Quantitative Demands,Burnout,and Allied Outcomes for Indian Police Officers

Theoretical Background and Hypotheses Formulation

For any given task, there is an optimal level of QD (van Veldhoven, 2014) that is necessary to challenge employees and to let them learn and thrive (Bakker et al., 2010). A low QD work context leads to ‘fatigue-like’ states (Nachreiner, 1999) such as reduced vigilance, satiation, and/or boredom. Conversely, when QD are above the optimal level, the employee has to manage tension levels to maintain performance (van Veldhoven, 2014). Following Hockey’s (1997) cognitive-energetical framework for stress and workload management, high QD induce extra efforts and energy for accomplishing work goals and for averting falling performance. This extra activation costs the employee physically and/or psychologically (e.g., in the forms of tiredness and irritability). In short term, the employee may recover (e.g., by taking a break, or performing less exhausting activities). However, sustained activation over a prolonged period makes recovery increasingly difficult. The inadequate recovery, then, starts taking a toll on employee’s vigour leading to an accumulation of fatigue-based symptoms (Eriksen et al., 1999). Accordingly, the long-term effects of sustained activation may be a gradual depletion of employee’s physical and/or mental resources, eventually leading to exhaustion (Bakker & Demerouti, 2007; Demerouti et al., 2001). The employee, alternatively, may resort to withdrawal behaviour (as a self-protective strategy to stop further energy depletion) that goes with a fading of dedication, eventually leading to cynicism.

In line with the above discussion, once the enthusiastic officers may become exhausted, detached from their work, and cynical (while dealing with high QD over a sustained period). Researchers have used the metaphor burnout to describe such psychological syndrome arising out of an awry relationship of employee with their work and a breakdown in adaption (Schaufeli, 2003). Thus, the study hypothesized:

H1:  QD is positively associated with burnout in officers.

As a result of ‘chronic workplace stress that has not been successfully managed’ (WHO, 2019), the burnout syndrome ‘develops gradually but may remain unnoticed for a long time by the individual involved. … is self-perpetuating because of inadequate coping strategies that are associated with the syndrome’ (Schaufeli & Enzmann, 1998, p. 36). The general version of the famous Maslach Burnout Inventory-General Survey (MBI-GS; Maslach et al., 1997) conceptualizes burnout syndrome with three dimensions—exhaustion, cynicism, and inefficacy. Exhaustion refers to the feelings of being overextended and depleted of one’s emotional and physical resources; cynicism refers to a negative, callous, or excessively detached response to various aspects of the job; and inefficacy refers to the feelings of incompetence and lack of achievement in work (Maslach & Leiter, 2007).

Inefficacy, however, is not a core dimension of burnout (Bakker et al., 2004; Green et al., 1991; Shirom, 1989). It may be considered as a consequence of burnout (Koeske & Koeske, 1989; Shirom, 1989), and/or as a reflection of a personality characteristic akin to self-efficacy (Cordes & Dougherty, 1993). As an alternative (to the most widely used MBI) measure of burnout, the Oldenburg Burnout Inventory (OLBI; Demerouti et al., 2003) conceptualizes burnout as a syndrome of work-related negative experiences, including feelings of exhaustion and disengagement from work. OLBI conceptualizes exhaustion as the consequence of intense physical, affective, and cognitive strain and refers disengagement (cf. cynicism) as distancing oneself from one’s work in general, work object, and work content (Demerouti & Bakker, 2008). Based on its negatively and positively framed items, OLBI represents the two ends of a continuum of work-related well-being: exhaustion versus vigour and disengagement versus dedication. Accordingly, OLBI can assess the core dimensions of burnout (by negatively framed items) and work engagement (by positively framed items) as well (González-Romá et al., 2006).

