Received: 29 November 2017 Revised: 3 July 2018 Accepted: 5 July 2018

DOI: 10.1111/1748-8583.12204

ORIGINAL ARTICLE

Are all aspects of lean production bad for workers?

An analysis of how problem‐solving demands
affect employee well‐being
Meng‐Long Huo1 | Peter Boxall2

1
Taylor's Business School, Taylor's University,
Subang Jaya, Selangor, Malaysia
2
Department of Management and
International Business, The University of
Auckland, Auckland, New Zealand
Correspondence
Meng‐Long Huo, Taylor's Business School,
Taylor's University, 1 Jalan Taylors, Subang
Jaya, Selangor 47500, Malaysia.
Email: m.huo@auckland.ac.nz
Abstract
This study is concerned with the debate around employee
well‐being in the environment of lean production. It applies
the job demands–resources model to examine the effects of
problem‐solving demands and job resources (training, par-
ticipation in decision‐making, and line manager support) on
employee engagement and exhaustion in a Chinese manu-
facturer. It examines previously untested interactions and
shows that these job resources created a “buffering effect”
in the relationship between problem‐solving demands and
exhaustion. It also shows a “coping effect” because the rela-
tionship between resources and engagement was strength-
ened as problem‐solving demands increased. Rather than
being uniformly positive or negative, the results suggest
that the overall impact of lean production on worker well‐
being is likely to depend on the ways in which managers
create scope for worker involvement in decision‐making,
target resources to the specific job demands, and adjust
resource levels to the degree of these demands.

KEYWORDS

employee well‐being, exhaustion, job demands–resources model,

lean production, problem‐solving demands, work engagement

1 | I N T RO DU CT I O N
The implementation of lean production is extensive in the world's manufacturing plants, but researchers remain
concerned that working conditions in lean settings lead to negative employee outcomes (e.g., Cullinane, Bosak,
Flood, & Demerouti, 2014; Delbridge, 2007; Vidal, 2007a). This includes a concern with work intensification and
its impacts on stress and also relates to the nature of the demands brought about by the typical lean implementation.
Vidal (2007a) argues that the lean principle of continuous improvement leads to an increased requirement for

Hum Resour Manag J. 2018;28:569–584. wileyonlinelibrary.com/journal/hrmj © 2018 John Wiley & Sons Ltd 569
570 HUO AND BOXALL

problem solving by employees. Similarly, Jackson and Mullarkey (2000) observe that problem‐solving demands are
particularly high for lean workers as a result of the elimination of buffer stocks—a fundamental principle of lean
production (Womack, Jones, & Roos, 1990). Under such circumstances, production output is more vulnerable to
fluctuation according to the quality of raw materials and variation in worker performance and machine efficiency.
Therefore, in order to maintain the stability of production flow, shop‐floor workers are required to shoulder greater
responsibility for anticipating and preventing problems that could interrupt production (Jackson & Mullarkey, 2000).
As argued by MacDuffie (1995), “workers must have both a conceptual grasp of the production process and the
analytical skills to identify the root cause of problems” because they are required in the lean production environment
to “identify and resolve problems as they appear on the line” (p. 201).
Prior research has generated mixed results on the relationship between problem‐solving demands and employee
well‐being. Outside the lean work environment, Schmitt, Zacher, and Frese (2012) have shown a connection between
problem‐solving demands and increased fatigue, whereas others have found nonsignificant results in terms of job sat-
isfaction, strain, anxiety, and depression (e.g., Holman, 2002). Inside the lean context, two contrasting perspectives
have emerged from previous studies. One proposes an inverse relationship between problem‐solving demands and
positive employee outcomes because dealing with this type of demand costs effort and consumes energy (Bouville
& Alis, 2014; Cullinane et al., 2014; Vidal, 2007b). For example, in a survey of machinists working in lean and nonlean
contexts in a British clothing manufacturer, Jackson and Mullarkey (2000) found that the level of problem‐solving
demand was significantly higher in lean teams and was related to greater job anxiety and depression. In contrast,
Cullinane, Bosak, Flood, and Demerouti (2013) develop the view that problem‐solving demands in lean production
are positive challenges that can be motivational for employees. However, no quantitative studies have tested this
proposition. This leaves a major gap in our knowledge regarding what kind of impact problem‐solving demands are
having on employee well‐being. Are they as bad as critics suggest or a positive source of motivation? If the process
of implementing lean methods is unstoppable and a high level of problem‐solving demand is inherent in such environ-
ments, how can management improve the outcomes for workers?
Located within the debate on the impact of lean demands on employee well‐being, we address these questions
in this paper, using survey data from a Chinese lean manufacturer of transportation equipment. Existing studies in
China, which is the world's largest contributor to manufacturing output, show evidence of a negative lean implemen-
tation‐to‐well‐being relationship (e.g., Chan, Chen, Xie, Wei, & Walker, 2014; Zhang, 2015). In order to enhance
worker outcomes in China, we need studies of the types of job resources that may counterbalance the lean job
demands being faced by the workforce. Our study contributes to this objective.
The paper's analysis is anchored in the job demands–resources (JD‐R) model (Demerouti, Bakker, Nachreiner, &
Schaufeli, 2001), which enables us to examine the factors that predict employee engagement and exhaustion and to
examine the interactions among them. In applying the JD‐R framework to the lean environment, we used a set of
workplace interviews to identify the prominent job characteristics in our specific setting and, thus, to customise
our hypotheses. Overall, the study contributes to the literature in two respects. First, empirical evidence is identified
for the first time to support Cullinane et al.'s (2013) prediction that problem‐solving demands under lean settings
have motivational properties, and our study advances this work by detecting complementary job resources necessary
to alleviate the health‐impairing potential of problem‐solving demands. Our findings support the argument that
research should move away from the idea that lean production, per se, has negative or positive well‐being implica-
tions and towards the question of how to implement it in ways that enhance its positive potentials and reduce its
negative possibilities (Hasle, Bojesen, Langaa Jensen, & Bramming, 2012). Second, we confirm previously untested
interaction effects between job resources and problem‐solving demands, which indicate how resources can enhance
the benefits of problem‐solving demands for workers in lean production. This result suggests that management has
valuable options to improve the impact of lean production on employee well‐being.
The paper is structured as follows. We begin with a review of the relevant literature and develop our hypotheses
through the theoretical framework of the JD‐R model. The research context and methods are then reported. This
leads to a description of the analytical procedures and results, followed by our discussion and conclusions.
HUO AND BOXALL 571

