health consequences for workers. Second, the study of labour market
experiences as they pertain across the life course—and older workers in particular—represents
an area requiring further attention. The American workforce is aging—those 50 and older will
account for 35 percent of the workforce by 2019 (Toossi 2009). Further research on the health
implications of job insecurity for this aging workforce is therefore warranted. Finally, I seek to
contribute to the theoretical and empirical literature on the stress process by integrating ideas and
concepts from the life course perspective. Stress researchers are increasingly turning to the life
course perspective as a means to contextualize the stress process; this dissertation seeks to
advance these efforts.
8
References Barling, Julian, and E. Kevin Kelloway. 1996. “Job Insecurity and Health: The Moderating Role
of Workplace Control.” Stress Medicine 12:253-259.
Burgard, Sarah A., Jennie Brand, and James S. House. 2009. “Perceived Job Insecurity and Worker Health in the United States.” Social Science & Medicine 69:777-785.
De Witte, Hans. 1999. “Job Insecurity and Psychological Health: Review of the Literature and Exploration of Some Unresolved Issues.” European Journal of Work and Organizational Psychology 8:155-177.
Elder, Glen H., Jr. 1998. “Life Course Theory and Human Development.” Sociological Analysis 1:1-12.
Erikson, Erik. H. 1998. The Life Cycle Completed. NY: Norton.
Farber, Henry S. 2010. “Job Loss and the Decline of Job Security in the United States.” Pp. 223-267 in Labour in the New Economy edited by Katherine Abraham, James Spletzer and Michael Harper. University of Chicago Press, Chicago.
Fenwick, Rudy. and Mark Tausig. 1994. “The Macroeconomic Context of Job Stress.” Journal of Health and Social Behavior 35:266–282.
Fullerton, Andrew S. and Michael Wallace. 2007. “Traversing the Flexible Turn: US Workers’ Perceptions of Job Security, 1977–2002.” Social Science Research 36:201-221.
Hirsch, Barry T., David A. MacPherson, and Melissa A. Hardy. 2000. “Occupational Age Structure and Access for Older Workers.” Industrial and Labour Relations Review 53: 401–418.
George, Linda K. 2007. “Life Course Perspectives on Social Factors and Mental Illness.” In W.R. Avison, J. D. Mcleod and B. A. Pescolsolido. Mental Health, Social Mirror (p.191-218). New York, Springer.
Greenhalgh, Leonard, and Zehava Rosenblatt. 1984. “Job Insecurity: Toward Conceptual Clarity.” Academy of Management Review 3:438–448.
Kalleberg, Arne L. 2009. “Precarious Work, Insecure Workers: Employment Relations in Transition.” American Sociological Review 74:1–22.
______. 2011. Good Jobs, Bad Jobs: The Rise of Polarized and Precarious Employment Systems in the United States, 1970s-2000s. New York: Russell Sage Foundation, American Sociological Association Rose Series in Sociology.
Lazarus, Richard S. and Susan Folkman. 1984. Stress, Appraisal, and Coping. New York: Springer.
Miech, Allen M. and Michael J. Shanahan. 2000. “Socioeconomic Status and Depression of the Life Course.” Journal of Health And Social Behavior 41:162-176.
Mirowsky, John, and Catherine E. Ross. 2003. Social Causes of Psychological Distress. Hawthorne, New York: Aldine De Gruyter.
9
Pearlin, Leonard I. 1999. “Stress and Mental Health: A Conceptual Overview.” Pp. 161–75 in Handbook for the Study of Mental Health: Social Contexts, Theories, and Systems, edited by Allan V. Horwitz and Teresa L. Scheid. Cambridge, England: Cambridge University Press.
Pearlin, Leonard I., Morton Lieberman, Elizabeth Menaghan, and Joseph Mullan. 1981. “The Stress Process.” Journal of Health and Social Behavior 22:337–56.
Pearlin, Leonard I. and Marilyn M. Skaff. 1996. “Stress and the Life Course: A Paradigmatic Alliance.” The Gerontologist 36:239-247.
Pearlin, Leonard I., Scott Schieman, Elena Fazio, and Stephen C. Meersman. 2005. “Stress, Health, and the Life Course: Some Conceptual Specifications.” Journal of Health and Social Behavior 46:205–19.
Sverke, Magnus, and Johnny Hellgren. 2002. “The Nature of Job Insecurity: Understanding Employment Uncertainty on the Brink of a New Millennium.” Applied Psychology 51:23–42.
Tausig, Mark, and Rudy Fenwick. 1999. “Recession and Well-being.” Journal of Health and Social Behavior 40:1-16.
Thoits, Peggy. 2006. “Personal Agency in the Stress Process.” Journal of Health and Social Behavior 47:309-323.
Toossi, Mitra. 2009. “Labour Force Projections to 2018: Older Workers Staying More Active.” Monthly Labour Review November 30–51.
Vallas, Steven P. 1999. “Rethinking Post-Fordism: The Meaning of Workplace Flexibility.” Sociological Theory 17:68−101.
Wheaton, Blair. 1983. “Stress, Personal Coping Resources, and Psychiatric Symptoms: An Investigation of Interactive Models.” Journal of Health and Social Behavior 24:100-24.
______. 1999. “Social Stress.” Pp. 277-300 in The Handbook of the Sociology of Mental Health, edited by C.S. Aneshensel and J.C. Phelan. Kluwer Academic/Plenum.
10
Chapter 2 Perceived Job Insecurity and Health Across the Life Course
Introduction Labour market changes in recent decades have exposed American workers to a growing sense of
uncertainty about their future employment (Kalleberg 2011). In response to increased global
competition and economic recessions, American firms have engaged in downsizing, plant
closures, and the replacement of full-time, permanent positions with fixed and short-term
employment contracts. This restructuring has contributed to the general decline in job security
over the last thirty years in the United States (Farber 2010). Researchers interested in the
implications of these changes for workers’ health have found that individuals who perceive their
job to be insecure are more likely to report anxiety, anger and depressive symptoms (Barling and
Kelloway 1996; De Witte 1999). Longitudinal research also reveals that prolonged exposure to
job insecurity reduces physical health, and in some cases has more detrimental health effects than
actual job loss (Burgard, Brand, and House 2009; Heaney, Israel, and House 1994; Lee et al.
2004). As a result, perceived job insecurity is widely accepted as an important contemporary
source of work stress (Sverke and Hellgren 2002).
According to the Stress Process Model, perceived job insecurity is considered a ‘chronic
stressor’, whose effects on health may take some time to be realized (Pearlin et al. 1981;
Wheaton 1999). Drawing from this model, I compare the effects of short-term and long-term
exposure to job insecurity on a variety of health outcomes. I suggest that the duration of the
experience may alter both the extent that the experience is damaging to health, as well as
influencing different aspects of mental and emotional health. While general mental health
penalties may be experienced only with sustained exposure to the threat of job loss, short-term
exposure to this threat may be associated with emotional problems, such as feelings of anger.
How the timing and specific health penalties of perceived job insecurity play out remains an
under-identified area in existing research.
11
An additional under-examined issue concerns whether the severity of perceived job insecurity’s
health penalties operate uniformly across the life course. While the Stress Process Model
acknowledges the role of social statuses in influencing exposure to stress, its treatment of stress
across the life course is lacking (Elder 1998; Pearlin and Skaff 1996). There is good reason,
however, to consider how job insecurity may operate uniquely across the life course as a stressor.
Since older workers who encounter job loss take longer to find reemployment and experience
greater wage penalties after rehire, the meaning of the threat of job loss—and the anxiety
experienced as a result—may change over the life course, as individuals encounter different
labour market constraints and opportunities, and acquire new statuses and associated obligations.
Thus, I examine age contingencies in the association between job insecurity and health. One
possibility is that the stress and subsequent health effects of job insecurity are stronger among
older workers, given that labour market penalties associated with job loss tend to be more acute
as workers age (Hirsch, Macpherson, and Hardy 2000). Alternatively, the health penalties of job
insecurity may peak during middle age when family obligations tend to be greatest and early
retirement is not a feasible solution to potential job loss. To-date, no research has considered age
contingencies in the stress of job insecurity using nationally representative survey data.
This paper addresses these weaknesses in the existing literature by integrating insights from the
life course perspective with the Stress Process Model to consider the temporal health penalties of
job insecurity across the life course. To do this, I use panel survey data following a national
sample of American workers in 2005 and 2007. Since the health effects of job insecurity may
operate over a longer-time frame that make their detection difficult when examined with cross-
sectional studies (Burgard et al. 2009), a longitudinal research design is therefore ideal for
comparing the health effects of episodic versus prolonged exposure to job insecurity. In the
following sections, I review the existing literature on job insecurity as a chronic stressor, before
discussing potential age contingencies in this relationship.
12
Literature Review
Perceived Job Insecurity and Health
Scholars have long demonstrated the impact of unemployment on health (Kessler, House and
Turner 1989; Schaufeli 1992; Tausig and Fenwick 1999). However, recent attention has focused
on the health consequences of the threat of unemployment, typically referred to as ‘perceived job
insecurity’ (Greenhalgh and Rosenblatt 1984). In one of the earliest studies of job insecurity,
Greenhalgh (1979) examined the experiences of New York state hospital workers facing
widespread downsizing, noting the deep levels of anxiety and trauma of workers who survived
the layoffs. Subsequent studies of firm closures and downsizing have demonstrated similar
findings regarding the experiences of remaining workers (Brockner 1990; Jick and Greenhalgh
1989).
Systematic economic and industrial restructuring since the 1980s has meant that job insecurity
has become an increasingly widespread experience that is not limited to workers surviving
downsizing. Firms’ efforts to improve efficiency by downsizing and relying more heavily upon
contingent work contracts have exposed workers from a wide variety of backgrounds to job
insecurity—those in blue-collar and white-collar work, and those with temporary or full-time
contracts (Kalleberg 2009; Polivka and Nardone 1989). In response to these economic trends,
survey researchers have developed self-reported measures of perceived job insecurity in order to
examine its prevalence and consequences in the wider workforce (Greenhalgh and Rosenblatt
1984). These questions examine either an individual’s worry or concern about their job security,
or their assessment of the likelihood that they will lose their job in the future. In comparison to
more objective indicators, such as firm layoffs and economic recessions, self-reported measures
of job insecurity are useful in that they are sensitive to individual differences in the subjective
interpretation of these events; despite this, research shows that perceived job insecurity is
correlated with more objective indicators of insecurity (Elman and O’Rand 2002).
According to the Stress Process Model, perceived job insecurity is considered a classic role
stressor that involves the anticipation of an undesirable event that is both uncontrollable and
unpredictable (Pearlin et al. 1981; Wheaton 1999). Here, the uncertainty surrounding future job
loss itself may function to reduce health (De Witte 1999; Lazarus and Folkman 1984). In contrast
13
to an actual event that an individual can directly observe and respond to—such as job loss—the
very ambiguity and uncertainty about the threat of job loss hinders an individual’s ability to
resolve or alleviate the threat. In fact, some suggest that perceived job insecurity might represent
a more potent stressor than job loss or episodes of unemployment (Burgard et al. 2009).
The Stress Process Model represents a useful framework to interpret job insecurity as a stressor
because of the model’s attention to how the duration of a stressor may influence the extent and
nature of its health consequences (Wheaton and Reid 2007). The model distinguishes between
different sources of stress, such as whether a stressor is experienced as an acute, event-based
strain, or whether it represents a durable, chronic strain. Job insecurity is generally framed as a
chronic stressor. As Wheaton (1999, p. 282) explains, chronic stressors “do not necessarily start
as an event but develop slowly and insidiously as continuing and problematic conditions in our
social environment or roles…” Here, the threat of job loss may be a contextual and ever-present
experience, given its potentially ambiguous and uncertain nature, where anxiety over a layoff
flows across work and nonwork roles. The implications of this for well-being are that health
penalties may occur more gradually for a chronic stressor than compared to the experience of
event-based stressors, such as the loss of a loved one or one’s job (Burgard et al. 2009).
Biological stress response theories, for example, describe how prolonged anxiety over an
imagined or real threat can generate an ‘allostatic load,’ whereby recurrent activation of the
body’s stress-response system gradually compromises autoimmune system functioning, reducing
resistance to other health problems or stressors (Seeman et al. 2001).
There is some empirical evidence that provides support for job insecurity as a chronic stressor.
Burgard and colleagues (2009) find stronger negative health effects for persistent job insecurity
in comparison to episodic experiences of insecurity in two longitudinal studies of American
workers. Similarly, researchers in the classic British Whitehall study found that civil servants
who experienced persistent job insecurity reported an increase in psychiatric morbidity and a
decline in self-reported health over a period of two years (Ferrie et al. 1995). Although
empirically unexamined, it is possible that short-term exposure may nevertheless have severe
penalties for certain aspects of health, such as emotional well-being. Psychological contract
theory (Rousseau 1995) suggests that the onset of job insecurity for some workers may represent
a break in the normative employer-employee relationship, which has been historically bound by
14
the mutual exchange of worker commitment for employer-guaranteed secure employment. The
violation of this norm may therefore result in negative emotional reactions, such as feelings of
anger and frustration that are experienced shortly after the onset of the violation.
Thus, based on the Stress Process Model and existing literature on the health consequences of the
threat of job loss, I consider job insecurity as a chronic role stressor, given its non-discrete nature
and its longer time frame in comparison to life events such as job loss (Wheaton 1999). The
negative health consequences of job insecurity should therefore be contingent on the duration of
the experience, with more profound effects resulting from prolonged exposure; short-term health
penalties may manifest themselves with respect to emotional aspects of mental well-being,
however.
Job Insecurity across the Life Course
Despite the growing body of knowledge on the link between job insecurity and health,
comparatively less is known about whether its health consequences are experienced uniformly
across age and the life course. There are various reasons to consider why age might be relevant
to the experience of job insecurity. First, research finds that job insecurity is growing in
prevalence among older age groups. In the last two decades, workers in their forties and older
have experienced a decline in average tenure length and an increase in job displacement rates
(Farber 2008). Traditionally protected from job loss due to their greater experience and
organizational tenure, economic restructuring has eroded the “life-time employment contract”
enjoyed by older workers, and in many cases actually penalized them (Kalleberg 1996; Salzman
1998). Their higher levels of human capital and once-advantageous group memberships (e.g.
union membership) have made them prime candidates for involuntary redundancy, as employers
adopt cost-cutting strategies that involve replacing well-paid older employees with cheaper,
younger workers (Fallick 1996).
Second, there is limited, but conflicting, evidence that age contingencies exist in the stress of job
insecurity. For example, a study by Kasl and colleagues (1975) that examined the health
experiences of men facing job loss in two manufacturing plants found that younger workers were
more likely to report health problems in the anticipatory stage before job loss. However, in a
meta-analysis of job insecurity studies, Cheng and Chan (2008) find evidence of an opposite
15
pattern: older age brought with it disproportionately higher health penalties as a result of job
insecurity.
The life course perspective represents a potentially useful framework for addressing the limited
and conflicting evidence on age contingencies. Life course researchers emphasize the importance
of taking into account the temporal and historical nature of life experiences, and how they may
be linked to later events and experiences in individuals’ lives (Elder 1998). Recently, stress
researchers have called for an allegiance with the life course perspective, arguing that a
consideration of the life course offers several advantages to the study of stress (Pearlin and Skaff
1996). I briefly outline these benefits and discuss their relevance to perceived job insecurity.
It is suggested that when applied to the experience of stress, the life course principles of timing
and trajectories illuminate how stressors may be contingent on contextual experiences that vary
across the life course, as well as individuals’ past and future life paths (Turner and Schieman
2007). The stress associated with a role transition like divorce, for example, may depend on the
past quality of the relationship being exited, as well as one’s future opportunities for meeting
new partners (Wheaton 1990). Additionally, acknowledging the historical context that a stressor
is experienced, and the historical period of a specific cohort of individuals encountering the
stressor, allows stress researchers to consider how historicity influences the consequences of
stressors for well-being. For example, divorce may entail less stigma and stress for younger
cohorts than compared to older cohorts who grew up in an era in which separation was less
common. Collectively, these observations suggest that life course timing and role trajectories, as
well as historical context, may serve to shape the meaning of stressors and an individual’s
perceptions of the resources they have to deal with stress (Pearlin and Skaff 1996).
Applying these principles to the threat of job loss, I suggest that 1) the life course timing of the
experience, 2) one’s current career trajectory and 3) cohort differences, may combine to shape
the meaning of this threat. The timing in which job loss occurs over the life course brings with it
varying prospects for reemployment, along with uneven financial consequences, given that
family obligations tend to vary over working life. An individual’s future career trajectory, such
as their proximity to retirement, may also serve to shape the inferred consequential nature of job
loss; if one is close to retirement, the threat may take on a different meaning compared to an
individual in the formative stages of their career. Finally, different cohorts of workers may react
16
to uncertainty over one’s continuing employment in different ways. Older workers, for whom job
security was a more normative work experience in their early working life (Farber 2010), may
view the contemporary threat of job loss differently from younger workers, who have grown up
in an era defined by labour market instability.
Based on these principles, I derive two hypotheses regarding differences in the experience of,
and meanings that individuals attach to the threat of job loss across the life course.
Stress of Aging Hypothesis
The ‘stress of aging’ hypothesis points to evidence that older workers who encounter job loss
experience more severe economic penalties. Older workers who are laid-off take longer to find
re-employment and encounter greater earnings losses in their new jobs (Hirsch et al. 2000).
Using panel data from the Survey of Income and Program Participation between 1996 and 2004,
Johnson and Mommaerts (2011) found that displaced older men (i.e. age 50-61) were around
forty-percent less likely to become re-employed each subsequent month after job loss. Upon
rehire, these workers experienced a wage loss of twenty-percent in comparison to the two-
percent loss experienced by their younger counterparts. While not as severe, displaced older
women encountered comparable penalties. Similar age-disparities in the consequences of job loss
have been documented elsewhere (Couch 1998; Couch, Jolly, and Placzek 2009). In part, the
disproportionate wage losses that older workers experience reflects their inability to translate
firm-specific human capital acquired in their previous job into equivalent wages upon rehire
(Becker 1962; Farber 1993). However, a portion of these wage losses and the longer search
periods after job displacement are a product of discriminatory age practices by employers who
often view older workers as less adaptable, difficult and costly to train (Bendick, Brown, and
Wall 1999; Rodriquez and Zavodny 2003). These challenges that older workers face in the
labour market should therefore shape the degree of threat that they attach to the risk of job loss.
Labour market constraints may not be the only factor that shapes the meaning of job loss for
older workers; in accordance with psychological contract theory, older workers should also view
job insecurity as a stronger violation of the normative employer-employee contract because of
their on-average longer tenure lengths with their employer (Farber 2008; Rousseau 1995).
Additionally, the formative stages of older workers’ careers—many of whom are from the baby
17
boomer generation—were characterized by high levels of job security, with often the expectation
of holding employment with a single employer for life (Farber 2010). These early work
experiences may have fostered beliefs about appropriate treatment from employers; beliefs that
imbue the meaning and experience of insecure employment as particularly unsettling within this
cohort. Younger generations of workers, in contrast, including those of generation ‘X’ and ‘Y,’
entered the labour market in the 1980s and onwards, when rates of job displacement and job
mobility were on an upward trend (Fullerton and Wallace 2007; Farber 2010). For these workers,
risk of job loss may represent a more ‘normal’ and acceptable feature of working life.
Thus, given the greater labour market penalties and challenges experienced by those seeking
employment in later-working life, along with cohort-specific expectations regarding job security,
I test a ‘stress of aging’ hypothesis that predicts that the stress and health effects associated with
perceived job insecurity should increase with age.
Stress of Middle Age Hypothesis
The ‘stress of middle age’ hypothesis, in contrast, suggests that the meaning of job insecurity
should be most salient during middle age, rather than late-working age. While both middle age
and older workers experience greater economic penalties due to job loss than those in the earlier
stages of their careers, the consequences of these penalties may be disproportionately higher
during middle age when family-related financial obligations tend to be greatest (Brenner 1973;
Erikson 1998). In contrast to middle age, individuals in the later stages of their careers have
fewer family dependants, and therefore may be better able to adjust to the economic penalties of
job loss. Indeed, other research finds that, regardless of family demands, older individuals are
better able to deal with challenges associated with economic hardship in general (Mirowsky and
Ross 2001).
The different career trajectories of middle age and older workers may also serve to imbue the
threat of job loss with different meanings and connotations. Many older workers who encounter
job loss may become discouraged and enter involuntary retirement. This is in part, of course,
because of the limited employment options for workers in later life (Szinovacz and Davey 2005),
but also because of their greater access to social security benefits and private pension plans that
may make early retirement feasible, if not desirable. Hence, for older workers the onset of job
18
insecurity and the threat of job loss may prompt thoughts about a premature transition into
retirement rather than feelings of anxiety.
It is also possible that cohort differences in expectations about job security create the most
frustration for middle age workers. As a group that were raised by a baby-boomer generation that
experienced job security and good promotion opportunities—job conditions that were
particularly enjoyed at middle age during the golden era of American capitalism—contemporary
middle age workers may be uniquely positioned in that they retain career aspirations and
expectations for their prime work years; but face a contemporary labour market that obstructs
these aspirations. It may be this generation, sandwiched between baby-boomers and young
workers, that experience the greatest incongruency between their job expectations and the reality
of the contemporary labour market; in other words, frustrations about job insecurity for this
group may arise out a combination of both cohort and age-related factors.
Consequently, an alternative hypothesis to the ‘stress of aging’ hypothesis predicts that the
experience of job insecurity is most harmful for health during middle age, due to cohort
differences in the meaning of job insecurity, and when the economic penalties associated with
job loss are most disruptive to meeting family obligations. The stress and health effects of
perceived job insecurity should therefore operate across age in a nonlinear fashion, with the
most severe effects experienced during middle age.
Job Insecurity and Absenteeism
While the Stress Process Model has been mainly used to examine the links between stress and
health, behavioral outcomes of relevance for health have also been assessed; for example
problem drinking (Wheaton 1999). Given the potential health consequences of job insecurity,
therefore, I briefly consider its relevance to work absenteeism—a behavior that may be a
consequent of reduced well-being, but also a predictor of poor health if workers with health
problems reduce absenteeism due to anxiety over job loss.
Two general perspectives inform understanding on absenteeism (Koslowsky 2009). The first
views absenteeism as an indicator of work withdrawal, where workers reduce attendance as a
result of unfavorable work attitudes, such as job dissatisfaction and low organizational
commitment (Hulin 1991). A second perspective considers absenteeism as a result of work stress
19
and burnout. Stress theories suggest that high levels of stress should result in psychosomatic
health problems that lead to increased absences due to sickness (Pearlin et al. 1981). In terms of
the links between perceived job insecurity and absenteeism, the former perspective frames job
insecurity as a likely antecedent of job dissatisfaction that should lead to withdrawal from work.
Workers who perceive their job to be insecure may feel disenfranchised and less committed to
their job, and become negligent in their attendance (Ashford, Lee and Bobko 1989; Cuyper and
De Witte 2006). Several studies have shown a subsequent increase in absenteeism rates in firms
that undergo downsizing, for example (Bourbonnais et al. 2005; Firns, Travaglione, and O’Neill
2006). The second perspective makes the same prediction, but for different reasons: as a role
stressor that reduces health, job insecurity should result in more frequent absences from work
(King 2000; Lim 1996).
An alternative possibility exists, however. Given the propensity for job insecurity to create
feelings of anxiety and trepidation about future job loss, it is possible that workers may improve
their attendance at work as part of an effort to demonstrate organizational commitment and ward
off job loss (Brockner 1990; Probst et al. 2007). As indirect of evidence of this possibility,
perceived job insecurity was found to be associated with increased sickness presenteeism—
attendance at work when one is ill—among workers in two Finnish and Canadian studies
(Caverley, Cunnigham, and MacGregor 2007; Heponiemi et al. 2010).
With respect to age differences in the consequences of job insecurity, informed by the ‘stress of
aging’ hypothesis, the greater anxiety over job loss experienced by older workers should result in
more health problems and less organizational commitment that lead to higher levels of
absenteeism. Alternatively, job insecurity may result in reduced absenteeism among older
workers as they attempt to reduce the likelihood of job loss. This latter prediction signals a
possible double-disadvantage for older workers in insecure employment. That is, even though
they may encounter more mental or physical health problems that require rest and recovery, they
may nevertheless feel pressured to attend work due to the threat of being laid off, potentially
exacerbating existing health issues. Finally, based on the prediction that the health effects of
insecurity peak during middle age, we may see either the greatest increase or decline in
absenteeism among middle age workers.