Burnout, which is defined as a ‘persistent, negative, work-related state of mind’ (Schaufeli & Enzmann, 1998, p. 36), diminishes opportunities for positive experiences at work and affects individuals’ health and performance (Schaufeli, 2017; Taris & Schaufeli, 2016). In the same (in relation to this study) policing population, Kumar and Narula (2020) have reported about burnout’s effect on certain positive and negative outcomes. The positive outcomes included job satisfaction (JS), commitment to the workplace (CW), organizational citizenship behaviour directed toward the organization (OCBO), organizational citizenship behaviour directed toward the individual (OCBI), and health. Burnout was associated negatively with these outcomes: large-sized effect on JS, CW, and health; medium-sized on OCBI; and small-sized on OCBO. The negative outcomes included work–family conflict (WFC), feeling to quit (FQ), physical aggression, and verbal aggression. Burnout had a large-sized positive effect on WFC and FQ; very small-sized (practically insignificant) positive effect on physical aggression; and no effect on verbal aggression (cf. Queirós et al., 2013).

Previous research in policing context has also reported the negative association of burnout with health, JS, organizational commitment, and OCB (Baka, 2015; Baruch-Feldman et al., 2002; Burke & Mikkelsen, 2006; Chiu & Tsai, 2006; Cropanzano et al., 2003; Kohan & Mazmanian, 2003; Martinussen et al., 2007; Rothmann, 2002; van Emmerik et al., 2005; Wang et al., 2014). Likewise, prior research has found a positive association between officers’ burnout and WFCs (Burke & Mikkelsen, 2006) and desire to quit the job (Cropanzano et al., 2003; Martinussen et al., 2007). Thus, in line with the hypothesis H1 and the previously reported association of burnout with the outcomes (Kumar & Narula, 2020), this study expected QD’s indirect (via burnout) effect to be negative for positive outcomes (i.e., JS, CW, OCBO, OCBI, and health) and positive for negative outcomes (i.e., WFC and FQ). Accordingly, hypotheses H2 was established as follows:

H2:  QD, via burnout, is negatively associated with (a) JS, (b) CW, (c) OCBO, (d) OCBI, and (e) health and positively associated with (f) WFC and (g) FQ.

Furthermore, the existing research on the link between QD and work-related attitudes, behaviours, and intentions has produced limited and inconsistent findings (van Veldhoven, 2014). Although a moderately high level of QD may challenge workers and trigger learning and skill development (Bakker et al., 2010), chronically high QD may have long-term harmful consequences for employee health, well-being, attitudes, and behaviours (Bowling & Kirkendall, 2012; Bowling et al., 2015). As discussed earlier, Indian police officers tend to face chronically high QD. Thus, this study expected QD’s direct association to be negative with positive outcomes and to be positive with negative outcomes, and established hypotheses H3 as:

H3:  QD is directly and negatively associated with (a) JS, (b) CW, (c) OCBO, (d) OCBI, and (e) health and positively associated with (f) WFC and (g) FQ.

Figure 1 illustrates the research model. Succinctly, the model incorporated burnout as a mediator in the relationship of QD with the outcomes—with an expectation from the link ‘QDburnout’ to be positive; and from both (direct and indirect) ‘QDoutcome’ links to be negative for positive outcomes, and positive for negative outcomes.

OPEN IN VIEWER

Figure 1.

Source: Authors’ own creation.

Methods

Measures and Scales

QD (Cronbach’s α = 0.66) was assessed with five items: four designed for the study and one item adapted from the Copenhagen Psychosocial Questionnaire [COPSOQ II] (Pejtersen et al., 2010). A sample item was ‘does your work require you to often work unpaid overtime’. Participants responded on a 5-point (4 = always to 0 = never) scale. Burnout (α = 0.79) was measured using eight negatively framed items of OLBI (Demerouti et al., 2003). An exhaustion item was ‘during my work, I often feel emotionally drained’ and a disengagement item was ‘sometimes I feel sickened by my work tasks’. Participants responded on a 4-point (3 = strongly agree to 0 = strongly disagree) scale.

JS (α = 0.84) and CW (α = 0.64) were respectively measured using four items and two items adapted from COPSOQ II (Pejtersen et al., 2010). A sample JS item was ‘regarding your work in general, how pleased you are with the overall working environment’, and the participants responded on a 5-point (0 = very satisfied to 4 = very dissatisfied) scale. A sample CW item was ‘to what extent do you think that your place of work is of great importance to you’, and the participants responded on a 5-point (4 = to a very large extent to 0 = to a very small extent) scale.

OPEN IN VIEWER

Table 1.