2 | E X A M I N I N G T H E E F F E C T O F P RO B L E M ‐ S O L V I N G D E M A N D S O N
W E L L‐ B E I N G U N D E R LE A N SE T T I N G S

Shah and Ward (2007, p. 791) refer to lean production as “an integrated socio‐technical system whose main objective
is to eliminate waste by concurrently reducing or minimizing supplier, customer, and internal variability.” Their
discussion emphasises two dimensions of lean production. On one level, lean production is a managerial philosophy
specifying guiding principles and goals for production, such as continuous improvement and waste elimination. On
the other, lean production encompasses a set of operating tools and techniques that can be directly observed, such
as those associated with total quality management, just‐in‐time inventory control, and work standardisation (Shah &
Ward, 2003). Although there are continuities with mass production, lean methods imply a shift in work practices and
in what is expected of the workforce. In a quantitative comparison of lean plants with traditional plants, Forza (1996)
found that the “lean production plants seem to use more teams for problem solving, to take employees' suggestions
more seriously, to rely more heavily on quality feedback both for workers and supervisors, to document production
procedures more carefully and to have employees able to perform a greater variety of tasks including statistical pro-
cess control” (p. 59). This was also demonstrated in a qualitative analysis of employee learning in an implementation
of lean production in which newly constituted work teams were trained in lean techniques and expected to engage in
problem solving to meet more ambitious work targets (Sterling & Boxall, 2013).
How the demands associated with lean implementations affect worker well‐being remains controversial (e.g.,
Bamber, Stanton, Bartram, & Ballardie, 2014). This includes the impact of problem‐solving demands. On the positive
side, Womack et al. (1990), whose landmark book helped to propagate lean production, consider them beneficial for
employee motivation due to their positive association with skill variety and skill utilisation. Cullinane et al. (2013) also
present a positive perspective, arguing that problem‐solving demands may function as challenge demands, promoting
employee work engagement. This type of work demand presents employees with opportunities to learn, achieve, and
display the kind of competence that tends to earn rewards (Crawford, LePine, & Rich, 2010).
Various studies by critics, however, identify problem‐solving demands as a detrimental lean job characteristic lead-
ing to negative well‐being outcomes in terms of increased job anxiety and depression (Jackson & Mullarkey, 2000) and
deteriorating health Bouville and Alis (2014). They argue that decentralisation of problem‐solving responsibilities is
likely to stress employees, particularly when they are not ready or willing to embrace these tasks (Vidal, 2007b).
In terms of employee outcomes, which of these competing positions is more credible? Or does the reality for
workers lie somewhere in between the two extremes, depending on the specifics of the implementation (Hasle
et al., 2012; Vidal, 2007a)? The study we report in this paper addresses this controversy.

3 | T HE J O B D E M A N DS – RE SO UR CE S M OD E L AN D T H E WE LL ‐ B EI NG
I M P L I C A T I O N S O F P R O B L E M ‐ S O L V I N G DE M A N D S

In the JD‐R framework, job resources promote employee work engagement (i.e., the motivational process), whereas
job demands predict exhaustion (i.e., the health‐impairing process) (Schaufeli & Taris, 2014). Job demands are the
“physical, social, or organizational aspects of the job that consume physical or mental effort and therefore induce
certain physiological and psychological costs (e.g. exhaustion)” (Demerouti et al., 2001, p. 501). Job resources refer
to “physical, psychological, social, or organizational aspects of the job that may do any of the following: 1) be
instrumental in accomplishing work goals; 2) reduce job demands; 3) foster individual growth and development”
(Demerouti et al., 2001, p. 501).
Recently, the JD‐R model has been revised to make a distinction between two types of job demand: job
hindrances and job challenges (Schaufeli & Taris, 2014). The former refers to demands that are energy‐depleting
obstacles to employee growth, learning, and goal accomplishment (Crawford et al., 2010). Examples of job hindrances
under lean settings include role overload (Huo & Boxall, 2017) and work pace and monitoring pressure (Cullinane
572 HUO AND BOXALL

et al., 2013). When faced with such hindrances, employees may experience diminished control and negative
emotions, making them less capable of successfully handling these demands and achieving work goals. Because of
these negative feelings, employees may be less motivated to apply themselves and may therefore resort to a passive,
disengaged style of coping (Crawford et al., 2010).
In contrast, challenge demands are both energy depleting and stimulating in nature (Van den Broeck, De Cuyper,
De Witte, & Vansteenkiste, 2010). Although dealing with job challenges consumes energy and may contribute to
exhaustion, employees are more likely to view them as opportunities to learn, achieve, and display the types of com-
petence that will bring rewards in the future (Crawford et al., 2010). Given these potential benefits, they should be
more willing to put effort into coping with challenge demands. Crawford et al.'s (2010) meta‐analysis reported that
such challenges were positively related to both work engagement and exhaustion.
In lean production environments, Cullinane et al. (2013) identify problem‐solving demands as job challenges,
predicting that they should relate positively to work engagement. Such demands should provide opportunities for
employees to utilise existing skills and develop new skills (de Treville & Antonakis, 2006). This contributes to the
fulfilment of psychological needs such as self‐efficacy, competence, and self‐worth (Bandura, 1997; Warr, 1989),
which in turn lead to positive motivational outcomes, including work engagement (Eatough & Spector, 2014). There
is no prior research investigating the relationship between problem‐solving demands and work engagement, but
studies of other types of employee outcomes in nonlean work environments, such as Zhou, Hirst, and Shipton's
(2012) study of 270 Chinese employees and their supervisors, find that problem‐solving demands are positively
associated with employee creative self‐efficacy and creativity. Accordingly, we hypothesise that:

Hypothesis 1. Problem‐solving demands as a form of challenge demand will be positively related to work
engagement.