20
In summary, competing predictions are tested regarding the tendency of job insecurity to result
in either increased or decreased absenteeism. Further, based on the ‘stress of aging’ hypothesis or
its alternative, the strength of these effects may either increase linearly with age, or conversely
peak during middle age.
Data and Methods To test the hypotheses described above, I analyze panel data from the Work, Stress and Health
study (WSH), which involved telephone interviews with adults in the 50 United States in 2005
and 2007. To obtain the original sample, a list-assisted random digit dialing (RDD) selection was
used and drawn proportionally from all 50 states from GENESYS Sampling Systems. Eligible
participants are 18 years of age or older and participating in the paid labour force. At Wave 1, 71
percent of eligible individuals were successfully interviewed yielding a sample of 1,800 adults.
At Wave 2 interviews, which occurred approximately 18-to-20 months after the initial interview,
1,286 of the original participants were successfully re-interviewed. Sample characteristics for the
WSH study are similar to the population estimates of the U.S. Census Bureau’s 2005 American
Community Survey (ACS). The working sample includes non-self employed workers between
25 and 65 years of age (N=956) working twenty hours or more.
I explain the rationale for the sample selection. I limit the sample to non self-employed workers
between the ages of 25 and 65. The measure of job insecurity that I use is designed to tap into
workers perceptions of the strength of the employer-employee psychological contract (Rousseau
1995); as such, even though self-employed workers may face future threats to their continuing
employment, this measure of insecurity is not applicable to workers who are not hired by an
employer. Additionally, as is the case with most studies of job insecurity, respondents between
the ages of 18-24 are omitted from the working sample, since this group are overridingly in
contingent jobs due their disproportionate participation in higher education (Polivka 1996).
Measures
Three outcomes are used to assess health. Psychological distress. To measure distress, I use an
amalgamation of items from the Center for Epidemiological Studies Depression Scale (CES-D,
21
see Radloff 1977; Ross and Mirowsky 1984) and the Kessler index of generalized psychological
distress (Kessler et al. 2002). Respondents were asked the following eight items: “In the last
seven days, on how many days have you . . . felt sad; felt like you just couldn’t get going; felt
unable to shake the blues; felt like everything was an effort; had trouble keeping your mind on
what you were doing; worried a lot about little things; felt anxious or tense; had trouble getting to
sleep or staying asleep.” Responses are coded in days per week from 0 to 7. The distress scale is
the mean response to the eight items. The scale has an alpha reliability of .84.
Feelings of anger. Respondents were asked: “In the last seven days, on how many days have you
. . . felt annoyed or frustrated; felt angry; felt very critical of others; yelled at someone or
something; lost your temper.” Responses are coded in days per week from 0 to 7. The anger scale
is the mean response to the five items and has an alpha reliability of .78 (see Schieman and Reid
2009).
Self-reported poor health. Respondents were asked: “At the present time, in general would you
say your health is…? (1) “poor”, (2) “fair”, (3) “good”, (4)“very good”, (5) “excellent.”
Responses are reverse coded, such that higher scores represent poorer self-rated health. Self-
rated health correlates strongly with more objective measures of health, such as physician’s
health evaluations (Idler and Kasl 1991).
Days Absent from Work. At wave 1 and 2 participants were asked: “In the last six months, how
many days have you been absent from work (not counting vacation)?” Days absent from work at
wave 1 and wave 2 are coded as continuous measures and logged to correct for skewness.
Perceived Job insecurity. Job insecurity is assessed with the following question: “How likely is it
that during the next couple of years will you lose your present job and have to look for a job with
another employer?” Response choices include: “not at all likely” (1), “somewhat likely” (2), and
“very likely” (3). This question has been used in several well-respected surveys including the
General Social Survey and the National Study of the Changing Workforce. I follow Burgard and
colleagues (2009) modeling strategy in order to assess temporal change in job insecurity. I create
three binary dummy variables that represent episodic (wave 1 or wave 2 only) and persistent job
insecurity (both waves), and contrast these with the absence of job insecurity at both time points.
Persistent insecurity is coded 1 if respondents report that they are ‘somewhat’ or ‘very likely’ to
22
lose their jobs at both interviews and coded 0 otherwise, (2) resolved insecurity is coded 1 if
respondents reported that they are ‘somewhat’ or ‘very likely’ to lose their job only at wave 1,
and coded 0 otherwise, and (3) new insecurity is coded 1 if respondents reported that they are
‘somewhat’ or ‘very likely’ to lose their job at wave 2 only, and coded 0 otherwise. The
reference category is the absence of reported job insecurity at both time points.
Age. To assess age contingencies I created age categories that reflect three stages of working life:
early (25-39), middle (40-54) and late working life (55-65). These categories are modeled as
dummy variables in all analyses, with ‘early age’ respondents omitted as the contrast category.
As is the case with most studies of job insecurity, respondents between the ages of 18-24 are
omitted from the working sample, since this group are overridingly in contingent jobs due their
disproportionate participation in higher education (Polivka 1996). In analyses not shown, I also
modeled age as a continuous variable and tested squared and quadratic age interactions with
perceived job insecurity. These analyses revealed similar results as those with age modeled as a
categorical variable. I decided to use the latter modeling strategy since the categories map onto
well-established stages in work and nonwork roles (Erikson 1998; Toossi 2009).
Prior health Problems and work conditions. In order to address the issue of health selection, I
control for participant health at wave 1 for each health outcome appropriate to the particular
analysis. Further, I include an indicator of health problems that participants reported having prior
to wave 1 that could select them into particular jobs that might potentially influence subsequent
health and job insecurity. Prior health problems is measured as a count of health problems that
participants reported being diagnosed with, including: high blood pressure, high cholesterol,
arthritis, stomach problems, and cancer.
In order to rule out the possibility that the effect of job insecurity on subsequent health is due to
some related change in participants’ work conditions between wave 1 and wave 2, I include
baseline measures of income, job pressures, job autonomy, and work hours, in addition to a
change measure indicating any increase/decrease or change in these conditions between wave 1
and wave 2. These conditions are measured in the following way:
23
Household income. Household income is assessed with the question: “For the complete year of
2004/2006, what was your total household income, including income from all of your paid jobs,
before taxes?”
Job pressures. Job pressures are assessed with responses to the following item: “How often do
the demands of your job exceed those doable in an 8-hour workday?” Response categories are
"never" (1), "rarely" (2), "sometimes" (3), and "frequently" (4).
Job autonomy. Participants were asked: “In your current job, how often does someone else
decide how you do your work?” Response categories are "never" (1), "rarely" (2), "sometimes"
(3), and "frequently" (4). I reverse coded the responses, such that higher responses indicate more
autonomy.
Work hours. Participants were asked: “How many hours do you work in a typical week at your
main job?” I use a continuous measure of work hours.
Between wave job loss. Participants were asked: “In the past 12 months, have you become
unemployed (laid off or fired)?” Yes coded (1) and no coded (0).
Job Sector (wave 1). I create dummy variables to indicate whether participants are employed by
either the government (1) or the private sector (0).
Education (wave 1). Education is coded as an ordinal variable with the following categories:
“less than high school,” “high school or GED,” “some college or associate degree,” “four-year
college degree,” and “graduate or professional degree.”
Gender. I use dummy codes for men (0) and women (1).
Marital Status (wave 1). Marital status is coded as “married (1) versus all other categories (0).
Children (wave 1). This is coded as the number of children under 18 living in the household,
ranging from no children to 4 or more children under age 18 living in the home.
Spouse works (wave 1). Participants who have a cohabitating partner working in the paid labour
force are coded 1 versus all others (0).
24
Race. I coded participants’ race as “white” (1) versus all other categories (0).
Plan of Analyses
In table 2.2, I use ordinary least squares (OLS) regression techniques and regress each of the
three health outcomes at wave 2—psychological distress, feelings of anger, and self-rated poor
health—on episodic and persistent job insecurity, adjusting for baseline levels of health. In each
of these analyses, I test age-by-job insecurity interactions to examine whether the association
between job insecurity and health is contingent on age. I then include marital and family statuses
to examine their contribution to any potential age-by-job insecurity contingencies.
In table 2.3, I regress days absent from work at wave 2 on episodic and persistent job insecurity,
adjusting for controls and absenteeism levels at the baseline. I then test age-by-job insecurity
interactions to examine whether the association between job insecurity and absenteeism is
contingent on age. In the final model, I include change in self-rated poor health to examine its
contribution to any potential age-by-job insecurity contingencies.
With regards to the choice of longitudinal design, a panel regression model with fixed-effects
specification is generally preferred to OLS approaches that rely upon the stringent assumption
that unmeasured factors are uncorrelated with included covariates (Allison 1999). However, I use
a lagged-outcome method (where a baseline measure of the health outcome is included as a
predictor) because participants whose exposure does not change over time would be removed
from analyses under a fixed effects design. This would therefore lead to the exclusion of those
with persistent insecurity from the analyses (Burgard et al. 2009).
Additionally, a common problem of longitudinal data analysis is that sample attrition can bias the
results; this is particularly problematic if baseline measures of the dependent measure predict
attrition (Mirowsky and Ross 2001). I therefore adjust for the likelihood of attrition in each of the
regression models. This adjustment takes into account the probability that each of the health
outcomes or independent variables predict attrition. Using probit regression analyses I regressed
“absent from the sample at wave 2” on all independent variables. I then transformed predictions
to obtain the Mills Inverse Ratio and included it in all of the analyses (see Appendix I). Although
adjustment for the hazard of attrition has little effect on estimates, I decided to retain it in all
regression analyses as a precaution (Mirwosky and Ross 2001).
25
Results
Descriptives
Table 2.1 reports descriptive statistics for all variables examined in the multivariate analyses at
wave 1 and wave 2. In comparison to those in middle age (age 40 to 54) and late working life
(age 55 to 65), those in early working life (age 25 to 39) reported higher levels of distress and
anger at wave 2, and fewer physical health problems prior to wave 1. Those in early working life
also have more children in the household. In comparison to early age workers, those in middle
age have more job autonomy, higher household income and report less absenteeism in the last six
months. Those in late working life reported poorer health at wave 1, less work hours, and are less
likely to have a spouse in the labour force. No statistically significant age differences exist with
respect to perceived job insecurity. The majority of respondents perceived no job insecurity at
either wave (72 percent). About eight percent reported job insecurity at wave 1, and a similar
percentage reported it at wave 2 only. Eleven percent reported persistent job insecurity at both
waves.
Multivariate Analyses
In Table 2.2, I present regression analyses with each of the health outcomes. I first regress
psychological distress at wave 2 on perceived job insecurity. In model 1, I find no evidence of a
statistically significant effect of either episodic or persistent job insecurity on distress. In model
2, however, the association between persistent job insecurity and distress is contingent on age;
the significant interaction between middle age and persistent insecurity indicates that in
comparison to those in early working life, persistent insecurity is associated with higher levels of
distress among middle age workers. The ‘late working age’ × persistent insecurity interaction is
not significant, indicating that age moderates the association between persistent insecurity and
distress in a nonlinear pattern. In model 3, I include marital and family statuses to examine
whether they contribute to the significant age × persistent insecurity interaction revealed in
model 2. Adjusting for marital status, children in the household and spousal working status does
not change either the size of the coefficient for the middle age × persistent insecurity interaction,
nor its statistical significance. Figure 2.1 illustrates this contingency. Additionally, it is worth
noting that I observe this interaction adjusting for changes in job conditions between wave 1 and
2, including job loss, changes in work hours, job pressures and job autonomy. Among these
26
control measures, I observe that job loss and increasing job pressures between wave 1 and 2 are
associated with higher levels of distress at wave 2.
Next, I regress feelings of anger at wave 2 on perceived job insecurity and find similar patterns
to those involving psychological distress. In model 1, the associations between episodic or
persistent job insecurity and feelings of anger are not statistically significant. Examining age ×
insecurity contingencies, however, reveals that middle age workers with persistent job insecurity
report more feelings of anger at wave 2. Early age workers and late age workers are unaffected
by episodic or persistent job insecurity. Further, the inclusion of marital and family statuses in
model 3 does not attenuate the middle age × persistent insecurity interaction. Figure 2.2
illustrates this contingency. Among the control measures, job loss between wave 1 and 2 is
associated with more feelings of anger—but these adjustments do not affect the age × insecurity
contingencies.
I then regress self-rated poor health at wave 2 on perceived job insecurity. Model 1 reveals no
significant main effects of episodic or persistent insecurity on poor health. However, the
significant interaction between middle age and persistent insecurity indicates that in comparison
to those in early working life, middle age workers with persistent insecurity report poorer health.
Further, this contingency is unaffected when marital and family statuses are included (model 3).
Figure 2.3 illustrates this contingency. It is also worth noting that job loss, changes in job
pressures and job autonomy do not influence self-rated health. In addition to testing age
contingencies in the associations between perceived job insecurity and each of the health
outcomes, I also tested a series of two-way and three-way interactions involving job insecurity
and several socio-demographic characteristics, including gender and socioeconomic status. These
analyses revealed no evidence of other significant interactions with perceived job insecurity.
Finally, in Table 3.1 I regress days absent from work at wave 2 on episodic and persistent job
insecurity, adjusting for controls and absenteeism at wave 1. Model 1 reveals no evidence of a
significant effect of either episodic or persistent job insecurity on days absent from work. In
model 2, I test for age interactions to see whether the association between job insecurity and
absenteeism is contingent on age. There is some evidence that among middle age workers,
persistent insecurity is associated with a decrease in the number days that they report being
absent from work (p<.10). In model 3, I control for change in self-rated health between wave 1
27
and 2. Adjusting for changes in health, the middle age × persistent insecurity interaction becomes
significant at the p<.05 level, indicating that self-rated health partially suppresses the association
between persistent insecurity and absenteeism for middle age workers. Since persistent job
insecurity leads to an increase in self-rated poor health for middle age workers (as shown in
Table 2.2), and self-rated poor health predicts more absenteeism, once I adjust for changes in
health between wave 1 and 2, persistent job insecurity is associated with a statistically significant
decrease in absenteeism for middle age workers. Hence, were it not for the tendency of persistent
job insecurity to result in health problems for middle age workers, these workers would display
an even larger decline in absenteeism at wave 2.
In analyses not shown, a several other of firm-level and individual-level measures were tested to
see if they influenced the association between job insecurity and absenteeism; including: firm
size, job sector and years of tenure. Previous research has shown these may shape withdrawal
from work (Berry, Lelchook and Clark 2011; Caverley et al. 2007). None of these measures were
associated with the dependent variable, or influenced its association with job insecurity,
however; and their inclusion did not significantly improve model fit.
Discussion The findings of this paper support recent research that demonstrates that chronicity matters when
it comes to the stress of job insecurity (Burgard et al. 2009). I observe a decline in health only
among workers reporting prolonged exposure to perceived job insecurity. Further, only middle
age workers experience a health penalty when confronted with persistent job insecurity.
Controlling for demographic characteristics, changes in work conditions and family structure,
discernable health effects of perceived job insecurity exist only for workers age 40 to 54. Finally,
the source of these health problems—anxiety over potential job loss—may push these workers to
reduce their absenteeism rates, posing a potential double-penalty when it comes to the negative
consequences of job insecurity. I discuss the implications of these findings with regards to
perceived job insecurity as a chronic stressor, and the changing labour market experiences for
middle age and mature workers.
28
Unemployment and job loss have been consistently shown to be harmful for health. No doubt
these are important sources of stress; yet they represent a comparatively less common experience
(with the exception of periods of economic recession such as the recent ‘Great Recession’) when
contrasted against the prevalence of insecure work (Kalleberg 2009). As an indication, seven
percent of participants in this study reported a job loss in the year prior to the interview; in
comparison, nineteen percent reported that they felt they were somewhat or very likely to lose
their job in the near future. Similar numbers have been documented elsewhere along with
evidence that suggests that these figures are part of a trend involving the growth of insecure
employment in the last four decades within the United States (Burgard et al. 2009; Farber 2010).
Organizational downsizing and lean production practices that have undermined the traditional
standard employment contract have undoubtedly contributed to this trend—justifying the need
for further research on job insecurity in order to better understand its health implications within
the contemporary labour force.
In response to these changes, research has demonstrated that job insecurity is potentially as
important a work stressor as job loss, with some evidence even suggesting that it may have more
severe health consequences (Burgard et al. 2009). While the findings of this study do not
necessarily confirm this latter claim, there is clear evidence that the perception of job insecurity
has tangible health effects. Based on the Stress Process Model and the life course perspective
(Elder 1998; Pearlin et al. 1981), this paper advances on prior research by revealing how the
association between insecurity and health is contingent on both the duration and life course
timing of insecurity as a stressor.
I find no evidence that episodic job insecurity impacts psychological distress, feelings of anger or
self-rated poor health. Participant health is affected neither by the recent onset nor past
experience of insecurity. It is only the persistent experience of insecurity that is consequential for
health, which more or less corroborates the findings of Burgard and colleagues (2009). In the
present study, I find that the health consequences of prolonged exposure to job insecurity are
remarkably consistent across health measures: feelings of distress, anger and self-rated
assessments of poor health all increase at wave 2 in response to persistent insecurity—though
these effects are contingent on age. It is important to underscore that these health penalties exist
over and above the impact of job loss, or changes in job conditions that may accompany a
29
decline in job security, and which could create a spurious association between insecurity and
health. The results also hold after controlling for prior health problems that could potentially self-
select individuals into less secure jobs and create the appearance of a causal link between
insecurity and health. As such, these results provide strong evidence for the temporal health
effects of job insecurity as a chronic stressor.
Pinpointing the specific processes through which insecurity impacts health is difficult given the
particular data available and the very nature of chronic stress. The uncertainty over one’s future
employment may have no start point for some workers who are restricted to precarious
employment opportunities, or it may be aroused sharply with a particular event such as an
employer notice of future downsizing. Regardless of the nature of its inception, the present
study’s findings suggest that the health effects of job insecurity are not immediately experienced,
but rather result from prolonged exposure to this stress. This may be, in part, because body and
psyche can cope and recover from episodic or transitory experiences of insecurity with little or
no health cost, and it is only when a critical point in the accumulation of anxiety over job loss is
passed that it becomes damaging to well-being. It is also possible that these time-lagged effects
may be a product of slowly adapting coping strategies (e.g. increased alcohol consumption or
smoking) that then cause future health problems. Future research should consider behavioral
responses to job insecurity over time as potential mediating links to well-being.
Informed by the life course perspective, it also appears that the timing of job insecurity
represents an important factor in determining its potency as a stressor. I find evidence that the
health penalties of job insecurity are greatest during middle age, when the consequences of job
loss are likely perceived by workers as most severe, and potentially incongruent with
expectations about acceptable work conditions. Only among those ages 40 to 54 does persistent
job insecurity lead to greater health problems; older workers in their mid fifties to mid sixties do
not report more health issues compared to younger workers. Here, it is possible that while labour
market penalties after a job loss increase with age, older workers’ proximity to retirement
tempers the fear and stress of job loss. Middle age workers have no such proximity to retirement,
however, and yet still may be vulnerable to age-discrimination practices and wage penalties after
rehire. Indeed, labour market research shows that job loss for middle age workers is increasingly
pushing them into lower paid and more precarious work arrangements—conditions that are
30
historically atypical for this age group (Elman and O’Rand 2002). Based on these labour market
trends, in addition to the greater family-related responsibilities that individuals tend to have
during middle age, it is perhaps unsurprising that the threat of job loss is met with the most
trepidation at this stage in the life course.
Unfortunately, I am not able to directly examine the specific meanings that workers attach to job
insecurity and the threat of job loss. Thus, I can only speculate that these age-contingent health
disparities are a result of the greater salience and anxiety that middle age workers associate with
insecurity. Controlling for marital and parental statuses does not explain the age contingent
effects of job insecurity, which would have provided some evidence for the argument that family
responsibilities play a role in shaping interpretations of and reactions to job insecurity. This may
be because the measures available inadequately tap into the financial demands of middle age
(information on childcare expenditure were unavailable, for example). Nonetheless, I am able to
examine individuals’ behavioral responses to insecurity—namely their attendance rate at work.
The fact that middle age workers are the only age group in which persistent job insecurity leads
to a reduction in absenteeism provides some evidence that these workers place greater
importance on the threat of job loss. This finding is important in itself, since the very factor that
accounts for the decline in health for these workers is also the force that pressures them to reduce
their work absences, when one might expect more sickness-related absenteeism to result. In fact,
it is possible that the attempt to show organizational commitment through increased work
attendance partially contributes to these workers’ poorer health.
One limitation of this study deserves comment; one that could potentially account for the age-
contingencies presented. The longitudinal design of the study was such that only those in the
paid labour force were re-interviewed at wave 2. It is possible that participants close to
retirement who reported job insecurity at wave 1 and who then lost their job before wave 2
simply chose to permanently leave the labour force. These individuals would therefore be
removed from the follow-up interview, resulting in a tendency to underestimate older workers
with persistent job insecurity, possibly obscuring the true association between insecurity and
health among this group. Similarly, it is also possible that older workers with job insecurity were
more likely to exit the labour force due to poor health, leaving only the healthiest and most
resilient older workers at wave 2. However, all analyses adjust for the hazard of attrition to rule
31
out these potentialities. Additionally, an examination of factors predicting attrition between wave
1 and wave 2 (see Appendix I, model 2) indicates that older workers reporting job insecurity at
wave 1 were not statistically more likely to be absent from wave 2 than any other age group.
Thus, I can be relatively comfortable in ruling out attrition as a potential explanation for the age
contingencies discovered.
Finally, it is worth noting that despite this study’s finding that the negative health effects of job
insecurity are encountered only during middle age, it is possible that older cohorts in the future
may experience similar problems. Given that precarious work is becoming more widespread for
middle age workers, as these individuals approach retirement in twenty years time, increasing
numbers of them may not have the option to retire early if faced with job loss—a situation that
many older workers already exist within. One would therefore expect that since future
generations of older workers will find it necessary to work longer to acquire the resources to
ensure a comfortable retirement, job insecurity might come to have similar health effects as those
reported for middle age workers in this study. Given that the American labour force is aging,
more research focused on the experiences of older workers is necessary to understand the health
implications of these important demographic and economic trends.
Conclusion
With the emergence of the ‘New Economy,’ the nature and dynamics of the contemporary labour
market have changed in recent decades—changes that have unevenly shifted economic
opportunities and inequalities across the American population. These changes have also brought
attention to new forms of work stress, such as the interpersonal demands of service work,
increased work-family dilemmas, and as examined in this paper, uncertainty over the stable
continuity of employment. At present, this latter stressor is somewhat unique in that it offers no
readily available solution in which to contend with it. Unlike job loss, no institutionalized
supports exist to alleviate the negative health effects of job insecurity, and its ambiguous nature
makes individual coping strategies difficult. Future research is therefore vital in order to better
understand it both as an economic trend and as a role stressor that presents serious health
problems for an aging American workforce.
32
References Allison, Paul D. 1999. Multiple Regression: A Primer. Thousand Oaks, CA: Pine Forge.
Ashford, Susan J., Cynthia Lee, and Philip Bobko. 1989. “Content, Causes, and Consequences of Job Insecurity: A Theory-based Measure and Substantive Test.” Academy of Management Journal 32:803-829.
Barling, Julian and E. Kevin. Kelloway. 1996. “Job Insecurity and Health: The Moderating Role of Workplace Control.” Stress Medicine 12:253-259.
Becker, Gary S. 1962. “Human Capital: A Theoretical Analysis.” The Journal of Political Economy 70:9-42.
Bendick, Marc, Lauren E. Brown, and Kennington Wall. 1999. “No Foot in the Door: An Experimental Study of Employment Discrimination Against Older Workers.” Journal of Aging and Social Policy 10:5–23.
Berry, Christopher M., Ariel M. Lelchook, and Malissa Clark. 2011. “A Meta-analysis of the Interrelationships between Employee Lateness, Absenteeism, and Turnover: Implications for Models of Withdrawal Behavior.” Journal of Organizational Behavior Published Online 21/08/2011.
