Increasing the probability of finding an interaction in work stress research: A two-wave longitudinal test of the triple-match principle

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Increasing the probability of finding an interactionin work stress research: A two-wave longitudinaltest of the triple-match principle

Sergio Chrisopoulos1*, Maureen F. Dollard1,Anthony H. Winefield1 and Christian Dormann21University of South Australia, Adelaide, Australia2Johannes Gutenberg University Mainz, Mainz, Germany

Research into work stress has attempted to identify job resources that can moderatethe effects of job demands on strain. The recently developed triple-match principle(TMP) proposes that job demands, resources, and strain can be conceptualized as beingcomposed of cognitive, emotional, and physical dimensions. When a psychologicalimbalance is induced by job demands, individuals activate corresponding resources toreduce the effects of the demands. A closer match occurs when the resources areprocessed in the same psychological domain as the demands. The further away from amatch, the less likely an interactive effect will become. Put simply, the likelihood offinding an interactive effect between job demands and job resources is greatest whendemands, resources, and strain are based on qualitatively similar dimensions (i.e.cognitive, emotional, and physical). For example, emotional support from colleagues islikely to buffer the effects of emotional demands on emotional exhaustion. The TMPwas tested in a sample of 179 Australian police officers in a two-wave longitudinal study.The likelihood of finding an interactive effect was related to the degree of matchbetween job demands, job resources, and strain with 33.3% of triple-match interactionssignificant, 22.2% when there was a double-match, and 0.0% when there was no match.These findings lend support to the TMP as a guiding framework, for research, toexplore possible interactive effects in work stress research, and for practice, to informinterventions matching resources to occupational demands, to offset strain.

Research in the specific area of occupational and organizational psychology,

occupational health psychology, identifies two broad components of the work

environment that are related to employee health and well-being: job demands and job

resources (Schaufeli & Bakker, 2004). Job demands refer to the degree to which the

* Correspondence should be addressed to Sergio Chrisopoulos, Work and Stress Research Group, School of Psychology,University of South Australia, City East Campus, GPO Box 2471, Adelaide, SA 5001, Australia(e-mail: [email protected]).

TheBritishPsychologicalSociety

17

Journal of Occupational and Organizational Psychology (2010), 83, 17–37

q 2010 The British Psychological Society

www.bpsjournals.co.uk

DOI:10.1348/096317909X474173

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

work environment contains stimuli that require sustained cognitive, emotional, and/or

physical effort (de Jonge & Dormann, 2006). Meeting these demands requires

considerable effort that is associated with high costs resulting in negative health effects

for the employee. In contrast, job resources refer to those aspects of work that may

reduce job demands, aid in achieving work goals, or stimulate personal growth,

learning, and development (Schaufeli & Bakker, 2004). Job resources can beconceptualized as a kind of energetic reservoir in the work environment that can be

tapped when the individual has to cope with job demands (de Jonge & Dormann, 2006;

see also Hobfoll, 1989, 2002).

In many cases, job demands cannot be reduced, therefore research has attempted to

identify job resources that may reduce the impact of job demands on subsequent strain.

More specifically, recent work stress theories have attempted to find the ideal match

between job demands and job resources. One such theory is the triple-match principle

(TMP; de Jonge & Dormann, 2003, 2006). The present study tests the theoreticalpredictions of the TMP by exploring the relation between cognitive, emotional, and

physical dimensions of job demands, job resources, and employee strain in an Australian

sample of frontline police officers.

Theoretical backgroundThe demand-control model (DC; Karasek, 1990; Karasek & Theorell, 1990) and the

extended demand–control–support model (DCS; Johnson & Hall, 1988) were perhapsthe first work stress theories to propose an interactive effect between job demands and

job resources. According to DCS theory, job control and social support (job resources)

help to moderate the effects of high job demands (operationalized as time pressure and

workload) on subsequent employee strain and well-being. Research exploring the

buffering hypothesis of the DC and DCS models has provided only modest support (for a

review, see de Lange, Taris, Kompier, Houtman, & Bonger, 2003; van der Doef & Maes,

1999). In their review, van der Doef and Maes found that non-supportive studies tended

to use broad conceptualizations of demands and control while supportive studies usedmore specific or focused measures.

More recently, the job demands–resources (JD–R) model (e.g. Demerouti, Bakker,

Nachreiner, & Schaufeli, 2001) proposed that job resources buffer the impact of job

demands on strain. By providing a heuristic framework the JD–R model proposes that

the presence of interactive effects of job demands and job resources depends on the

specific nature of the job. The implication of this is that any job resource that relates

specifically to the type of occupation under consideration could potentially buffer

the effects of job demands on strain. However, while the JD–R model helps to focus theorigin of job resources it does not predict the specific nature of job resources that are

more likely to act as potential buffers as well as the underlying processes.

The triple-match principleAccording to the matching hypothesis (Cohen & Wills, 1985), when the type of available

job resource (e.g. emotional resource) corresponds to the existing job demand (e.g.

emotional demand) then the resource is likely to mitigate the effects of the demand

resulting in less strain (de Jonge, Dormann, & van den Tooren, 2008). This scenariois referred to as a double-match of common kind (de Jonge & Dormann, 2003, 2006).

The concept of a double-match was extended by Frese (1999) who proposed that strain

should also be considered as a source of match in the stressor–resource–strain triad.

18 Sergio Chrisopoulos et al.

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

More specifically, Frese argued that social dysfunction would be most affected by social-

type resources, such as social support, than non-social resources. Put another way,

social resources are more likely to act as a buffer for social dysfunction than for physical

or psychological strain. This match was labelled double-match of an extended kind.

The demand-induced strain compensation model (DISC model; de Jonge & Dormann,

2003, 2006; de Jonge et al., 2008) extends the notion of a double-match of extendedkind to include a match between either demands or resources on the one hand, and

strain on the other. Therefore, social stressors (e.g. interactions with customers) or to

social resources (e.g. support from colleagues) are more likely to be related to social

dysfunction than less social areas of strain.

