See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/258200282

Morale and Workplace Performance

Article  in  Work and Occupations · August 2006

DOI: 10.1177/0730888406290054

CITATIONS READS

59 8,582

2 authors, including:

Stephen Frenkel
UNSW Sydney
136 PUBLICATIONS   4,602 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Garment Supply Chain Governance Project View project

All content following this page was uploaded by Stephen Frenkel on 09 July 2021.

The user has requested enhancement of the downloaded file.

 Work and Occupations
Volume 33 Number 3
August 2006 335-361
© 2006 Sage Publications
Morale and Workplace 10.1177/0730888406290054
http://wox.sagepub.com
Performance hosted at
http://online.sagepub.com

David L. Weakliem
University of Connecticut
Stephen J. Frenkel
Australian Graduate School of Management

This study considers the relationship between morale and workplace produc-
tivity in a representative sample of Australian workplaces. It focuses on three
questions: the shape of the relationship, whether the effects of morale are
contingent on other factors, and the paths by which any effect takes place.
The results show that morale influences productivity in an approximately
linear fashion. The effect of morale on productivity appears to be larger when
management regards product quality as important and attempts to develop a
corporate ethic and culture. Morale is associated with greater work effort, but
the relationship between work effort and productivity becomes stronger at
higher levels of morale. Thus, part of the influence of morale on productiv-
ity is a matter of increasing the effectiveness of workers’ efforts.

Keywords: morale; productivity; work effort

D espite managerial efforts at monitoring and supervision, nearly all

employees retain some control over how hard and effectively they work.
Hence, workers’ values, attitudes, or beliefs potentially affect their produc-
tivity. In a recent study, Bewley (1999) found that managers “believed almost
universally that morale had an important effect on productivity, though the
effect might be hard to measure” (p. 48). Moreover, firms often refrain from
actions that would produce an immediate gain in profits, such as cutting wages
during periods of high unemployment, because of concern about potential
damage to morale (Bewley, 1999; Blinder & Choi, 1990; Campbell &
Kamlani, 1997).

Authors’ Note: Please direct correspondence to David L. Weakliem, Department of

Sociology, University of Connecticut, Unit 2068, Storrs, CT 06269. A previous version of this
article was presented at the Emerging Futures of the Workplace Symposium, Australian
Graduate School of Management, June 2002. We thank the participants in this symposium
and the reviewers and editor of this journal for valuable comments on previous drafts.

335
336 Work and Occupations

The relationship between morale and productivity was a major topic in

the early days of organizational research (Haire, 1954). Since the 1950s,
researchers have tended to turn away from morale, focusing on behavior or on
objective features of organizations. The newer work, however, still relies on
assumptions about the relationship between subjective states and behavior.
For example, some observers suggest that productivity can be enhanced by a
“high road” approach, whose features include “employment security, team-
work, employee participation in decision making, and a relatively egalitarian
status system” (Frenkel, 2003, p. 141; also see Appelbaum, Bailey, Berg, &
Kalleberg, 2000). This claim rests on the assumption that workers will per-
form more effectively when they feel secure and respected. Others point to the
importance of “employee citizenship behavior,” which is used to cover a range
of practices that make the organization work more effectively (Hodson, 2001;
Organ, 1988). As Hodson (2002) notes, “The implicit model of organizational
productivity and effectiveness . . . is one in which technical factors of produc-
tion and organizational leadership must be supplemented by worker effort and
enthusiasm to reach optimal or even competitive levels” (p. 70). Accounts of
citizenship behavior also rest on assumptions about the subjective factors
affecting effort and enthusiasm. For example, Hodson (2002) argues for the
importance of reciprocity: Workers will respond to fair and predictable treat-
ment from management by willingness to put more effort into the job.
These general assumptions are plausible because there is a substantial
body of evidence suggesting that people with more positive feelings about
their job and employer tend to be more productive (Judge, Thoresen, Bono,
& Patton, 2001). However, several important questions about the relation-
ship between workers’ feelings and performance remain unanswered. With
the exception of Harter, Schmidt, and Hayes (2002), nearly all previous
research on this topic deals with the individual rather than the workplace
level. The relationship may be stronger at the workplace level because of
spillover effects in which an individual’s morale affects the productivity of
coworkers. For example, workers with low morale might become involved
in disputes with coworkers or managers, forcing them to divert time and
energy from their own jobs. On the other hand, some of the individual-level
association may reflect the influence of other individual qualities, particu-
larly aspects of personality, on both productivity and morale (Judge et al.,
2001). Also, there is uncertainty about the scope of the relationship between
workers’ feelings and performance. Perrow (1986) holds that “in many jobs
there is no room for high performance” (p. 87). In his view, most jobs are
designed to require only a basic minimum of performance, which can be
obtained by rewards and punishments, so that in most cases there is no
 Weakliem, Frenkel / Effects of Workplace Morale 337