Bourbonnais, Renée, Chantal Brisson, Michel Vézina, Benoît Masse, and Caty Blanchette. 2005. “Psychosocial Work Environment and Certified Sick Leave Among Nurses during Organizational Changes and Downsizing.” Relations Industrielles 60:483–509.
Brenner, M. H. 1973. Mental Illness and the Economy. Cambridge: Harvard University Press.
Brockner, J. 1990. “Scope of Justice in the Workplace: How Survivors React to Co-Worker Layoffs.” Journal of Social Issues 46:95-106.
Burgard, Sarah A., Jennie Brand, and James S. House. 2009. “Perceived Job Insecurity and Worker Health in the United States.” Social Science & Medicine 69:777-785.
Cappelli, Peter. 2008. Talent on Demand: Managing Talent in an Age of Uncertainty. Boston, MA: Harvard Business School Press.
Caverley Natasha, J. Barton Cunnigham, James N. MacGregor. 2007. “Sickness Presenteeism, Sickness Absenteeism, and Health following Restructuring in a Public Service Organization.” Journal of Management Studies 44:304 –319.
Cheng, Grand. H. L. and Darius K. S. Chan. 2008. “Who Suffers More from Job Insecurity? A Meta-Analytic Review.” Applied Psychology 57: 272–303.
Couch, Kenneth A. 1998. “Late Life Job Displacement.” The Gerontologist 3:7-17.
Couch, Kenneth A., Nicholas A. Jolly, and Dana W. Placzek. 2009. “Earnings Losses of Older Displaced Workers: A Detailed Analysis with Administrative Data.” Research on Aging 31:17–40.
De Witte, Hans. 1999. “Job Insecurity and Psychological Health: Review of the Literature and Exploration of Some Unresolved Issues.” European Journal of Work and Organizational Psychology 8:155-177.
33
De Cuyper, Nele. and Hans De Witte. 2006. “The Impact of Job Insecurity and Contract Type on Attitudes, Well-being and Behavioral Reports: A Psychological Contract Perspective. Journal of Occupational and Organizational Psychology 79:395– 409.
Elder, Glen H., Jr. 1998. “Life Course Theory and Human Development.” Sociological Analysis 1:1-12.
Elman, Cheryl, and Angela M. O’Rand. 2002. “Perceived Job Insecurity and Entry into Work-Related Education and Training Among Adult Workers.” Social Science Research 31:49-76.
Erikson, Erik. H. 1998. The Life Cycle Completed. NY: Norton.
Fallick, Bruce. C. 1996. “A Review of the Recent Empirical Literature on Displaced Workers.” Industrial and Labour Relations Review 50:5–16.
Farber, Henry S. 1993. “The Incidence and Costs of Job Loss: 1982-91.” Brookings Papers on Economic Activity: Microeconomics 73-119.
______. 2008. “Short(er) Shrift: The Decline in Worker-Firm Attachment in the United States.” Pp. 10–37 in Laid Off, Laid Low: Political and Economic Consequences of Employment Insecurity, edited by K. S. Newman. New York: Columbia University Press.
______2010. “Job Loss and the Decline of Job Security in the United States.” Pp. 223-267 in Labour in the New Economy edited by Katherine Abraham, James Spletzer and Michael Harper. University of Chicago Press, Chicago.
Ferrie, Jane Elizabeth, Martin John Shipley, Michael Gideon Marmot, Stephen Alfred Stansfeld, and George Davey Smith. 1995. "Health Effects of Anticipation of Job Change and Nonemployment: Longitudinal Data from the Whitehall II Study." British Medical Journal 311:1264-1269.
Firns, Ian, Anthony Travaglione, and Grant O’Neill. 2006. “Absenteeism in Times of Rapid Organizational Change.” Strategic Change 15:113–128.
Fullerton, Andrew S. and Michael Wallace. 2007. “Traversing the Flexible Turn: US Workers’ Perceptions of Job Security, 1977–2002.” Social Science Research 36:201-221.
Greenhalgh, Leonard. 1979 “Job Security and the Disinvolvement Syndrome: An Exploration of Patterns of worker Behaviour under Conditions of Anticipatory Grieving over Job Loss. Unpublished doctoral dissertation, Cornell University.
Greenhalgh, Leonard, and Zehava Rosenblatt. 1984. “Job insecurity: Toward Conceptual Clarity.” Academy of Management Review 3:438–448.
Heaney, Catherine, Barbara A. Israel, and James S. House. 1994. “Chronic Job Insecurity Among Automobile Workers: Effects on Job Satisfaction and Health.” Social Science & Medicine 38:1431–1437.
Heponiemi, Tarja, Marko Elovainio, Jaana Pentti, Marianna Virtanen, Hugo Westerlund, Pekka Virtanen, Tuula Oksanen, Mika Kivimäki, Jussi Vahtera. 2010. “Association of Contractual and Subjective Job Insecurity With Sickness Presenteeism Among Public Sector Employees.” Journal of Occupational and Environmental Medicine 52:830-835.
34
Hirsch, Barry T., David A. MacPherson, and Melissa A. Hardy. 2000. “Occupational Age Structure and Access for Older Workers.” Industrial and Labour Relations Review 53:401–418.
Hulin, C. L. 1991. “Adaptation, Persistence, and Commitment in Organizations.” In M. D. Dunnette, and L. M. Hough (Eds.), Handbook of Industrial and Organizational Psychology (2nd ed., Vol. 2, pp. 445–505). Palo Alto, CA: Consulting Psychologists Press.
Idler, Ellen L. and Stanislav Kasl. 1991. “Health Perceptions and Survival: Do Global Evaluations of Health Status Really Predict Mortality?” Journal of Gerontology 46:S55-S65.
Jick, Todd D. and Leonard Greenhalgh. 1989. "Survivor Sense Making and Reactions to Organizational Decline: Effects of Individual Differences." Management Communication Quarterly 2:305-327.
Koslowsky, M. 2009. “A Multi-level Model of Withdrawal: Integrating and Synthesizing Theory and Findings.” Human Resource Management Review 19:283–303.
Johnson, Richard W. and Corina Mommaerts. 2011. “Age Differences in Job Loss, Job Search, and Reemployment.” Washington, DC: The Urban Institute.
Kalleberg, Arne L. 1996. "Changing Contexts of Careers: Trends in Labor Market Structures and Some Implications for Labor Force Outcomes." Pp.343-58 in Generating Social Stratification: Toward a New Research Agenda, edited by Alan C. Kerckhoff. Boulder, CO: Westview Press.
______. 2009. “Precarious Work, Insecure Workers: Employment Relations in Transition.” American Sociological Review 2009 74:1–22.
______. 2011. Good Jobs, Bad Jobs: The Rise of Polarized and Precarious Employment Systems in the United States, 1970s-2000s. New York: Russell Sage Foundation, American Sociological Association Rose Series in Sociology.
Kasl, Stanislav V., Susan Gore, and Sidney Cobb. 1975. “The Experience of Losing a Job: Reported Changes in Health, Symptoms and Illness Behavior.” Psychosomatic Medicine 37:106–122
Kessler, Ronald C., J. Blake Turner, and James S. House. 1987. “Unemployment, Reemployment, and Emotional Functioning in a Community Sample.” American Sociological Review 54:648–657.
Kessler, Ronald C, G. Andrews, L. J. Colpe, E. Hiripi, D. K. Mroczek, S-L. T. Normand, E. E. Walters, and A. M. Zaslavsky. 2002. “Short Screening Scales to Monitor Population Prevalences and Trends in Non-specific Psychological Distress.” Psychological Medicine 32:959–976.
King, James E. 2000. “White-Collar Reactions to Job insecurity and the Role of the Psychological Contract: Implications for Human Resource Management.” Human Resource Management 39:79–91.
Lazarus, Richard S. and Susan Folkman. 1984. Stress, Appraisal, and Coping. New York: Springer.
35
Lee, Sunmin, Graham A. Colditz, Lisa F. Berkman, and Ichiro Kawachi. 2004. “Prospective Study of Job Insecurity and Coronary Heart Disease in US Women.” Annals of Epidemiology 14:24–30.
Lim, Vivien K. G. 1996. “Job Insecurity and its Outcomes: Moderating Effects of Work-based and Nonwork-based Social Support.” Human Relations 49:171–194.
Mirowsky, John, and Catherine E. Ross. 2001. “Age and the Effect of Economic Hardship on Depression.” Journal of Health and Social Behavior 42:132–50.
Pearlin, Leonard I., Morton Lieberman, Elizabeth Menaghan, and Joseph Mullan. 1981. “The Stress Process.” Journal of Health and Social Behavior 22:337–56.
Pearlin, Leonard I. and Marilyn M. Skaff. 1996. “Stress and the Life Course: A Paradigmatic Alliance.” The Gerontologist 36:239-247.
Polivka, Anne E. and Thomas Nardone. 1989. “On the Definition of Contingent Work.” Monthly Labour Review 112:9–14.
Polivka, Anne E. 1996. “A Profile of Contingent Workers.” Monthly Labour Review 119:10–21.
Probst Tahira M, Susan M. Stewart, Melissa Gruys, and Bradley W. Tierney. 2007. “Productivity, Counterproductivity and Creativity: The Ups and Downs of Job Insecurity.” Journal of Occupational and Organizational Psychology 80:479–497.
Radloff, L. S. 1977. “The CES-D Scale: A Self-Report Depression Scale for Research in the General Population.” Applied Psychological Measurement 1:385-401.
Rodriguez, D. and M. Zavodny 2003. “Changes in the Age and Education Profile of Displaced Workers.” Industrial and Labour Relations Review 56:498-510.
Rousseau, Denise M. 1995. Psychological Contracts in Organizations: Understanding Written and Unwritten Agreements. Thousand Oaks, CA.
Ross, Catherine E. and John Mirowsky. 1984. “Components of Depressed Mood in Married Men and Women: The Center for Epidemiological Studies’ Depression Scale.” American Journal of Epidemiology 119:997-1004.
Salzman, H. 1998. "Restructuring and Skill Needs: Will Firms Retrain?" Pp. 125- 140 in: The Changing Educational Quality of the Workforce, edited by R. Zemsky and P. Capelli. Annals of the American Academy of Political and Social Science Thousand Oaks: SAGE.
Schaufeli, W. 1992. “Unemployment and Mental Health in Well and Poorly Educated Schoolleavers.” Pp. 253–271 in On the Mysteries of Unemployment: Causes, Consequences and Policies, edited by C. Verhaar and L. Jansma. Dordrecht, The Netherlands: Kluwer Academic Publishers.
Schieman, Scott, and Sarah Reid. 2009. “Job Authority and Health: Unraveling the Competing Suppression and Explanatory Influences.” Social Science and Medicine 69:1616-1624.
Seeman, Teresa E, Bruce S. McEwen, John W. Rowe, and Burton H. Singer. “Allostatic Load as a Marker of Cumulative Biological Risk: MacArthur Studies of Successful Aging.” Proceedings of the National Academy of Sciences of the United States of America 98:4770-4775.
36
Sverke, Magnus, and Johnny Hellgren. 2002. “The Nature of Job Insecurity: Understanding Employment Uncertainty on the Brink of a New Millennium.” Applied Psychology 51:23–42.
Szinovacz, Maximiliane E. and Adam Davey. 2005. “Predictors of Perception of Involuntary Retirement.” The Gerontologist 45:36-47.
Tausig, Mark, and Rudy Fenwick. 1999. “Recession and Well-being.” Journal of Health and Social Behavior 40:1-16.
Toossi, Mitra. 2009. “Labour Force Projections to 2018: Older Workers Staying More Active.” Monthly Labour Review November 30–51.
Turner, Heather A. and Scott Schieman. 2007. “Stress Processes Across the Life Course: Introduction and Overview.” Pp. 1-15 in Stress Processes Across the Life Course: Advances in Life Course Research, edited by Heather A. Turner and Scott Schieman. Elsevier.
Wheaton, Blair. 1990. “Life Transitions, Role Theories and Mental Health.” American Sociological Review 55:209–223.
______. 1999. “Social Stress.” Pp. 277-300 in The Handbook of the Sociology of Mental Health, edited by C.S. Aneshensel and J.C. Phelan. Kluwer Academic/Plenum.
Wheaton, Blair and Sarah Reid. 2008. “The Role of Timing vs. Duration in the Cumulative Work History Effects of Job Exits and Nonemployment on Women’s Mental Health.” Stress Processes across the Life Course: Advances in Life Course Research 13:193–232. Emerald GroLimited.
37
TABLE 2.1 Means and Proportions of all Study Variables Early Work
Age (25-39) Middle Age (40-54)
Late Work Age (55-65) Total
Health outcomes (Wave 2) Psychological Distress 1.734 1.532* 1.264** 1.555 Anger 1.941 1.477*** 1.100*** 1.560 Self-rated Poor Health 2.313 2.369 2.435 2.373 Wave 1 and earlier health Psychological Distress 1.944 1.744* 1.442** 1.772 Anger 1.979 1.611*** 1.116*** 1.632 Self-rated Poor Health 2.254 2.372 2.438* 2.339 Health Problems Prior to Wave 1 .356 .625*** .767*** .5648 Absenteeism Days Absent from Work at Wave 1 3.794 3.428 3.068 3.471 Days Absent from Work at Wave 2 4.599 3.253* 4.589 3.965 Perceived job insecurity Perceived Insecurity at Wave 1 .199 .200 .180 .193 Perceived Insecurity at Wave 2 .202 .185 .170 .195 Episodic and Persistent Insecurity Absent Insecurity .702 .721 .751 .722 New (Wave 2 only) .098 .078 .068 .083 Resolved (Wave 1 only) .088 .080 .068 .081 Persistent (Wave 1 and Wave 2) .111 .118 .112 .115 Job conditions (Wave 1) Work Hours (Weekly) 43.303 44.699 38.022* 42.323 Job Pressures 3.033 2.991 2.879 2.976 Job Autonomy 2.356 2.552* 2.555 2.469 Employed in Public Sector .255 .282 .241 .265 Objective job insecurity Job Loss between Wave 1 & 2 .078 .071 .040 .069 Socio-demographics (Wave 1) Female .643 .569* .634 .608 Married .594 .627 .578 .611 Spouse Works .529 .529 .317*** .491 Children in the Household 1.512 .792** .254*** .914 White .702 .775* .787 .751 Household Income 53264.38 65498.45*** 50551.12 59046.68 Education 3.722 3.699 3.831 3.707 N 295 450 211 956 Comparison with the “Early Work Age” group is statistically significant at * p < .05; ** p < .01; *** p < .001 (two–tailed test).
38
TABLE 2.2 Regression of Psychological Distress, Anger, and Self-rated Poor Health at wave 2 on Perceived Job Insecurity, Age, Interactions, Job Conditions and Controls (N=956)
Psychological Distress Anger Self-rated Poor Health Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3
Perceived Job Insecurity New Job Insecurity a .086 -.031 -.024 .089 .017 .025 .103 .017 .032 Resolved Job Insecurity a .003 .130 .146 .036 -.282 -.252 .008 -.130 -.137 Persistent Job Insecurity a .213 -.225 -.226 .104 -.392* -.368* .152 -.146 -.141 Age Age 40 to 54 b .284 .186 .198 -.343* -.413*** -.467** .005 -.095 -.010 Age 55 to 65 b .233 .207 .181 -.407* -.415** -.518** .028 -.031 .021 Insecurity × Age Interactions Age 40 to 54 × New a ——
.264
.243
——
.070
.078
——
.145
.151 Age 40 to 54 × Resolved a
——
-.135
-.159
——
.390
.361
——
.252
.262 Age 40 to 54 × Persistent a
——
.737**
.742**
——
1.110***
1.098***
——
.420*
.424* Age 55 to 65 × New a ——
.010
.000
——
.174
.127
——
.111 .072 Age 55 to 65 × Resolved a ——
-.344
-.354
——
.694 .673 ——
.121
.127 Age 55 to 65 × Persistent a ——
.459
.486
——
.311
.336
——
.310
.308 Job Experiences/Conditions
Job Loss between Wave 1 & 2 .391* .383* .382* .292* .292* .270 .109 .103 .103 Employed in Public Sector .205* .208* .196* .094 .088 .068 .048 .053 .052 Work Hours (Wave 1) -.006 -.006 -.007 .002 .001 .001 .001 .001 .001 Change in Work Hours .003 .002 .002 .006 .007 .007 -.000 -.000 -.000 Job Pressures (Wave 1) .021 .025 .024 .014 .013 .018 .047 .046 .047 Change in Job Pressures .140** .146** .146** .056 .063 .061 .035 .034 .034 Job Autonomy (Wave 1) -.110* -.114* -.112* -.067 -.070 -.062 .019 .017 .017 Change in Job Autonomy -.061 -.067 -.063 -.054 -.048 -.047 .036 .033 .035 Controls Married c ——
——
.081 ——
——
-.065 ——
——
.057 (Continued on Next Page)
39
Table 2.2 Continued Psychological Distress Anger Self-rated Poor Health Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3
Children in the Household ——
——
.026
——
——
.081**
——
——
.013 Spouse Works ——
——
-.163
——
——
-.047
——
——
.042 Household Income (Wave 1) -.000 -.000 -.000 .000 .000 .000 -.000 -.000 -.000
Change in Household Income -.000 -.000 -.000 -.000 -.000 -.000 .000 .000 .000 Female .244** .248** .267** -.036 -.028 -.042 .046 .043 .066 White -.039 -.037 -.035 -.016 -.007 .016 -.103 -.100 -.101 Education .013 .010 .020 -.009 -.012 -.001 -.047 -.050 -.025 Wave 1 and earlier Health Health Problems prior to W1 .083 .084 .081 .089 .091 .087 .130*** .125*** .127*** Health Outcome at Wave 1 .419*** .417*** .414*** .473*** .481*** .478*** .533*** .535*** .522*** Hazard of Attrition -4.595** -4.955** -4.954* .387 .428 .644 -.100 .001 -.113 Constant 3.787 3.852 4.014 .768 .836 .583 1.116 1.122 1.691 R2 .361 .368 .373 .308 .319 .325 .388 .393 .394 * p < .05; ** p < .01; *** p < .001 (two–tailed test). a Compared to insecure at neither wave b Compared to age 25 to 39. c Compared to Never Married and Previously Married Note: Unstandardized regression coefficients are shown in the table.
40
Table 2.3 Regression of Days Absent from Work at wave 2 (logged) on Perceived Job Insecurity, Age, Interactions, Job Conditions and Controls (N=956)
Model 1 Model 2 Model 3 Perceived Job Insecurity New Job Insecurity a .106 .144 .121 Resolved Job Insecurity a -.056 -.052 -.030 Persistent Job Insecurity a -.004 .117 .127 Age Age 40 to 54 b -.008 .012 -.087 Age 55 to 65 b .076 .091 -.040 Insecurity × Age Interactions Age 40 to 54 × New a ——
-.086
-.097 Age 40 to 54 × Resolved a ——
.052
.021 Age 40 to 54 × Persistent a ——
-.210
-.257* Age 55 to 65 × New a ——
.002
.032 Age 55 to 65 × Resolved a ——
-.127
-.144 Age 55 to 65 × Persistent a ——
-.109
-.171 Job Experiences/Conditions
Job Loss between Wave 1 & 2 .144* .147* .132 Employed in Public Sector .084* .083* .073 Work Hours (Wave 1) .001 .001 .001 Change in Work Hours -.001 -.001 -.001 Job Pressures (Wave 1) .012 .013 .013 Change in Job Pressures -.011 .014 -.021 Job Autonomy (Wave 1) -.050* -.049* -.043 Change in Job Autonomy -.025 -.026 -.030 Controls Married c -.128* -.133* -.192** Children in the Household .021 .022 .020 Spouse Works -.003 -.003 -.009 Household Income (Wave 1) .000 .000 .000 Change in Household Income .000 .000 .000 Female .045 .040 .007 White .018 .018 .031 Education .026 .026 .009 Wave 1 and Changes in Health Days Absent at Wave 1 .215*** .217*** .208*** Self-rated Poor Health at Wave 1 ——
——
.095*** Change in Self-rated Poor Health ——
——
.112*** Hazard of Attrition -1.175 -1.085 .177 Constant 1.896 1.832 1.523
R2 .145 .148 .176 * p < .05; ** p < .01; *** p < .001 (two–tailed test). a Compared to insecure at neither wave b Compared to age 25 to 39. c Compared to Never Married and Previously Married Note: Unstandardized regression coefficients are shown in the table.
41
FIGURE 2.1 Age Differences in the Effect of Job Insecurity on Psychological Distress
Note: Predicted values shown above are derived from Model 3 of Table 2.2, with wave 2 distress as the dependent variable. All control variables are held constant at their respective means.
1.2
1.4
1.6
1.8
2.0
2.2
2.4
Absent Resolved New Persistent
Psyc
holo
gica
l Dis
tress
(Wav
e 2)
Perceived Job Insecurity
Ages 25 - 39 Ages 40 - 54 Ages 55 - 65
42
FIGURE 2.2 Age Differences in the Effect of Job Insecurity on Anger
Note: Predicted values shown above are derived from Model 3 of Table 2.2, with wave 2 anger as the dependent variable. All control variables are held constant at their respective means.
1.0
1.2
1.4
1.6
1.8
2.0
2.2
Absent Resolved New Persistent
Ange
r (W
ave
2)
Perceived Job Insecurity
Ages 25 - 39 Ages 40 - 54 Ages 55 - 65
43
FIGURE 2.3 Age Differences in the Effect of Job Insecurity on Self-rated Poor Health
Note: Predicted values shown above are derived from Model 3 of Table 2.2, with wave 2 self-rated poor health as the dependent variable. All control variables are held constant at their respective means.
2.0
2.1
2.2
2.3
2.4
2.5
2.6
Absent Resolved New Persistent
Self-
rate
d Po
or H
ealth
(Wav
e 2)
Perceived Job Insecurity
Ages 25 - 39
Ages 40 - 54
Ages 55 - 65
44
Chapter 3 Control in the Face of Uncertainty: The Sense of Personal Control
and the Link between Job Insecurity and Distress
Introduction The growth of uncertain and unstable employment in recent decades has drawn increased
attention from health researchers interested in the implications of this trend for workers’ well-
being (Kalleberg 2011). This research has moved beyond an investigation of the health
consequences of unemployment (Kessler, House and Turner 1989; Tausig and Fenwick 1999),
toward a consideration of how a wider range of negative employment experiences influence
health, especially the threat of job loss and uncertain future employment—key features of what is
often referred to as ‘job insecurity’ (Burgard, Brand, and House 2009; De Witte 1999). In some
cases, the threat of job loss has been found to exert similar health penalties as unemployment
(Burgard et al. 2009; Heaney, Israel, and House 1994; Lee et al. 2004). Consequently, job
insecurity is considered an important contemporary source of chronic stress and a salient
predictor of poor health (Sverke and Hellgren 2002).
Given the current and rather persistent strife in the American and Canadian economies, job
insecurity has become quite normative for many people. The question becomes: How can
workers cope with the potential health consequences? One set of potential answers is based on
the ideas of the Stress Process Model (Pearlin et al. 1981): The stress of job insecurity may be
buffered by the availability of certain coping resources. In this paper, I focus on one particular
resource—the sense of control—a personal resource that may help to mitigate the stress
associated with uncertain, unpredictable, and tenuous employment experiences (Pearlin et al.
2005). While sense of control has been shown to buffer the impact of a variety of role stressors
on health (Avison 2001; Wheaton 1983), it has received only limited empirical attention with
respect to the stress of job insecurity. Given that economic uncertainty is one of the key
explanations for why job insecurity so stressful, a strong sense of agency and personal control
may alter the meaning and perceived threat attached to this uncertainty.
While sense of control may moderate the association between job insecurity and distress, it may
also represent a key mediating mechanism through which job insecurity reduces well-being.
45
Perceived control over one’s life may foster protective beliefs and behaviors that lessen the stress
of job insecurity, but the very uncertain and controllable nature of a threatened employment
situation may come to erode one’s sense of control, which is an important component of good
mental health (Ross and Sastry 1999). To-date, I know of no study that has examined these
mediating and moderating possibilities involving sense of personal control, job insecurity and
health in a population-based sample of workers from a broad range of social statuses and
occupational backgrounds.