The central proposition of the DISC model, the TMP (de Jonge & Dormann, 2003,

2006; de Jonge et al., 2008), combines both of these types of matches to predict the

optimal match between job demands, job resources, and job-related strain. The

theoretical underpinning of the DISC model is drawn from the notion of functionalhomeostatic regulation. Applied to organizational settings, when a psychological

imbalance is induced by job demands, self-regulation processes are activated in order to

cope with the demands and to consequently restore some form of balance (Pomaki &

Maes, 2002; Vancouver, 2000). de Jonge et al. (2008) therefore argue that when a

psychological imbalance is induced by job demands, individuals activate functional,

corresponding job resources to reduce the effects of the demand. A match occurs when

external resources provide a similar function to that of an internal resource when

combating stress. In other words, individuals deal with stressors using easily availableinternal resources that are qualitatively similar to the nature of the demand. If these

resources are either depleted or not available then a demand for matching external

resources is created, which may be of similar use. If matching external resources are not

available or if they are depleted then individuals will search for other resources. For

example, dealing with emotionally demanding customers has been shown to be strongly

related to emotional exhaustion (e.g. Dormann & Zapf, 2004; Zapf, Seifert, Schmutte,

Mertini, & Holz, 2001). Emotional demands may best be matched by internal emotional

self-regulation strategies (cf. Grandey, 2000). When the individual lacks these internalresources, external resources such as emotional support from colleagues may also be

useful. When these resources are not available or are depleted, individuals will then

search for other resources (e.g. informational support on how to best deal with

demanding customers). Put simply, the TMP predicts that the strongest interactive

effects of job demands and job resources would be observed when demands, resources,

and strain are based on qualitatively similar psychological dimensions. For example, an

emotional resource is most likely to buffer the effects of emotional demands on

emotional exhaustion.Inherent within the TMP is the multidimensionality of concepts. Job demands, job

resources, and strain are conceptualized as being composed of three psychological

dimensions: cognitive, emotional, and physical. As described by de Jonge and Dormann

(2006): cognitive demands impinge primarily on the brain processes involved in

information processing; emotional demands relate to the effort required to conform to

organizationally desired emotions during interpersonal interactions (cf. Morris &

Feldman, 1996); and physical demands are primarily related to the musculoskeletal

system (i.e. physical aspects of behaviour).Similarly job resources can have three distinct dimensions: cognitive-informational

(e.g. colleagues providing informational support or control at work); emotional (e.g.

colleagues providing emotional support); and physical (e.g. providing instrumental aide).

The triple-match principle 19

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Finally, strains can be cognitive (e.g. professional efficacy is cognitively laden);

emotional (e.g. emotional exhaustion); and physical (e.g. somatic complaints including

head aches). Put together, an interaction is most likely to occur when all three

components (job demands, job resources, and strain) are based on qualitatively similar

domains. This does not exclude the possibility of a non-match interaction as these have

been found throughout the literature. Instead, the TMP poses that the probability offinding an interaction increases as the level of match between demands, resources, and

strain increases (de Jonge & Dormann, 2006).

Empirical studies conducted to test the TMP show that results in general have been

supportive (de Jonge, et al., 2008). More specifically, 8 out of 11 TMP studies reviewed

showed evidence in support of the TMP. The most notable support for the TMP is

provided by de Jonge and Dormann (2006) in their longitudinal study of two groups of

health care workers in The Netherlands. Combining the results of both samples, of the

possible 6 triple-match interactions, 2 were found to be significant (33.33%). There werealso 2 out of 12 possible double-match of common kind interactions (16.7%) and 4 out of

24 double-match of extended kind interactions (16.7%). No non-match interactions

were found.

The present studyThe present study applies the TMP to a longitudinal sample of Australian frontline policeofficers. Previous research into police stressors has focused on identifying specific

aspects of policing that are demanding, including violence in policing (e.g. Kop,

Euwema, & Schaufeli, 1999), critical incidents (e.g. Liberman et al., 2002), and more

generally, occupational and organizational stressors (e.g. Brown & Campbell, 1990;

Burke, 1993; Collins & Gibbs, 2003; Greller, Parsons, & Mitchell, 1997; Hart, Wearing, &

Headey, 1995). Only one study that we are aware of found interactive effects of job

demands and job resources on strain (Kaufmann & Beehr, 1989).

In their cross-sectional study of police officers in a mid-western state in the USA,Kaufmann and Beehr (1989) tested interactive effects of job demands (operationalized

as a combination of under-utilization of skills, quantitative workload, and job future

ambiguity), and social support (tangible, emotional, and instrumental support) on strain

(operationalized as job satisfaction, boredom, workload dissatisfaction, and depression).

Both job demands and strain were parcelled into single-item constructs. Of the seven

possible interactions only one interaction was found between job demands (cognitively

laden), instrumental support (physically laden), and strain (cognitively laden). Hence, a

double-match of extended kind was found. However, the interaction found implieda reverse buffering effect.

The present study therefore extends the work of de Jonge and Dormann (2006) by

exploring the interactive effects of job demands, job resources, and job-related strain in

a two-wave longitudinal sample of Australian police officers. More specifically, we

investigated whether matching psychological dimensions of demands, resources, and

strain (in terms of emotional, cognitive, and physical) results in a greater likelihood of

finding an interactive effect. In addition, we extend the findings by using a different set

of measures to those used in their study with the exception of emotional exhaustion. Byusing a different set of scales we hope to strengthen the previous findings, and rule out

effects due to method effects.

In line with the predictions of the TMP, we generally hypothesize that the likelihood

of finding an interaction between job demands, resources, and strain increases with the

20 Sergio Chrisopoulos et al.

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

number of matching variables. To achieve this, we chose strain variables that reflected

the three psychological dimensions. For emotional strain we chose emotional

exhaustion, the main component of burnout (Maslach, 1993). Cognitive strain was

operationalized as reduced professional efficacy (Schaufeli, Leiter, Maslach, & Jackson,

1996), reflecting a reduced sense of esteem at work. Finally, physical strain was

operationalized as the presence of physical symptoms such as back or neck pain. Usingthese measures for strain, two general hypotheses were posed:

Hypothesis 1: Job resources moderate the relation between job demands and health/well-beingoutcomes (i.e. reduced professional efficacy, emotional exhaustion, and physical complaints).There is a weaker association between job demands and job-related strain for police officerswith high job resources than for officers with low job resources.

Hypothesis 2: A moderating effect is more likely to be found when there is a triple-matchbetween job demands, job resources, and job-related strain, than when there is a double-matchor non-match between demands, resources, and strain.

Method

Sample and procedureThe present study was a two-wave panel survey conducted on Australian frontlinepolice officers employed in the state of Victoria (1-year time lag). A random sample

(N ¼ 3; 000) was drawn from the Police Association members database. The sample

comprised all sergeants and senior sergeants, and 50% of probationary constables,

constables, senior constables, and leading senior constables working in Victorian

metropolitan and regional uniform police stations. For reasons of confidentiality, the

research packs were mailed to each member’s home address by the Police Association.

As researchers did not have access to the mailing list a unique matching code was

generated based on the first three letters of the respondent’s mother’s maiden name andthe respondent’s year of birth. The survey packs were administered on two occasions,

12 months apart, to the same initial sample. At Time 1, 519 officers returned a

completed questionnaire (response rate of 17.3%) and at Time 2, 456 officers responded

(15.2%). The final sample (respondents to both panels) consisted of 179 respondents

(34.5% of the initial sample).