In addition to the main effects of challenge demands on work engagement and exhaustion, the JD‐R model pro-
poses two interaction hypotheses. One refers to the “buffering effect,” which argues that job resources alleviate the
impact of job demands (challenge demands in this study) on exhaustion (Bakker, Demerouti, Taris, Schaufeli, & Schreurs,
2003). The rationale for this argument is that workers who have sufficient job resources at their disposal are likely to
handle daily job demands better (Bakker & Demerouti, 2014). In lean settings, the knowledge and skill necessary for
problem solving are often fostered through a management investment in training resources (de Treville & Antonakis,
2006; Sterling & Boxall, 2013). To the extent that employees are not overloaded, participation in decision‐making in
continuous improvement programmes also acts as a resource, offering opportunities for learning, teamwork, and work
planning (Brännmark & Holden, 2013). Although such forms of lean participation are rarely designed to foster individual
autonomy and often restrict it (Vidal, 2007a), they can help employees to analyse and cope with problem‐solving
demands. In addition, Forza (1996) argues that managerial support is an important job resource to offset the detrimental
impact of problem‐solving demands. The support of the direct supervisor, in allocating tasks, providing technical advice,
and listening to concerns, is likely to moderate the health‐impairing effect of problem‐solving demands. The theoretical
rationale for this argument is provided by the job demand–control–support model (Karasek & Theorell, 1990), which
proposes that social support buffers the negative impact of job demands. Although there is no empirical evidence for
the interaction effect of job resources and problem‐solving demands in predicting exhaustion, it is clear that we should
hypothesise that:

Hypothesis 2. Job resources (i.e., training, employee participation in decision‐making, and line manager
support) moderate the effect of problem‐solving demands on exhaustion such that the effect will be
reduced in the presence of high rather than low job resources.

The second interactive effect between demands and resources as outlined in the JD‐R model argues that job
resources are more strongly associated with engagement when job demands are high (the “coping effect”) (Bakker,
Hakanen, Demerouti, & Xanthopoulou, 2007). In other words, it is under stressful conditions that employees will
HUO AND BOXALL 573

more likely need job resources to cope with existing demands and to reduce stress (Demerouti & Bakker, 2011). The
rationale draws on conservation of resources theory (Hobfoll, 2001), which argues that people weigh reward and
punishment differently and in such a way that resource loss (i.e., stemming from a job demand) is perceived as signif-
icantly more salient than resource gain. As a result, people are more sensitive to work conditions that bring about
resource losses than those that lead to resource gains. This perspective explains why job resources gain their greatest
motivating potential when employees are faced with high job demands (Bakker et al., 2007). For example, in a study
of Finnish teachers, Bakker et al. (2007) found that the relationship between job resources (e.g., job control and
supervisor support) and work engagement is strongest when teachers are confronted with high levels of pupil
misconduct (the indicator of job demands). We expect the same principle to apply in a lean context in respect of
problem‐solving demands. This leads to

Hypothesis 3. Problem‐solving demands moderate the effect of job resources on work engagement such
that the effect will be increased given high rather than low problem‐solving demands.

4 | R E S E A R C H C O N T E X T A ND M E T H O D S

Our study was conducted in a Chinese transportation equipment manufacturer employing 453 front‐line workers.
The company switched to the lean philosophy from traditional manufacturing principles associated with mass
production in 2008 in order to reduce production waste and improve process efficiency. The primary lean practices
applied in the company include total quality management, just‐in‐time inventory control, total preventative mainte-
nance, continuous improvement, and work standardisation.
The data were gathered in 2015 following a two‐stage process. In Stage 1, we conducted semistructured
interviews with 10 front‐line workers. The objective was to guide the design of the questionnaire survey in the next
stage so that the most salient job characteristics could be identified and measured. These interviews confirmed
that workers now faced a higher level of problem‐solving demands as a result of lean implementation, but they
saw problem‐solving opportunities as a way of using their skills and gaining recognition from management. They also
told us that they had been offered the three types of job resource referred to above: training on lean production
techniques, the personal support of their line manager, and opportunities for participation in decision‐making. The
last of these is akin to the “consultative participation” observed by Vidal (2007a), which falls short of extensive
empowerment but nonetheless serves as a job resource by improving employee voice in operating procedures and
working conditions.
Stage 2 involved the administration of a self‐report questionnaire entitled “The Employee Experience of Lean
Production.” The survey consisted of existing measurement scales with direct reference to lean practices (e.g., quality
circles and lean production training). Back‐translation techniques were utilised for survey translation. The surveys
were distributed during work hours by the first author and two HR professionals to the entire personnel at their
preshift meetings using the paper‐and‐pen method. All of the 453 front‐line workers were invited to participate
and asked to return the survey within 2 weeks' time. Upon completion, questionnaires were placed by participants
in a sealed envelope and submitted to either one of the two HR staff, who then relayed the questionnaires to the
first author. Three hundred and seventy‐one workers returned the survey, a response rate of 82%. Respondents
were primarily male (93%), and their average organisational tenure was 7 years (SD = 5.7). In terms of qualifications,
29.9% had only completed high school, 66.1% had obtained a technical or vocational qualification, and 4% held a
Bachelor's degree.
Given that two of the measurement scales (i.e., line manager support and problem‐solving demands) were
modified in this study, we decided to cross‐validate the factorial validity of these scales by using a second, indepen-
dent worker sample (n = 357) from a fast‐moving consumer goods (FMCG) manufacturer with extensive use of lean
practices. The objective was to strengthen the reliability and generalisability of research findings (Hakanen, Bakker, &
Schaufeli, 2006).
574 HUO AND BOXALL

4.1 | Measures for the independent variables

Unless noted otherwise, all measurement items were scored on a “1 = strongly disagree” to “7 = strongly agree” scale
and are listed in Appendix A.
Lean production‐related training was measured with a three‐item scale from Cullinane et al. (2014) (example
item: “The company provides me with adequate technical training in lean production”). Employee participation in
decision‐making was gauged using a four‐item scale from Huo and Boxall's (2018) study (example item: “My manager
actively seeks my ideas for improving working conditions”). Line manager support was measured using six items from
an eight‐item scale designed by Rhoades and Eisenberger (2002) (example item: “My manager really cares about my
well‐being”). Because studies have consistently found that negatively phrased items exhibit low factor loadings and
cause problems for interpretation (e.g., Piccolo & Colquitt, 2006), we removed the two negatively phrased items.
Problem‐solving demands were measured on a five‐item scale by Jackson, Wall, Martin, and Davids (1993) (example
item: “I regularly encounter problems in my job that I have not met before”).