To answer these questions, I examine data from two national samples of American and Canadian
workers in 2005 and 2011, respectively. I test whether sense of personal control both mediates
and moderates the association between perceived job insecurity and psychological distress.
Additionally, given that some research finds age differences in the health consequences of job
insecurity (Cheng and Chan 2008), I consider whether the propensity for sense of control to
buffer the stress of job insecurity varies across the life course. In the following section I discuss
possible age contingencies in more detail as part of a review of the literature on the linkages
between job insecurity, health, and the sense of personal control.
Literature Review
Job Insecurity and Health
In recent decades, organizational downsizing and the growth of nonstandard work arrangements
have led to a decline in job security and the erosion of the guarantee of long-term, stable
employment in the United States (Kalleberg 2011). Much of the research examining these trends
and their consequences for well-being focuses on the extent that workers consider their job as
insecure, often referred to as “perceived job insecurity.” Greenhalgh and Rosenblatt (1984:438)
define perceived job insecurity as a: “perceived powerlessness to maintain desired continuity in a
threatened job situation.”
The association between job insecurity and health can be interpreted through Jahoda’s (1982)
latent deprivation model and the Stress Process Model (Pearlin et al. 1981). Given that
employment provides access to valued material, psychological and social resources, the
46
anticipation of job loss and the future unavailability of these resources represent a potential role
stressor, particularly when this ‘anticipated’ future event is uncontrollable and unpredictable
(Lazarus and Folkman 1984; Wheaton 1999). In fact, it is suggested that it is the unpredictability
and uncertainty associated with this stressor that makes it so stressful—potentially making it a
more potent stressor than job loss or episodes of unemployment (Burgard et al. 2009). These
theoretical explanations are validated by a variety of qualitative studies of workers in firms that
have downsized, and for whom the prospect of future job loss is likely (Greenhalgh and
Rosenblatt 1984; Jick and Greenhalgh 1989). Recurring themes in these studies were workers’
reports of the stress associated with the uncertainty over their employment situation, and their
feelings of helplessness in the face of impending job loss. In accordance with this qualitative
evidence, a growing body of quantitative research has demonstrated that job insecurity has
tangible consequences for health. Perceptions of job insecurity are associated with increased
feelings of depression, anxiety and poor self-rated health (Burgard et al. 2009; Ferrie et al. 1995).
Given the growth of contingent work arrangements in the last few decades (e.g. temporary and
contract work), some scholars contend that job precariousness represents a critical public health
concern of potentially epidemic proportions (Fischer and Thayer 2006).
Does the Sense of Control Buffer the Stress of Job insecurity?
What factors might counteract or protect workers from the health effects of job insecurity?
According to the Stress Process Model (Pearlin et al. 1981), the availability of certain coping
resources may mitigate the harmful consequences of job insecurity. A feature of the self-concept,
the sense of personal control represents one such resource. Similar to constructs such as self-
efficacy and mastery, the sense of control refers to the degree to which that an individual
believes they have control over the things that happen in their life (Turner and Roszell 1994;
Mirowsky and Ross 2003; Pearlin and Schooler 1978). Research has demonstrated that a strong
sense of control fosters good physical and mental well-being, encouraging flexible coping
strategies and resilience in response to adverse or challenging events (Turner and Roszell 1994;
Ross and Mirowsky 1989).
Research and theory in the mental health literature often frames sense of control as a ‘buffering
resource’ with “the capacity to hinder, prevent, or cushion the development of the stress process
and its outcomes.” (Pearlin 2006:405). With regards to stress associated with the threat of job
47
loss, this buffering effect may operate in two related ways. First, individuals with a high sense of
control are more likely to perceive potential job loss as less threatening and of less consequence.
Given that one of the major purported mechanisms through which job insecurity manifests itself
as a stressor is through the imagined consequences of job loss (Sverke and Hellgren 2002),
individuals with a low sense of control may be particularly vulnerable since they are more likely
to consider the worst case scenarios associated with the event. In contrast, a high sense of control
likely fosters a more positive outlook that downplays the negative aspects of job loss, while up-
playing the positive possibilities—viewing a layoff as an opportunity for job mobility, for
example. Thus, a high sense of control is likely to negate a key deleterious feature of job
insecurity—anxiety over an uncertain future—because it engenders a belief that one can deal
with and overcome unforeseen eventualities such as job loss (Pearlin and Aneshensel 1986).
Additionally, individuals with a high sense of control are more likely to pursue strategies that
reduce the impact of future job loss—such as retraining or proactively looking for more secure
employment—further reducing its perceived threat and future health consequences. In a small
longitudinal study of dual-earner couples facing job loss (N=79), Sweet and Moen (2012) find,
for example, that those who actively prepared in advance of the layoff reported less severe health
problems and emotional challenges after the event. Thus, as a social-psychological trait that taps
into one’s level of perceived agency, the sense of control should influence the subjective
meanings that individuals assign to the threat of job loss, as well as their behavioral responses to
the threat.
While other empirical research documents how the sense of control buffers the effects of a
variety of life stressors on health (Avison 2001; Pearlin et al. 2007; Wheaton 1983), there is only
limited evidence of it acting as a buffer in the face of job insecurity. Based on a sample of
Swedish female nurses, Naswall, Sverke and Hellgren (2005) find job insecurity was more
strongly associated with poor psychological health among those who attributed outcomes in their
life to forces outside of their control. Most research, however, utilizes either control tied to one’s
job, or job-specific measures of efficacy or control, rather than a general sense of control.
Barling and Kelloway (1996), for example, find that perceptions of workplace control moderated
the impact of perceived job insecurity on both blood pressure and physical health among a small
sample of coal miners. Similarly, Schreurs et al. (2010) find that high levels of job autonomy
48
weaken the effect of job insecurity and health. The same authors found no evidence of a
moderating effect of job self-efficacy, however.
Despite the paucity of evidence on the subject, some research is indicative of the mechanisms
through which the benefits of the sense of control may be operate: one study of layoff survivors
found that those with a high sense of mastery were more likely to engage in control-oriented
coping strategies in response to the threat of job loss, including actions such as discussing their
situation with their supervisor (Armstrong-Stassen 1993). These behaviors may not remove the
threat of job loss, but they may nonetheless serve to reduce the anxiety that one attaches to the
threat (Kinnunen, Mauno and Siltaloppi 2010). Collectively, then, the existing empirical
literature is suggestive of the resource-aspects of the sense of control, though there remains an
absence of larger representative studies that have validated it as a resource within the wider
labour force.
Sense of Control as a Mediator
Research and theory recognizes that while personality traits and coping skills may buffer and
weaken the health effects of stressors, these resources are also susceptible to the conditions
whose effect they moderate (Pearlin et al. 2007). Since the sense of control is not a fixed aspect
of personality but rather one that is shaped by objective, social conditions (Skaff, Pearlin, and
Mullan 1996), it is plausible that job insecurity may itself influence perceptions of personal
control, that then subsequently impact health.
Several classic studies have demonstrated how the self-concept, and the sense of control in
particular, is influenced by the social organization of work (Kohn and Schooler 1983; Seeman
1967). Research has also revealed how unemployment exerts a strong negative effect on feelings
of mastery (Avison 2001; Broman, Hamilton, and Hoffman 2000). Given the centrality of work
in most peoples’ lives, the inability or powerlessness to prevent the loss of employment strikes at
the very heart of the belief that one has control over the things that happen in their life.
Moreover, unlike job loss, an event and role transition that an individual can directly observe and
respond to, the very ambiguity and uncertainty surrounding the possibility of being laid off may
hinder the ability to resolve and come to terms with the threat (Burgard et al. 2009). Without
such recourse, then, it is very likely that one will experience a decreased sense of control in their
49
life. And, since the link between a low sense of control and poorer health has been well-
established (Mirowsky and Ross 1989; Wheaton 1983), I expect one of primary mechanisms
through which job insecurity reduces well-being is through its propensity to reduce perceptions
of personal control. Sense of control should therefore partially mediate the association between
job insecurity and health.
Age Contingencies
Emerging evidence suggests that the processes involving sense of control, job insecurity and
health may vary in their nature and extent across the life course (Cheng and Chan 2008). I
therefore also consider whether age, as a key life course marker, influences 1) health levels
associated with job insecurity, and 2) the propensity of sense of control to buffer the stress of job
insecurity.
A life course approach to stressors and their consequences for well-being suggests that both
stressful events and nonevents potentially vary in their meaning and impact across salient stages
in individuals’ lives (Elder 1998; Sweet and Moen 2011); in the case of job insecurity, across life
course stages as they align with one’s career (i.e. early, mid or late career). Thus, the stage in
one’s career at which job insecurity is encountered may influence the level of threat that an
individual attaches to possible job loss, their perception of available resources to deal with the
threat, and the extent that this threat conflicts with expectations about acceptable work
conditions—factors that should shape the degree that job insecurity is distressing. Specifically,
several age- and cohort-patterned factors appear relevant, including one’s current occupational
status and experience, expectations regarding future career and nonwork trajectories (e.g.
retirement plans), standards about acceptable or desirable working conditions, as well as the
presence of nonwork demands, such as family-related financial obligations. Since the existence
and extent of these factors are both age-specific and cohort-specific, we should therefore see age
differences in the stress associated with job insecurity.
Some suggest, for example, that the stress of job insecurity should be most harmful during
middle age when workers are most dependent on steady, full-time employment in order to meet
family responsibilities (Fullerton and Wallace 2007; László et al 2010). Here, the risk of job loss
may carry greater significance in comparison to younger workers who have less financial
50
responsibilities, as well as older workers whose proximity to retirement age may reduce the
perceived threat of job loss (Brenner 1973; Erikson 1998). Additionally, in comparison to
workers entering the labour market, middle age workers have generally accumulated higher
levels of occupational prestige and status; the onset of job insecurity may be particularly
incongruent to these workers’ beliefs about appropriate treatment from their employer. From this
view, then, which I label the stress of middle age, the negative health effects of precarious and
uncertain employment should be greatest during middle age.
Alternatively, other researchers have argued that the health penalties of job insecurity may
increase linearly with age (Cheng and Chan 2008). Given that older workers tend to experience
more severe economic penalties when job loss occurs—older workers who are laid-off, for
example, take longer to find re-employment and encounter greater earnings losses in their new
jobs (Hirsch, MacPherson, and Hardy, 2000)—the perceived consequences of job loss may
engender greater feelings of anxiety among workers in the later stages of their careers. Moreover,
older workers may consider job insecurity a particularly jarring experience given that this
cohort—many of whom are of the baby-boomer generation—enjoyed high levels of job security
earlier in their careers (Farber 2010). Finally, since workers are now delaying retirement due to
insufficient retiree plans and social security benefits (Toossi 2009), older workers facing the
prospect of job loss may perceive it with the same trepidation as their younger counterparts.
From this view, then, the stress of job insecurity may be greatest among older workers; a view I
label as the stress of aging hypothesis.
Data and Methods
To test the hypotheses described above, I analyze data from two national surveys of workers in
the United States and Canada: The Work, Stress and Health study (WSH) involved telephone
interviews with adults in the 50 United States in 2005 and 2007. To obtain the sample, a list-
assisted random digit dialing (RDD) selection was used and drawn proportionally from all 50
states from GENESYS Sampling Systems. Eligible participants are 18 years of age or older and
participating in the paid labour force. Study participants received a $30 gift card for completing
the interview Seventy-one percent of eligible individuals were successfully interviewed yielding
51
a sample of 1,800 adults. The working sample includes non-self employed between 25 and 65
years of age (N=1,219), working twenty hours or more a week.
The Canadian Work Stress and Health study (CAN-WSH), is a nationally representative sample
of the Canadian labour force. Interviews were conducted by telephone between January and July
2011. To be eligible to participate in the study, individuals had to be: (1) residing in Canada; (2)
18 years of age or older; (3) currently working at a paid job or operated an income-producing
business; (4) employed in the civilian labour force; and 5) living in a non-institutional residence.
Calls were made to a regionally stratified unclustered random probability sample generated by
random-digit-dial methods. Study participants received a $20 gift card for completing the
interview. The final full sample was 6,005. The response rate was approximately 40 percent.2
The working sample includes non-self employed workers between 25 and 65 years of age
(N=3,790), working twenty hours or more a week.1
I briefly explain the rationale for the sample selection. I limit the sample to non self-employed
workers between the ages of 25 and 65. The measure of job insecurity that I use is designed to
tap into workers perceptions of the strength of the employer-employee psychological contract
(Rousseau 1995); as such, even though self-employed workers may face future threats to their
continuing employment, this measure of insecurity is not applicable to workers who are not hired
by an employer. Additionally, as is the case with most studies of job insecurity, respondents
between the ages of 18-24 are omitted from the working sample, since this group are
overridingly in contingent jobs due their disproportionate participation in higher education
(Polivka 1996).
Measures
Psychological distress. In the WSH study I use an amalgamation of items from the Center for
Epidemiological Studies Depression Scale (CES-D, see Radloff 1977; Ross and Mirowsky 1984)
and the Kessler index of generalized psychological distress (Kessler et al. 2002). Respondents
were asked the following eight items: “In the last seven days, on how many days have you . . .
felt sad; felt like you just couldn’t get going; felt unable to shake the blues; felt like everything
was an effort; had trouble keeping your mind on what you were doing; worried a lot about little
things; felt anxious or tense; had trouble getting to sleep or staying asleep.” Responses are coded
52
in days per week from 0 to 7. The distress scale is the mean response to the eight items. The
scale has an alpha reliability of .84.
In the CAN-WSH study I use 7 items of generalized psychological distress adapted from the
Kessler index (K10, see Kessler et al. 2002). These items ask about the frequency of the
following symptoms in the past month: “anxious or tense,” “nervous,” “worry a lot about little
things,” “had trouble keeping your mind on what you were doing,” “restless or fidgety,” “sad or
depressed,” and “hopeless.” Response choices are “all of the time” (1), “most of the time” (2),
“some of the time” (3), “a little of the time” (4), and “none of the time” (5). I reverse-coded these
responses and averaged the items such that higher scores indicate greater levels of distress (α =
.83).
Perceived job insecurity. In both the WSH and CAN-WSH studies perceived job insecurity is
assessed with the following question: “How likely is it that during the next couple of years will
you lose your present job and have to look for a job with another employer?” Response choices
for the CAN-WSH study include: “not at all likely” (1), “not too likely” (2), “somewhat likely”
(3), and “very likely” (4). Response choices for the WSH study include: “not at all likely” (1),
“somewhat likely” (2), and “very likely” (3). This question has been used in several well-
respected surveys including the General Social Survey and the National Study of the Changing
Workforce. Perceived job insecurity is coded 1 if respondents reported that they are ‘somewhat’
or ‘very likely’ to lose their job, and coded 0 otherwise.
Sense of personal control. Mirowsky and Ross's (2003) 2 x 2 index of the sense of personal
control is used in the WSH study. It asks participants to report their level of agreement or
disagreement with the eight statements; there are two statements in each of the four categories.
Statements that measure the level that individuals claim control over good outcomes include: (1.
"I am responsible for my own successes," and (2. "I can do just about anything I really set my
mind to." Items that measure claims of control over bad outcomes include: (3. "My misfortunes
are the result of mistakes I have made," and (4. "I am responsible for my failures." Items that
assess the extent that individuals deny control over good outcomes are: (5. "The really good
things that happen to me are mostly luck," and (6. "There's no sense planning a lot--if something
good is going to happen it will." The last two statements measure the denial of control over bad
outcomes: (7. "Most of my problems are due to bad breaks," and (8. "I have little control over the
53
bad things that happen to me." Responses to statements 1 through 4 are coded "strongly
disagree" (-2), "disagree" (-1), "neutral" (0), "agree" (1), and "strongly agree (2). Responses to
statements 5 through 8 are coded "strongly disagree" (2), "disagree" (1), "neutral" (0), "agree" (-
1), and "strongly agree (-2). I averaged responses; higher scores indicate a greater sense of
control (α = .52).
In the CAN-WSH study participants are asked the extent they agree with the following four
statements: “You have little control over the things that happen to you,” “There is really no way
you can solve some of the problems you have,” “You often feel helpless in dealing with
problems of life,” “Sometimes you feel that you are being pushed around in life,” “You can do
just about anything you really set your mind to.” Responses choices include: “strongly agree”
(1), “agree” (2), “disagree” (3), and “strongly disagree” (4). Responses for the fourth statement
were reverse-coded, and then summed and averaged with the remaining three statements; higher
scores indicate a greater sense of control (α = .60).
Age. To assess age contingencies I contrast those age 25-34 (the reference category) with the
following age groups: middle age workers (two categories for those age 35-44 and those age 45-
54) and older workers (those age 55-65). I split middle age workers into two categories in order
to consider the different family and career experiences of those in early-mid life and those
approaching late-working life (Toossi 2009).
Work Conditions
Personal income. In both the WSH and CAN-WSH studies, one item assesses total personal
earnings in the previous year—from all sources. I used the following categories: “$25,000 or
less,” “$25,001 to $50,000,” “$50,001 to $75,000,” “$75,001 to $100,000,” “$100,001 to
$125,000,” and “More than $125,000.” I use the modal category ($25,001 to $50,000) as the
omitted reference group.
Job pressures. Participants in the WSH study were asked: “How often do the demands of your
job exceed those doable in an 8-hour workday?” Response categories are "never" (1), "rarely"
(2), "sometimes" (3), and "frequently" (4). In the CAN-WSH study, three items ask about the
frequency of the following pressures in the past three months: “Felt overwhelmed by how much
54
you had to do at work?” “Have to work on too many tasks at the same time?” “The demands of
your job exceeded the time you have to do the work?” Response choices are coded: “never” (1),
“rarely” (2), “sometimes” (3), “often” (4), and “very often” (5). I averaged the items; higher
scores indicate more job pressure (α = .85).
Job autonomy. In the WSH study participants were asked: “In your current job, how often does
someone else decide how you do your work?” Response categories are "never" (1), "rarely" (2),
"sometimes" (3), and "frequently" (4). Responses were reverse coded, such that higher responses
indicate more autonomy. In the CAN-WSH study, three items assess job autonomy. Participants
were asked the extent that they agree or disagree with the following statements: “I have the
freedom to decide what I do on my job,” “It is basically my own responsibility to decide how my
job gets done,” and “I have a lot of say about what happens on my job.” Response choices are
coded “strongly disagree” (1), “somewhat disagree” (2), “somewhat agree” (3), and “strongly
agree” (4). I averaged responses to create the index; higher scores reflect more autonomy (α =
.78).
Work hours. Participants were asked: “How many hours do you work in a typical week at your
main job?” I use a continuous measure of work hours.
Plan of Analyses
Table 3.1 presents descriptive statistics for all measures in both samples. In table 3.2 (CAN-
WSH) and table 3.3 (WSH), using ordinary least squares regression I regress psychological
distress on job insecurity and the sense of personal control, adjusting for controls. I follow a
strategy to test whether the association between job insecurity and distress is mediated and
moderated by sense of control, and whether these relationships operate differently across age. In
model 1, I examine the association between distress and job insecurity. In model 2, I include
sense of control to examine its contribution to the association between insecurity and distress
observed in model 1. Then, in model 3, I test a job insecurity x sense of control interaction to
examine whether the association between insecurity and distress is moderated by sense of
control. Models 4 and 5 test age contingencies involving the associations between job insecurity,
sense of control and distress. In model 4, I include insecurity x age interactions to examine
whether the association between job insecurity and distress is contingent on age. Finally in model
55
5, I test whether the two-way interaction involving insecurity and sense of control examined in
model 3 is contingent on age. To do this, I test a series of three-way interactions involving
insecurity, sense of control and age. All models include the following controls: gender, marital
status, number of children younger than age 18 in the household, education, occupation and job
sector.
Results
Descriptives
Table 3.1 reports descriptive statistics for all measures in both samples. In the CAN-WSH and
WSH studies, older workers (55-65) report lower levels of distress in comparison to young
workers (25-34). In the CAN-WSH study, older workers are the least likely to think that they
may lose their job in the next two years (17%), while those 35-44 (23%) are the most likely to
report the possibility of future job loss. Similar age-patterns exist in the WSH study, but they are
not statistically significant. In the CAN-WSH study, young workers report a higher sense of
control than those age 45-54 and 55-65. No statistically significant age differences exist in the
WSH study with regards to sense of control.
Multivariate Analyses
CAN-WSH Sample
In Table 3.2, I present regression analyses with the CAN-WSH sample. As shown in model 1,
job insecurity is associated with higher levels of distress. Among the controls, the following
groups report lower levels of distress: men, the married, those with higher incomes, and those
reporting fewer job pressures and more job autonomy. In model 2, sense of control is associated
with lower levels of distress. The inclusion of sense of control substantially weakens this
association, though it remains statistically significant—indicating that part of the mechanism
through which job insecurity leads to more distress, operates through its tendency to decrease
feelings of personal control. This mediation effect is statistically significant (p<.001).
Additionally, adjusting for sense control, older individuals report lower levels of distress;
evidence that sense of control suppresses an association between older age and distress: were it
not for their lower levels of personal control, older workers would report even less distress.
56
I then examine whether sense of control moderates the association between job insecurity and
distress. In model 3, the ‘job insecurity × sense of control’ interaction is statistically significant.
Workers reporting higher levels of personal control are less affected by the stress of job
insecurity; in fact, at the highest levels of personal control, job insecurity is not associated with
higher distress. Conversely, among those reporting a low sense of personal control, job insecurity
is associated with the highest levels of distress. An incremental r-square test reveals that the
improvement in r-square between models 2 and 3 is significant (F=5.86, p<.05). In models 4 and
5, I find no evidence of age contingencies in the association between job insecurity and distress.
Moreover none of the three-way interactions in model 6 involving age, sense of control and job
insecurity are statistically significant—indicating that the sense of control buffers the stress of
job insecurity uniformly across the life course.
WSH Sample
In Table 3.3, I present regression analyses with the WSH sample. In model 1, job insecurity is
associated with more psychological distress. Additionally, older workers report statistically
significant lower levels of distress. Among the control variables, service workers, those with low
wages and more job pressures report higher levels of distress. By contrast, the following groups
report lower levels of distress: men, the highly educated, those with no children in the household,
and non-married individuals.
In model 2, sense of control is associated with lower levels of distress. As in the CAN-WSH
sample, the inclusion of sense control weakens (and thus partly mediates) the association
between insecurity and distress, though to a much lesser degree. This mediation effect is
statistically significant (sobel test: p<.05). In model 3, I find that sense of control moderates the
association between job insecurity and distress. In comparison to those with low levels of
personal control, workers with a high sense of control report less distress when they encounter
job insecurity. An incremental r-square test reveals that the improvement in r-square between
models 2 and 3 is significant (F=5.48, p<.05).
Model 4, reveals evidence that the association between job insecurity and distress is contingent
on age. The significant interaction between job insecurity and workers age 35-44 indicates that,
in comparison to younger workers, job insecurity is associated with higher levels of distress
57
during the early stages of middle age. The ‘insecurity × 45-54’ and ‘job insecurity × 55-65’
interactions are not statistically significant, suggesting that the stress of job insecurity is
nonlinear across age. In model 5, none of the three-way interactions involving age, sense of
control and job insecurity are statistically significant.
Given research showing gender differences in the ways men and women navigate work and
family roles across the life course (Moen 1992; Winslow 2005), for both the CAN-WSH and
WSH samples, I also tested for gender contingencies in the focal associations and interactions
between job insecurity, sense of control, and distress; however these analyses yielded no
statistically significant results.
Discussion This paper makes several contributions that advance understanding on the links between the
sense of control and the stress of job insecurity—an increasingly common source of work stress
for many North Americans. Based on two national samples of American and Canadian workers, I
show that, despite acting as a key resource for ameliorating the deleterious health consequences
of job insecurity, the sense of control also represents an indirect route through which job
insecurity influences health. Additionally, I also document age-contingencies in the stress of
employment precariousness, an area requiring further examination, given an aging workforce in
both the United States and Canada.
With the ripples of the recent American financial crisis continuing to influence workers, firms,
and stock markets, the stress of job loss is a health issue that is often raised with regards to the
economic downturn. At the peak of the recession, close to one in ten were unemployed in the
United States, with almost half of these individuals experiencing long-term unemployment,
lasting over twenty-six weeks. This dire state of the American labour market has resulted in the
prolonged hardship for many workers and their families—a well-documented source of stress
(Mirowsky and Ross 2001). However, the negative mental and physical health consequences
generated by the recession are most likely not restricted to those workers who experienced
layoffs, but also many surviving workers whose own job situation has been made increasingly
58
tenuous. My results support this point, showing that even the threat of job loss—whether real or
imagined—can be damaging to one’s health (Burgard et al. 2009).