room for workers to exercise much choice. Similarly, Grunberg, Anderson-

Connolly, and Greenberg (2000) note that some workers may be unable to
afford to express negative sentiments. They find that lower level workers
who have close contact with layoffs respond with increased work effort and
suggest that “with less power and fewer options, they may have to swallow
hard and toe the line” (Grunberg et al., p. 25).
As a result of recent research, we have learned a good deal about the
relationships between various objective features of organizations and organiza-
tional performance. We also know a good deal about influences on workers’
attitudes and perceptions. However, much less is known about whether and
how these attitudes and perceptions influence organizational performance. The
present research is designed to fill this gap by examining the relationship
between workers’ morale and productivity using the Australian Workplace
Industrial Relations Survey (AWIRS), a representative sample of workplaces
with more than 20 employees. An important feature of the AWIRS is that it
obtains information from both managers and a sample of workers at each work-
place. Reports of productivity are taken from the managerial survey, whereas
estimates of workplace morale are obtained from the employee survey.
Although morale is a traditional term in organization research, observers
such as Roethlisberger (1941) criticized it as vague, and recent researchers
have proposed a variety of more specific terms involving aspects of the job and
organization. However, Organ (1997) notes that there is an empirical tendency
for all kinds of positive or negative attitudes to go together—for example,
workers who are more satisfied with their jobs also tend to feel better about
their coworkers and management. Thus, even if it is possible to distinguish
among more narrowly defined constructs, it is reasonable to think of a general
factor underlying all of them, and Organ (1997) suggests that morale be used
to refer to this factor. This study follows his suggestion of using morale as a
general term for positive feelings about the job. Making distinctions among the
effects of more specific subjective factors is a task for future research.
However, rather than including all aspects of the job, we limit morale to feel-
ings about the “prescribed activities of the group” (Manning, 1991, p. 455),
because these activities are immediately relevant to productivity. The concept
of “organizational commitment” introduced by Porter, Steers, Mowday, and
Boulian (1974; see also Lincoln & Kalleberg, 1990) is similar to this sense of
morale, but the usual ways of measuring organizational commitment involve
not only general support for the organization and its goals but also statements
about intention to remain with the work organization and willingness to work
hard. Morale, as defined here, is further in the background: a general orienta-
tion that may influence intentions and ultimately behavior.
338 Work and Occupations

Theory and Hypotheses

Most research and theory suggests that there is at least some positive
relationship between morale and productivity. However, there is uncertainty
about the shape of the relationship, with some accounts implying diminish-
ing returns. Also, many discussions suggest that the degree to which morale
influences productivity will differ, depending on various characteristics of
the job or organization. Finally, there is uncertainty about the paths through
which morale influences productivity. The most straightforward possibility
is that workers with higher morale are willing to work harder, but several
other avenues have been suggested. This section develops hypotheses relat-
ing to these three general areas.