4.2 | Measures for the outcome variables

In line with the JD‐R model, work engagement and exhaustion were selected as the well‐being indicators. This makes
it possible to compare our findings with other studies using the JD‐R model in lean settings (e.g., Cullinane et al.,
2014). In addition, both constructs are recognised antecedents to employee health and attitudinal and behavioural
outcomes. For example, exhaustion has been shown to be predictive of ill health, reduced work capability, and lower
organisational commitment (Hakanen et al., 2006) and job satisfaction (Skaalvik & Skaalvik, 2010). Similarly, work
engagement has been found to be associated with better health, performance, and commitment and lower turnover
intentions (Halbesleben, 2010).
As its core subdimensions, the “vigour” and “dedication” subscales of the shortened Utrecht Work Engagement
Scale were adopted to measure engagement (e.g., Demerouti, Mostert, & Bakker, 2010). Each scale comprises three
items. Examples are “When I get up in the morning, I feel like going to work” (vigour) and “My job inspires me”
(dedication). Participants answered on a “1 = never” to “7 = always (everyday)” scale. The “exhaustion” subscale
was derived from the Oldenburg Burnout Inventory (Demerouti, Bakker, Vardakou, & Kantas, 2003). We adopted
the four items Hu and Schaufeli (2011) used to gauge exhaustion in the Chinese setting (example item: “After work
I usually feel worn out and weary”).

4.3 | Measures for the control variables

On the basis of prior research (e.g., Brough et al., 2013), we included age, gender, and organisational tenure as
controls.

4.4 | Statistical analysis

Structural equation modelling using the Mplus statistical software (version 7.4) was adopted for hypothesis testing.
Step 1 examined the factorial validity of all of the measurement scales. To achieve this, we first utilised the primary
sample of data to determine the factor structures of the measurement scales and then cross‐validated these struc-
tures using the second sample collected from the FMCG company. At Step 2, we drew on the primary sample data
to test the hypothesised structural model with “job resources” specified as a second‐order latent variable reflected by
three first‐order latent variables (i.e., training, line manager support, and employee participation in decision‐making),
in line with previous JD‐R studies. Similarly, work engagement was operationalised as a second‐order latent variable.
The remaining first‐order latent variables relating to problem‐solving demands and exhaustion were measured at the
item level. The three control variables including age, gender, and tenure were modelled as observed variables. Results
HUO AND BOXALL 575

of Harman's single‐factor test showed that common method variance is unlikely to have had a substantial effect on
the current research findings (Podsakoff, MacKenzie, & Podsakoff, 2012).

5 | RESULTS

5.1 | Descriptive statistics

Table 1 displays the means, standard deviations, correlations, and reliability coefficients of the latent variables used in
the study. All of the scales demonstrate good composite reliability (>0.70). As anticipated, job resources (i.e., training,
employee participation, and line manager support) are positively correlated with engagement and negatively
correlated with exhaustion. Problem‐solving demands are positively correlated with engagement but not, however,
significantly correlated with exhaustion.

5.2 | Testing factorial validity

Using the primary sample data, we separately fitted one‐factor congeneric confirmatory factor analysis (CFA) models
for each of the study variables. In every step of the one‐factor CFA model testing, if the chi‐square statistic did not
meet the minimum accepted standards, changes were made to the model by dropping problematic items, one at a
time, providing that the changes made substantive sense. Specifically, we removed one item (“my manager considers
my goals and values”) from the line manager support scale because it showed a high cross‐factor correlated measure-
ment error. In addition, one item (“I regularly have to solve problems which have no obvious correct answer”) was
removed from the problem‐solving demands scale due to its low factor loading (<0.40). All of the remaining items
were found to be adequate indicators of their intended constructs and thus were retained in their respective models.
We then conducted a CFA with these remaining items by specifying a full four‐factor structure (job resources,
problem‐solving demands, work engagement, and exhaustion) using both the primary sample and the validation
sample from the FMCG company. Results showed a good model fit in both samples (in the primary sample:
χ2(288) = 617.016, CFI = 0.946, root mean square error of approximation [RMSEA] = 0.055, and standardised root
mean square residual [SRMR] = 0.062; in the validation sample: χ2(288) = 622.824, CFI = 0.929, RMSEA = 0.059,
and SRMR = 0.079; Hu & Bentler, 1999). All of the items significantly loaded onto their intended factors and their
factor loadings were uniformly higher than 0.40 in both samples (results available from the authors). These facts
strongly support the factorial validity of the measurement scales used in this study.

TABLE 1 Means, standard deviations, composite reliability, and correlations between latent variables (N = 371)
Study variables M SD 1 2 3 4 5 6

1. Problem‐solving demands 5.28 0.94 (0.72)

2. Line manager support 5.01 1.17 0.445*** (0.91)
3. Training 4.84 1.29 0.324*** 0.557*** (0.86)
4. Employee participation 4.58 1.03 0.438*** 0.638*** 0.539*** (0.90)
5. Exhaustion 4.57 1.20 −0.042 −0.016 −0.160* −0.106 (0.77)
6. Work engagement 4.30 1.22 0.523*** 0.452*** 0.557*** 0.497*** −0.425*** (0.92)

Note. Composite reliability coefficients are on the diagonal in parentheses.

*p < 0.05.
**p < 0.01.
***p < 0.001.
576 HUO AND BOXALL

TABLE 2 Intercorrelations among the four latent variables estimated in the full measurement model
Study variables 1 2 3

1. Job resources
2. Problem‐solving demands 0.540
3. Engagement 0.643 0.525
4. Exhaustion −0.109 −0.044 −0.425

5.3 | Testing convergent and discriminant validity

To examine convergent and discriminant validity, the full four‐factor CFA model was specified and tested using the
primary sample. As noted above, the model fit was adequate, and all items underlying a common factor had substan-
tial standardised factor loadings (>0.40). This fact indicated the presence of convergent validity (Kline, 2011). Table 2
displays estimated correlations for all possible pairs of latent variables. All are well below the cut‐off value of 0.90
(Kline, 2011), implying that we can infer good discriminant validity.