While the data sources that I examine in this paper do not speak directly to the experiences of
workers during the recession, these results inform understanding on the stress that many
employed workers faced during this period, and likely continue to face. I find consistent results
showing that job insecurity has negative health consequences for workers north and south of the
American-Canadian border. Further, roughly one in five workers reported job insecurity in both
samples, indicating that even in non-recession periods, this stressor is not restricted to a small
minority of workers—patterns that have been documented elsewhere as part of a growing trend
since the 1970s (Fullerton and Author 2007). Thus, the findings of this paper contribute to a
growing body of research that suggests that insecure employment should be treated as a very real
and prevalent work stressor in the contemporary labour market. Additionally, I extend research
in this area by revealing how the stress of job insecurity may be buffered by a high sense of
personal control.
In both American and Canadian samples I find that workers with a strong sense of control over
their life are less affected by the stress of job insecurity; in fact, in the CAN-WSH sample, job
insecurity is not associated with distress for workers with the highest levels of perceived control.
I can only surmise at the reasons for this buffering effect, but workers who see outcomes in the
life as under their control likely view job loss as an event that they can resolve or at least
adequately cope with. Workers with low levels of personal control, in contrast, are likely to
envision worst-case scenarios that could result from being laid off; thus experiencing high levels
of anticipatory anxiety. These findings are consistent with other research (Ross and Sastry 1999)
that has shown the value of a high sense of control in the face of chronic stress—in this particular
case, sense of control acts as a personal resource for reducing the stress associated with chronic
uncertainty.
In addition to the buffering role of the sense of control, I also find evidence that suggests that it
represents a key mechanism through which job insecurity increases distress. In the CAN-WSH
sample, once I adjust for sense of control, the association between job insecurity and distress is
reduced by a quarter. This mediation effect indicates that part of the reason why job insecurity is
associated with higher distress is because of its tendency to reduce one’s sense of control, a
59
personality trait that is important for health in the face of the daily challenges and stress of
everyday life.
The dual role of the sense of control as both a mediator and moderator of the association between
job insecurity and health can be interpreted through what others have referred to as structural
amplification. Structural amplification describes the direct and indirect processes through which
social conditions “undermine the personal attributes that otherwise would moderate their
undesirable consequences” (Ross 2011:290). Formally, it exists when a moderator of an
association between a social condition and outcome also results from the social condition. Other
research has found evidence of structural amplification when examining the effect of
neighborhood disorder on feelings of mistrust. Ross (2001) describes how neighborhood disorder
exerted a direct positive impact on feelings of mistrust, while also leading to a reduced sense of
personal control, which then amplified the association between disorder and mistrust. I find
evidence of this process with respect to the stress of job insecurity. While a high sense of control
buffers the association between job insecurity and distress, my findings suggests that it is also
undermined by the very experience of job insecurity—thus amplifying the stressful nature of the
experience. I am unable to speak to the specifics regarding how this amplification process
operates, such as whether it operates over the short-term or alternatively with long-term exposure
to job insecurity. Further analysis with longitudinal data is therefore necessary to address this
issue.
Age Contingencies
My results also serve to support emerging evidence that the stress of job insecurity varies
according to life course position (Cheng and Chan 2009). Calls to combine the insights of the
Stress Process Model and life course theory have drawn increased attention to the timing of
stressors and the ways in which stressful role transitions intersect with other contextual factors,
such as the quality of the role under duress, and family structure/composition (Elder 1998; Sweet
and Moen 2011). Due to the data limitations, the analyses in this paper are unable to take a
lifelong view of stressors and their health implications over the long-term; but the findings
revealed here do at least lend support to many of these life course considerations. Specifically, I
contend that age-disparities in the stress of job insecurity are in part a function of an individual’s
current position and trajectory in their career, along with the nature and extent of family-related
60
demands coinciding with that position. Informed by this argument, I derived two specific age-
related hypotheses; one hypothesis predicting that the stress of job insecurity is greater among
older workers, and an alternative hypothesis predicting that this stress should be greatest during
middle age.
I find some support for the stress of middle age hypothesis, but only among participants in the
WSH study. The greater distress that American workers in the early stages of middle age (35-44)
experience as a result of job insecurity may be due to their exposure to family-related financial
responsibilities that makes the threat of job loss a more significant stressor for this group. Middle
age workers approaching late working life (45-54), and older American workers (55-65) facing
job insecurity, by contrast, do not report more distress than workers in the youngest group (25-
34). The absence of a significant interaction for older workers may be explained by their
proximity to retirement; though the fact that those age 45-54 do not report more distress as a
result of job insecurity is perhaps surprising, given that they likely have similar family
obligations as those age 35-44, as well as facing increasing labour market constraints (i.e.
increased age discrimination). These results do indicate, however, that the least affected by job
insecurity are those in formative stage of their careers; for this group, the stress of job insecurity
may be tempered by the fact that job mobility and instability are traditionally more common and
expected early on in one’s career—not to mentioned being now normative features of the
contemporary labour market that these workers have grown-up in. Additionally, the fact that
many younger workers are increasingly delaying family formation (Mathews and Hamilton
2009) means that this age group likely has fewer financial dependants to provide for, potentially
reducing the perceived danger of a layoff. Finally, in both studies I find no evidence of
statistically significant three-way interactions between age, sense of control and job insecurity;
thus all age groups equally experience the protective benefits of a strong sense of control.
American-Canadian Comparisons
Since this paper analyzes studies from two countries at two different time-points, it is important
to consider the similarities and differences in economic and labour market conditions that may
account for consistencies and inconsistencies between the results of the two analyses. The free-
market focused economies of Canada and the US share many similarities, particularly when
contrasted with northern European countries that have more extensive welfare state protections
61
and other labour market regulation systems (Kahn 2012). The consequences of job loss may be
more ominous for workers in North America because of the absence of such protections; an
explanation that may account for the similar health experiences reported by WSH and CAN-
WSH survey participants who encountered job insecurity. Nevertheless, labour market
differences do exist. Canada has generally stronger labour market protections and higher union
density than the United States, for example (Kahn 2012). Additionally, Canadian workers who
are laid-off also have greater access to health insurance compared to American workers. This
greater access to health care may account for the absence of age contingencies in analyses of the
Canadian study. For American workers who are laid off, the loss of health care coverage may
pose economic problems that particularly problematic during middle age when family
obligations are highest. Their children may still receive coverage, but the threat of losing one’s
personal coverage may nevertheless create greater anxiety during this stage, given that personal
health costs may disrupt one’s ability to provide for their family. In contrast, Canadian workers’
continued access to health coverage might dampen this anxiety among middle age workers with
families—possibly resulting in more similar health penalties over the life course.
Another important difference is the timing of the two studies. The American WSH study was
conducted prior to the recession during a period of relative economic stability, while the CAN-
WSH study was conducted after the recession, but nevertheless during a period of global
economic and financial uncertainty. It is perhaps, then, surprising that the increased uncertainty
of this period did not lead to greater perceptions of insecurity in the CAN-WSH study. However,
with a more detailed examination of responses to the insecurity survey item, differences between
participants’ experiences across the two studies emerge. Specifically, while a similar proportion
of workers in each study reported that they were ‘somewhat likely’ to lose their job in the next
two years, almost twice the proportion of Canadian participants reported that they were ‘very
likely’ to lose their job (8 percent), relative to American workers (4.5 percent). The measurement
strategy used in this paper thus obscures these differences. It should be noted that other analyses
with perceived insecurity modeled as an ordinal measure revealed similar results as those
presented here. In order to examine age contingencies, the dichotomous measure was preferred,
given the low cell sizes that were created from interacting age with each of the ordinal levels of
insecurity.
62
Several limitations of this research deserve brief mention. While the hypothesized linkages
between the focal variables are consistent with prior theory and qualitative evidence, I am unable
to test the causal ordering between job insecurity, sense of control, and distress. Longitudinal
analyses are therefore required to more accurately verify my claims. Additionally, while the
measure of job insecurity is well-established and commonly used, future research would benefit
from the inclusion of a more comprehensive set of questions that tap into both subjective and
objective components of employment precariousness—such as recent organizational downsizing,
or measures that more deeply probe one’s sense of worry regarding potential job loss. Despite
this, I believe that asking workers about the likelihood of future job loss is a concise and succinct
strategy for assessing job security, and one that avoids confounding personality and mental
health traits (e.g. neuroticism) with the appraisals of security—an issue that survey items that
measure the degree that one expresses worry about job loss may fall prey to.
Conclusion The sense of control has been shown to be a critical psychological disposition that explains a
considerable proportion of mental health disparities between individuals and groups (Pearlin et
al. 2007). A primary mechanism that accounts for its health benefits is the propensity for a strong
sense of control to reduce the deleterious impact of role stressors on well-being by fostering
resilience and proactive behaviors in the face of these stressors. In the case of this paper, sense of
control is shown to alleviate the stress of an uncertain and potentially damaging future event: the
loss of employment. At the same time, however, exposure to this uncontrollable and uncertain
threat may also come to erode one’s sense of control, thus serving to amplify the stress of job
insecurity. This structural amplification effect warrants further thought and empirical study, since
it may represent a useful framework for understanding the long-term links between job insecurity
and well-being.
63
Notes
1. There were approximately 27,000 cases that are classified as “refusals,” but information about
whether or not these cases would have been eligible to participate is unavailable. A substantial
percentage simply “hung up” immediately. It is estimated that roughly 40 percent of these cases
would have been eligible to participate in the study. Despite the response rate being lower than
desirable levels, comparisons with the Canadian working population show that the sample is
nationally representative. Additionally, weighted analyses revealed very similar results as those
presented here; thus providing no evidence of sample-bias.
64
References Armstrong-Stassen, Marjorie. 1993. “Survivors’ Reactions to a Workforce Reduction.” Revue
Canadienne des Sciences de Administration 10:334-344.
Avison, William R. 2001. “Unemployment and its Consequences for Mental Health.” Pp. 177–200 in Restructuring Work and the Life Course, edited by Victor W. Marshall, Walter Heinz, Helga Krueger, and Anil Verma. Toronto, Canada: University of Toronto Press.
Barling, Julian, and E. K. Kelloway. 1996. “Job Insecurity and Health: The Moderating Role of Workplace Control.” Stress Medicine 12:253-259.
Brenner, M. Harvey. 1973. Mental Illness and the Economy. Cambridge, MA: Harvard University Press.
Broman, Clifford L., V. Lee Hamilton, and William S. Hoffman. 2000. Stress and Distress among the Unemployed: Hard Times and Vulnerable People. New York: Kluwer Academic/Plenum Publishers.
Burgard, Sarah A., Jennie Brand, and James S. House. 2009. “Perceived Job Insecurity and Worker Health in the United States.” Social Science & Medicine 69:777-785.
Cheng, Grand. H. L. and Darius K. S. Chan. 2008. “Who Suffers More from Job Insecurity? A Meta-Analytic Review.” Applied Psychology 57: 272–303.
De Witte, H. 1999. “Job Insecurity and Psychological Health: Review of the Literature and Exploration of Some Unresolved Issues.” European Journal of Work and Organizational Psychology 8:155-177.
Elder, Glen H., Jr. 1998. “Life Course Theory and Human Development.” Sociological Analysis 1:1-12.
Erikson, Erik. H. 1998. The Life Cycle Completed. NY: Norton.
Ferrie, Jane Elizabeth, Martin John Shipley, Michael Gideon Marmot, Stephen Alfred Stansfeld, and George Davey Smith. 1995. "Health Effects of Anticipation of Job Change and Nonemployment: Longitudinal Data from the Whitehall II Study." British Medical Journal 311:1264-1269.
Fischer, Joachim E. and Julian F. Thayer. 2006. “Invited Commentary: Tapping the Tip of the Iceberg.” American Journal of Epidemiology 163:888-890.
Greenhalgh, Leonard, and Zehava Rosenblatt. 1984. “Job insecurity: Toward Conceptual Clarity.” Academy of Management Review 3:438–448.
Heaney, Catherine, Barbara A. Israel, and James S. House. 1994. “Chronic Job Insecurity Among Automobile Workers: Effects on Job Satisfaction and Health.” Social Science & Medicine 38:1431–1437.
Hirsch, Barry T., David A. MacPherson, and Melissa A. Hardy. 2000. “Occupational Age Structure and Access for Older Workers.” Industrial and Labour Relations Review 53: 401–418.
65
Jahoda, Marie. 1982. Employment and Unemployment: A Social Psychological Analysis. New York: Cambridge University Press.
Jick, Todd D. and Leonard Greenhalgh. 1989. "Survivor Sense Making and Reactions to Organizational Decline: Effects of Individual Differences." Management Communication Quarterly 2:305-327.
Kahn, Lawrence M. 2012. “Labour Market Policy: A Comparative View on the Costs and Benefits of Labour Market Flexibility.” Journal of Policy Analysis and Management 31:94-110.
Kalleberg, Arne. 2011. Good Jobs, Bad Jobs: The Rise of Polarized and Precarious Employment Systems in the United States, 1970s-2000s. New York: Russell Sage Foundation, American Sociological Association Rose Series in Sociology.
Kessler, Ronald C., J. Blake Turner, and James S. House. 1987. “Unemployment, Reemployment, and Emotional Functioning in a Community Sample.” American Sociological Review 54:648–657.
Kessler, Ronald C, G. Andrews, L. J. Colpe, E. Hiripi, D. K. Mroczek, S-L. T. Normand, E. E. Walters, and A. M. Zaslavsky. 2002. “Short Screening Scales to Monitor Population Prevalences and Trends in Non-specific Psychological Distress.” Psychological Medicine 32:959–976.
Kinnunen, Ulla, Saija Mauno, and Marjo Siltaloppi. 2010. “Job Insecurity, Recovery and Well-being at Work: Recovery Experiences as Moderators.” Economic and Industrial Democracy 31:179–194.
Kohn, Melvin L. and Carmi Schooler. 1983. Work and Personality: An Inquiry into the Impact of Social Stratification. Norwood, N.J. Ablex Pub. Corp.
Lazarus, Richard S. and Susan Folkman. 1984. Stress, Appraisal, and Coping. New York: Springer.
Lee, Sunmin, Graham A. Colditz, Lisa F. Berkman, and Ichiro Kawachi. 2004. “Prospective Study of Job Insecurity and Coronary Heart Disease in US Women.” Annals of Epidemiology 14:24–30.
Mirowsky, John, and Catherine E. Ross. 2001. “Age and the Effect of Economic Hardship on Depression.” Journal of Health and Social Behavior 42:132–50.
______. 2003. Social Causes of Psychological Distress. Hawthorne, New York: Aldine De Gruyter.
Naswall, K., M. Sverke and J. Hellgren. 2005. ‘The Moderating Role of Personality Characteristics on the Relationship between Job insecurity and Strain.” Work and Stress 19:37–49.
Pearlin, Leonard I. 1999. “The Stress Process Revisited: Reflections on Concepts and Their Interrelationships.” Pp. 395-415 in Handbook of the Sociology of Mental Health, edited by C. S. Aneshensel and J. C. Phelan. New York: Kluwer Academic/Plenum Publishers.
Pearlin, Leonard I., and Carmi Schooler. 1978. “The Structure of Coping.” Journal of Health and Social Behavior 19:2–21.
66
Pearlin, Leonard I., Morton Lieberman, Elizabeth Menaghan, and Joseph Mullan. 1981. “The Stress Process.” Journal of Health and Social Behavior 22:337–56.
Pearlin, Leonard I. and Carol Aneshensel. 1986. "Coping and Social Supports: Their Functions and Applications." Pp. 53-74, Applications of Social Science to Clinical Medicine and Health, edited by L.H. Aiken and D. Mechanic. New Brunswick, NJ: Rutgers University Press
Pearlin, Leonard I., Scott Schieman, Elena Fazio, and Stephen C. Meersman. 2005. “Stress, Health, and the Life Course: Some Conceptual Specifications.” Journal of Health and Social Behavior 46:205–19.
Pearlin, Leonard I., Kim B. Nguyen, Scott Schieman, and Melissa A. Milkie. 2007. “The Life-Course Origins of Mastery among Older People.” Journal of Health and Social Behavior 48:164-179.
Polivka, Anne E. 1996. “A Profile of Contingent Workers.” Monthly Labour Review 119:10–21.
Radloff, L. S. 1977. “The CES-D Scale: A Self-Report Depression Scale for Research in the General Population.” Applied Psychological Measurement 1:385-401.
Ross, Catherine E. 2011. “Collective Threat, Trust, and the Sense of Personal Control.” Journal of Health and Social Behavior 52:287-296.
Ross, Catherine E. and John Mirowsky. 1984. “Components of Depressed Mood in Married Men and Women: The Center for Epidemiological Studies’ Depression Scale.” American Journal of Epidemiology 119:997-1004.
______. 1989. "Explaining the Social Patterns of Depression: Control and Problem Solving-or Support and Talking?" Journal of Health and Social Behavior 30:206-19.
Ross, Catherine E., and Jaya Sastry. 1999. “The Sense of Personal Control: Social-Structural Causes and Emotional Consequences.” Pp. 369–94 in Handbook of the Sociology of Mental Health, edited by Carol S. Aneshensel and Jo C. Phelan. New York: Kluwer.
Rousseau, Denise M. 1995. Psychological Contracts in Organizations: Understanding Written and Unwritten Agreements. Thousand Oaks, CA.
Schreurs, Bert, Hetty van Emmerik, Guy Notelaers, and Hans De Witte. 2010. “Job Insecurity and Employee Health: The Buffering Potential of Job Control and Job Self-efficacy.” Work and Stress 24:56-72.
Seeman, Melvin. 1967. “On the Personal Consequences of Alienation in Work.” American Sociological Review 32:273-285.
Seeman, Teresa E, Bruce S. McEwen, John W. Rowe, and Burton H. Singer. 2001. “Allostatic Load as a Marker of Cumulative Biological Risk: MacArthur Studies of Successful Aging.” Proceedings of the National Academy of Sciences of the United States of America 98:4770-4775.
Skaff Marilyn M., Leonard Pearlin, Joseph T. Mullan. 1996. “Transitions in the Caregiving Career: Effects on Sense of Mastery.” Psychology and Aging 11:247–257.
Sweet, Stephen and Phyllis Moen. 2012. “Dual-Earners Preparing for Job Loss: Agency, Linked Lives, and Resilience.” Work and Occupations 39:35–70.
67
Sverke, Magnus, and Johnny Hellgren. 2002. “The Nature of Job Insecurity: Understanding Employment Uncertainty on the Brink of a New Millennium.” Applied Psychology 51:23–42.
Tausig, Mark, and Rudy Fenwick. 1999. “Recession and Well-being.” Journal of Health and Social Behavior 40:1-16.
Toossi, Mitra. 2009. “Labour Force Projections to 2018: Older Workers Staying More Active.” Monthly Labour Review November 30–51.
Turner, R. Jay, and Patricia Roszell. 1994. “Personal Resources and the Stress Process.” Pp. 179–210 in Stress and Mental Health: Contemporary Issues and Prospects for the Future, edited by William. Avison and Ian. H. Gotlib. New York: Plenum Press.
Wheaton, Blair. 1983. "Stress, Personal Coping Resources, and Psychiatric Symptoms: An Investigation of Interactive Models." Journal of Health and Social Behavior 24:100-24.
______. 1999. “Social Stress.” Pp. 277-300 in The Handbook of the Sociology of Mental Health, edited by C.S. Aneshensel and J.C. Phelan. Kluwer Academic/Plenum.
68
TABLE 3.1 Means and Proportions of all CAN-WSH Study Variables
Age Categories Total
25-34 35-44 45-54 55-65
Psychological Distress 2.213 2.205 2.215 2.083* 2.168
Perceived Job Insecurity .210 .235 .214 .171* .211
Sense of Personal Control 2.989 2.967 2.994 2.943 2.959
Female .597 .599 .638 .601 .610
$25,000 or less .155 .132 .113
* .140 .131
$25,001 to $50,000 .434 .334 .338 .314 .352
$50,001 to $75,000 .257 .239 .226 .221 .235
$75,001 to $100,000 .087 .168 .170 .172 .153
$100,001 to $125,000 .025 .043
* .050
* .044
* .042
More than $125,000 .022 .040* .053
* .059
* .044
Job Pressures 3.104 3.218* 3.187 3.097 3.162
Job Autonomy 2.795 2.882 2.835
* 2.905
* 2.823
Work Hours (Weekly) 40.346 40.085 40.070 43.303 40.277
Employed in Public Sector .325 .363 .409* .405
* .379
Executive .088 .108 .114 .094 .103
Professional .364 .364 .310* .325 .336
Technical .132 .168
* .156 .179 .158
Sales .055 .055 .057 .056 .058
Administrative .100 .076 .120 .104 .101
Service .116 .096 .116 .109 .115
Production .141 .130 .141 .124 .124
Education 5.239 5.245 4.945* 4.479
* 5.058
Married .412 .581* .542
* .561
* .531
Children in the Household .920 1.434* .610
* .123
* .804
N 757 1017 1325 691 3790
Comparison with the “25-34” group is statistically significant at * p < .05; ** p < .01; *** p < .001 (two–
tailed test).
69
TABLE 3.2 Means and Proportions of all WSH Study Variables
Age Categories Total
25-34 35-44 45-54 55-65
Psychological Distress 2.033 1.900 1.860 1.667* 1.877
Perceived Job Insecurity .192 .217 .200 .201 .205
Sense of Personal Control .852 .871 .79 .800 .830
Female .640 .520
** .585 .631 .589
$25,000 or less .382 .254
* .223
* .255
* .275
$25,001 to $50,000 .415 .439 .379 .372 .402
$50,001 to $75,000 .133 .161 .234* .217
* .187
$75,001 to $100,000 .049 .078 .096
* .079 .077
$100,001 to $125,000 .006 .017 .029 .054 .023
More than $125,000 .013 .049 .043 .029 .035
Job Pressures 3.019 3.104 2.995 2.983 3.028
Job Autonomy 2.408 2.453 2.557
* 2.493 2.482
Work Hours (Weekly) 42.284 45.520* 46.557
* 42.561 44.516
Employed in Public Sector .221 .257 .256 .261 .247
Professional .267 .329 .312 .326 .301
Technical .392 .300
* .361 .368 .355
Craft .045 .095* .061 .054 .066
Service .133 .115 .133 .133 .128
Labor .105 .112 .104 .101 .106
Education 5.301 5.271 5.291 5.384 5.309
Married .516 .604* .608
* .577 .579
Children in the Household 1.051 1.280* .649
* .605
* .826
White .683 .716 .804* .786
* .749
N 306 346 393 239 1284
Comparison with the “25-34” group is statistically significant at * p < .05; ** p < .01; *** p < .001 (two–
tailed test).