Sociodemographic Characteristics of the Sample (n = 1,223)

Baseline Characteristic	f	Percentage
Designation
Constable	419	34.3
Head Constable	349	28.5
Assistant sub-inspector	296	24.2
Sub-inspector	072	05.9
Inspector	015	01.2
Gender
Male	874	71.5
Female	261	21.3
Category#
Scheduled Caste (SC)	196	16.0
Backward Class (BC)	295	24.1
General (unreserved)	620	50.7
Job experience
≤15 years	579	47.4
>15 years	423	34.5
Age
≤35 years	566	46.3
>35 and ≤50 years	459	37.5
>50 years	120	09.8
Educational qualification
Matriculation	248	20.3
Senior secondary	333	27.2
Graduation and higher	557	45.5
Marital status
Married	896	73.3
Unmarried	226	18.5

Source: Authors’ own compilation.

Notes: Percentages do not add to 100 (and f not to 1,223) because of missingness.

^# In India, category is a caste-based social stratification. Haryana government has provision for 20% reservation to Scheduled Caste (SC) candidates and 27% to Backward Class (BC) candidates in state police force.

OCBO (α = 0.61) and OCBI (α = 0.78) were measured using four items adopted from Lee and Allen’s (2002) study. A sample OCBO item was ‘how often you offer ideas to improve the functioning of the police department’ and OCBI item was ‘do you willingly give your time to help others who have work-related problems’. Participants responded on a 5-point (4 = always to 0 = never) scale.

Health (α = 0.85) was measured using the Patient-Reported Outcomes Measurement Information System (PROMIS^®), global physical health (GPH-2), and global mental (GMH-2) health scales (Hays et al., 2017). A GPH-2 item was ‘in general, how would you rate your physical health’ and GMH-2 item was ‘in general, how would you rate your satisfaction with your social activities and relationships’. Participants responded on a 5-point (4 = excellent to 0 = poor) scale.

WFC (α = 0.83) was measured by adapting a 4-item scale from COPSOQ II (Pejtersen et al., 2010). An item was ‘do your friends or family tell you that you work too much’. Using a 4-point scale, participants responded to three items with anchors ranging from 3 (yes, certainly) to 0 (no, not at all) and to one item with anchors ranging from 3 (yes, often) to 0 (no, never). A 5-item scale (α = 0.72) designed for the study was used to assess FQ. An item was ‘to what extent, do you think you are not in the job/profession of your dream’. Using a 5-point scale, participants responded to two items with anchors ranging from 4 (to a very large extent) to 0 (to a very small extent); to two items with anchors ranging from 4 (yes, certainly) to (no, not at all); and to one item with anchors ranging from 4 (always) to 0 (never).

Although most of the items/scales were adapted from previously validated and reliable scales, the same were re-assessed in reference to the Indian context (Kumar, 2020; Kumar & Narula, in press). The negatively framed items were reverse-coded such that the higher the score, the higher the level of the phenomenon. All the scales (except QD [α = 0.66], CW [α = 0.64], and OCBO [α = 0.61]) had high reliabilities. Although the generally agreed upon lower limit for coefficient α is 0.70, ‘there are limited grounds for adopting such a heuristic’ (Taber, 2018, p. 1288). Alpha values below 0.70 are realistically expectable when dealing with psychological constructs (Kline, 1999), and values may be acceptable down to 0.60 (Hair et al., 2014). Schmitt (1996) has also concluded ‘there is no sacred level of acceptable or unacceptable level of alpha. In some cases, measures with (by conventional standards) low levels of alpha may still be quite useful’ (p. 353). Overall, the scales used in this study seemed sufficiently reliable.

Results

Table 2 presents the intercorrelations, and Table 3 summarizes the results of hypothesis tests. As expected in hypothesis H1, QD was positively associated (0.537, p < 0.001) with burnout (R² = 0.288, p < 0.001). Furthermore, all the seven models (see Table 3) predicting the outcomes were significant (p < 0.001), with effect size ranging from very large ( f² = 1.151 for FQ) to medium ( f² = 0.093 for OCBO). As Table 3 presents, indirect effect of QD was significant for all the seven outcomes (95% CI did not straddle zero) in the expected direction, whereas the direct effect was significant (p < 0.001) in the expected direction for only three (i.e., JS, WFC, and FQ) out of the seven outcomes. Thus, the results supported hypotheses H2 (H2a to H2g), H3a, H3f, and H3g. Both, indirect and direct, effects of QD for JS (β_indirect = -0.224, β_direct = -0.112), WFC (β_indirect = 0.213, β_direct = 0.325), and FQ (β_indirect = 0.347, β_direct = 0.140) were in the same direction (suggesting partial mediation of burnout in QD’s relationship with JS, WFC, and FQ).