5.4 | Testing the main effect of problem‐solving demands on well‐being

The structural model as shown in Figure 1 (control variables are not shown for clarity's sake) tested the main effects
of problem‐solving demands on work engagement and exhaustion. This model provided a good fit to the data:
χ2(354) = 706.542, CFI = 0.942, RMSEA = 0.052, and SRMR = 0.060. Both job resources (β = 0.508, p < 0.001)
and problem‐solving demands (β = 0.234, p < 0.01) were positively related to work engagement. Thus, Hypothesis 1
was supported. In contrast, the main effect of job resources on exhaustion was negative but not statistically signifi-
cant (β = −0.124, p > 0.05). This effect could be significant in a larger sample. Alternatively, this may indicate the pres-
ence of an interaction effect between resources and problem‐solving demands, which makes the main effect of
resources on exhaustion unstable. In addition, none of the personal variables (age, gender, and organisational tenure)
was a significant predictor of either of the outcomes. The predictors in the model jointly accounted for 47.1% of the
variance in engagement and 13.4% of the variance in exhaustion.

FIGURE 1 The structural model testing the main effects of problem‐solving demands on well‐being as outlined in
the job demands–resources model. Control variables: Age, gender, and tenure were included as control variables. All
of the path coefficients are standardised. **p < 0.01; ***p < 0.001. n.s.: not significant
HUO AND BOXALL 577

TABLE 3 The interaction between resources and problem‐solving demands in predicting exhaustion
Path coefficients Unstandardised effect p value 95% LL 95% UL

The main effect of resources −0.191 0.195 −0.481 0.109

The main effect of problem‐solving demands −0.048 0.847 −0.644 0.382
Resources/problem‐solving demands interaction −0.986 0.011 −1.829 −0.301

Note. LL: lower limit of confidence interval; UL: upper limit of confidence interval.

5.5 | Testing the interaction effects between resources and problem‐solving demands
The full structural model involving the interaction terms was estimated to test the interaction effects between job
resources and problem‐solving demands. In line with Cheung and Lau (2015), we employed bias‐corrected
confidence intervals based on 1,000 bootstrap samples as the criterion judging the statistical significance of the
interaction effects. Tables 3 and 4 display the results for the estimated main and interaction effects. It is clear from
Table 3 that the relationship between the lean resources/problem‐solving demands interaction and exhaustion is
significant (unstandardised path coefficient = −0.986, 95% confidence interval [CI] [−1.829, −0.301]). This confirms
Hypothesis 2 that lean job resources buffer the effect of problem‐solving demands on exhaustion.
Table 4 reports the finding that the path coefficient relating the lean resources/problem‐solving demands inter-
action term to engagement is significant (unstandardised path coefficient = 0.397, 95% CI [0.112, 0.817]). Therefore,
Hypothesis 3, which posited that higher levels of job resources would be more strongly associated with engagement
when problem‐solving demands are high (a “coping effect”), was confirmed.
To visualise the patterns of the two significant interaction effects, we graphed the moderation plots. Figure 2
displays the moderation of the effect of problem‐solving demands on exhaustion by job resources at three different
levels (Aiken, West, & Reno, 1991). Simple slope analysis was conducted to determine whether the slopes of the

TABLE 4 The interaction between resources and problem‐solving demands in predicting engagement
Path coefficients Unstandardised effect p value 95% LL 95% UL

The main effect of resources 0.781 0.000 0.476 1.148

The main effect of problem‐solving demands 0.601 0.017 0.194 1.214
Resources/problem‐solving demands interaction 0.397 0.021 0.112 0.817

Note. LL: lower limit of confidence interval; UL: upper limit of confidence interval.

FIGURE 2 Moderation plot for the effect of problem‐solving demands on exhaustion moderated by job resources
[Colour figure can be viewed at wileyonlinelibrary.com]
578 HUO AND BOXALL

TABLE 5 The conditional effects of problem‐solving demands on exhaustion at various levels of job resources
Three levels of the moderator Job resources Unstandardised effect 95% BCLL 95% BCUL

+1 SDR 0.806 −0.842 −1.893 0.011

Mean R 0.000 −0.048 −0.644 0.382
−1 SDR −0.806 0.746 0.186 1.321

Note. +1 SDR: one standard deviation above the mean; −1 SDR: one standard deviation below the mean; R: value of the mod-
erator job resources; BCLL: lower limit of bias‐corrected bootstrap confidence interval; BCUL: upper limit of bias‐corrected
bootstrap confidence interval.

three regression lines in Figure 2 significantly vary from zero (see Table 5). The pattern of these three regression lines
lends strong support for the buffering effect of job resources. When job resources are low, at one standard deviation
below the mean (−1 SD), the effect of problem‐solving demands on exhaustion is significantly positive (0.746, 95% CI
[0.186, 1.321]). When job resources increase to a higher level (the mean value), the impact of problem‐solving
demands on exhaustion reduces (−0.048, an effect that is not significantly different from zero, 95% CI [−0.644,
0.382]). This implies that the level of job resources can change the strength of the problem‐solving demands‐to‐
exhaustion relationship, which is a clear buffering effect.
Figure 3 shows the interaction plot for the relationship between job resources and engagement as moderated by
problem‐solving demands. As the level of problem‐solving demands increases, the positive job resources‐to‐engage-
ment relationship becomes increasingly strong. Simple slope analysis shows that all of the three regression slopes in
Figure 3 are statistically significant (Table 6). This lends strong support to the coping effect as predicted in
Hypothesis 3.