70
TABLE 3.3 CAN-WSH: Regression of Psychological Distress on Focal Independent Variables
(N=3,790)
Model 1 Model 2 Model 3 Model 4 Model 5
Job Insecurity and Personal Control
Perceived Job Insecurity .189***
.107***
.410**
.085 .392
Sense of Personal Control — -.580**
-.556***
-.580***
-.568***
Job Insecurity × Personal Control — — -.107* — -.108
Age Contingencies
35 – 44 years-old a -.003 -.032 -.032 -.050 -.146
45 – 54 years-old a -.039 -.089
** -.089
** -.091 -.052
55 – 65 years-old a -.051 -.120
** -.118
** -.117 -.232
Job Insecurity × 35 – 44 — — — .075 .344
Job Insecurity × 45 – 54 — — — .009 -.107
Job Insecurity × 55 – 65 — — — -.019 -.409
35– 44 × Personal Control — — — — .032
45 – 54 × Personal Control — — — — -.013
55 – 65 × Personal Control — — — — .039
Job Insecurity × 35– 44 × Personal Control — — — — -.094
Job Insecurity × 45 – 54 × Personal Control — — — — .039
Job Insecurity × 55 – 65 × Personal Control — — — — .140
Income
$25,001 or less b .064 .051 .050 .050 .048
$50,001 to $75,000 b -.077
** -.030 -.031 -.030 -.030
$75,001 to $100,000 b -.115
** -.013 -.015 -.014 -.013
$100,001 to $125,000 b -.166
** -.087 -.087 -.086 -.083
More than $125,000 b -.146
** -.026 -.026 -.027 -.024
Work Conditions
Work Hours (Weekly) .000 .001 .001 .001 .001
Job Pressures .218***
.151***
.150**
.151**
.150***
Job Autonomy -.129***
-.066**
-.065**
-.066**
-.064**
Employed in Public Sector .020 .030 .030 .030 .031
Occupation
Executive c .024 0.054 .052 .053 .052
Technical c -.001 -.010 -.009 -.010 -.008
Continued on next page
71
Table 3.3 Continued
Model 1 Model 2 Model 3 Model 4 Model 5
Sales c .007 .013 .013 .011 .008
Administrative c -.005 -.040 -.041 -.041 -.040
Service c .034 .016 .016 .016 .015
Production c -.003 -.029 -.028 -.030 -.030
Demographics
Women .085**
.123**
.122**
.120**
.120**
Education -.002 .006 .005 .005 .004
Married -.100**
-.064**
-.064**
-.063**
-.062**
Children in the Household .085 .011 .010 .011 .011
Constant 1.837
3.526 3.452
3.531
3.495
R2
.187 .352 .358 .358 .360
* p < .05; ** p < .01; *** p < .001 (two–tailed test). a Compared to 25 – 34 year-olds.
b Compared to $25,001 to $50,000.
c Compared to professionals.
72
TABLE 3.4 WSH Sample: Regression of Psychological Distress on Focal Independent Variables
(N=1,284)
Model 1 Model 2 Model 3 Model 4 Model 5
Job Insecurity and Personal Control
Perceived Job Insecurity .555***
.523***
.909***
.118 .324
Sense of Personal Control — -.462***
-.347***
-.461***
-.399
Job Insecurity × Personal Control — — -.490* — -.255
Age Contingencies
35 – 44 years-old a -.095 -.098 -.094 -.230 -.333
45 – 54 years-old a -.121 -.181 -.191 -.258
* -.166
55 – 65 years-old a -.367
** -.421
** -.423
** -.529
** -.776
*
Job Insecurity × 35 – 44 — — — .651* .115
Job Insecurity × 45 – 54 — — — .397 -.112
Job Insecurity × 55 – 65 — — — .536 .305
35– 44 × Personal Control — — — — 1.058
45 – 54 × Personal Control — — — — .501
55 – 65 × Personal Control — — — — .748
Job Insecurity × 35– 44 × Personal Control — — — — -.443
Job Insecurity × 45 – 54 × Personal Control — — — — -.234
Job Insecurity × 55 – 65 × Personal Control — — — — -.266
Income
$25,00 or less b .257
* .205 .198 .205 .196
$50,001 to $75,000 b -.178 -.150 -.147 -.151 -.148
$75,001 to $100,000 b .035 .076 .090 .078 .104
$100,001 to $125,000 b -.114 -.113 -.108 -.097 -.100
More than $125,000 b -.039 .019 .025 .007 .012
Work Conditions
Work Hours (Weekly) -.001 .001 .001 .001 .001
Job Pressures .211***
.213***
.216***
.215***
.220***
Job Autonomy -.048 -.043 -.042 -.042 -.039
Employed in Public Sector -.035 -.008 -.011 -.007 -.007
Occupation
Service c .342
* .335
* .340
* .330
* .342
*
Technical c .089 .056 .060 .051 .052
Continued on next page
73
Table 3.4 Continued
Model 1 Model 2 Model 3 Model 4 Model 5
Craft c .252 .216 .226 .232 .237
Labor c .115 .059 .075 .094 .121
Demographics
Women .541***
.543***
.553***
.556***
.566***
Education -.074**
-.065**
-.066**
-.065**
-.069**
Married -.252**
-.251**
-.251**
-.261**
-.261**
Children in the Household .070* .078
* .078
* .080
* .078
*
White .160 .170 .168 .175 .177
Constant 1.303 1.411
1.134
1.464
1.101
R2
.125 .141 .146 .148 .150
* p < .05; ** p < .01; *** p < .001 (two–tailed test). a Compared to 25 – 34 year-olds.
b Compared to $25,001 to $50,000.
c Compared to professionals.
74
Chapter 4 The Consequences of Job Insecurity and Job Degradation for the
Sense of Personal Control
Introduction The sense of personal control is the generalized belief that one has control over meaningful
events and circumstances in their life. It is considered an important psychological disposition that
is associated with good psychological and physical well-being (Mirowsky and Ross 2003;
Wheaton 1983). While a subjective appraisal, the sense of control is nevertheless a learned
expectation, and one that is shaped by patterned social-structural experiences over the life course
(Ross and Sastry 1999). As such, it is considered a key-mediating link through which social
conditions influence mental health (Seeman 1959). In this paper I consider how the social
organization of work—specifically the organization of insecure work—shapes the sense of
control, and how this process potentially operates uniquely across age, a key social status.
Scholars have long demonstrated how job conditions, such as autonomous, interesting and self-
directed work, influence feelings of mastery and self-efficacy (Kohn and Schooler 1983; Ross
and Mirowsky 1992; Schieman and Plickert 2008). However, with economic and organizational
restructuring contributing to the proliferation of insecure and precarious work arrangements in
recent decades, the degree of control and certainty over one’s continuing employment represents
a particularly important and contemporary work experience that may influence the sense of
control (Kalleberg 2011). The proportion of American workers who perceive their jobs as
insecure has risen steadily since the 1970s (Fullerton and Wallace 2007), and with the recent
‘Great Recession,’ job insecurity has become an issue attracting widespread interest. Since
employment serves as a key source of economic livelihood, as well as identity, it seems plausible
that the threat of job loss, and the uncertainty surrounding the event, could constitute a chronic
stressor that may challenge the extent that an individual believes they exert control over their life
(Sverke and Hellgren 2002; Wheaton 1999).
While insecure employment may directly influence beliefs about personal control, it may also do
so as part of a broader process involving deteriorating job quality. Insecurity researchers have
argued that the loss of desirable job conditions represents an overlooked dimension of job
75
insecurity, and suggest the importance of distinguishing between insecurity about the continuity
of one’s job from insecurity over valued aspects of one’s job (Greenhalgh and Rosenblatt 1984;
Sverke and Hellgren 2002). Given that many workers surviving organizational downsizing and
restructuring tend to experience not only a decline in job security, but also job degradation in the
form of increased job pressures and reduced decision-making latitude (Burke and Nelson 1998;
Quinlan and Bohle 2009), part of the association between job insecurity and a decreased sense of
control may be accounted for by changing access or exposure to these job resources and demands
(Bakker and Geurts 2004). To address this issue, I consider whether a decline in job security is
associated with a change in workers’ reports of decision-making latitude and job pressures over
time, and whether any changes in these job conditions mediate the potential association between
job insecurity and sense of control.
In order to these issues, I analyze two waves of panel data drawn from a national study of
American workers in 2005 and 2007. As part of these analyses, I consider whether the potential
associations between job insecurity, job quality and the sense of control are contingent on two
temporal factors: 1) the duration of job insecurity, and 2) its life course timing. I outline the
rationale for examining these contingencies, as well as the links between insecurity and sense of
control in the following section.
Literature Review
The Sense of Personal Control
The belief that one can effectively control important aspects of their life is informed by a variety
of constructs, including mastery and self-efficacy at one end of the continuum, and
powerlessness, fatalism and external locus of control on the other (Turner and Roszell 1994;
Mirowsky and Ross 2003; Pearlin and Schooler 1978). Here, I focus on a particular construct
that shares considerable conceptual overlap with all of these: the sense of personal control.
Individuals with a high or internal locus of personal control believe that the things that happen in
their life are mostly of their doing rather than the result of chance or outside forces. These beliefs
are shaped through socialization and personal experience, as an individual comes to realize the
extent that their actions result in desirable versus undesirable outcomes (Wheaton 1980). Interest
76
in the sense of control originates in part from the fact that it represents a useful concept for
illuminating the ways in which the self is socially constituted (Ross and Sastry 1999). However,
it has also drawn interest from stress researchers who have shown it to be a key personal resource
for coping with the deleterious consequences of stressors (Mirowsky and Ross 1989).
With regards to the antecedents and formative elements of personal control, one relevant social
domain is the workplace. Beyond the beneficial value of employment itself (Mirowsky and Ross
1992), the workplace constitutes an important patterned source of opportunity and constraint that
may influence feelings of mastery and personal control. A considerable body of scholarship
informs knowledge of the relationship between the social organization of work and personal
control, from Marx ([1852] 1983), who saw work activity as a key source of identity and self-
actualization, but one that was increasingly estranged under capitalist relations, to Kohn and
Schooler (1983), who considered specific occupational sources of powerlessness (see also
Seeman 1967; Schieman 2002; Schieman and Plickert 2008). Collectively, this research
highlights how the degree that one is able to exercise self-direction in the pacing, nature, and
timing of work are important predictors of high personal control.
A related issue that has received increased attention is the decline in job security in the United
States in recent decades (Hollister 2011; Kalleberg 2011). However, while scholars have
documented the mental and physical health consequences of insecure employment, less research
has considered the implications for the sense of control. Indeed, while stress process researchers
have suggested that personal control represents a resource for reducing the stress associated with
job insecurity (Pearlin 1989; Pearlin Schieman, Fazio, and Meersman 2005), few have
considered it as an outcome of job insecurity. I discuss existing theoretical and empirical work on
this subject in the following section.
Job Insecurity and Feelings of Powerlessness
Decreasing job security in recent decades is one aspect of a broader change in paid work that is
often described in terms of the erosion of the post-WWII psychological contract between
employers and employees (Rousseau 1995). The decline of this contract, which specifies the
exchange of worker commitment and loyalty for secure, well-paid employment with
advancement opportunities, has been driven by an interlinked set of economic and political
77
changes that have led employers to increasingly prioritize flexible labour (Kalleberg 2011). This
prioritization is evident in the growing popularity of organizational strategies designed to achieve
numerical and functional flexibility, such as downsizing and the hiring of short or fixed-term
contracts (Hollister 2011). Consequently, scholars have become interested in workers’ responses
to these changes, and in particular, their reactions to growing uncertainty about their continuing
employment (for a review, see Sverke and Hellgren 2002).
Job insecurity has been defined in various ways, including “an individual’s expectations about
continuity in a job situation” (Davy et al. 1997:323), and one’s “powerlessness to maintain
desired continuity in a threatened job” (Greenhalgh and Rosenblatt 1984:438). A common theme
among these definitions is that it represents a subjective appraisal of the threat to one’s
immediate employment. Despite its subjective nature, research demonstrates that workers’
perceptions of insecure employment represent a relatively accurate indicator of more objective
measures of insecurity (Elman and O’Rand 2002).
There are various reasons to expect a link between job insecurity and the sense of control.
Research documents that job loss and unemployment are associated with reduced personal
control (Avison 2001; Broman, Hamilton, and Hoffman 2000), presumably because of the loss of
valued material, psychological and social resources attached to employment (Jahoda 1982).
Thus, it seems likely that the threat and anticipation of job loss may have similar consequences,
as individuals realize they are not in control of their continuing access to these resources, and
generalize this lack of control to other aspects of their life. The chronic, non-discrete nature of
job insecurity may also function to reduce perceptions of control (Wheaton 1999). In contrast to
an actual event that an individual can directly observe and respond to—such as job loss—the
very ambiguity and uncertainty about the threat of job loss hinders an individual’s ability to
resolve or alleviate the threat (De Witte 1999; Lazarus and Folkman 1984). The lack of options
to resolve this threat, coupled with its unpredictability, may therefore result in feelings of
powerlessness and low perceptions of control. Further, the absence of secure, long-term
employment may reduce an individual’s tendency to engage in certain actions that require long-
term planning, such as starting a family or purchasing a home—behaviors that may otherwise
foster positive beliefs about personal control (Hollister 2011).
78
While research has shown that a variety of work-related stressors are associated with lower
perceptions of personal control (Christie and Barling 2009; Schieman and Plickert 2008), there is
a dearth of strong empirical evidence regarding the effects of job insecurity. Waters and Moore
(2002) found that those experiencing frequent layoffs were more likely to believe that outcomes
in their life were a result of powerful others; however the generalizability of the authors’ findings
was limited by their small convenience-based sample (n=129). Moreover, larger cross-sectional
studies of North American and European workers that have examined the association between
insecurity and control report conflicting results (Lau and Knardahl 2004; Schieman 2002).
Added to this, there is a relative absence of research utilizing repeated observations of insecurity
and personal control over time—research designs that are better able to establish evidence of
causality. One notable exception is a study by Kinnunen, Feldt, and Mauno (2003) who found a
negative association between job insecurity and self esteem based on a one-year longitudinal
study of a non-representative sample of Finnish workers (n=456).
Despite the absence of consistent quantitative evidence on the subject, qualitative studies of
workers facing downsizing and the prospect of future job loss are suggestive of a link between
job insecurity and the sense of control. Greenhalgh (1979) examined the experiences of New
York state hospital workers who survived downsizing, for example, and noted how these workers
reacted to their situation in a manner similar to a traumatic event. Recurring themes are evident
in other studies that discuss the feelings of powerlessness and helplessness that workers facing
future job loss experienced (Greenhalgh and Rosenblatt 1984; Jick and Greenhalgh 1989). Based
on existing quantitative and qualitative research, therefore, I expect that workers reporting job
insecurity will also report a reduced sense of personal control.
Job Degradation as a Mediator
When workers experience job insecurity, is likely a work condition that is not encountered in
isolation. Some research finds that workers who survive organizational downsizing and
restructuring not only perceive less job security but also increased job pressures and reduced job
control (Burke and Nelson 1998; Quinlan and Bohle 2009). Exposure to these conditions, if
prolonged, may come to further erode the sense of control. Indeed, while the majority of research
on job insecurity has focused on the antecedents or consequences of perceptions of insecurity,
Greenhalgh and Rosenblatt’s (1984) original definition of insecurity considers it a
79
multidimensional construct that is based on the interaction between subjective characteristics and
objective job conditions. Using this broader conceptualization of job insecurity, I consider two
aspects of job degradation that may be relevant to declining job security, and that may serve to
either mediate or compound its effect on the sense of control: work intensification and reduced
decision-making latitude. That is, I seek to tease out the specific processes through which
insecurity influences personal control; either through the threat and uncertainty of job loss on the
one hand, and potential job degradation on the other.
Work intensification represents a workplace condition that is relevant to a broad section of
workers, both in secure and insecure jobs. Nevertheless, it is considered a hallmark experience of
workers in firms that downsize or restructure their organization (Kalleberg 2011). These
pressures may result from technical or organizational changes that increase the pace of work to
levels that exceed an individual’s ability to adequately complete tasks; or due to multiple or
conflicting demands from coworkers, managers or customers. The human consequences of
excessive time, task or relational work demands are well documented. Scholars find that
excessive job pressures are associated with higher levels of psychological distress and physical
health problems (Christie and Barling 2009; Schieman 2002). At the same time, excessive job
pressures signify a degree of powerlessness regarding an individual’s ability to adequately
complete work activities or meet job demands. Job pressures may therefore reduce feelings of
mastery in the work role, and potentially attitudes about personal control in general.
Similarly, organizational restructuring that is tied to declining job security may also lead to a
reduction in the control afforded to workers and the degree that they are able to make decisions
at work—job conditions that are consistently shown to be associated with high personal control
(Kohn and Schooler 1983). Moreover, the symbolic loss of a valued job feature such as decision-
making latitude may signify a loss of status and standing in the workplace, which may also serve
to reduce feelings of value, self-efficacy and personal control.
Based on these ideas and existing research, then, I expect that job insecurity should be associated
with a decline in job quality, in the form of excessive job pressures and reduced decision-making
latitude; and, since these conditions should be associated with lower sense of control, they should
subsequently mediate part of the overall association between job insecurity and personal control.
80
Contingencies: Duration and Life Course Timing
Research and theory suggest that the link between job insecurity and sense of personal control
may be contingent on: 1) the duration of exposure to job insecurity, and 2) life course timing. I
briefly discuss the theoretical and empirical rationale for these contingencies in turn.
Duration. If job insecurity does reduce perceptions of personal control, what is the temporal
nature of this association? Two scenarios appear plausible. One possibility is that given the
relatively stable nature of the sense of control as a feature of the self-concept (Mirowsky and
Ross 2003), the effect of job insecurity may only become evident with persistent, long-term
exposure. The threat of job loss and the pervading sense of insecurity may reduce beliefs about
personal control once they are seen as durable and intractable; this view of the consequences of
job insecurity is closest to its conceptualization as a chronic role stressor (Wheaton 1999). An
alternative possibility, however, is that the stable presence of insecurity has less or no impact on
personal control, since individuals may adjust to “environment constants” and come to see them
as normative (Mirowsky and Ross 2001). From this view, feelings of mastery and personal
control may be downwardly malleable only to a certain level, and after which, job insecurity may
have no further influence. Indeed, individuals may potentially regain a sense of control as they
come to reinterpret or downplay the threat of job loss as unimportant in their lives. The impact of
recently experienced job insecurity may therefore exert a stronger influence on personal control
than persistent insecurity. I test each of these hypotheses, which I label as the persistent and
novelty effects of job insecurity, respectively.
Life Course Timing. The effect of job insecurity on the sense of personal control may also vary
across age. Population studies of adults document that the sense of control varies across the life
course, peaking during middle age, before declining with older age (Mirowsky 1995), and there
is evidence that the deleterious effects of life stressors on personal control may be greater during
later life (Pearlin et al. 2007). A meta-analysis of the health consequences of job insecurity, for
example, found that older workers were more affected than younger workers (Cheng and Chan
2008). The disproportionate health effects across age may be because of the greater threat and
consequences that older workers associate with the threat of job loss—older workers who are
laid off face significant labour market challenges in comparison to younger workers (Hirsch et al.
2000). Or, it may be due to a cohort effect in which younger workers who have grown up in an
81
era of high employment instability may be more acclimatized and accepting to precarious work
conditions than their older counterparts (Farber 2010). Similarly, younger workers at the
beginning of their careers may be more accepting of the job degradation that often accompanies
decreased job security. In contrast, older workers, due to their greater experience and job status,
may perceive a decline in job autonomy or an increase in work pressures as more unexpected and
disagreeable, which may be subsequently more detrimental to their sense of control. As such, job
insecurity may represent a more stressful, debilitating and unfamiliar experience for older
workers (Schieman 2002). These potentialities suggest that the extent that job insecurity leads to
a lowered sense of control should increase with age.
Data and Methods To test the hypotheses described above, I analyze panel data from the Work, Stress and Health
study (WSH), which involved telephone interviews with adults in the 50 United States in 2005
and 2007. To obtain the original sample, a list-assisted random digit dialing (RDD) selection was
used and drawn proportionally from all 50 states from GENESYS Sampling Systems. Eligible
participants are 18 years of age or older and participating in the paid labour force. At Wave 1, 71
percent of eligible individuals were successfully interviewed yielding a sample of 1,800 adults.
At Wave 2 interviews, which occurred approximately 18-to-20 months after the initial interview,
1,286 of the original participants were successfully re-interviewed. Sample characteristics for the
WSH study are similar to the population estimates of the U.S. Census Bureau’s 2005 American
Community Survey (ACS).
The working sample includes non-self employed workers, between the ages of 25 and 65, who
participated in both waves of the survey, worked twenty hours or more a week, and who did not
change or lose their job between interviews (N=775). I briefly explain the rationale for the
sample selection. The measure of job insecurity that I use is designed to tap into workers
perceptions of the strength of the employer-employee psychological contract (Rousseau 1995);
as such, even though self-employed workers may face future threats to their continuing
employment, this measure of insecurity is not applicable to workers who are not hired by an
employer. Additionally, as is the case with most studies of job insecurity, respondents between
82
the ages of 18-24 are omitted from the working sample, since this group are overridingly in
contingent jobs due their disproportionate participation in higher education (Polivka 1996).
Finally, I restrict the sample to those who had the same employer between interviews because I
am specifically interested in examining how job insecurity is associated with changes in job
quality; that is, changes in job quality for a specific, single job. Including in the sample those
who changed jobs would not allow for such an examination.
Measures
Sense of personal control. Mirowsky and Ross's (2003) 2 x 2 index of the sense of personal
control is used. It asks participants to report their level of agreement or disagreement with the
eight statements; there are two statements in each of the four categories. Statements that measure
the level that individuals claim control over good outcomes include: (1. "I am responsible for my
own successes," and (2. "I can do just about anything I really set my mind to." Items that
measure claims of control over bad outcomes include: (3. "My misfortunes are the result of
mistakes I have made," and (4. "I am responsible for my failures." Items that assess the extent
that individuals deny control over good outcomes are: (5. "The really good things that happen to
me are mostly luck," and (6. "There's no sense planning a lot--if something good is going to
happen it will." The last two statements measure the denial of control over bad outcomes: (7.
"Most of my problems are due to bad breaks," and (8. "I have little control over the bad things
that happen to me." Responses to statements 1 through 4 are coded "strongly disagree" (-2),
"disagree" (-1), "neutral" (0), "agree" (1), and "strongly agree (2). Responses to statements 5
through 8 are coded "strongly disagree" (2), "disagree" (1), "neutral" (0), "agree" (-1), and
"strongly agree (-2). I averaged responses; higher scores indicate a greater sense of control (α =
.55).
Perceived Job insecurity. Job insecurity is assessed with the following question: “How likely is it
that during the next couple of years will you lose your present job and have to look for a job with
another employer?” Response choices include: “not at all likely” (1), “somewhat likely” (2), and
“very likely” (3). This question has been used in several well-respected surveys including the
General Social Survey and the National Study of the Changing Workforce. I follow Burgard and
colleagues (2009) modeling strategy in order to assess temporal change in job insecurity. I create
three binary dummy variables that represent episodic (wave 1 or wave 2 only) and persistent job
83
insecurity (both waves), and contrast these with the absence of job insecurity at both time points.
Persistent insecurity is coded 1 if respondents report that they are ‘somewhat’ or ‘very likely’ to
lose their jobs at both interviews and coded 0 otherwise, (2) resolved insecurity is coded 1 if
respondents reported that they are ‘somewhat’ or ‘very likely’ to lose their job only at wave 1,
and coded 0 otherwise, and (3) new insecurity is coded 1 if respondents reported that they are
‘somewhat’ or ‘very likely’ to lose their job at wave 2 only, and coded 0 otherwise. The
reference category is the absence of reported job insecurity at both time points.
Excessive job pressures. The following question is used to measure excessive job pressures: “In
the past 30 days, has anyone at work made too many demands on you?” If participants reported
yes, I then asked about the role-set source: “Was it a supervisor, someone you supervise,
customer/client, coworker, or someone else at work?” Participants could choose any source and
describe its frequency: (1) rarely, (2) sometimes, or (3) frequently. Individuals who reported no
one are the reference category. Participants were then asked a follow-up question: “How often do
the demands of your job exceed those doable in an 8-hour workday?” Response choices are (0)
never, (1) rarely, (2) sometimes, and (3) frequently. Responses to these items were standardized
and averaged to create the job pressure index.
Decision-making latitude. Two items assess decision-making latitude: “How often do you make
decisions on what needs to be done?” and “How often do you have the chance to solve
problems?” Response choices are (1) never, (2) rarely, (3) sometimes, and (4) frequently. I
averaged the items and higher scores indicate more decision-making latitude (α = .66).
Age. Age is modeled as a continuous variable.
Work hours. Participants were asked: “How many hours do you work in a typical week at your
main job?” I use a continuous measure of work hours.
Personal income. Personal income is assessed with the question: “For the complete year of 2004,
what was your personal income, including income from all of your paid jobs, before taxes?”
Job sector. I create dummy variables to indicate whether participants are employed by either the
government (1) or the private sector (0).
84
Occupation. To assess occupation, I use respondents’ reported job titles and main duties of the
“main job at which you worked last week” to code responses into five categories in accordance
with the Bureau of Labour Statistics codes: professional (managerial and professional specialty
occupations), administrative (technical, sales, and administrative support occupations), service
(service occupations), craft (precision production, craft, and repair occupations), and labour
(operators or labourers). In regression analyses, craft is the reference category.