To explore whether attrition may have biased the results, an independent-sample

t test was performed on all study variables comparing respondents to both waves to

those that only responded to the first wave. No significant differences were found.The demographic details of the initial sample were not available to the research team,

thus a comparison could only be performed with information on the Victorian police

force as a whole. According to the Australian Institute of Criminology (2003) constables

(including probationary constables and senior constables) comprised 78% of all sworn

police officers in Victoria (excluding officers above the rank of senior sergeant). This

compares to 36% in the current study. In terms of gender, the proportion of females by

rank, the sample was comparable to the population. Therefore, the present sample was

representative by gender but was overrepresented by sergeants.The final panel sample comprised 161 male and 18 female officers who were aged

between 20 and 64 years (M ¼ 43:4, SD ¼ 7:9) and held the following ranks: constable

(N ¼ 13, 7.3%); senior constable (N ¼ 39, 21.8%); leading senior constable (N ¼ 12,

6.7%); sergeant (N ¼ 89, 49.7%); and senior sergeant (N ¼ 26, 14.5%).

The triple-match principle 21

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

MaterialsJob demands were assessed using the Demand-Induced Strain Questionnaire (DISQ 1.0;

de Jonge et al., 2003), which has been well-validated in several studies (e.g. de Jonge &

Peeters, 2009; van de Ven, Vlerick, & de Jonge, 2008; van den Tooren & de Jonge, 2008).

The three subscales used addressed cognitive, emotional, and physical demands. Items

included: ‘making complex decisions’ (cognitive); ‘dealing with people (i.e. the public,colleagues, or supervisors) who have unrealistic expectations’ (emotional); and

‘performing physically strenuous tasks’ (physical). Respondents were asked to indicate

how often their work required then to undertake each of the tasks rated on a five-point

scale (1 ¼ very rarely=never, 5 ¼ very often=always). Scale reliabilities for the present

study were 0.87 for cognitive demands, 0.83 for emotional demands, and 0.73 for

physical demands.

Job resources were also assessed using the DISQ 1.0 (de Jonge et al., 2003). The

DISQ uses items that assess the cognitive, emotional, and physical resources that closelymatch the respective components of autonomy and support. Cognitive resources were

assessed using five items such as ‘The opportunity to vary complex tasks with simple

tasks’ and ‘Help when solving complex tasks’. Emotional resources were assessed using

five items such as ‘Emotional support from peers and supervisors’ and ‘To be able to

defuse emotionally heated interactions with others’. Finally, physical resources were

assessed using five items such as ‘Help when tackling physically strenuous tasks’ and

‘The opportunity to take a break when things get physically strenuous’. Respondents

were asked to indicate whether they had enough of each resource to allow them toperform their job effectively rated on a five-point scale (1 ¼ very rarely=never,

5 ¼ very often=always). Scale reliabilities for the present study were 0.76 for cognitive

demands, and 0.77 for both the emotional demands and physical demands scales.

Cognitive strain was assessed using the professional efficacy subscale of the Maslach

Burnout Inventory – General Survey (Schaufeli et al., 1996). The subscale consisted of

six items such as ‘I can effectively solve the problems that arise in my work’. All items

were rated on a seven-point scale (0 ¼ never, 6 ¼ always). Low levels of professional

efficacy reflect cognitive strain, therefore to assist in the analyses the items werereversed scored. This way, high scores on this scale reflect high levels of cognitive strain.

The reliability coefficient for cognitive strain in the present study was 0.84 for both

waves of data, and the test–retest coefficient was .63.

Emotional strain was assessed using the emotional exhaustion subscale of the

Maslach Burnout Inventory – General Survey (Schaufeli et al., 1996). The emotional

exhaustion subscale consisted of five items such as ‘I feel emotionally drained from my

work’. All items were rated on a seven-point scale (0 ¼ never, 6 ¼ always). The

reliability coefficient for emotional strain in the present study was 0.93 for both waves ofdata, and the test–retest coefficient was .67.

Physical strain was derived from a well-validated questionnaire on physical health

symptoms (de Jonge, Mulder, & Nijhuis, 1999). The derived scale consisted of three

items: ‘muscle pains/aches’, ‘back or neck pain’, and ‘tiredness’. Respondents were

asked to indicate how often they had suffered from these problems in the last month,

rated on a five-point scale (0 ¼ very rarely=never, 4 ¼ very often=always). The reliability

coefficient for physical strain in the present study was 0.83 at Time 1 and 0.82 at Time 2,

and the test–retest coefficient was .79.In addition to the above items, the demographic variables of gender, age, and rank

(1 ¼ constable, 2 ¼ sergeant) were also included. Descriptive statistics for all study

variables are displayed in Table 1.

22 Sergio Chrisopoulos et al.

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Table

1.Descriptive

statistics

ofstudyvariables(N

¼179)

Variables

Mean

SD1

23

45

67

89

10

11

12

13

14

15

1.Rank

0.65

0.48

–

2.Gender

0.10

0.30

2.28**

–

3.Age

43.35

7.85

.51**

2.33**

–

4.Cognitivedem

ands

4.15

0.53

.22**

2.09

.03

(.87)

5.Cognitiveresources

2.09

0.61

.07

.03

.00

2.26**

(.76)

6.Physicaldem

ands

3.06

0.61

2.16*

2.04

2.14

.28**

2.38**

(.73)

7.Physicalresources

2.19

0.70

.03

.11

.05

2.24**

.74**

2.34**

(.77)

8.Emotionaldem

ands

3.80

0.61

.11

2.09

.04

.53**

2.41**

.33**

2.34**

(.83)

9.Emotionalresources

2.10

0.70

2.11

.20*

2.07

2.13

.69**

2.15

.62**

2.37**

(.77)

10.CognitiveoutcomeT1

1.46

0.86

2.09

2.01

2.11

2.02

2.27**

.25**

2.27**

.12

2.25**

(.84)

11.CognitiveoutcomeT2

1.40

0.84

2.01

2.05

2.08

2.06

2.20*

.23**

2.19*

.12

2.24**

.63**

(.84)

12.Physicaloutcomes

T1

2.09

1.06

2.09

2.07

2.09

.23**

2.41**

.32**

2.31**

.23**

2.35**

.25**

.25**

(.83)

13.Physicaloutcomes

T2

2.19

1.02

.01

.00

.03

.17*

2.22**

.23**

2.11

.19*

2.21**

.21**

.22**

.67**

(.82)

14.EmotionalexhaustionT1

2.68

1.24

2.01

2.05

2.05

.32**

2.49**

.29**

2.40**

.35**

2.45**

.34**

.31**

.71**

.54**

(.93)

15.EmotionalexhaustionT2

2.73

1.30

2.03

2.11

2.07

.32**

2.47**

.28**

2.42**

.36**

2.45**

.37**

.41**

.65**

.54**

.79**

(.93)

Note.Reliability

coefficients

(Cronbach’salpha)

aredisplayedin

parentheses

alongthediagonal.T1,Time1;T2,Time2;*p

,:05;**p,

:01(two-tailed).

The triple-match principle 23

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Analytical strategyWe adopted the same analytical strategy as that of de Jonge and Dormann (2006). All

analyses were conducted with structural equation modelling (LISREL 8; Joreskog &

Sorbom, 1993) to cross the three content domains (cognitive, emotional, and physical)

simultaneously with the three conceptual domains (demands, resources, and strain).