FIGURE 3 Moderation plot for the effect of job resources on engagement moderated by problem‐solving demands
[Colour figure can be viewed at wileyonlinelibrary.com]

TABLE 6 The conditional effects of job resources on engagement at various levels of problem‐solving demands
Three levels of the moderator Problem‐solving demands Unstandardised effect 95% BCLL 95% BCUL

+1 SDp 0.509 0.983 0.584 1.428

Mean p 0.000 0.781 0.476 1.148
−1 SDp −0.509 0.579 0.317 0.918

Note. +1 SDp: one standard deviation above the mean; −1 SDp: one standard deviation below the mean; p: value of the mod-
erator problem‐solving demands; BCLL: lower limit of bias‐corrected bootstrap confidence interval; BCUL: upper limit of
bias‐corrected bootstrap confidence interval.
HUO AND BOXALL 579

6 | DISCUSSION

This study is the first to provide empirical support to Cullinane et al.'s (2013) prediction that problem‐solving
demands under lean settings can bring motivational benefits for employees and should be considered as a type of
job challenge. Our findings also support the ongoing evolution of the JD‐R model by adding evidence to the
argument that not all job demands are energy depleting and detrimental for employees (Crawford et al., 2010;
Schaufeli & Taris, 2014). In this study, problem‐solving demands functioned as challenges, showing a positive linkage
to engagement, which suggests that they can help fulfil employee needs for learning, competence, and growth. This
result stands in contrast to previous studies that report a negative relationship between problem‐solving demands
and employee well‐being (e.g., Bouville & Alis, 2014; Jackson & Mullarkey, 2000). Despite the positive impact of
problem‐solving demands on engagement identified in this study, its effect size is smaller than that of the association
between job resources and engagement, implying that work engagement is more responsive to changes in the level
of resources than the degree of problem‐solving demands.
A major objective of our study was to test hypotheses relating to the interaction between job resources and
problem‐solving demands in predicting employee well‐being. The results supported both of the moderation
hypotheses. One is the buffering effect of job resources in the problem‐solving demands‐to‐exhaustion relation-
ship, suggesting that challenge demands are less likely to produce exhaustion when appropriate job resources
are provided. This is the first such finding in a lean production context. The other interaction hypothesis validated
in this study has to do with the strengthening relationship between resources and engagement when problem‐
solving demands are high (“the coping effect”). This finding implies that it is under more stressful conditions that
employees will most appreciate relevant job resources provided by management (Demerouti & Bakker, 2011). In
the context of an implementation of lean production, then, the results show the importance of matching resources
to the specific demands being faced. In our case, resources of training, opportunities for participation in decision‐
making, and line manager support are making a difference to the ability of workers to rise to the challenge of the
problem‐solving demands.
The results in respect of problem‐solving demands suggest they may have quite a different impact than role
overload or increases in work intensity under lean production (Huo & Boxall, 2017). The former have the potential
to foster better skill utilisation and make work more interesting, whereas the latter can generate fatigue and pose
a risk to well‐being (Koukoulaki, 2014).

6.1 | Limitations and directions for future research

This is a cross‐sectional study of a single case, limiting causal interpretations. A logical next step is testing the
model through a longitudinal design, especially one in which the “technical and organisational context” and the
dynamics of implementation are closely observed (Hasle et al., 2012, p. 846). Moreover, research would benefit
from examining whether it is the nature or the magnitude of problem‐solving demands that makes the greatest
impact on well‐being and the possible mediation mechanisms involved. It is possible, for example, that a high level
of demand for participation in decision‐making might trigger role overload, leading to worsened employee health
(Brännmark & Holden, 2013). In addition, future studies could investigate the effects of lean job characteristics on
a fuller range of well‐being variables, such as job satisfaction and employee absence and sickness rates. It would
also be useful to investigate the possibility of three‐way interactions involving job resources, problem‐solving
demands, and employee work orientations because individual orientations to work may affect the relationship
between problem‐solving demands and job satisfaction (Vidal, 2007b). Lastly, the interaction effects between
resources and lean‐specific hindrance demands (e.g., role overload and work pace) were not examined in this
study. Future efforts should be dedicated to identifying the job resources that can best minimise the health‐
impairing effects of hindrance demands.
580 HUO AND BOXALL

7 | C O N CL U S I O N S

The results of this study suggest that resources such as relevant training, line manager support, and employee partic-
ipation in decision‐making, along with the challenges posed by problem‐solving demands, can enhance employee
engagement in the environment of lean production. Its primary contribution lies in testing previously unexamined
interactions between resources and problem‐solving demands. The verification of the “buffering” effect highlights
the importance of a careful matching of resources to the type of demand faced by the production workforce. The
confirmation of the “coping” effect shows the increasing value of relevant job resources as demands increase. These
results imply that not all job demands associated with lean production are detrimental for employee well‐being.
Demands of a challenging nature may enhance work motivation through creating work that employees find more ful-
filling (Cullinane et al., 2013). Problem‐solving activities are more intellectually engaging than highly repetitive aspects
of the job. In contrast, hindrance demands under lean settings (e.g., role overload and monitoring responsibilities)
remain of concern in terms of their impacts on health (Huo & Boxall, 2017). How lean manufacturers can ameliorate
these types of demand—or adjust the balance between challenge and hindrance demands—remains a key issue.
Where does this leave us in terms of the great debate over the human impacts of lean production? Our study
implies that lean production contains within it the seeds of both positive and negative outcomes for workers. In
any particular implementation, the overall impact on well‐being depends on how management fosters the aspects
that can enhance well‐being and addresses the potential threats (Conti, Angelis, Cooper, Faragher, & Gill, 2006; Hasle
et al., 2012). In the specific implementation we assessed, job challenges to do with problem solving showed their
positive value in fostering employee engagement, suggesting that they can help employees to reach more of their
potential. In addition, management successfully provided a range of resources that ameliorated the impact of these
challenges on exhaustion and that were directly supportive of greater engagement. Rather than being uniformly
positive or negative, the overall impact of lean production on worker well‐being is likely to depend on the ways in
which managers create scope for worker involvement in decision‐making, target resources to the specific job
demands, and adjust resource levels to the degree of these demands.

ACKNOWLEDGEMEN T

We thank Professor Gordon W. Cheung for his statistical advice. This research did not receive any specific grant from
funding agencies in the public, commercial, or not‐for‐profit sectors.

CONF LICT OF INT E RE ST

The authors declare that they have no conflict of interest.