Gender. I use dummy codes for men (0) and women (1).
Race. I coded participants’ race as “white” (1) versus all other categories (0).
Marital status. Marital status is coded as “married (1) versus all other categories (0).
Children. This is coded as the number of children under 18 living in the household, ranging from
no children to 2 or more children under age 18 living in the home.
Education. Education is coded as an ordinal variable with the following categories: “less than
high school,” “high school or GED,” “some college or associate degree,” “four-year college
degree,” and “graduate or professional degree.”
Plan of Analyses
My analytical strategy is designed to examine the following: the association between perceived
job insecurity and the sense of personal control, the extent that decline in job quality (excessive
job pressures, reduced decision-making latitude) contributes to this association, and whether the
associations between insecurity and the sense of control are contingent on age. To this end, I first
examine whether perceived job insecurity is associated with an increase in job pressures and
reduced decision-making latitude between wave 1 and wave 2. In table 4.2, I use ordinary least
squares (OLS) regression techniques and regress job pressures (and decision-making latitude) at
wave 2 on episodic and persistent job insecurity, adjusting for baseline (i.e. wave 1) levels of job
pressures (decision-making latitude) and controls.
In model 1 of table 4.3, I use ordinary least squares (OLS) regression techniques and regress
sense of personal control on episodic and persistent job insecurity, adjusting for baseline levels
of personal control, and controls. I then enter change in job pressures in model 2, and change in
85
decision-making latitude in model 3, to examine their respective contribution to any association
between job insecurity and sense of personal control revealed in model 1. I then test age-by-job
insecurity interactions in model 4 to examine whether the association between job insecurity and
sense of personal control is contingent on age.
With regards to the choice of longitudinal design, a panel regression model with fixed-effects
specification is generally preferred to OLS approaches that rely upon the stringent assumption
that unmeasured factors are uncorrelated with included covariates (Allison 1999). However, I use
a lagged-outcome method (where a baseline measure of job quality/sense of personal control is
included as a predictor) because participants whose exposure does not change over time would
be removed from analyses under a fixed effects design. This would therefore lead to the
exclusion of those with persistent insecurity from the analyses (Burgard et al. 2009).
Additionally, a common problem of longitudinal data analysis is that sample attrition can bias the
results; this is particularly problematic if baseline measures of the dependent measure predict
attrition (Mirowsky and Ross 2001). I therefore adjust for the likelihood of attrition in each of the
regression models. This adjustment takes into account the probability that each of the outcomes
or independent variables predict attrition. Using probit regression analyses I regressed “absent
from the sample at wave 2” on all independent variables. I then transformed predictions to obtain
the Mills Inverse Ratio and included it in all of the analyses. Although adjustment for the hazard
of attrition has little effect on estimates, I decided to retain it in all regression analyses as a
precaution (Mirwosky and Ross 2001).
Results
Descriptives
Table 4.1 reports descriptive statistics for all variables examined in the multivariate analyses at
wave 1 and wave 2. The majority of participants report an absence of job insecurity at both
waves (seventy-four percent), while twelve percent reported job insecurity at both waves. Seven
percent reported job insecurity at wave 2 but not at wave 1, and another seven percent reported
insecurity at wave 1 but not at wave 2.
86
Multivariate Analyses
In Table 4.2, I present regression analyses with excessive job pressures and decision-making
latitude as dependent variables. I first regress job pressures at wave 2 on perceived job insecurity.
In model 1a, I find that episodic (new) job insecurity is associated with higher levels of job
pressures at wave 2; no statistically significant association exists between job pressures and
persistent or resolved job insecurity. Among the control measures, work hours are positively
associated with job pressures at wave 2.
I then regress decision-making latitude at wave 2 on perceived job insecurity. Model 1b reveals
that persistent job insecurity is associated with lower levels of decision-making latitude at wave
2. No statistically significant association exists between episodic (new or resolved) job insecurity
and decision-making latitude. Among the control measures, women and older workers report
lower levels of decision-making latitude at wave 2.
In Table 4.3 I regress the sense of personal control on perceived job insecurity and changes in
job quality. Model 1 reveals a statistically significant association between new job insecurity and
personal control at wave 2, though this association is significant at the p<.10 level. No significant
association exists between personal control and persistent or resolved job insecurity. In model 2,
I include excessive job pressures at wave 2 to examine whether job pressures mediates the
association between new insecurity and personal control revealed in model 1. While the
inclusion of job pressures at wave 2 weakens the association between new insecurity and
personal control and reduces it to non-significance, sobel tests reveal no evidence of a
statistically significant mediation effect. Moreover, job pressures at wave 2 are not associated
with personal control. In model 3, with job pressures omitted, I include decision-making latitude
at wave 2 and find no evidence that it is associated with personal control, or that it mediates the
association between new job insecurity and control.
Finally, in model 4 I examine whether the association between job insecurity and personal
control is contingent on age. The interaction between new insecurity and personal control is not
significant, indicating that the association between new insecurity and personal control does not
vary across age. However, the statistically significant interaction between persistent insecurity
and personal control reveals that the association between persistent job insecurity and personal
87
control is contingent upon age. Specifically, while younger workers with persistent job insecurity
have similar levels of personal control as those reporting an absence of job insecurity, the
tendency for persistent insecurity to be associated with lower levels of personal control begins at
the age of thirty and increases with age; though the difference only becomes notable for older
workers. An incremental r-square test reveals that the improvement in r-square between models 3
and 4 is significant (F=7.24, p<.01). Figure 4.1 illustrates this contingency.
Attrition
The longitudinal design of the study was such that only those in the paid labour force were re-
interviewed at wave 2. It is possible therefore that attrition may account for the association
between job insecurity and sense of control, as well as the age contingencies that I find.
Individuals reporting job insecurity at wave 1, for example, may be more likely to experience
subsequent job loss, resulting in their absence from wave 2. All analyses therefore adjust for the
hazard of attrition to rule out these possibilities. Additionally, an examination of factors
predicting attrition between wave 1 and wave 2 (see Appendix I, model 2) shows that attrition is
largely random with respect to the baseline focal measures. One other related possibility
concerns the decision to restrict the analytical sample to workers who did not change employers
between interviews; it is possible that those with a high sense of control who reported job
insecurity at the first interview were more likely to look for and find another more secure job by
the second interview, resulting in their absence from the analytical sample. Consequently, the
possible omission of these individuals from the analytical sample might create the false
appearance of a negative association between job insecurity and sense of control, since the
number of individuals with job insecurity and low sense of control would be overestimated.
However, analyses (not shown) that include workers who changed employers between interviews
demonstrated similar findings as those presented here, ruling out this possibility.
88
Discussion
This paper makes three key contributions to scholarship on the social-psychological linkages
between a contemporary work experience and personality, namely by: 1) documenting the
association between job insecurity and sense of control in a large, national sample of workers, 2)
revealing how this association is contingent on the duration of job insecurity and the age it is
experienced, and 3) assessing the extent that job insecurity is associated with job degradation,
and whether this degradation accounts for the association between insecurity and sense of
control.
Given the decrease in job security in recent decades, and particularly in the light of the recent
recession, scholars have become interested in documenting the implications of these trends with
respect to workers’ health. However an issue that has been less considered is whether job
insecurity may also shape important psychological dispositions that are fundamental to health
and individual functioning. We know much about how autonomous and self-directed work
conditions are associated with personality dispositions like the sense of personal control; but
what of the uncertainty that permeates the daily work experiences of individuals who live under
the continual threat of job loss? The chronic nature of this threat, its ambiguous nature, and the
relative lack of options to resolve it may present serious challenges to the belief that one
commands control over their life. The findings of this paper support the conceptualization of job
insecurity as a chronic work stressor that leads to feelings of powerlessness; however, as I
discuss, the extent that it does so depends on several key temporal factors.
Duration and Timing
My findings provide evidence that the effect of job insecurity on the sense of control is
contingent on both the duration and timing of the experience. I find no evidence that the past
experience of job insecurity has any lagged or long-term consequences for personal control. In
contrast, I find partial support for both the novelty and persistent hypotheses. Workers
experiencing the recent threat of job loss report lower levels of personal control relative to those
without job insecurity—suggestive of the possibility that the sense of control is a psychological
disposition that is reactive over relatively short time-spans. I also find, however, that persistent
job insecurity is associated with reduced personal control, rejecting the notion that individuals
89
eventually adapt to environmental constants, where ever-present uncertainty about job loss is the
constant.
A caveat exists here, however. The association between persistent job insecurity and reduced
personal control is evident only among middle age and older workers. This suggests that younger
workers may in some manner adapt to prolonged exposure to job insecurity. I do not have access
to these individuals’ prior work histories, but workers in their twenties who report job insecurity
at both waves of the survey likely have encountered no other type of employment situation in
their career. Job insecurity may therefore be a normative experience for these individuals; likely
they have either already adapted to their uncertain economic situation, or they may not connect
such powerlessness to their general sense of control. Here, these possibilities speak to a
potentially important issue regarding the meaning of stressors and the ways in which individuals’
prior experiences and beliefs may influence reactions to stress (Wheaton 1999). The fact that
older workers experience the greatest reduction in personal control in response to the threat of
job loss suggests that this threat holds a different set of meanings for these individuals; that is,
they may perceive it as either more consequential for their careers or family, or it is more
incongruent with their prevailing expectations about the employer-employee psychological
contract. Unfortunately, I am unable to determine the specific meanings that older workers attach
to job insecurity, or the ways in which they are consequential for personal control. It may be that
both worries about the chances of rehire, as well as their stronger expectations about job security,
contribute to the disproportionate penalties that they experience with respect to their sense of
personal control.
The above explanation why younger workers may be unresponsive to prolonged job insecurity
notwithstanding, it is still surprising that older workers do not incur greater penalties from
recently experienced job insecurity. One might expect to find similar age contingencies with
respect to new and persistent job insecurity. Additional analyses of curvilinear age contingencies
did not shed any light on this issue. More research is therefore necessary to address this question.
Are these findings suggestive that young workers who are now entering a labour market
characterized by high instability and mobility will be unaffected by the threat of job loss, since
they have experienced nothing else? Perhaps not. With only two waves of interviews, two years
apart, I have a relatively short snapshot of workers’ lives. A longer view of the life course is
90
necessary—a key tenet of the life course perspective (Elder 2003)—in order to fully consider
how beliefs about personal control may be influenced by insecure employment experiences. Such
an approach would allow for an investigation of how job insecurity influences decisions about
family formation, home purchases, and other significant life events that may either serve to foster
or undermine the sense of control. Thus, an accurate appraisal of the consequences of the recent
recession for young workers—a group whose employment opportunities have been hardest hit—
would require follow-up interviews with these individuals over the next decade.
The Contribution of Job Degradation
Traditional conceptualizations of job insecurity define it as a perceived powerlessness with
regards to control over one’s continuing employment. However, broader conceptualizations of
the construct acknowledge that the insecurity or loss of valued job conditions may also be
relevant (Sverke and Hellgren 2002). Declining job security and job quality is therefore likely
experienced in tandem, as employers place greater demands on surviving workers, or restructure
in ways that reduce worker autonomy. My findings demonstrate evidence of this: workers in the
sample who reported a decline in job security also reported greater job pressures and less
autonomy at wave 2—though these associations depend on the duration of job insecurity. I find
that the recent—but not persistent—experience of job insecurity is associated with an increase in
excessive job pressures. That excessive job pressures are not associated with persistent job
insecurity is somewhat surprising, but it may be that individuals come to adjust to the changes in
pressures that result from organizational restructuring or downsizing.
While it is only the recent experience of job insecurity that I find to be associated with increased
job pressures, I find the opposite with regard to decision-making latitude; only workers reporting
persistent insecurity reported a decline in decision-making latitude at work. Here, it may be that
decision-making latitude is a job resource that is eroded more slowly, as organizations engage in
work-redesigns that emphasize efficiency at the expense of worker autonomy.
With respect to these findings, one possibility that is worth noting is that workers with job
insecurity may perceive an increase in job pressures or a reduction in job latitude without an
actual objective change in job conditions. This could explain why new job insecurity is
associated with job pressures, but not persistent insecurity. As an individual comes to consider
91
job insecurity an accepted feature of their job situation, the tendency to appraise more job
pressures may also decline. It is harder, however, to see how this explanation might account for
changes in decision-making latitude, which are associated with persistent insecurity only. If
participants’ reports of job autonomy were purely a result of insecurity shaping perceptions about
decision-making latitude, one would expect to find an association between the recent onset of job
insecurity and latitude. Thus, it would seem more plausible that reports of changing job quality
are—at least in part—the result of objective changes in workers’ jobs.
Despite the links between job insecurity and job degradation, I find no evidence that changes in
job pressures or decision-making latitude account for the association between insecurity and
sense of control. Indeed, I find no evidence that changes in job pressures or job latitude are
associated with personal control. One possibility is that personal control is reactive only with
longer-term exposure to such job conditions. Kohn and Schooler’s (1983) longitudinal study of
the association between occupational self-direction and sense of control examined changes over
a ten-year period; the two-year gap between interviews in the present study may be inadequate to
identify an association. In contrast, the threat of job loss may result in a more immediate and
stronger feeling of powerlessness since it challenges an important aspect of identity—
employment. Moreover, although I do not directly test this, it may be that in contexts of threat
and economic uncertainty, that the importance of valued job conditions for the sense of control
diminish in importance, as workers become more focused on maintaining employment, rather
than the quality of their employment.
One limitation of this study deserves mention. The lagged-dependent design method utilized
does not provide a strict test of temporal ordering between the focal measures of interest. That is,
it is possible that individuals with a low sense of control are more likely to perceive that their job
is insecure, for example, rather than the reverse association hypothesized in this paper.
Moreover, it is possible that the relationship between insecurity and the sense of control is
actually reciprocal. Christie and Barling (2009), for example, find evidence that individuals
reporting high level of work stressors subsequently reported lower levels of personal control; but
they also find that low levels of personal control subsequently predicted higher levels of work
stressors—leading them to the conclusion that the nature of stress and stress recovery is a
dynamic process, and one that is better explained in terms of positive or negative control spirals.
92
I am unable to replicate Christie and Barling’s (2009) study because of the absence of multi-item
measures that are required in the latent curve modeling technique that they use, but in analyses
not shown, I find no evidence that changes in sense of control predict perceived job insecurity at
wave 2. A logistic regression with perceived job insecurity at wave 2 regressed as the
independent variable on change in sense of control revealed no evidence that sense of control
significantly predicted job insecurity. Thus, while I cannot rule out the possibility that the
relationship between the focal measures of interest is dynamic and reciprocal, these additional
analyses support a causal argument in which perceptions of job insecurity influence beliefs about
personal control, rather than the opposite.
Conclusion It has been suggested that the anticipation of an undesirable event, like job loss, can be more
distressing than the event itself (Lazarus and Folkman 1984). This may be in part because the
possibility of a future event does not lend itself well to coping behaviors that are typically only
useful when the event actually occurs, but also because the perceived absence of available coping
strategies generates feelings helplessness and powerlessness—naturally unpleasant and
distressing feelings. Moreover, if the situation that one lacks control over is durable and concerns
a particularly salient role, such as employment, it is possible that these feelings may become
generalized as part of a personality trait that undermines instrumental and proactive behaviors
necessary to deal with the daily hassles and stresses of life. The findings of this paper with
respect to the sense of personal control are suggestive of this latter indirect and long-term process
that may connect job insecurity to well-being. Future research should seek to further examine
how insecure employment relations may shape personal control over the long-term, and the
subsequent effects on workers’ health over the life course.
93
References Allison, Paul D. 1999. Multiple Regression: A Primer. Thousand Oaks, CA: Pine Forge.
Avison, William R. 2001. Unemployment and its consequences for mental health. In V.W. Marshall, Krueger, & A. Verma (Eds.), Restructuring work and the life course (pp. 177-200). Toronto: University of Toronto Press.
Bakker, Arnold B. and Sabine A. E. Geurts. 2004. “Toward a Dual-Process Model of Work-Home Interference.” Work and Occupations 31:345–66.
Broman, Clifford L., V. Lee Hamilton, and William S. Hoffman. 2000. Stress and Distress among the Unemployed: Hard Times and Vulnerable People. New York: Kluwer Academic/Plenum Publishers.
Burgard, Sarah A., Jennie Brand, and James S. House. 2009. “Perceived Job Insecurity and Worker Health in the United States.” Social Science & Medicine 69:777-785.
Burke, Ronald J. and Debra Nelson. 1998. “Downsizing, Restructuring and Privatization: a North American Perspective.” Pp. 21–54 M.K. Gowing, J.C. Quick, J.D. Kraft (Eds.), The New Organizational Reality: Downsizing, Restructuring and Revitalization, American Psychological Association, Washington DC.
Cheng, Grand. H. L. and Darius K. S. Chan. 2008. “Who Suffers More from Job Insecurity? A Meta-Analytic Review.” Applied Psychology 57: 272–303.
Christie Amy M. and Julian Barling. 2009. “Disentangling the Indirect Links Between Socioeconomic Status and Health: The Dynamic Roles of Work Stressors and Personal Control.” Journal of Applied Psychology 94:1466-1478.
Davy, Jeanette A., Angelo J. Kinicki, and Christine L. Scheck. 1997. “A Test of Job Security’s Direct and Mediated Effects on Withdrawal Cognitions.” Journal of Organizational Behavior 18:323–349.
De Witte, H. 1999. “Job Insecurity and Psychological Health: Review of the Literature and Exploration of Some Unresolved Issues.” European Journal of Work and Organizational Psychology 8:155-177.
Elder, Glen, H. J. 2003. “The Emergence and Development of Life Course Theory.” Pp. 3-19 of Handbook of the Life Course, edited by Jeylan T. Mortimer and Michael J. Shanahan. New York: Kluwer Academic/Plenum Publishers.
Elman, Cheryl, and Angela M. O’Rand. 2002. “Perceived Job Insecurity and Entry into Work-Related Education and Training Among Adult Workers.” Social Science Research 31:49-76.
Farber, Henry S. 2010. “Job Loss and the Decline of Job Security in the United States.” Pp. 223-267 in Labour in the New Economy edited by Katherine Abraham, James Spletzer and Michael Harper. University of Chicago Press, Chicago.
Feldt, Taru, Ulla Kinnunen, and Saija Mauno. 2000. “A Mediational Model of Sense of Coherence in the Work Context: A One-Year Follow-up Study.” Journal of Organizational Behavior 21:461–476.
94
Fullerton, Andrew S. and Michael Wallace. 2007. “Traversing the Flexible Turn: US Workers’ Perceptions of Job Security, 1977–2002.” Social Science Research 36:201-221.
Greenhalgh, Leonard. 1979 “Job Security and the Disinvolvement Syndrome: An Exploration of Patterns of worker Behaviour under Conditions of Anticipatory Grieving over Job Loss. Unpublished doctoral dissertation, Cornell University.
Greenhalgh, Leonard, and Zehava Rosenblatt. 1984. “Job insecurity: Toward Conceptual Clarity.” Academy of Management Review 3:438–448.
Hirsch, Barry T., David A. MacPherson, and Melissa A. Hardy. 2000. “Occupational Age Structure and Access for Older Workers.” Industrial and Labour Relations Review 53: 401–418.
Hollister, Matissa N. 2011. “Employment Stability in the U.S. Labour Market: Rhetoric vs. Reality.” Annual Review of Sociology 37:305-324
Jahoda, Marie. 1982. Employment and Unemployment: A Social Psychological Analysis. New York: Cambridge University Press.
Jick, Todd D. and Leonard Greenhalgh. 1989. "Survivor Sense Making and Reactions to Organizational Decline: Effects of Individual Differences." Management Communication Quarterly 2:305-327.
Kalleberg, Arne. 2011. Good Jobs, Bad Jobs: The Rise of Polarized and Precarious Employment Systems in the United States, 1970s-2000s. New York: Russell Sage Foundation, American Sociological Association Rose Series in Sociology.
Kohn, Melvin L. and Carmi Schooler. 1983. Work and Personality: An Inquiry into the Impact of Social Stratification. Norwood, N.J. Ablex Pub. Corp.
Lau, B., and S. Knardahl. 2008. “Perceived Job Insecurity, Job Predictability, Personality, and Health.” Journal of Occupational and Environmental Medicine 44:172–181.
Lazarus, Richard S. and Susan Folkman. 1984. Stress, Appraisal, and Coping. New York: Springer.
Marx, Karl. 1983[1852]. “From the Eighteenth Brumaire of Louis Bonaparte.” The Portable Karl Marx. Eugene Kamenka, editor. Penguin.
Mirowsky, John. 1995. “Age and the Sense of Control.” Social Psychology Quarterly 58:31–43.
Mirowsky, John, and Catherine E. Ross 1992. "Age and Depression." Journal of Health and Social Behavior 33:187-205.
______. 2001. “Age and the Effect of Economic Hardship on Depression.” Journal of Health and Social Behavior 42:132–50.
______. 2003. Social Causes of Psychological Distress. Hawthorne, New York: Aldine De Gruyter.
Pearlin, Leonard I., and Carmi Schooler. 1978. “The Structure of Coping.” Journal of Health and Social Behavior 19:2–21.
Pearlin, Leonard I. 1983. “Role Strains and Personal Stress.” Pp. 3–32 in Psychosocial Stress: Trends in Theory and Research, edited by Howard Kaplan. New York: Academic Press.
95
Pearlin, Leonard I., Scott Schieman, Elena Fazio, and Stephen C. Meersman. 2005. “Stress, Health, and the Life Course: Some Conceptual Specifications.” Journal of Health and Social Behavior 46:205–19.
Pearlin, Leonard I., Kim B. Nguyen, Scott Schieman, and Melissa A. Milkie. 2007. “The Life-Course Origins of Mastery among Older People.” Journal of Health and Social Behavior 48:164-179.
Polivka, Anne E. 1996. “A Profile of Contingent Workers.” Monthly Labour Review 119:10–21.
Quinlan, Michael and Philip Bohle. 2009. “Overstretched and Unreciprocated Commitment: Reviewing Research on the Occupational Health and Safety Effects of Downsizing and Job Insecurity.” International Journal of Health Services 39:1-44.
Ross, Catherine E. and John Mirowsky. 1989. "Explaining the Social Patterns of Depression: Control and Problem Solving-or Support and Talking?" Journal of Health and Social Behavior 30:206-19.
Ross, Catherine E., and Jaya Sastry. 1999. “The Sense of Personal Control: Social-Structural Causes and Emotional Consequences.” Pp. 369–94 in Handbook of the Sociology of Mental Health, edited by Carol S. Aneshensel and Jo C. Phelan. New York: Kluwer.
Rousseau, Denise M. 1995. Psychological Contracts in Organizations: Understanding Written and Unwritten Agreements. Thousand Oaks, CA.
Schieman, Scott. 2002. “Socioeconomic Status, Job Conditions, and Well-Being: Self-Concept Explanations for Gender-Contingent Effects.” The Sociological Quarterly 43:627–46.
Schieman, Scott and Gabriele Plickert. 2008. “How Knowledge is Power: Explaining the Association between Education and the Sense of Control.” Social Forces 87:153–84.
Seeman, Melvin. 1959. “On the Meaning of Alienation.” American Sociological Review 24:783–91.
Seeman, Melvin. 1967. “On the Personal Consequences of Alienation in Work.” American Sociological Review 32:273-285.
Sverke, Magnus, and Johnny Hellgren. 2002. “The Nature of Job Insecurity: Understanding Employment Uncertainty on the Brink of a New Millennium.” Applied Psychology 51:23–42.
Turner, R. Jay, and Patricia Roszell. 1994. “Personal Resources and the Stress Process.” Pp. 179–210 in Stress and Mental Health: Contemporary Issues and Prospects for the Future, edited by William. Avison and Ian. H. Gotlib. New York: Plenum Press.
Waters, Lea E. and Kathleen A. Moore. 2002. “Self-esteem, Appraisal and Coping: A Comparison of Unemployed and Re-employed People.” Journal of Organizational Behaviour 23:593-604.
Wheaton, Blair. 1980. "The Sociogenesis of Psychological Disorder: An Attributional Theory." Journal of Health and Social Behavior 21:100-24.
Wheaton, Blair. 1983. "Stress, Personal Coping Resources, and Psychiatric Symptoms: An Investigation of Interactive Models." Journal of Health and Social Behavior 24:100-24.