OPEN IN VIEWER

Table 2.

Means, Standard Deviations, and Intercorrelations for Study Variables

Variable	M	SD	1	2	3	4	5	6	7	8	9
1. QD	2.31	0.49		0.523***	–0.344***	–0.241***	0.053ǂ	0.072*	0.211***	0.536***	0.497***
2. Burnout	0.94	0.39	0.537***		–0.480***	–0.484***	–0.217***	–0.322***	–0.436***	0.582***	0.719***
3. JS	2.12	0.79	–0.336***	–0.477***		0.525***	0.354***	0.292***	0.577***	–0.476***	–0.602***
4. CW	2.88	0.75	–0.233***	–0.472***	0.511***		0.629***	0.490***	0.430***	–0.332***	–0.699***
5. OCBO	2.94	0.46	0.059*	–0.209***	0.344***	0.624***		0.755***	0.395***	–0.094**	–0.341***
6. OCBI	3.12	0.60	0.057*	–0.307***	0.287***	0.484***	0.758***		0.327***	–0.098**	–0.275***
7. Health	2.03	0.79	–0.205***	–0.441***	0.572***	0.418***	0.383***	0.328***		–0.431***	–0.503***
8. WFC	2.03	0.70	0.538***	0.571***	–0.459***	–0.309***	–0.074**	–0.080**	–0.420***		0.616***
9. FQ	0.93	0.88	0.487***	0.722***	–0.587***	–0.691***	–0.334***	–0.278***	–0.493***	0.604***

Source: Authors’ own compilation.

Notes: Values above/below the diagonal are intercorrelations before/after adjusting for participants’ sociodemographic characteristics (i.e., age, gender, category, marital status, educational qualification, designation, and job experience).

QD = Quantitative demands, JS = job satisfactions, CW = commitment to the workplace, OCBO = organizational citizenship behaviour directed toward the organization, OCBI = organizational citizenship behaviour directed toward the individual, WFC = work–family conflict, FQ = feeling to quit.

^***p < 0.001, ^**p < 0.01, ^*p < 0.05, and ^ǂ p < 0.10 (2-tailed).

For four (i.e., CW, OCBO, OCBI, and health) of the outcomes, QD’s direct effect was in opposite direction (i.e., positive) to that of the indirect effect (see Table 3). These results pointed out inconsistent (cf. Baron & Kenny, 1986) mediation where the mediator burnout acted more like a suppressor variable (Fritz & MacKinnon, 2008; Kenny, 2018; Mackinnon et al., 2000). In the full model (when controlled for burnout), QD displayed a reverse (i.e., from -r to + β, albeit insignificant) effect for CW and health and a stronger (i.e., β > r) effect for OCBO and OCBI. Besides, burnout displayed a stronger (i.e., β > r) effect for the mentioned outcomes.

OPEN IN VIEWER

Table 3.