Therefore, the analysis consisted of 12 variables (3 £ demographics, 3 £ demands,3 £ resources, and 3 £ strain). Furthermore, interaction effects between demands and

resources were also included but because of the number of possible interactions

(3 demands £ 3 resources £ 3 strain ¼ 27 possible interactions) we decided to follow

the approach used by de Jonge and Dormann (2006) and split the analysis into two

according to the theoretical assumptions.

The first analysis included all triple-match interactions (e.g. cognitive demands

£ cognitive resources ! cognitive strain) and all double-match of common kind (e.g.

cognitive demands £ cognitive resources ! emotional strain). Therefore, the firstanalysis looked at 9 of the possible 27 interactions.

In the second analysis, we included the double-match of extended kind interactions

(e.g. cognitive demands £ emotional resources ! cognitive strain) and the non-match

interactions where no variable matched (e.g. cognitive demands £ emotional

resources ! physical strain). Therefore in the second analysis, we tested 12 double-

match interactions of extended kind and an additional 6 non-match interactions.

Lagged interaction effects were analysed such that all study variables and

multiplicative interactions at Time 1 were used to predict strains at Time 2. Followingde Jonge and Dormann (2006) the structural models were fully saturated because

residuals among the outcome variables were allowed to correlate. Consequently, we

could not report fit indices as saturated models always have a perfect fit.

Results

Table 2 presents the results obtained from simultaneously testing triple-match effects of

demands–resources–strains and double-match effects of common kind of demands–

resources. Unstandardized coefficients (B), standard errors (SE), T values, and

standardized coefficients (b) are presented in the table. Independent variables are

shown in the first column. The three forms of strain analysed were reduced professional

efficacy (cognitive strain), emotional exhaustion (emotional strain), and physical

symptoms (physical strain). The proportion of variance (R2) explained by the interactive

model was .44 for cognitive strain, .53 for physical strain, and .69 for emotionalexhaustion.

For cognitive strain (reduced professional efficacy) the results showed one

significant interaction effect (p , :05), between cognitive demands and cognitive

resources (a triple-match). This interaction was plotted using the method described by

Aiken and West (1991). Values of the predictor variables were plotted at one standard

deviation above and below the mean to create two regression lines: one for high levels of

job resources and one for low levels of resources. To explore whether the slope of the

regressions lines were significantly different from zero, post hoc analyses wereconducted, as described by Aiken and West (1991). The regression lines representing

the interaction between cognitive demands and cognitive resources in predicting

cognitive strain are shown in Figure 1. The figure shows that cognitive strain was not

associated with cognitive demands at low levels of cognitive resources (21 SD; simple

24 Sergio Chrisopoulos et al.

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Table

2.Lagged

structuralequationmodelsofreducedprofessionalefficacy,em

otionalexhaustion,andphysicalsymptomswithtriple-m

atch

anddouble-m

atch

‘commonkind’interactions(N

¼179)

Reducedprofessionalefficacy

2Emotionalexhaustion2

Physicalsymptoms2

Variable

BSE

Tb

BSE

Tb

BSE

Tb

Controlvariables

Gender

20.01

.17

20.08

0.00

20.26

.20

21.30

20.06

0.25

.19

1.35

0.08

Age

20.01

.01

20.97

20.07

0.00

.01

0.06

0.00

0.01

.04

1.36

0.09

Rank

0.20

.12

1.60

0.12

20.07

.15

20.46

20.02

20.09

.14

20.69

20.05

Dem

ands

andresources

Cognitivedem

ands

20.11

2.06

21.77

20.13

0.10

.08

1.29

0.07

0.04

.07

0.54

0.04

Cognitiveresources

0.08

.09

0.95

0.10

0.06

.10

0.63

0.05

0.13

.09

1.37

0.13

Physicaldem

ands

0.07

.06

1.38

0.01

0.04

.07

0.59

0.03

20.02

.06

20.33

20.02

Physicalresources

0.02

.07

0.32

0.03

20.03

.09

20.35

20.02

0.02

.08

0.27

0.02

Emotionaldem

ands

0.05

.06

0.67

0.05

0.03

.07

0.41

0.02

0.02

.07

0.34

0.02

Emotionalresources

20.06

.07

20.91

20.08

20.10

.09

21.11

20.08

0.00

.08

0.00

0.00

Time1outcom

evariables

Cogn.strain

(reducedefficacy)Time1

0.49

.06

7.95**

0.51

0.17

.07

2.28*

0.11

0.04

.07

0.60

0.04

Emo.exhaustionTime1

0.06

.06

1.11

0.10

0.60

.07

8.84**

0.57

0.12

.06

1.79

0.14

Phys.symptomsTime1

0.01

.07

0.14

0.01

0.17

.08

2.23*

0.14

0.63

.07

8.71**

0.65

Interactioneffects

Cogn.dem

ands£cogn.resources

20.10

.05

21.98*

20.14

T20.18

.06

22.95**

20.16

D0.07

.06

1.26

0.08

DEmo.dem

ands£em

o.resources

0.07

.05

1.47

0.10

D20.03

.06

20.57

20.03

T0.13

.05

2.41*

0.15

DPhys.dem

ands£phys.resources

20.07

.04

21.68

20.11

D0.11

.05

2.20*

0.11

D20.05

.05

20.99

20.06

T

Note.Cogn.,cognitive;em

o.,em

otional;phys.,physical;T,triple-m

atch;D,double-m

atch.*p

,:05;**p,

:01(two-tailed).

The triple-match principle 25

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

slope test t ¼ 20:47, p ¼ ns). In contrast, in the presence of high cognitive resources

(measured at Time 1), levels of cognitive strain at Time 2 decrease with increasing levelsof Time 1 cognitive demands (þ1 SD; simple slope test t ¼ 2:61, p ¼ :01). It must be

noted that the combination of low levels of cognitive demands and high levels of

cognitive resources was associated with higher levels of cognitive strain. At high levels

of demands, however, the reverse occurs, the level of cognitive strain decreases to its

lowest level.

No other triple-match interaction was found. In terms of double-match of common

kind (i.e. where demands and resources match) three interactions were found. First, a

double-match was found between cognitive demands and cognitive resources onemotional exhaustion measured 1 year later. Figure 2 shows that in the presence of high

cognitive resources (þ1 SD) the relationship between cognitive demands and emotional

exhaustion was not significant (simple slope test: t ¼ 0:58, p ¼ ns). In contrast, in the

presence of low cognitive resources, increasing levels of cognitive demands were

associated with higher levels of emotional exhaustion (þ1 SD; simple slope test:

Figure 1. Triple-match interaction between cognitive demands, cognitive resources, and cognitive strain.

Figure 2. Double-match interaction of the common kind between cognitive demands, cognitive

resources, and emotional exhaustion.