Overall Relationship
Some accounts suggest that increases from low to moderate levels of
morale will have more effect on productivity than increases from moderate to
high levels. Workers with sufficiently low morale can respond by poor work-
manship, wasting resources, or even sabotage. However, workers with unusu-
ally high morale may not be able to do much to enhance productivity. Perrow
(1986) argues that many jobs do not offer much scope for exceptionally good
performance and observes that special effort by workers may reduce produc-
tivity from management’s point of view. For example, workers who try to
provide customers with the best possible service will spend more time on
each one and hence serve fewer customers than workers who aim for merely
adequate service. A second reason for diminishing returns is that measures to
increase morale often have costs (Shister, 1950). Inexpensive measures such
as using reasonable care in selection of supervisors could increase morale but
raising it to unusually high levels might require expensive measures such as
keeping workers on the payroll during periods of low demand. These argu-
ments suggest that the effects of morale on productivity are nonlinear: dimin-
ishing and possibly even becoming negative at higher levels of morale.

Interactions
The idea that the effects of morale are contingent on other factors has a
prominent place in the literature. Table 1 lists plausible interactions that can
be considered using the AWIRS data; although many variables are listed,
the hypotheses follow from a small number of basic principles. First, it is
often suggested that morale will have more effect where workers have more
 Weakliem, Frenkel / Effects of Workplace Morale 339

Table 1
Variables in Hypothesized Interactions With Morale
Autonomy

1. Workers’ average rating of “How much influence or input do you have about. . . the
type of work you do?”
2. Workers’ average rating of “How much influence or input do you have about. . . how
you do your work?”
3. Workers’ average rating of “how much influence or input do you have about. . . the
pace at which you do your job?”
4. Proportion of workers who are union members.
5. Number of union delegates at workplace per worker.
6. Proportion of nonmanagerial workers who are professionals or semiprofessionals.
7. Proportion of nonmanagerial workers who are operatives or laborers.
8. Agreement that “management would choose quality improvements over price reductions.”
9. Choosing “quality of product,” “quality of workforce,” or “response to consumer
needs” as the most important factor for competitive success.

Management Strategy

1. Agreement that the “organisation devotes considerable resources towards having a

corporate ethic and culture.”
2. Agreement that the organization “devotes considerable resources to the management
of this workplace’s human resources.”

Size

1. Logarithm of number of workers in the workplace.

2. Logarithm of the number of workers in all the organization’s workplaces in Australia.
3. Whether organization includes multiple workplaces.

Characteristics of Workers

1. Education (adjusted for occupation)

autonomy, in the sense of control over the production process (Shister,

1950; Perrow, 1986).1 At one extreme, if discipline and measurement of
output were costless, there would be no room for morale to have any effect:
Managers could simply specify the amount that must be produced, and
workers would be obliged to produce that amount. At the other extreme, if
individual output could not be measured or workers could not be disci-
plined, workers would have complete discretion: Dissatisfied workers
could reduce their output to zero or even hinder the work of others. In actual
work settings, autonomy is a matter of degree, so the effect of morale on
productivity can be expected to increase with the degree of autonomy.
340 Work and Occupations

There are a number of ways that autonomy can be defined and mea-
sured. As Slichter (1920/1961) noted, autonomy can result either from for-
mal organization or from workers’ position in the day-to-day process of
production. The AWIRS contains a relatively direct measure of autonomy
in everyday tasks—employees’ ratings of how much “control or influence”
they have over various aspects of their work. The three aspects listed in
Table 1 are selected, because they involve the process of production and
hence are most closely related to autonomy on the job.
Several other variables, although not measures of autonomy, are poten-
tially relevant because they may influence autonomy. Unionization may
enhance autonomy by giving workers protection against dismissal and
other forms of discipline. Autonomy is generally greater in skilled jobs,
because they generally require more judgment on the part of the workers.
In the AWIRS, managers are asked to classify their workforce into several
groups, so average skill can be measured by the reported proportion of
skilled manual workers and professionals. These variables affect many
aspects of the organization; so any interactions involving them could be
interpreted in other ways. Hence, the results for these variables should be
taken in conjunction with those for the direct measure.
Finally, the difficulty of measuring output enhances autonomy, because
workers whose output is hard to measure can reduce their effectiveness
with less risk of detection. Quality of output is generally more difficult to
measure than quantity, because quality problems may not become apparent
until after the product is in the hands of the consumer. Hence, morale might
have more effect on productivity when quality is an important factor in
competitive success. The AWIRS survey contains two questions involving
the importance of quality: whether management would choose quality
improvements over cost reductions and whether quality is the crucial factor
for competitive success with the workplace’s major product or service.
A second general factor that may affect the importance of morale is the
organization’s general approach to management. Kalleberg (2003) notes
that firms facing similar competitive pressures may have a choice of strate-
gies. One approach is to attempt to build commitment to the organization
and create a sense that workers and management are part of a team. An
alternative is to avoid any sense of lasting commitments and motivate work-
ers by a combination of immediate rewards and punishments. One would
expect morale to have more effect on productivity in firms that followed the
first strategy. It should be noted that this hypothesis is not equivalent to the
claim that morale will be higher when management follows a strategy of
this type. Rather, the hypothesis is that in such firms, the state of morale
 Weakliem, Frenkel / Effects of Workplace Morale 341

