The reciprocal relationship between job insecurity and depressive symptoms: A latent transition analysis

• DOI: http://doi.org/10.1002/job.2250
• Date: 01 November 2018
• Published date: 01 November 2018
• Author: Tinne Vander Elst,Guy Notelaers,Anders Skogstad

SPECIAL ISSUE ARTICLE

The reciprocal relationship between job insecurity and

depressive symptoms: A latent transition analysis

Tinne Vander Elst

1,2

*|Guy Notelaers

3

*|Anders Skogstad

3

1

Knowledge, Information and Research Center

(KIR), IDEWE, Leuven, Belgium

2

Occupational and Organisational Psychology

and Professional Learning, KU Leuven, Leuven,

Belgium

3

Department of Psychological Science,

University of Bergen, Bergen, Norway

Correspondence

Tinne Vander Elst, Knowledge, Information

and Research Center (KIR), IDEWE,

Interleuvenlaan 58, 3001 Leuven, Belgium; or

Occupational and Organisational Psychology

and Professional Learning, KU Leuven, Leuven,

Belgium.

Email: tinne.vanderelst@idewe.be

Funding information

Rikstrygdeverket, Grant/Award Number:

2004/04102‐013/2095; Kommunenes

Sentralforbund, Grant/Award Number:

044034; Næringslivets Hovedorganisasjon,

Grant/Award Number: S‐2413

Summary

Previous studies on the relationship between job insecurity and depressive symptoms have

mainly focused on the stressor‐to‐strain effect from job insecurity to depressive symptoms, on

rather secure and healthy employees, and on rank‐order relationships. This is not entirely in line

with stress theories suggesting intraindividual and reciprocal relationships between high levels of

stressors and strain. In reply, this study investigated whether high levels of job insecurity were

related to subsequent high levels of depressive symptoms, and vice versa. Cross‐lagged dual pro-

cess latent Markov model analysis with 3‐wave data (time lags of 2 and 3 years) from a represen-

tative sample of the Norwegian working force (N= 2,539) revealed 5 latent states of job

insecurity and 6 latent states of depressive symptoms. As hypothesized, a reciprocal relationship

between the “high job insecurity”state and the “depressed”state was found: Previously highly

job‐insecure employees were more likely to be depressed at the next measurement point

(OR = 42.54), and employees labeled as depressed were more likely to experience high job

insecurity later on (OR = 69.92). This study contributes to stress theory by demonstrating that

stressors and strain may relate differently depending on the level of stressor and strain

experienced.

KEYWORDS

depressive symptoms, dual processlatent Markov model analysis, jobinsecurity, reciprocal

relationships,stress theories

1|INTRODUCTION

Job insecurity, defined as the perceived threat to lose the current job

in the future, is an important work stressor (Heaney, Israel, & House,

1994; Sverke, Hellgren, & Näswall, 2002). It is considered to be a

demanding aspect of the work situation that depletes employees'

energy and decreases their mental and physical health in the long run

(De Witte, Vander Elst, & De Cuyper, 2015). Job insecurity has, for

instance, been related to elevated levels of emotional exhaustion, more

depressive feelings and anxiety, higher blood pressure, and headaches

(for an overview, see De Witte et al., 2015; for meta‐analytical results,

see Cheng & Chan, 2008; Sverke et al., 2002). Although traditionally

scholars have considered job insecurity as a work stressor leading to

decreased health, suggesting a stressor‐to‐strain effect (Taris &

Kompier, 2014), recent studies have advanced a reciprocal relationship

between job insecurity and health (De Cuyper, Mäkikangas, Kinnunen,

Mauno, & Witte, 2012; Vander Elst, Van den Broeck, De Cuyper, &

De Witte, 2014). This corresponds with stress theories applying a

process conceptualization of stress (e.g., Hobfoll, 2001; Lazarus &

Folkman, 1984) in which stressors and strain may continuously

influence each other. We follow this innovative line of research by

investigating the reciprocal relationship between job insecurity and

depressive symptoms. Depressive symptoms reflect concomitants of

a clinical depressive syndrome, including dysphoric mood and affect

(e.g., withdrawal of life interest, lack of motivation, and loss of vital

energy), as well as feelings of hopelessness and futility and other

cognitive and somatic correlates (Derogatis, Lipman, Rickels,

Uhlenhuth, & Covi, 1974).

In addition, we believe there is a misfit between stress theories

(e.g., Hobfoll, 2001; Lazarus & Folkman, 1984) and many empirical

tests on the relationship between job insecurity and health/depressive

symptoms. First, stress theories (e.g., Hobfoll, 2001; Lazarus &

Folkman, 1984) suggest that the relationship between stressors and

strain may be different depending on the perceived level of stress

*

Joint first authors

Received: 19 August 2016 Revised: 18 October 2017 Accepted: 23 October 2017

DOI: 10.1002/job.2250

J Organ Behav. 2018;39:1197–1218. © 2017 John Wiley & Sons, Ltd.wileyonlinelibrary.com/journal/job 1197

and that the relationship is particularly high when high levels of the

stressor (e.g., job insecurity) are experienced. While appraisal theory

advances that events will only lead to strain when an individual per-

ceives a loss or threat (Lazarus, 1999), the primacy of resource loss in

conservation of resources theory (COR theory; Hobfoll, 2001)

suggests that resource loss or threat is disproportionally more salient

—and thus more strongly relates to strain—than positive events. To

our knowledge, no studies have previously taken into consideration

the possibility that job insecurity and strain (e.g., depressive symptoms)

may relate differently depending on the experienced level of job

insecurity and/or strain. In addition, although stress theories suggest

that job insecurity may particularly result in strain upon a certain level

of job insecurity, so for higher levels of job insecurity, most empirical

tests on the job insecurity–depressive symptoms relationship fail to

grasp high levels of job insecurity. Mostly, this relationship was

investigated using average estimates in linear regression models

among rather job‐secure instead of job‐insecure employees, neglecting

individuals scoring high on job insecurity. Moreover, scholars identify-

ing a high job insecurity group often did not use statistical criteria to

distinguish highly job‐insecure employees from other (less job‐

insecure) employees (e.g., Boya, Demiral, Ergor, Akvardar, & De Witte,

2008; Ferrie, Shipley, Stansfeld, & Marmot, 2002; Rugulies, Thielen,

Nygaard, & Diderichsen, 2010), which possibly have led to

biased results.