96
Wheaton, Blair. 1999. “Social Stress.” Pp. 277-300 in The Handbook of the Sociology of Mental Health, edited by C.S. Aneshensel and J.C. Phelan. Kluwer Academic/Plenum.
97
TABLE 4.1 Means and Proportions of all Study Variables Sense of Personal Control Wave 1 .850 Wave 2 .872 Perceived Insecurity % reporting at Wave 1 .194 % reporting at Wave 2 .180 Episodic and Persistent Insecurity % Absent Insecurity .740 % New (Wave 2 only) .065 % Resolved (Wave 1 only) .075 % Persistent (Wave 1 and Wave 2) .120 Work Conditions Excessive Job Pressures Wave 1 .073 Wave 2 .116 Decision-making Latitude Wave 1 3.620 Wave 2 3.548 Work Hours (Weekly) 44.125 Personal Income 45884.82 Employed in Public Sector .276 Socio-demographics Female .608 Married .652 Children in the Household .982 White .745 Education 3.709 N 775
98
TABLE 4.2 Regression of Excessive Job Pressures and Decision-making Latitude at Wave 2 on Perceived Job Insecurity and Controls (N=775)
Excessive Job Pressures
Decision-making Latitude
Perceived Job Insecurity
New Job Insecurity a 0.181+ -0.101
Resolved Job Insecurity a 0.106 0.005
Persistent Job Insecurity a 0.031 -0.115*
Other Work Conditions
Work Hours (Weekly) 0.010*** 0.001
Employed in Public Sector -0.008 -0.035
Occupation
Labor b -0.273* -0.089
Administrative b -0.178** -0.053
Service b -0.258** -0.167**
Craft b -0.149 -0.044
Personal Income 0.000 0.000
Demographics
Female 0.034 -0.134**
Age -0.008 -0.010**
Education 0.011 0.018
Married -0.030 -0.041
Children in the Household 0.012 -0.011
White -0.057 -0.009
Wave 1
Excessive Job Pressures 0.428*** —
Decision-making Latitude — 0.537***
Constant -.441 .870
R2 .261 .339
+ p < .10; * p < .05; ** p < .01; *** p < .001 (two–tailed test). a Compared to absent job insecurity b Compared to professionals.
99
TABLE 4.3 Regression of Sense of Personal Control at wave 2 on Perceived Job Insecurity, Age, Interactions, Job Conditions and Controls (N=775)
Model 1 Model 2 Model 3 Model 4 Perceived Job Insecurity New Job Insecurity a -0.088+ -0.076 -0.040 -0.017 Resolved Job Insecurity a -0.059 -0.057 0.340 0.306 Persistent Job Insecurity a -0.042 -0.037 0.313 0.338+ Changes in Job Quality Excessive Job Pressures (Wave 1) -0.016 -0.005 -0.015 -0.017 Change in Excessive Job Pressures — -0.026 — 0.306 Decision-making Latitude (Wave 1) 0.024 0.007 0.019 0.338 Change in Decision-making Latitude — 0.033 — -0.017 Age -0.001 -0.001 0.001 0.001 Insecurity × Age Interactions Age × New a — — -0.001 -0.001 Age × Resolved a — — -0.009 -0.008 Age × Persistent a — — -0.008* -0.008* Other Work Conditions Work Hours (Weekly) -0.001 -0.001 -0.001 0.000 Employed in Public Sector -0.018 -0.017 -0.021 -0.020 Occupation Labor b -0.095 -0.099 -0.090 -0.093 Administrative b 0.016 0.013 0.023 0.019 Service b -0.140** -0.142** -0.128** -0.130** Craft b -0.021 -0.023 -0.009 -0.011 Personal Income 0.000 0.000 0.000 0.000 Demographics Education 0.039** 0.038** 0.033* 0.032* Married -0.010 -0.009 -0.020 -0.020 Children in the Household 0.041* 0.042* 0.036 0.036 White 0.066 0.063 0.049 0.046 Wave 1 Sense of Control at Wave 1 0.552*** 0.547*** 0.544*** 0.539*** Constant 0.663 0.608 0.418 0.362 R2 .331 .351 .388 .388
+ p < .10; * p < .05; ** p < .01; *** p < .001 (two–tailed test). a Compared to absent job insecurity b Compared to professionals.
100
FIGURE 4.1 Age Contingencies in the Association between Job Insecurity and Sense of Control
0.2
0.4
0.6
0.8
1
1.2
1.4
25 28 31 34 37 40 43 46 49 52 55 58 61 64
Sen
se o
f Con
trol
Age
Persistent Insecurity
New Insecurity
Resolved Insecurity
Absent
101
Chapter 5 Conclusions and Discussion
Overview of Contribution and Themes
Declining Job Security
Social, economic and political changes in the last three decades have led to a diversification of
employment relations and work experiences in the United States (Kalleberg 2009). The ‘standard
employment relation’ has been increasingly replaced by nonstandard work arrangements that
deviate from the full-time, permanent employment contract (Vosko 2006). These labour market
trends defy generalization with respect to workers’ experiences in the workplace. Job quality has
risen for some segments of the labour force—evident in the growth of professional
occupations—and conversely declined for others (e.g. the spread of precarious work) (Kalleberg
2011). However, one trend that has been widespread and universal across the labour force is the
decline in job security over the last three decades (Fullerton and Wallace 2007). Few groups
have been untouched by systematic economic restructuring; restructuring that has led firms to
externalize onto workers the uncertainty of contemporary market capitalism, in the form of
flexible and insecure employment relations.
How workers experience and react to this increased risk and uncertainty has been the central
issue addressed by this dissertation. Existing research documents the deleterious health
consequences of perceived job insecurity as a stressor (Sverke and Hellgren 2002), but questions
remain about the conditions under which the threat and uncertainty of job loss reduces health.
This dissertation has sought to understand how the duration of the experience of job insecurity,
as well as its timing in the life course, represents key factors in the relationship between
perceived job insecurity and health, in addition to the role of personality in these processes.
Guided by the Stress Process Model and the life course perspective (Elder 1998; Pearlin et al.
1981), I reveal how the stress of job insecurity is greatest when the experience is prolonged and
encountered during middle age, as well as being most detrimental for those with a low or
external locus of personal control. I also begin to ‘unpack’ the mechanisms through which job
insecurity influences individuals by examining whether job degradation accounts for the negative
health consequences of insecure work.
102
The Duration and Life Course Timing of Job Insecurity
Previous research on the stress of perceived job insecurity has failed to adequately assess how
the timing and duration of the experience represent potential factors that shape its health
penalties. The first paper of this dissertation addresses this omission by examining the effect of
short-term and prolonged exposure to job insecurity on mental and physical health, and the
extent that this is contingent on life course position. Analyses of a national and longitudinal
survey of American workers reveal a decline in mental and physical health among middle age
workers reporting prolonged exposure to perceived job insecurity. I also find that the likely
source of these health penalties—anxiety over the threat of job loss—may reduce absenteeism
among middle age workers, as they attempt to consolidate their position with their employer. The
findings of the paper thus support the conceptualization of perceived job insecurity as a chronic
stressor, while providing evidence of the theoretical value of the life course perspective with
regard to the study of the stress process. Here, I argue that that the meaning of job insecurity and
the threat of job loss vary across the life course with changing labour market and family
experiences, along with different expectations about acceptable work conditions. Specifically,
middle age may represent the ‘perfect storm’ of high family financial commitments combined
labour market vulnerability and the expectation of secure employment—a combination that
imbues the threat of job loss with the most danger, frustration and anxiety.
The Sense of Control and the Link between Job Insecurity and Distress
The second paper in this dissertation makes several contributions that advance understanding
regarding the links between the sense of control and the stress of job insecurity. Drawing on two
national samples of American and Canadian workers, I show that a high sense of personal
control buffers the deleterious health consequences of job insecurity. However, the sense of
control is also shown to mediate the impact of job insecurity on psychological distress, indicating
that one of the mechanisms through which job insecurity decreases mental well-being is through
its propensity to erode the extent that one believes they exert control over their life. Individuals
with a lowered sense of control subsequently experience higher levels of distress because of a
reduced efficacy in dealing with stress in everyday life. In addition, the large sample size of the
American and Canadian surveys allow for a more detailed investigation of age contingencies in
the stress of job insecurity. These analyses provide mixed results and only partial support for the
103
stress of middle age hypothesis. In the American WSH study, workers age 35-44 report the
highest levels of distress in response to job insecurity. No age contingencies are found in the
Canadian CAN-WSH study however. The paper concludes with a consideration of structural
amplification as an explanation of the dual mediating and moderating role that the sense of
control plays with regard to the stress of job insecurity. Structural amplification represents a
potentially useful concept for understanding the longer-term links between job insecurity and
health.
Examining the Sense of Control as an Outcome of Job Insecurity
The third paper advances on the previous paper with a more detailed examination of how
perceived job insecurity may shape the sense of personal control, and how this process may be
contingent on the duration and life course timing of the experience. The paper also assesses the
extent that job degradation, which is often associated with declining job security, is responsible
for the lowered personal control reported by workers facing the threat job loss.
The study of the connections between personality and the social organization of work has a rich
heritage, driven in large part by the classic work of Kohn and Schooler in the 1970s and 1980s
that examined the links between personality and occupational self-direction. This third paper
advances on this work by considering a contemporary feature of workplace organization—
insecure work—and how it is associated with changes in the sense of control over time. I find
that both recently experienced and prolonged job insecurity reduce the sense of control, but this
latter effect exists only among middle age and older workers. I speculate that the different role of
age in the effects of short-term and prolonged insecurity on the sense of control may be reflective
of a normalizing process for younger workers exposed to prolonged job insecurity. These
workers may, in some manner, adapt to prolonged job insecurity, particularly if they have had no
other experience with job security; they may simply not consider such powerlessness as
important or relevant to their general sense of control over their life. These findings again point
to the role of the meaning that individuals attach to stressors—meanings that may serve to reduce
or exacerbate the potency of a particular stressor. The fact that older workers experience the
greatest reduction in personal control in response to the threat of job loss suggests that this threat
holds a different set of meanings for these individuals; that is, they may perceive it as either more
consequential for their career or family, or it may be more incongruent with their prevailing
104
expectations about the employer-employee psychological contract (Rousseau 1995). The value of
integrating the life course approach into the study of stress is apparent here, since the meanings
of job loss and job insecurity likely depend on social circumstances that vary over working life,
along with perceptions of whether job insecurity considered normative or unusual.
Finally the potential role of job degradation (rising job pressures and reduced decision-making
latitude) in the association between job insecurity and personal control is assessed). Analyses
reveal that perceptions of job insecurity are associated with reduced decision-making latitude and
increased job pressures—albeit differently across age—but changes in job quality do not account
for the tendency of job insecurity to reduce personal control. I conclude from these results that
declining job security likely comes packaged with a general decline in job quality, but it is the
threat of job loss that primarily serves to reduce the sense of personal control.
Summary Collectively, these papers support previous arguments that chronicity is a necessary requirement
for certain stressors to impact health (Wheaton 1999). This appears especially true for perceived
job insecurity, which may primarily begin to influence health through feelings of anxiety. As
these feelings become prolonged, they may constitute an allostatic load on the individual that
gradually compromises their mental and physical resistance to stress in general, resulting in the
gradual accumulation of health problems rather than immediate penalties. A parallel process that
may account for the longer-term health effects of job insecurity is through its tendency to
influence personality traits that are important to well-being. The association between job
insecurity and a declining sense of control revealed in paper three suggests support for this
possibility. According to the Stress Process Model, the sense of control is considered an
important resource for helping individuals overcome the daily challenges of life by promoting
active coping strategies to resolve these challenges. The erosion of personal control due to job
insecurity may represent the first stage in a longer process in which individuals become
increasingly vulnerable to stress as their repertoire of coping strategies diminish.
The consistent relevance of timing in the stress of job insecurity is matched by a set of findings
across of each of the three papers that demonstrate the importance of age and life course position.
105
All papers show evidence that the penalties of job insecurity peak at specific stages in the life
course. In each paper, I have discussed the role of labour market constraints and family demands
experienced in middle and later working life that may shape the consequences of job loss and the
meaning of job insecurity. Additionally, I have suggested that younger workers may consider the
experience as normative—particularly if they have encountered nothing else in their short
working life. Middle age and older workers, in contrast, may consider job insecurity a more
unexpected and distressing experience; however, it may be middle age workers, in particular,
because of their expectations about job security at the prime stage of working life, who are most
vulnerable to the stress of employment uncertainty. Collectively, my findings suggest that
duration and life course timing operate together in order to shape the meaning and stress of job
insecurity—highlighting the value of a theoretical alliance between Stress Process and life course
researchers (Pearlin and Skaff 1996).
Limitations Several limitations of this dissertation deserve mention. While the measure of perceived job
insecurity is commonly used in survey research, the inclusion of a more comprehensive set of
questions that assess individuals’ perceptions of job security is desirable. A single item indicator
of job insecurity does not allow for assessment of reliability. Despite this, I believe that asking
workers about the likelihood of future job loss is a concise and succinct strategy for assessing job
insecurity, and one that avoids confounding personality and mental health traits (e.g.
neuroticism) with the appraisals of security. Future research, however, would benefit from
multiple measures of job insecurity, along with measures that allow for greater fidelity in
responses (e.g. job security measured on a scale of 1-100). This would make available a range of
alternative multivariate techniques, including fixed-effects panel regression and structural
equation modeling. The former technique would enable a more rigorous examination of changes
in health to changes in predictors, thereby removing the effects of all time-stable characteristics,
while the latter technique would allow for a more comprehensive empirical modeling of the
hypothesized linkages between insecurity and health (Allison 1999). With a single-item measure
of perceived job insecurity, these techniques are unavailable.
106
A second important weakness of this dissertation is the inability to directly examine the
meanings that workers attach to job insecurity. I have argued that these meanings may shape the
anxiety associated with the threat of job loss, and that they should vary according to life course
position. I therefore rely on age as a proxy for life course position and as an indirect indicator of
the meanings that individuals of different ages attach to job insecurity. Assumptions about the
different meanings that individuals attach to job loss are informed by prior research and theory,
but it would be desirable to assess, with the use of survey items, the specific consequences that
individuals attach to future job loss, or their views about early retirement, for example. In-depth
interviews would also provide more detailed information on how individuals interpret and
negotiate the experience of job insecurity, and how this intersects with social statuses and family
structure. This would allow for a teasing-out of cohort-effects from age-effects in the experience
of job insecurity, for example.
A number of limitations are worthy of mention with regards to the integration of the life course
perspective with the Stress Process Model, such as the inability to comprehensively examine
gender contingencies and the absence of a thorough application of key life course principles. I
discuss these in more detail in the following section under ‘Future Research,’ but it is important
to acknowledge these limitations. Unfortunately, sample-size limitations in the longitudinal
WSH study limit the ability to adequately perform subgroup analyses for men and women. The
relatively limited number of participants that reported each of the job insecurity experiences
meant that there were low cell-sizes for examining three-way interactions between insecurity,
age and gender. A larger sample was available in the form of the CAN-WSH study, but this was
limited by its cross-sectional design. As future waves of data are collected as part of this ongoing
study, more in-depth and longitudinal subgroup analyses will become possible. Finally, a key
emphasis of the life course perspective is to consider a long view of individuals’ lives. With two
waves of survey data, eighteen months apart, this was obviously not attainable. Only with further
data points, or the use or retrospective data, can this goal be achieved. I discuss this issue in more
detail in the next section on future research.
107
Future Research The integration of the life course perspective with the Stress Process Model offers several
advantages for examining the contingent nature of stressors, and in the case of this dissertation,
the stress of job insecurity. While this dissertation has been guided by these principles, I discuss
how future research could expand on this research in order to offer a more comprehensive life
course analysis of the experience of job insecurity. First, a longer view of individuals’ lives is
necessary in order to examine how the experience of job insecurity is actively negotiated over
time, and how it may influence later life outcomes that have implications for health. How might
uncertainty over one’s continuing employment influence marital decisions or home purchases?
These are important life decisions that may be delayed until secure employment is achieved, but
whose delay may have consequences for health and well-being. In particular, this extended view
would allow for the investigation of whether more serious physical health problems develop over
time. Similarly, changes in personality traits like the sense of personal control—that change
slowly over the life course—may be more noticeable when a longer view of workers’ lives is
adopted. Following workers over a period of five or ten years may therefore shed light on
whether a prolonged history of job insecurity results in the accumulation of problems, or whether
individuals develop resistance to uncertainty, either due to a process of normalization or active
renegotiation of the experience.
Given that men and women navigate work and family roles across the life course in different
ways, another area warranting further research is a consideration of how the experience of job
insecurity may shaped by, or contingent on gender (Winslow 2005). Men and women navigate
distinct career paths due to gendered meanings about work and family roles that serve to
constrain women’s career decisions—decisions that are further challenged by a social
organization of working time that considers continuous, full-time employment as normative and
expected throughout adulthood (Moen 1992). Given a continuing gender division household
labour (Hochschild 1989), the career paths of middle age women, in particular, often diverge
noticeably from their male counterparts. Moen and Sweet (2004), for example, describe a
dominant neotraditional strategy enacted by dual-earner households, which specifies that
husbands have the ‘main’ career job and wives work in less demanding jobs with shorter hours in
108
order to deal with domestic and family responsibilities. Thus, given gender differences in the
nature and level of labour force participation across the life course, there may be gender
differences the extent that the stress of job insecurity is contingent on life course position.
Specifically, it is possible that job insecurity may be more stressful for men because of their
higher levels of labour force participation on one hand, and continuing gender role stereotypes
that equate masculinity with the “good provider” (Christiansen and Palkovitz 2001). Future
research should therefore build off the age contingencies revealed in this dissertation by
considering how the stress of job insecurity may be influenced by the intersection of gender and
age as key social statuses.
Another issue deserving of further investigation is the extent that regional labour market
differences influence both the extent that an individual perceives job insecurity, and the health
consequences of those perceptions. Research has shown that local labour market conditions can
have important consequences for individual and household economic outcomes (Cohen and
Huffman 2003; Heckman et al. 1999). Regional unemployment rates have been shown to
influence the effect of specific instances of job loss on individual health, for example (Turner
1995). Similar regional effects may therefore exist with regard to the health penalties of job
insecurity. High unemployment rates may increase the anxiety that individuals attach to the
threat of job loss if they perceive few alternative employment opportunities in the event of being
laid-off. Conversely, a high unemployment rate may make job insecurity a more normative
experience, and thus less distressing. Additionally, rural-metropolitan differences may shape the
meanings that individuals attach to job insecurity. Job loss in a rural setting may require
individuals to commute further or move in order to find new work for example; causing greater
disruption than for individuals living in metropolitan areas. Further research is necessary to
examine these possibilities.
Finally, since the findings of this study suggest differences in the experience of job insecurity
across social statuses, future research should investigate how other groups of workers may
respond uniquely to the threat of job loss. Certain advantaged groups, such as professional
workers, for example, may not be worried about a lack of job security because they are rich in
human capital and have ‘market power’ (Kalleberg 2011). Here, then, it is important to
distinguish between job insecurity and ‘employment’ insecurity. The health penalties of
109
perceived job insecurity may be low or non-existent among workers who believe they can easily
attain new employment in the event of job loss. As job insecurity becomes an increasingly
common experience for workers, future research should consider whether it is encountered
uniquely across different social statuses and occupations with respect to its health penalties.
110
References Cohen, Philip N, and Matt L. Huffman. 2003. "Occupational Segregation and the Devaluation of
Women's Work Across U.S. Labour Markets." Social Forces 81:881-907.
Christiansen, Shawn L. and Rob Palkovitz. 2001. “Why the ‘Good Provider’ Role Still Matters: Providing as a Form of Paternal Involvement.” Journal of Family Issues 22:84–106.
Elder, Glen H., Jr. 1998. “Life Course Theory and Human Development.” Sociological Analysis 1:1-12.
Fullerton, Andrew S. and Michael Wallace. 2007. “Traversing the Flexible Turn: US Workers’ Perceptions of Job Security, 1977–2002.” Social Science Research 36:201-221.
Heckman, James, Hidehiko Ichimura, Jeffrey Smith, and Petra Todd. 1999. “Characterizing Selection Bias Using Experimental Data.” Econometrica 66:1017-1098.
Hochschild, Arlie. 1989. The Second Shift. New York: Viking.
Kahn, Lawrence M. 2012. “Labour Market Policy: A Comparative View on the Costs and Benefits of Labour Market Flexibility.” Journal of Policy Analysis and Management 31:94-110.
Kalleberg, Arne L. 2009. “Precarious Work, Insecure Workers: Employment Relations in Transition.” American Sociological Review 74:1–22.
______. 2011. Good Jobs, Bad Jobs: The Rise of Polarized and Precarious Employment Systems in the United States, 1970s-2000s. New York: Russell Sage Foundation, American Sociological Association Rose Series in Sociology.
Moen, Phyllis, and Stephen Sweet. 2004. “From ‘Work-Family’ to ‘Flexible Careers’: A Life Course Reframing.” Community, Work and Family 7:209-226.
Pearlin, Leonard I., Morton Lieberman, Elizabeth Menaghan, and Joseph Mullan. 1981. “The Stress Process.” Journal of Health and Social Behavior 22:337–56.
Pearlin, Leonard I. and Marilyn M. Skaff. 1996. “Stress and the Life Course: A Paradigmatic Alliance.” The Gerontologist 36:239-247.
Rousseau, Denise M. 1995. Psychological Contracts in Organizations: Understanding Written and Unwritten Agreements. Thousand Oaks, CA.
Sverke, Magnus, and Johnny Hellgren. 2002. “The Nature of Job Insecurity: Understanding Employment Uncertainty on the Brink of a New Millennium.” Applied Psychology 51:23–42.
Turner, J. B. 1995. “Economic Context and the Health Effects of Unemployment” Journal of Health & Social Behavior 36:213–229.
Vosko, Leah F. 2006. Precarious employment: Towards An Improved Understanding of Labour Market Insecurity. In L. F. Vosko (Ed.), Precarious Employment: Understanding Labour market insecurity in Canada (pp. 1–33). Montreal and Kingston: McGill-Queen’s University Press.
111
Wheaton, Blair. 1999. “Social Stress.” Pp. 277-300 in The Handbook of the Sociology of Mental Health, edited by C.S. Aneshensel and J.C. Phelan. Kluwer Academic/Plenum.
Winslow, Sarah. 2005. “Work-family Conflict, Gender, and Parenthood, 1977-1997.” Journal of Family Issues 26:27–55.
112
Appendix I Predicting the Hazard of Attrition I follow a procedure used by Mirowsky and Ross (2001) to correct for potential attrition related bias. If baseline health influences attrition, unobserved changes in health may also. Controlling for the hazard of attrition corrects for this possibility. The appendix table below presents probit regression analyses with absent from wave 2 regressed on health outcomes, insecurity and controls at wave 1 (model 1). None of the baseline health measures predict attrition from the sample at wave 2. The low pseudo R-square of .063 also shows that attrition is largely random with respect to the model’s baseline measures. In model 2, age-by-insecurity interactions are added to examine whether older workers with job insecurity are more likely to be absent from the sample at wave 2. Both interactions are non-significant, indicating that older workers with job insecurity are equally as likely to present at wave 2 as younger workers. APPENDIX TABLE: Probit Regression of Absent from the Sample at Time 2 on Time 1 Health Outcomes, Perceived Job Insecurity and Age
Model 1 Model 2 Distress .030 .027 Anger -.021 -.018 Self-rated Health -.032 -.033 Sense of control -.026 -.023 Age 40 to 54 .009** .011** Age 55 to 65 -.312** -.311** Perceived Job Insecurity -.014 .080 Employed in Public Sector .004 .004 Work Hours .001 .001 Job Pressures .009 .009 Job Autonomy .090** .090** Married -.295*** -.292*** Children in the Household -.026 -.031 Spouse Works -.013 -.012 Household Income -.000 -.000 Female -.141 -.140 White -.216** -.214** Education -.103*** -.104*** Insecurity × Age Interactions Age 40 to 54 × Job Insecurity ——
-.007 Age 55 to 65 × Job Insecurity ——
.001 Constant .237 .216
R2 .063 .064 * p < .05; ** p < .01; *** p < .001 (two–tailed test).
top related