will have more effect on productivity, for better or for worse. A firm that
tries to cultivate a sense of community may still suffer from low morale
because of lack of management skill, a history of poor labor relations, or
other factors, and the hypothesis predicts that such firms will have low pro-
ductivity. Conversely, high morale will not result in high productivity
unless the company follows a strategy that takes advantage of its potential.
As Whyte (1955) puts it:

If workers are well satisfied with their jobs and think well of management,
then high productivity does not necessarily follow. People may simply be
happy to be members of the organization and have no urge to contribute to its
goal of production. (p. 1)

The AWIRS contains two questions about management’s general strategy:

whether “this organisation devotes considerable resources toward having a
corporate ethic and culture” and whether the organization “devotes consider-
able resources to the management of this workplace’s human resources.”
Although it would be desirable to measure the way that management imple-
ments its reported principles, it is difficult to do so, given that there are
numerous possibilities, many of which involve day-to-day actions rather than
the existence of formal programs. Hence, the analysis in this article rests on
the assumption that there is some correlation between management’s stated
goals and actions. If any evidence of an interaction is found, identification of
the relevant objective conditions will be a task for future research.
A third general factor is size, which has a substantial effect on many
aspects of organizations. It is possible to make arguments for either an
increasing or decreasing effect of morale as size increases. On one hand,
large workplaces tend to have a more bureaucratic organization, with stan-
dardized systems of measurement, reward, and punishment. To the extent
that such systems are effective, workers will have less autonomy in larger
workplaces. It is also possible that in smaller workplaces, workers have
more sense of personal connection to the firm, providing them with extra
motivation to reward or punish their employer. Both of these considerations
imply that morale will be more important in smaller workplaces. On the
other hand, supervision may be easier in smaller workplaces, because the
owners or top managers can observe workers directly rather than relying on
the reports of subordinates. In this case, workers in small firms will have
less autonomy, potentially reducing the effects of morale. Because size may
be defined in several ways, none of which is clearly superior, we consider
several alternatives listed in Table 1.
342 Work and Occupations

Finally, the effect of morale might also depend on workers’ orientation

toward authority. If workers have a strong general sense of obligation to fol-
low orders, their work effort will be less dependent on morale. Although
the AWIRS does not contain direct measures of attitude toward authority, it
does have information educational attainment. Researchers such as Kohn
and Schooler (1969) have found that more educated people are consider-
ably more inclined to question authority.2 Hence, morale may be more
important in which the average educational level of the workforce is higher.