26 Sergio Chrisopoulos et al.

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

t ¼ 4:54, p , :01). The next two double-match interactions between physical demands

and physical resources on emotional exhaustion (Figure 3) and between emotional

demands and emotional resources on physical symptoms (Figure 4), resemble reversed

buffering effects. More specifically, at low levels of physical resources (21 SD), high

levels of physical job demands tend to be associated with low levels of emotional

exhaustion (simple slope test: t ¼ 1:87, p , :10). At high levels of physical resources(þ1 SD), there was no relationship between physical job demands and emotional

exhaustion (simple slope test: t ¼ 1:08, p ¼ ns). In terms of the double-match between

emotional demand and emotional resources on physical strain, increasing levels of

emotional demands were associated with high levels of physical strain in the presence of

high levels of emotional resources (þ1 SD; simple slope test: t ¼ 2:52, p , :05). There

was no relationship between emotional demands and physical strains when emotional

resources were low (21 SD; simple slope test: t ¼ 0:40, p ¼ ns).

The next analysis investigated the non-match interactions as well as the interactionsinvolving double-match of extended kind, where strains match job demands or where

strains match job resources. Table 3 shows that five significant interactions were found.

Of these, two were found for cognitive strain; a double-match of extended kind between

cognitive demands and physical resources; and a non-match between emotional

demands and physical resources.

The double-match, illustrated in Figure 5 shows that in the presence of high physical

resources (þ1 SD) cognitive strain decreases with increasing levels of cognitive

demands (simple slope test: t ¼ 2:27, p , :05). No relationship was found betweenphysical demands and cognitive strain when physical resources were low (21 SD;

simple slope test: t ¼ 20:37, p ¼ ns). For the non-match interaction, Figure 6 suggests a

reversed buffering interaction, however, when a slope test was performed neither

regression line was significant (þ1 SD; simple slope test: t ¼ 20:50, p ¼ ns; 21 SD;

t ¼ 20:66, p ¼ ns).

For emotional exhaustion only one significant interaction was found and that was for

a double-match of extended kind between cognitive demands and emotional resources.

Figure 7 shows that the interaction was in the expected direction. More specifically, thecombination of low emotional resources and high cognitive demands were associated

Figure 3. Double-match interaction of the common kind between physical demands, physical

resources, and emotional exhaustion.

The triple-match principle 27

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

with high levels of emotional exhaustion (simple slope test: t ¼ 5:18, p , :01).

In contrast, in the presence of high levels of emotional resources there was no

relationship between emotional exhaustion and cognitive demands (simple slope test:

t ¼ 0:76, p ¼ ns).

Finally, for physical symptoms, two interactions were found. Figure 8 shows a

reversed effect of demands and resources. The combination of high emotional demands

and high cognitive resources were associated with high levels of physical symptoms(simple slope test: t ¼ 2:01, p , :05). Further, in the presence of low cognitive

resources, there was no relationship between emotional demands and physical

symptoms (simple slope test: t ¼ 1:05, p ¼ ns). The second interaction for physical

symptoms was a double-match of extended kind between physical demands and

physical strain (Figure 9). Under conditions of high cognitive resources, there was no

relationship between physical demands and physical strain (simple slope test: t ¼ 0:93,

p ¼ ns). However, at low levels of cognitive resources increasing levels of physical

demands were associated with higher levels of physical strain (simple slope test:t ¼ 3:18, p , :01).

To summarize findings, of the 27 possible interactions we found 9 significant

interactions. Of these, 5 were in the predicted direction, in that the association between

high job demands and adverse health/well-being would be weaker in the presence of

high levels of job resources than when job resources were low (Hypothesis 1). However,

we also found the presence of 4 reverse interactions whereby high levels of job

resources were associated with higher levels of strain, in each case a non-triple-match

scenario.In terms of whether moderating effects were more likely to be found with increasing

level of match between job demands, resources, and strain (Hypothesis 2) we found 1

out of 3 significant triple-match interactions, 1 (and 2 reversed) out of 6 significant

double-match of common kind interactions, 3 out of 12 significant double-match of

extended kind, and 0 (and 2 reversed) out of 6 non-match interactions. Table 4 shows

that when valid interactions were taken into account (i.e. interactions that conform to

general stress theory), the ratio of valid interactions decreased as the match between job

demands, resources, and strain departed from the triple-match.

Figure 4. Double-match interaction of the common kind between emotional demands, emotional

resources, and physical symptoms.

28 Sergio Chrisopoulos et al.

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Table

3.Lagged

structuralequationmodelsofreducedprofessionalefficacy,em

otionalexhaustion,andphysicalsymptomswithnon-m

atch

ordouble-m

atch

‘extended

kind’interactions(N

¼179)

Reducedprofessionalefficacy

2Emotionalexhaustion2

Physicalsymptoms2

Variable

BSE

Tb

BSE

Tb

BSE

Tb

Controlvariables

Gender

20.07

.17

20.43

20.03

20.18

.20

20.86

20.04

0.21

.19

1.10

0.06

Age

20.01

.01

20.96

20.07

0.00

.01

0.16

0.01

0.01

.01

1.20

0.08

Rank

0.22

.12

1.76

0.13

20.04

.14

20.31

20.02

20.09

.13

20.70

20.05

Dem

ands

andresources

Cognitivedem

ands

20.11

.06

21.64

20.12

0.08

.08

0.71

0.04

0.08

.07

1.10

0.08

Cognitiveresources

0.05

.08

0.54

0.06

0.05

.10

0.48

0.04

0.15

.09

1.60

0.15

Physicaldem

ands

0.08

.06

1.35

0.09

0.04

.07

0.53

0.03

20.01

.06

20.24

20.02

Physicalresources

0.08

.07

1.02

0.09

20.03

.09

20.40

20.03

0.02

.08

0.25

0.02

Emotionaldem

ands

0.05

.06

0.87

0.07

0.05

.07

0.66

0.04

20.02

.07

20.34

20.02

Emotionalresources

20.09

.07

21.23

20.11

20.09

.09

21.05

20.07

20.01

.08

20.15

20.01

Time1outcom

evariables

Cogn.strain

(reducedefficacy)Time1

0.49

.06

7.98**

0.50

0.16

.07

2.27*

0.11

0.02

.07

0.30

0.02

Emo.exhaustionTime1

0.08

.06

1.30

0.11

0.58

.07

8.46**

0.56

0.11

.06

1.77

0.14

Phys.symptomsTime1

0.00

.07

0.06

0.00

0.20

.08

2.51*

0.15

0.63

.07

8.62**

0.65

Interactioneffects

Cogn.dem

ands£em

ot.resources

0.04

.07

0.60

0.06

D20.23

.08

22.78**

20.19

D0.03

.08

0.35

0.03

NCogn.dem

ands£phys.resources

20.17

.08

22.15*

20.23

D0.03

.09

0.28

0.02

N0.00

.08

0.01

0.00

DEmo.dem

ands£cogn.resources

20.04

.07

20.57

20.06

D20.13

.09

21.52

20.12

D0.26

.08

3.28**

0.32

NEmo.dem

ands£phys.resources

0.17

.08

2.07*

0.23

N0.11

.10

1.18

0.10

D20.10

.09

21.08

20.11

DPhys.dem

ands£cogn.resources

20.06

.05

21.24

20.10

D0.00

.06

0.03

0.00

N20.15

.06

22.72*

20.20

DPhys.dem

ands£em

o.resources

20.08

.06

21.26

20.09

N0.12

.07

1.79

0.10

D0.07

.07

0.98

0.07

D

Note.Cogn.,cognitive;em

o.,em

otional;phys.,physical;D,double-m

atch;N,non-m

atch;*p

,:05;**p,

:01(two-tailed).