Additionally, most studies on the job insecurity–strain relationship

have used a between‐person analysis tapping into differences

between individuals and rank‐order relationships. Although these stud-

ies have contributed to our knowledge on the relationship between job

insecurity and strain outcomes (e.g., depressive symptoms), they may

not grasp intraindividual processes over time as described in stress

theories (Hamaker, Kuiper, & Grasman, 2015).

The main goal of the current study is to offer a test of the job

insecurity–depressive symptoms relationship that corresponds with

dynamic theories on the reciprocal stressor–strain relationship

(Hobfoll, 2001; Lazarus & Folkman, 1984) and to overcome the

above disadvantages of previous studies, by (a) testing the reciprocal

relationship between job insecurity and depressive symptoms, (b)

examining whether this relationship may differ for different levels

of job insecurity and depressive symptoms, and (c) testing these rela-

tionships as intraindividual processes over time by means of a

within‐person (instead of a between‐person) research analysis.

Hereby, building on appraisal theory (Lazarus & Folkman, 1984)

and COR theory (Hobfoll, 2001), we particularly predict that high

states of job insecurity are related to future high states of depressive

symptoms, and vice versa. For this purpose, we conducted a cross‐

lagged dual process latent Markov model analysis (LMM analysis;

Vermunt & Magidson, 2009) using three‐wave repeated‐measure-

ment data (with time lags of 2 and 3 years between subsequent

measurements) from a representative sample of 2,539 Norwegian

employees. LMM analysis enables us to discern latent states of high

job insecurity and high depressive symptoms and to investigate their

mutual impact within persons over time. Hence, by means of LMM

analysis, we will study whether highly job‐insecure employees at pre-

vious measurement occasions are more likely to be depressed in the

future, and vice versa.

1.1 |The reciprocal relationship between job

insecurity and depressive symptoms

Stress theories present stress as an ongoing, intraindividual process, in

which stressors and strain may continuously affect each other (Lazarus,

1999; Taris & Kompier, 2014). Following appraisal theory (Lazarus &

Folkman, 1984), situational appraisals, coping, stress complaints, and

emotion may change from situation to situation and over time and

must be regarded as processes. In this respect, Lazarus (1999) stated

that linear stimulus–response formulations are “really much too simple

to reflect the complex events taking place in mind, emotion, and action,

and the multiple directions of cause and effect”(p. 22), pointing at the

existence of reciprocal relationships. Additionally, in his COR theory,

Hobfoll (2001) highlights change in the stress experience, for instance,

by introducing the notion of “loss spirals,”in which continued losses of

resources may have an accelerated effect on further losses. On the

basis of the process conceptualization of stress, we may thus predict

a bidirectional or reciprocal relationship between stressors and strain.

More specifically, we propose that job insecurity and depressive symp-

toms predict each other in either direction: Job insecurity may predict

subsequent depressive symptoms, and simultaneously, depressive

symptoms may predict subsequent job insecurity.

The impact of job insecurity on depressive symptoms can specifi-

cally be understood as follows. Building on appraisal theory (Lazarus &

Folkman, 1984), job insecurity (i.e., a primary threat appraisal) is pre-

dicted to result in increased powerlessness to deal with the threatened

job situation (i.e., a secondary control appraisal): Employees do not

know whether they will lose their job in the future, which may hamper

feelings of being in control over the stressful situation (Vander Elst

et al., 2014). Feelings of decreased control may in turn evoke strain.

Next, COR theory (Hobfoll, 2001) states that individuals try to foster

and gain resources, defined as all objects, conditions, personal charac-

teristics, and energies that are valuable for survival. Although the avail-

ability of resources may stimulate individuals' well‐being, a threat to

resources (e.g., job insecurity as the threat to the current job) may lead

to strain.

Multiple cross‐sectional studies have demonstrated a positive

association between job insecurity and depressive symptoms (e.g.,

Boya et al., 2008; D'Souza, Strazdins, Lim, Broom, & Rodgers, 2003;

Meltzer et al., 2010). Meltzer et al. (2010), for example, demonstrated

that job insecurity was positively related to feelings of depression

among respondents of the third national survey of psychiatric morbid-

ity in Great Britain. Similarly, on the basis of 77 studies, Cheng and

Chan (2008) offered meta‐analytical evidence for the association

between job insecurity and the more general outcome of psychological

health (including depressive symptoms), r= .20; 95% CI [−0.18, −0.28].

The results regarding the effect of job insecurity on depressive symp-

toms deriving from prospective studies are less univocal (e.g., Burgard,

Brand, & House, 2009; Ferrie et al., 2002; Mattiasson, Lindgarde,

Nilsson, & Theorell, 1990; Mohr, 2000; Rugulies, Bultmann, Aust, &

Burr, 2006). Burgard et al. (2009), for instance, found that persistent

perceived job insecurity predicted depressive symptoms among men,

after controlling for prior levels of depressive symptoms, whereas

Mohr (2000) could not establish an effect of job insecurity on depres-

sive symptoms.

1198 VANDER ELST ET AL.