Paths of Influence
There are several channels through which morale may affect productiv-
ity. The most straightforward possibility is that higher morale leads to more
work effort, which in turn increases productivity. Workers with higher
morale will generally enjoy their work more and will also feel more deter-
mination to carry out the task even when the work itself is disagreeable.
An alternative path is suggested by recent discussions of “organizational
citizenship behavior”—activity that makes the organization work more
smoothly and effectively (Hodson, 2001; Organ, 1988). Some types of orga-
nizational citizenship behavior require little or no extra effort—for example,
giving good advice to a coworker or bringing a problem to the attention
of management. Such actions will enhance productivity by increasing the
effectiveness of coworkers’ efforts. Leibenstein (1975) observes that effort
alone will not necessarily increase productivity and may even reduce it—for
example, a worker who devotes effort to seeing that rules are followed to the
letter may make it more difficult for other workers to carry out their jobs.
At low levels of morale, workers’ efforts may be predominantly defensive,
directed toward protecting themselves against criticism or advancing their
own status at the expense of their coworkers. At higher levels of morale,
in contrast, workers will direct their efforts toward accomplishing the goals
of the organization. Also, Bakker and Guerts (2004) suggest that work can
be characterized by a state of “flow” or being “fully concentrated and
engrossed in one’s work, whereby time passes quickly” (p. 350). Workers in
this state are likely to perform more effectively, because there will be less
struggle and wasted effort. High morale may increase productivity by being
associated with greater levels of “flow” rather than greater effort.
A final way that morale may enhance productivity is by facilitating
agreement between workers and employers. There is always a good deal of
uncertainty about the effects of any proposed change to an organization.
Where morale is low, workers may suspect that any changes proposed by
 Weakliem, Frenkel / Effects of Workplace Morale 343

the employer are intended to reduce their wages, job security, or quality of
work. Where morale is high, they may be more willing to believe that pro-
posed changes will work to their benefit, making it easier for firms to adopt
the most effective practices and equipment.

Data and Measurement

The AWIRS was conducted by the Department of Employment and

Workplace Relations in 1995. The survey was given in a random sample of
about 2,000 workplaces with 20 or more employees, with the probability
of selection proportional to the size of the workforce (Morehead, Steele,
Alexander, Stephen, & Duffin, 1997). The AWIRS includes several separate
questionnaires. General information on the workplace was collected from
the senior workplace manager, whereas information on employment relations
was collected from the senior manager directly responsible for the area. In
some cases, especially in smaller workplaces, the general management and
employee relations questionnaires were filled out by the same person. Finally,
employees were randomly selected to receive a questionnaire concerning var-
ious characteristics of their jobs and their attitudes toward management. The
number of employees interviewed depended on the size of the workplace and
ranged from 1 to 77. The median number of workers interviewed was 9, and
the interquartile range was 6 to 14. Because of the official sponsorship, the
response rates to all parts of the survey were very high.
The combination of a representative national sample of workplaces and
a reasonably large number of respondents per workplace is an important
advantage of the AWIRS. There are a few other surveys based on national
samples of workplaces or employers, notably the National Organizations
Survey (NOS) in the United States and the Workplace Employment
Relations Survey (WERS) series in Great Britain. The NOS includes only
one respondent per workplace and consequently is not suitable for our pur-
poses. The WERS began in 1980 but did not include employee respondents
until 1998. The 1998 WERS is similar to the AWIRS in its general design
and content, but many of the specific questions are different, so that it is not
possible to use both in the same study. We chose to rely on the AWIRS
because of its superior measures of morale.
The unit of analysis is the workplace, because productivity is measured at
that level. We take the level of morale as given and do not attempt to explain
why average levels might differ among workplaces. Because morale may be
influenced by both individual-level and workplace-level characteristics, an
344 Work and Occupations

analysis that sought to explain differences in morale would involve a multi-

level model. The analysis of productivity differences, however, can be con-
ducted entirely at the workplace level, although morale and several other
independent variables are workplace averages of variables measured at the
individual level.