ORCID

Meng‐Long Huo http://orcid.org/0000-0003-4583-6592

RE FE R ENC ES
Aiken, L., West, S., & Reno, R. (1991). Multiple regression: Testing and interpreting interactions. Thousand Oaks, CA: Sage
Publications.
Bakker, A., and Demerouti, E. (2014). Job demands–resources theory, in P. Chen and C. Cooper (eds.), Work and well‐being: A
complete reference guide (vol. 3), Hoboken, NJ: John Wiley & Sons.
Bakker, A., Demerouti, E., Taris, T., Schaufeli, W., & Schreurs, P. (2003). A multigroup analysis of the job demands–resources
model in four home care organizations. International Journal of Stress Management, 10(1), 16–38.
Bakker, A., Hakanen, J., Demerouti, E., & Xanthopoulou, D. (2007). Job resources boost work engagement, particularly when
job demands are high. Journal of Educational Psychology, 99(2), 274–284.
Bamber, G., Stanton, P., Bartram, T., & Ballardie, R. (2014). Human resource management, lean processes and outcomes for
employees: Towards a research agenda. International Journal of Human Resource Management, 25(21), 2881–2891.
HUO AND BOXALL 581

Bandura, A. (1997). Self‐efficacy: The exercise of control. New York, NY: Freeman & Co.
Bouville, G., & Alis, D. (2014). The effects of lean organizational practices on employees' attitudes and workers' health:
Evidence from France. International Journal of Human Resource Management, 25(21), 3016–3037.
Brännmark, M., & Holden, R. (2013). Packages of participation: Swedish employees' experience of lean depends on how they
are involved. IIE Transactions on Occupational Ergonomics and Human Factors, 1(2), 93–108.
Brough, P., Timms, C., Siu, O. L., Kalliath, T., O'Driscoll, M. P., Sit, C. H., & Lu, C. Q. (2013). Validation of the job demands–
resources model in cross‐national samples: Cross sectional and longitudinal predictions of psychological strain and work
engagement. Human Relations, 66(10), 1311–1335.
Chan, A., Chen, Y., Xie, Y., Wei, Z., & Walker, C. (2014). Disposable bodies and labor rights: Workers in China's automotive
industry. Working USA: The Journal of Labor & Society, 17(4), 509–529.
Cheung, G., & Lau, R. (2015). Accuracy of parameter estimates and confidence intervals in moderated mediation
models: A comparison of regression and latent moderated structural equations. Organizational Research Methods,
1, 1–24.
Conti, R., Angelis, J., Cooper, C., Faragher, B., & Gill, C. (2006). The effects of lean production on worker job stress. Interna-
tional Journal of Operations & Production Management, 26(9), 1013–1038.
Crawford, E., LePine, J., & Rich, B. (2010). Linking job demands and resources to employee engagement and burnout:
A theoretical extension and meta‐analytic test. Journal of Applied Psychology, 95(5), 834–848.
Cullinane, S., Bosak, J., Flood, P., & Demerouti, E. (2013). Job design under lean manufacturing and its impact on employee
outcomes. Organizational Psychology Review, 3, 44–61.
Cullinane, S., Bosak, J., Flood, P., & Demerouti, E. (2014). Job design under lean manufacturing and the quality of working
life: A job demands and resources perspective. International Journal of Human Resource Management, 25(21),
2996–3015.
de Treville, S., & Antonakis, J. (2006). Could lean production job design be intrinsically motivating? Contextual, configura-
tional, and levels‐of‐analysis issues. Journal of Operations Management, 24(2), 99–123.
Delbridge, R. (2007). HRM and contemporary manufacturing. In P. Boxall, J. Purcell, & P. Wright (Eds.), The Oxford handbook
of human resource management (pp. 405–427). Oxford, UK: Oxford University Press.
Demerouti, E., & Bakker, A. (2011). The job demands–resources model: Challenges for future research. The Journal of Individ-
ual Psychology, 37(2), 1–9.
Demerouti, E., Bakker, A., Nachreiner, F., & Schaufeli, W. (2001). The job demands–resources model of burnout. Journal of
Applied Psychology, 86(3), 499–512.
Demerouti, E., Bakker, A., Vardakou, I., & Kantas, A. (2003). The convergent validity of two burnout instruments: A
multitrait–multimethod analysis. European Journal of Psychological Assessment, 19(1), 12–23.
Demerouti, E., Mostert, K., & Bakker, A. (2010). Burnout and work engagement: A thorough investigation of the indepen-
dency of both constructs. Journal of Occupational Health Psychology, 15(3), 209–222.
Eatough, E., & Spector, P. (2014). The role of workplace control in positive health and wellbeing. In P. Chen, & C. Cooper
(Eds.), Work and wellbeing: A complete resource guide (pp. 91–110). Oxford, UK: Wiley‐Blackwell.
Forza, C. (1996). Work organization in lean production and traditional plants: What are the differences? International Journal
of Operations & Production Management, 16(2), 42–62.
Hakanen, J., Bakker, A., & Schaufeli, W. (2006). Burnout and work engagement among teachers. Journal of School Psychology,
43(6), 495–513.
Halbesleben, A. (2010). A meta‐analysis of work engagement: Relationships with burnout, demands, resources, and conse-
quences. In A. Bakker, & M. Leiter (Eds.), Work engagement: A handbook of essential theory and research (pp. 102–117).
New York, NY: Psychology Press.
Hasle, P., Bojesen, A., Langaa Jensen, P., & Bramming, P. (2012). Lean and the working environment: A review of the
literature. International Journal of Operations & Production Management, 32(7), 829–849.
Hobfoll, S. (2001). The influence of culture, community, and the nested‐self in the stress process: Advancing conservation of
resources theory. Applied Psychology, 50(3), 337–421.
Holman, D. (2002). Employee wellbeing in call centres. Human Resource Management Journal, 12(4), 35–50.
Hu, L., & Bentler, P. (1999). Cut‐off criteria for fit indexes in covariance structure analysis: Conventional criteria versus new
alternatives. Structural Equation Modelling: A Multidisciplinary Journal, 6(1), 1–55.
Hu, Q., & Schaufeli, W. (2011). The convergent validity of four burnout measures in a Chinese sample: A confirmatory factor‐
analytic approach. Applied Psychology, 60(1), 87–111.
582 HUO AND BOXALL