The triple-match principle 29

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Discussion

The TMP (de Jonge & Dormann, 2003, 2006; de Jonge et al., 2008) proposes that the

likelihood of finding an interactive effect of job demands and job resources is greatestwhen demands, resources, and strain are processed in the same psychological domain

(i.e. cognitive, emotional, or physical). The present study provides support for this

notion. Further, the present paper extends the findings of de Jonge and Dormann (2006)

in two ways. First, in their study, two groups of Dutch health care workers were

investigated whereas the present study replicated their findings in a sample of Australian

police officers suggesting that the principle holds in two separate occupations and in

two separate countries providing some support for the generalizability of the theory.

Second, we used a different set of measures (with the exception of emotionalexhaustion) with similar results adding strength to the TMP.

Figure 6. Non-match interaction between emotional demands, physical resources, and cognitive

strain.

Figure 5. Double-match interaction of the extended kind between cognitive demands, physical

resources, and cognitive strain.

30 Sergio Chrisopoulos et al.

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

In terms of the types of interactions found, the current study found one out of three

tested triple-match interactions – the interaction between cognitive demands and

cognitive resources on reduced professional efficacy (cognitive strain). This finding

supports the notion that when tasks are cognitively demanding, the availability of

cognitive resources (both in the form of informational support and cognitive autonomy)

negates the effects of high levels of demands on cognitive strain (low levels of

professional efficacy). Put another way, the combination of high cognitive demands andhigh cognitive resources is associated with increased levels of professional efficacy – a

finding predicted by the active learning hypothesis of the DC model (Karasek &

Theorell, 1990) and the balance principle of the DISC model (de Jonge & Dormann,

2003, 2006; de Jonge et al., 2008). According to the balance principle, optimal

conditions for active learning, growth, creativity, and performance exist when

there is a balanced mixture of high job demands and corresponding job resources

(de Jonge et al., 2008).

Figure 8. Non-match interaction between emotional demands, cognitive resources, and physical symptoms.

Figure 7. Double-match interaction of the extended kind between cognitive demands, emotional

resources, and emotional exhaustion.

The triple-match principle 31

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

In terms of a double-match, 4 out of 18 double-match interactions were found that

were in the predicted direction (i.e. excluding the 2 interactions that were in the

opposite direction to that predicted). Of these, 1 was a double-match of common kind

between cognitive demands, cognitive resources, and emotional exhaustion; and 3 were

double-match of extended kind between cognitive demands, physical resources, andcognitive strain; cognitive demands, emotional resources, and emotional exhaustion; and

physical demands, cognitive resources, and physical symptoms. These findings are in line

with the matching hypotheses proposed by Cohen and Wills (1985) and Frese (1999).

More specifically, strain is mitigated when there is a match between job demands and job

resources, or when strain is matched by either job demands or resources.

It is possible that all triple-match interactions should be significant. A lack of triple-

match especially in the emotional domain was surprising and different from previous

tests of the TMP (de Jonge & Dormann, 2006; de Jonge, Le Blanc, Peeters, & Noordam,2008). In their study, de Jonge and Dormann (2006) found a triple-match between

emotional demands/resources/strain in one sample of health care providers and a triple-

match between physical demands/resources/strain in another. Similarly, de Jonge

et al. (2008) found a triple-match for emotional demands/resources/strain in a cross-

sectional sample of aged care providers. One possible explanation for the difference in

findings is related to the samples examined. It is generally accepted that health care

providers face more emotionally demanding situations than many other occupations.

Figure 9. Double-match interaction of the extended kind between physical demands, cognitive

resources, and physical symptoms.

Table 4. Summary of analyses of interaction effects with different patterns of match

Interaction patternValidinteractions

Reversedinteractions

Testedinteractions

Ratio of validinteractions/interactionstested (%)

Triple-match 1 0 3 33.3Double-match (both kinds) 4 2 18 22.2Double-match (common kind) 1 2 6 16.7Double-match (extended kind) 3 0 12 25.0Non-match 0 2 6 0.0

32 Sergio Chrisopoulos et al.

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Additionally, seeking emotional support when needed could be expected to be

appropriate for health care workers. Police work may be as equally demanding emotionally.

However, in contrast police officers are generally expected to be more ‘macho’ and

therefore may not seek emotional support from colleagues and supervisors. Along these

lines, while emotional resources may be available, the officers may not access them for

fear that doing so may reflect a weakness in their character. This cultural influence maybe the reason why emotional resources did not offset the impact of emotional demands

on emotional resources. The lack of triple-match for policing in the physical domain

may simply be that demands in the physical domain were less pressing for police officers.

In terms of the nature of the interactions, the current study found that while the

pattern was consistent to that expected for high levels of demands, a different picture

emerged at the lower end of the spectrum. For nearly all interactions, at low levels of

demands, high levels of resources were associated with higher levels of strain

compared to low levels of resources. In other words, while the presence of resourcesmitigates the relationship between high demands and level of strain, having too many

resources when demands are low may also be detrimental to officer well-being. One

possible explanation is provided by Warr’s (1987) Vitamin model. In essence, this

model suggests that while the presence of job resources is beneficial, in some cases

excessive levels of resources may be harmful. One example is job autonomy. High

levels of job autonomy may be harmful to employee mental health as it may imply

uncertainty, difficulty in decision making and high responsibility. Combined with low

job demands, high levels of job resources may lead to strains associated with lack ofopportunity to actually utilize resources.

The notion of excessive job resources resulting in more strain may help explain the

4 interactions that were contrary to predictions, otherwise referred to as reverse

buffering or enhancer effects. In the present study, enhancer effects were found for

2 out of 18 double-match interactions and for 2 out of 6 non-match interactions. More

specifically, the combination of high levels of physical demands and high levels of

physical resources were associated with higher levels of emotional exhaustion than

when resources were low. Likewise, the combination of high levels of emotionaldemands and resources were associated with higher levels of physical strain than when

resources were low. Enhancer effects have also been reported in other studies

(e.g. Beehr, 1976; Frese, 1999; Kaufmann & Beehr, 1989). Other suggestions for the

possible mechanisms that could produce enhancer effects include where discussing

the situation with supportive others could result in the perception of matters being

worse than they originally first thought (LaRocco, House, & French, 1980), and in the

situation where the source of the support is also the source of the demand (for a review,

see Kaufmann & Beehr, 1989).Kaufmann and Beehr (1989) argue, in line with stress adaptation theories that

suggest that individuals adapt to environmental stressors by selecting appropriate

strategies, that it is possible that individuals may select informational support when

dealing with job demands and strain that are highly related. A closely related view is

presented by Frese (1999) who suggests that the emotional support component of

social support does not change the objective stress situation, particularly when the

stressor is physical in nature. Applied to the TMP, it could be argued that when a job

resource is available that does not adequately match the psychological processesassociated with the job demand then there is a risk of increasing strain (reverse

buffering). The finding that the only non-match interactions found in the present study

were reverse buffering interactions lends support to this idea.