Results of Regression Analyses for Burnout and the Outcomes

		Burnout	JS	CW	OCBO	OCBI	Health	WFC	FQ
Direct effect
QD	β	0.537***	–0.112***	0.028ns	0.240***	0.312***	0.045ns	0.325***	0.140***
	SEHC4	0.025	0.030	0.032	0.033	0.029	0.029	0.029	0.024
	t	21.908	–3.685	0.903	7.308	10.695	1.537	11.393	5.799
Burnout	β		–0.417***	–0.487***	–0.338***	–0.474***	–0.465***	0.397***	0.647***
	SEHC4		0.030	0.031	0.033	0.032	0.031	0.028	0.023
	t		–13.782	–15.872	–10.257	–15.010	–14.819	14.064	28.329
	R 2	0.288***	0.237***	0.223***	0.085***	0.163***	0.196***	0.401***	0.535***
	F (HC4)	479.943	179.968	167.406	55.445	118.658	129.949	452.699	775.568
Cohen’s f2		0.404	0.311	0.287	0.093	0.195	0.244	0.669	1.151
Indirect effect of QD
	β		–0.224	–0.262	–0.181	–0.255	–0.250	0.213	0.347
	SEboot		0.019	0.020	0.019	0.021	0.021	0.018	0.019
95% PCI	Lower limit		–0.262	–0.303	–0.221	–0.297	–0.291	0.180	0.310
	Upper limit		–0.187	–0.223	–0.144	–0.215	–0.211	0.248	0.386
Percentage of total effect			66.67%	Ȉ–m	Ȉ–m	Ȉ–m	Ȉ–m	39.59%	71.25%

Source: Authors’ own compilation.

Notes: HC4 = heteroscedasticity-consistent standard error (Cribari-Neto), PCI = percentile confidence interval (from 10,000 bootstrap samples).

QD = quantitative demands, JS = job satisfactions, CW = commitment to the workplace, OCBO = organizational citizenship behaviour directed toward the organization, OCBI = organizational citizenship behaviour directed toward the individual, WFC = work–family conflict, FQ = feeling to quit.

Ȉ-m = inconsistent mediation (opposite sign of direct and indirect effects).

^***p < 0.001 and ^ns p > 0.10.

Discussion

This study pertained to the police force of an Indian state, Haryana. The police force of the state (designated as Haryana Police) is struggling with an acute shortage of officers: on an average, only 39,818 officers (vs. the sanctioned strength of 54,204 officers) were available on ground for the years 2001–2018 (Kumar, 2020). The understaffing, coupled with the increasing crime rate (that recorded a compound annual growth rate of 4.82% from the year 2001 to 2018; Kumar, 2020), has made the police force incessantly overburdened and quantitatively demanding. This study empirically examined the ramifications of QD for burnout experience and allied outcomes for the officers.

In line with earlier studies in policing context (Wolter et al., 2019), the study revealed QD’s positive association with burnout, signifying that the officers who take their job quantitatively demanding are more likely to experience burnout. As expected in hypotheses H2a, H2f, and H2g, the study found burnout responsible for carrying over two-third of QD’s negative effect on officers’ JS, about two-fifth of the positive effect on WFC, and slightly less than three-fourth of the positive effect on FQ. These findings were parallel to the findings of earlier studies vis-à-vis mediation of burnout in job demands’ relationships with JS (Baruch-Feldman et al., 2002; Martinussen et al., 2007; Rothmann, 2002; Wang et al., 2014), with WFC (Hall et al., 2010), and with turnover intention (Schaufeli & Bakker, 2004).

Furthermore, the findings revealed officers’ burnout to be the sole factor responsible for QD’s negative effect on their CW, extra-role behaviours (OCBO and OCBI), and experience of health. More specifically, QD of policing job associate undesirably with the mentioned outcomes only via officers’ burnout, though not directly. The findings, albeit surprising, were in line with the notion that ‘job demands are not necessarily negative’ (Schaufeli & Bakker, 2004, p. 296). A plausible explanation for the ‘not-undesirable’ effect of QD may be attributed to the conservation of resources (COR; Hobfoll, 1989) theory, and social exchange theory (SET).

When confronted with high QD, an officer tends to follow a performance protection strategy (Hockey, 1997) and puts in extra effort and energy. In due course, the officer is likely to experience resource loss threat or actual resource loss (Chen et al., 2015; Hobfoll, 1989, 2002; Hobfoll et al., 2015). As resource gain becomes more salient (Hobfoll, 2002) in situations of resource loss, the officer may strive for a resourceful environment (to offset resource loss) under demanding situations. SET argues that reciprocal interdependences between parties generate obligations to repay (Cropanzano & Mitchell, 2005; Emerson, 1976)—in plain words, when ‘A’ helps ‘B’, ‘B’ feels obliged and tends to repay to ‘A’. Thus, with a theoretical rationale in SET, the officer is more likely to engage himself/herself in positive behaviours, such as in extra role helping behaviours (OCBO and OCBI), to supplement his/her performance protection strategy and resource gain efforts. Furthermore, the officer is less likely to let the QD adversely affect his/her CW and health. Apparently, this will be more conducive to his/her performance protection strategy, efforts to generate obligations to repay, and resource gain efforts.