Huo, M., & Boxall, P. (2017). Lean production and the well‐being of the frontline manager: The job demands–resources
model as a diagnostic tool in Chinese manufacturing. Asia Pacific Journal of Human Resources, 55(3), 280–297.
Huo, M., & Boxall, P. (2018). Instrumental work values and responses to HR practices: A study of job satisfaction in a Chinese
manufacturer. Personnel Review, 47(1), 60–73.
Jackson, P., & Mullarkey, S. (2000). Lean production teams and health in garment manufacture. Journal of Occupational Health
Psychology, 5(2), 231–245.
Jackson, P., Wall, T., Martin, R., & Davids, K. (1993). New measures of job control, cognitive demand, and production respon-
sibility. Journal of Applied Psychology, 78(5), 753–762.
Karasek, R., & Theorell, T. (1990). Healthy work. New York, NY: Basic Books.
Kline, R. (2011). Principles and practice of structural equation modelling. New York, NY: Guilford Publications.
Koukoulaki, T. (2014). The impact of lean production on musculoskeletal and psychosocial risks: An examination of
sociotechnical trends over 20 years. Applied Ergonomics, 45(2), 198–212.
MacDuffie, J. (1995). Human resource bundles and manufacturing performance: Organizational logic and flexible production
systems in the world auto industry. ILR Review, 48(2), 197–221.
Piccolo, R., & Colquitt, J. (2006). Transformational leadership and job behaviours: The mediating role of core job character-
istics. Academy of Management Journal, 49(2), 327–340.
Podsakoff, P., MacKenzie, S., & Podsakoff, N. (2012). Sources of method bias in social science research and recommenda-
tions on how to control it. Annual Review of Psychology, 63, 539–569.
Rhoades, L., & Eisenberger, R. (2002). Perceived organizational support: A review of the literature. Journal of Applied Psychol-
ogy, 87(4), 698–714.
Schaufeli, W., & Taris, T. (2014). A critical review of the job demands–resources model: Implications for improving
work and health. In G. Bauer, & O. Hammig (Eds.), Bridging occupational, organizational and public health.
Dordrecht, Netherlands: Springer.
Schmitt, A., Zacher, H., & Frese, M. (2012). The buffering effect of selection, optimization, and compensation strategy use on
the relationship between problem solving demands and occupational well‐being: A daily diary study. Journal of
Occupational Health Psychology, 17(2), 139–149.
Shah, R., & Ward, P. (2003). Lean manufacturing: Context, practice bundles, and performance. Journal of Operations
Management, 21(2), 129–149.
Shah, R., & Ward, P. (2007). Defining and developing measures of lean production. Journal of Operations Management, 25(4),
785–805.
Skaalvik, E., & Skaalvik, S. (2010). Teacher self‐efficacy and teacher burnout: A study of relations. Teaching and Teacher
Education, 26(4), 1059–1069.
Sterling, A., & Boxall, P. (2013). Lean production, employee learning and workplace outcomes: A case analysis through the
ability–motivation–opportunity framework. Human Resource Management Journal, 23(3), 227–240.
Van den Broeck, A., De Cuyper, N., De Witte, H., & Vansteenkiste, M. (2010). Not all job demands are equal: Differentiating
job hindrances and job challenges in the job demands–resources model. European Journal of Work and Organizational
Psychology, 19(6), 735–759.
Vidal, M. (2007a). Manufacturing empowerment? ‘Employee involvement’ in the labour process after Fordism. Socio‐
Economic Review, 5, 197–232.
Vidal, M. (2007b). Lean production, worker empowerment, and job satisfaction: A qualitative analysis and critique. Critical
Sociology, 33, 247–268.
Warr, P. (1989). Work, unemployment and mental health. Oxford, UK: Oxford University Press.
Womack, J., Jones, D., & Roos, D. (1990). The machine that changed the world. New York, NY: Rawson Associates.
Zhang, L. (2015). Lean production ‘with Chinese characteristics’: A case study of China's automobile industry. International
Journal of Sociology, 45(2), 152–170.
Zhou, Q., Hirst, G., & Shipton, H. (2012). Promoting creativity at work: The role of problem‐solving demand. Applied Psychol-
ogy, 61(1), 56–80.

How to cite this article: Huo M‐L, Boxall P. Are all aspects of lean production bad for workers? An analysis of
how problem‐solving demands affect employee well‐being. Hum Resour Manag J. 2018;28:569–584. https://
doi.org/10.1111/1748‐8583.12204
HUO AND BOXALL 583

APPENDIX A
M E A S U R E M E N T SC A LE S

Work engagement
1 When I get up in the morning, I feel like going to work.
2 At my work, I feel bursting with energy.

3 At my job I feel strong and vigorous.

4 My job inspires me.
5 I am enthusiastic about my job.
6 I am proud of the work that I do.

Exhaustion
1 After work I usually feel worn out and weary.
2 There are days when I feel tired before I arrive at work.
3 After work, I tend to need more time than in the past in order to relax and feel better.
4 During my work, I often feel emotionally drained.

Lean production‐related training

1 The company provides me with adequate technical training in lean production.
2 The company provides me with adequate team skills training (e.g., communication and interpersonal skills).
3 The company provides me with adequate quality training.

Line manager support

1 My manager really cares about my well‐being.

2 My manager cares about my opinions.

3 My manager is willing to help if I need a special favour.
4 Help is available from my manager when I have a problem.
5 My manager would forgive an honest mistake on my part.

Employee participation in decision‐making

1 I am encouraged to suggest ideas for improving standard operating procedures.

2 My manager actively seeks my ideas for improving working conditions.

3 My manager welcomes my ideas when I speak up in quality circle meetings.
4 My manager actively seeks my opinion when we encounter a production problem.
584 HUO AND BOXALL

Problem‐solving demands
1 I am regularly required to deal with problems which are difficult to solve.
2 I regularly have to use my production knowledge to help prevent problems arising in my job.

3 The problems I deal with regularly require a thorough knowledge of production in my area.
4 I regularly encounter problems in my job that I have not met before.