The triple-match principle 33

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

LimitationsOne limitation of this longitudinal study was the relatively low response rate. One

reason for this is that questionnaires were mailed out by the Police Association to

members on their database. Unfortunately, it could not be determined who was an

active frontline officer and within the scope of the study (i.e. gazetted to a uniform

police station). For example, approximately 10% of responses were from officers whowere not gazetted to a station (excluded from the sample), as assessed by an initial

screening item. Consequently, the reported response rate is an underestimation of the

actual response rate. Furthermore, it has been suggested that Australian police officers

are less likely to respond to questionnaires unless accompanied by letter of introduction

from a senior administrator (Hart et al., 1995). In their study, Hart and colleagues

achieved a similar response rate (23%) in a validation study when the questionnaire was

not accompanied by a letter from a senior police administrator, compared to 53% in a

sample that included a letter of introduction. Our survey was accompanied by a letter ofsupport by the Police Association rather than the police force. In addition, anecdotally it

has been suggested that this particular occupational group has been over-surveyed in

the past few years.

In terms of representativeness, sergeants were overrepresented in the present study.

The result of this may be that the current findings reflect sergeants rather than police

officers as a whole. This may explain why only a triple-match was found for cognitive

demands/resources/strain as sergeants’ roles tend to be more managerial than frontline.

Replication studies with more representative samples are required to establish whetherthis is the case.

The implications are that the current findings may be specific to the present sample.

Further research is required to replicate these findings in more representative samples.

However, it should be noted, however, that the research was concerned with

relationships between variables rather than differences between levels of variables, the

sample was matched from T1 to T2, and the general pattern of findings of this study

reflect those of de Jonge and Dormann (2006), overall providing some credibility to the

stability and the generalizability of the findings.Another limitation of the study is that it relied on self-report outcome measures,

potentially increasing the likelihood of common method bias. This could be reduced in

future research by obtaining more objective measures such as days absent from work or

measures derived from other sources like supervisors or colleagues (cf. de Jonge &

Peeters, 2009). Unfortunately, this level information was not available in the current study.

There are several strengths to the current study worth noting. A methodological

strength was the use of longitudinal modelling which enabled us to analyse lagged

relationships between job demands, resources, and strain. Second, we used structuralequation modelling in which the three different outcome measures were analysed

simultaneously. By regressing each type of strain at Time 2 on all types of strain at Time 1,

we were able to eliminate the conceptual overlap and eliminate chance capitalization

(i.e. reduction of Type I errors). Finally, temporal separation of variables and interaction

findings suggest common method bias may not be such an issue.

Implications and future researchFuture research, especially for policing, may benefit from assessing the psychological

components of more specific job demands that relate specifically to tasks performed by

police. Previous police stress research usually involves a ‘laundry list’ of stressors that

34 Sergio Chrisopoulos et al.

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

are combined into global constructs. By separating these into the three psychological

dimensions, and matching the appropriate resources and strain, the probability of

identifying appropriate buffering resources is increased. For example, corresponden-

ce/paperwork may be matched by more cognitive resources such as informational

support than emotional resources like emotional support. As argued by Cohen and Wills

(1985), the use of global measures that tap the general availability of resources withoutaccessing specific resources is likely to result in main effects without buffering effects.

In terms of the overall findings of the present study, in addition to addressing the

need to match resources to demands and strain in terms of psychological processing,

the present findings highlight the differences in the nature of resources that are relevant

in different occupations. For example, de Jonge and Dormann (2006) found that in aged

care workers, emotional resources were the most relevant resource, possibly due to

nature of their work. In aged care residences, employees are often exposed to residents

that have serious medical conditions and are sometimes exposed to death and dying.In terms of emotional labour, these employees then have to regulate their emotions to

present the organizationally required emotions of a caring caregiver (cf. Morris &

Feldman, 1996). Consequently, emotional resources are crucial in these instances, and

availability and access may be not problematic for these workers.

In contrast, for policing, cognitive resources presented as the most important

resource. The measure used in the current study resembles a combination of

informational support and cognitive autonomy. In the case of policing, officers are often

exposed to situations that require quick decisions (Patterson, 2003). The availability ofappropriate information and the ability to access this information is crucial, especially

when decisions involve life or death situations. This may be through appropriate

training in handing demanding situations or through mentoring programmes whereby

less experienced officers are given access to more experience colleagues or supervisors.

The fact that emotional resources seemed less important due to their lack of

amelioration potential was unexpected. Therefore, additional research is required to

understand occupationally specific dynamics that may occur when resources are

available and required but not accessed due to cultural influences, and the implicationsof this for the TMP. Matching of quantity in addition to quality may also be indicated as an

additional line of research to further explicate reverse buffering findings.

In conclusion, the findings lend support to the TMP as a guiding framework, for

research, to explore possible interactive effects in work stress research, and for practice,

to inform interventions matching resources to occupational demands, to offset strain.

Acknowledgements

This research is funded by the Australian Research Council (ARC grant number: LP0562310) and

Industry partners – the Police Association. We would like to thank Jan de Jonge for his invaluable

advice and input in the preparation of this manuscript.

References

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions.

Newbury Park, CA: Sage.

Australian Institute of Criminology (2003). The composition of Australia’s police services as at

June 2002. Canberra: Author.

The triple-match principle 35

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Beehr, T. A. (1976). Perceived situational moderators of the relationship between subjective role

ambiguity and role strain. Journal of Applied Psychology, 61, 35–40.

Brown, J. M., & Campbell, E. A. (1990). Sources of occupational stress in the police. Work and

Stress, 4, 305–318.

Burke, B. J. (1993). Work–family stress, conflict, coping, and burnout in police officers. Stress

Medicine, 9, 171–180.

Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis. Psychological

Bulletin, 98, 310–357.