Thus, the direct effect of QD on OCBO, OCBI, CW, and health seems explicable. However, an ‘unwanted’ variance shared by burnout caused the indirect effect of QD (for OCBO, OCBI, CW, and health) to be negative. The opposite nature of direct and indirect effects, by cancelling each other, made the total effect of QD to be less than the direct effect (for OCBO and OCBI) and to be in opposite direction to that of the direct effect (for CW and health). In other words, when burnout (in MLR) took away the ‘unwanted’ shared variance, QD’s positive effect for OCBO and OCBI emerged stronger, and negative effect for CW and health disappeared.

Practical Implications

This study potentially contributes to knowledge, especially, in understanding how and when QD, apparently negative, may have beneficial effects. In addition, the study findings may have key practical implications for policymakers and police administrators. Firstly, the study highlighted QD as a potential target for intervention. QD are largely controllable by structural interventions (e.g., by adopting scientific human resource planning and by increasing the supply of officers). Thus, the foremost practical suggestion in this regard is to systematize the human resource planning process to replace the existing ad-hoc or emergent process primarily dependent on the whim of the government (Model Police Manual, Vol. 3, n.d.).

Secondly, the study findings can be of value for administrators to comprehend the pathways by which QD adversely affects the ‘outcomes’. In this regard, the study identified burnout as the venue for targeted interventions. Interventions, such as counselling, stress management training sessions, and so forth, may address burnout. Reduced burnout, then, may alleviate the detrimental impacts of QD. However, as Maria et al. (2018) have discussed, a review of existing stress management interventions (Patterson et al., 2014) found only small-sized effect, revealing that the existing interventions were not so effective. Thus, future research is needed to develop effective interventions to address burnout (Maria et al., 2018).

Limitations and Future Directions

Although this study is novel in the context of Indian policing, its potential limitations necessitate for a caveat while interpreting the results. Firstly, the study used non-probability sampling, which may limit the generalizability of the findings. However, the data were collected with due care for neutrality and control for bias; besides, the sample was large and plausibly representative. Secondly, the self-reported data and cross-sectional design may raise concerns about common method variance and conclusions drawn about causality. Although common method variance was not problematic, causal inferences may be arguable. The observed associations were consistent with previous works on burnout and allied issues; however, it is likely that reverse may be the case. For instance, officers who experience burnout may take their work more quantitatively demanding, and/or officers having, say, lesser levels of JS may develop burnout feeling. Thus, future research in Indian context might use longitudinal designs for more conclusive inferences about the causal relationships.

Thirdly, perhaps the risk of ‘omitted variables’ bias remained in the study. There may be other potentially demanding aspects (such as emotional demands, cognitive demands, role ambiguity and conflict, and political interference) and positive aspects (such as support, trust, justice, and reward and recognition programmes) that may be included in future research. Besides, the role of individual differences (such as personality variables and/or personal resources) in the model may be considered.

Finally, the study results need a caveat for generalization to populations in different contexts, particularly those other than policing context. With regard to the applicability of the results to police work, regional, socio-economic, and cultural differences need to be considered; for instance, despite the similarities in the laws and acts governing police structure and functions across Indian states, it is likely that perceptions about QD and their impacts differ for officers in states diverse from Haryana on socio-economic and cultural fronts, and/or in crime rate. Taken together, this study results are applicable to the police work in Haryana and appear to be applicable, but with caveats, to the police work in other Indian states. Thus, future research may examine the model in different contexts, such as in states with lower/higher crime rate (vs. Haryana), in socio-economically and culturally different (vs. Haryana) states, and in other occupations (such as lawyers, judges, forensic laboratory staff, and doctors) associated with law enforcement set-up.

Author Note

This article is based on the data collected and prepared for a doctoral dissertation by Anil Kumar. Portions of the findings were presented at the 5th International Management Conference, Advances in Management through Research, Innovation & Technology (AMRIT), organized by Fortune Institute of International Business, New Delhi, India. We have no conflict of interest to disclose.