Collins, P. A., & Gibbs, A. C. C. (2003). Stress in police officers: A study of the origins, prevalence

and severity of stress-related symptoms within a county police force. Occupational Medicine,

53, 256–264.

de Jonge, J., & Dormann, C. (2003). The DISC model: Demand-induced strain compensation

mechanisms in job stress. In M. F. Dollard, A. H. Winefield, & H. R. Winefield (Eds.),

Occupational stress in the service profession (pp. 43–74). London: Taylor & Francis.

de Jonge, J., & Dormann, C. (2006). Stressors, resources, and strains at work: A longitudinal test of

the triple-match principle. Journal of Applied Psychology, 91, 1359–1374.

de Jonge, J., Dormann, C., & van den Tooren, M. (2008). The demand-induced strain compensation

model: Renewed theoretical considerations and empirical evidence. In K. Naswall, J. Hellgren,

& M. Sverke (Eds.), The individual in the changing working life (pp. 67–87). Oxford:

Cambridge University Press.

de Jonge, J., Dormann, C., van Vegchel, N., von Nordheim, T., Dollard, M. F., & Cotton, S. (2003).

DISQ 1.0: The DISC questionnaire english version 1.0. Utrecht: Utrecht University.

de Jonge, J., Le Blanc, P. M., Peeters, M. C. W., & Noordam, H. (2008). Emotional job demands and

the role of matching job resources: A cross-sectional survey study among health care workers.

International Journal of Nursing Studies, 45(10), 1460–1469.

de Jonge, J., Mulder, M. J. G. P., & Nijhuis, F. J. N. (1999). The incorporation of different demand

concepts in the job demand-control model: Effects on health care professionals. Social Science

and Medicine, 48(9), 1149–1160.

de Jonge, J., & Peeters, M. C. W. (2009). Convergence of self-reports and coworker reports of

counterproductive work behavior: A cross-sectional multi-source survey among health care

workers. International Journal of Nursing Studies, 46, 699–707.

de Lange, A. H., Taris, T. W., Kompier, M. A. J., Houtman, I. L. D., & Bonger, P. M. (2003). The very

best of the millennium: Longitudinal research and the demand–control–(support) model.

Journal of Occupational Health Psychology, 8, 282–305.

Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The job demands–resources

model of burnout. Journal of Applied Psychology, 86, 499–512.

Dormann, C., & Zapf, D. (2004). Customer-related social stressors and burnout. Journal of

Occupational Health Psychology, 9, 61–82.

Frese, M. (1999). Social support as a moderator of the relationship between work stressors and

psychological dysfunctioning: A longitudinal study with objective measures. Journal of

Occupational Health Psychology, 4, 179–192.

Grandey, A. A. (2000). Emotional regulation in the workplace: A new way to conceptualize

emotional labor. Journal of Occupational Health Psychology, 5, 95–110.

Greller, M. M., Parsons, C. K., & Mitchell, D. R. D. (1997). Additive effects and beyond:

Occupational stressors and social buffers in a police organization. In J. C. Quick, L. R. Murphy,

& J. J. Hurrell (Eds.), Stress and wellbeing at work: Assessments and interventions for

occupational mental health (pp. 33–47). Washington, DC: APA.

Hart, P. M., Wearing, A. J., & Headey, B. (1995). Police stress and well-being: Integrating

personality, coping and daily work experiences. Journal of Occupational and Organiz-

ational Psychology, 68, 133–156.

Hobfoll, S. E. (1989). Conservation of resources: A new attempt at conceptualizing stress.

American Psychologist, 44, 513–524.

36 Sergio Chrisopoulos et al.

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Hobfoll, S. E. (2002). Social and psychological resources and adaptation. Review of General

Psychology, 6, 307–324.

Johnson, J. V., & Hall, E. M. (1988). Job strain, work place social support and cardiovascular

disease: A cross-sectional study of a random sample of Swedish population. American Journal

of Public Health, 78, 1336–1342.

Joreskog, K. G., & Sorbom, D. (1993). LISREL 8: User’s Reference Guide. Chicago Scientific

Software International.

Karasek, R. A. (1990). Job demands, job decision latitude, and mental strain: Implications for job

redesign. Administrative Science Quarterly, 24, 285–308.

Karasek, R. A., & Theorell, T. (1990). Healthy work: Stress, productivity and the reconstruction of

working life. New York: Basic Books.

Kaufmann, G. M., & Beehr, T. A. (1989). Occupational stressors, individual strains, and social

supports among police officers. Human Relations, 42, 185–197.

Kop, N., Euwema, M., & Schaufeli, W. (1999). Burnout, job stress and violent behaviour among

Dutch police officers. Work and Stress, 13, 326–340.

LaRocco, J. M., House, J. S., & French, J. R. P. (1980). Social support, occupational stress, and

health. Journal of Health and Social Behavior, 21, 202–218.

Liberman, A. M., Best, S. R., Metzler, T. J., Fagan, J. A., Weiss, D. S., & Marmar, C. R. (2002). Routine

occupational stress and psychological distress in police. Policing: An International Journal of

Police Strategies and Management, 25, 421–439.

Maslach, C. (1993). Burnout: A multidimensional perspective. In W. B. Schaufeli, C. Maslach, &

T. Marek (Eds.), Professional burnout: Recent developments in theory and research

(pp. 19–32). Washington, DC: Taylor & Francis.

Morris, J. A., & Feldman, D. C. (1996). The dimensions, antecedents, and consequences of

emotional labour. Academy of Management Review, 21, 986–1010.

Patterson, G. T. (2003). Examining the effects of coping and social support on work and life stress

among police officers. Journal of Criminal Justice, 31, 215–226.

Pomaki, G., & Maes, S. (2002). Predicting quality of work life: From work conditions to self-

regulation. In E. Gullone & R. A. Cummins (Eds.), The universality of subjective well-being

indicators (pp. 151–173). Dordrecht: Kluwer Academic.

Schaufeli, W. B., & Bakker, A. B. (2004). Job demands, job resources, and their relationship with

burnout and engagement: A multi-sample study. Journal of Organizational Behavior, 25,

293–315.

Schaufeli, W. B., Leiter, M. P., Maslach, C., & Jackson, S. E. (1996). TheMaslach Burnout Inventory –

General Survey test manual. Palo Alto, CA: Consulting Psychologists Press.

Vancouver, J. B. (2000). Self-regulation in organizational settings: A tale of two paradigms.

In M. Boekaerts, P. R. Pintrich, & M. Zeidner (Eds.), Handbook of self-regulation

(pp. 303–341). San Diego, CA: Academic Press.

van de Ven, B., Vlerick, P., & de Jonge, J. (2008). The interplay of job demands, job resources and

cognitive outcomes in informatics. Stress and Health, 24, 375–382.

van den Tooren, M., & de Jonge, J. (2008). Managing job stress in nursing: What kind of resources

do we need? Journal of Advanced Nursing, 63, 75–84.

van der Doef, M., & Maes, S. (1999). The job demand-control (-support) model and psychological

well-being: A review of 20 years of emprical research. Work and Stress, 13, 87–114.

Warr, P. (1987). Work, unemployment and mental health. Oxford: Clarendon Press.

Zapf, D., Seifert, C., Schmutte, B., Mertini, H., & Holz, M. (2001). Emotional work and job stressors

and their effects on burnout. Psychology and Health, 16, 527–545.

Received 4 June 2008; revised version received 26 June 2009

The triple-match principle 37