Productivity
In a sample covering workplaces producing different goods and services,
it is difficult to employ objective measures of productivity. Measures based
on physical output, such as those used by Appelbaum et al. (2000), are not
feasible. Tomaskovic-Devey and Skaggs (1999) use financial data to con-
struct measures of productivity in a diverse sample of workplaces. However,
the AWIRS does not contain financial data, and for reasons of confidential-
ity, it is not possible to identify the companies and add information from
outside sources. Hence, we use a subjective measure of productivity from a
question in the general management survey: “In your opinion, how does the
level of labour productivity here compare with your major competitors?”
Respondents chose answers from a 5-point scale, ranging from a lot lower
to a lot higher. This measure of productivity is also used in the recent stud-
ies of Moreton (1999) and Addison and Belfield (2001), which are based on
the WERS. Although a subjective judgment necessarily involves some mea-
surement error, managers have access to the relevant data and strong incen-
tives to get accurate knowledge about productivity.
There is a clear tendency for managers to give their own company the
benefit of the doubt: Nearly half think that their productivity is higher than
their competitors, and only 12% think that it is lower. However, for the pur-
poses of explaining differences in productivity, the important question is
not whether there is a tendency toward “optimism,” but whether any indi-
vidual differences in this tendency are correlated with the independent vari-
ables. A uniform tendency to evaluate performance favorably will affect the
intercept but will have no impact on the estimated effects of the indepen-
dent variables. Individual differences in optimism that are uncorrelated
with the independent variables will also have no effect on the estimated
effects of independent variables, because they can be regarded as part of the
regression error term (Machin & Stewart, 1990). There may be some sys-
tematic differences in managers’ optimism, but they are likely to depend on
characteristics of the managers themselves. Hence, they will not affect the
estimated effects of variables that are measured by workers’ reports—such
as our key variable, morale.
 Weakliem, Frenkel / Effects of Workplace Morale 345

Morale
As discussed above, we define morale as positive attitudes toward the
“prescribed activities of the group” (Manning, 1991, p. 455). Because it is
management that determines those activities, morale in this sense is largely
a matter of attitudes toward management.3 The AWIRS employee survey has
three questions concerning management at the workplace: whether it “does
its best to get along with employees,” whether it “can be trusted to tell things
the way they are,” and satisfaction with “the way management treat you and
others here.” It is possible to make conceptual distinctions between these
variables—for example, one can imagine a management that is trustworthy
but also rigid and authoritarian. However, all attitudes go together in prac-
tice, with correlations of between .70 and .75 at the workplace level.4 Hence,
we use the sum of the three items as an overall index of morale. Values range
from 3 to 9, with larger numbers indicating greater morale.

Control Variables
Economic theorists have proposed a number of production functions
relating output to capital and labor inputs. However, these do not provide
much guidance for the selection of control variables, because they include
constants that might represent the effects of a variety of factors, such as the
state of technology, the quality of management, or the organization of pro-
duction. The most relevant empirical works are those of Moreton (1999)
and Addison and Belfield (2001), who analyze productivity in a wide vari-
ety of workplaces. The choice of control variables is guided by their find-
ings and general considerations of plausibility:

1. Age: Older workplaces may be less well adapted to new conditions and
may have older equipment. On the other hand, firms that are less pro-
ductive are more likely to go out of business, so that the firms that have
survived longer will tend to be among the more efficient ones. In this
case, age is not a true cause of productivity but a proxy for variables that
contribute to productivity and are not separately measured. Age is mea-
sured by two variables, one for time operating at the current location, and
one for time that the firm has been producing its major product.
2. Introduction of new plant equipment or office machinery in the past 2
years: Plants with more up-to-date equipment are expected to be more
productive.
3. Size of workplace and of company: There may be economies or disec-
onomies of scale.
346 Work and Occupations

4. Multiplant firm: Workplaces that are part of a multiplant firm may benefit
from the ability to specialize, suggesting that they will be more productive.
On the other hand, multiplant firms may find it more difficult to coordinate
the activity of different plants, which could reduce productivity.
5. Unionization rate: Many economists have argued that unions reduce
productivity by restricting management’s power to organize production
in the most efficient way. On the other hand, Freeman and Medoff
(1984) argue that unions provide regular channels for workers to voice
their concerns and resolve grievances that would otherwise produce
turnover or covert resistance. Their account suggests that unionization
will generally increase productivity.
6. Number of union representatives in the workplace: This is an alterna-
tive measure of union strength.5
7. Number of unions: The existence of multiple unions may lead to disrup-
tion because of demarcation disputes and rivalry for members. Moreton
(1999) finds evidence that multiple unionism reduces productivity.
8. Workers’ average education: Productivity can be expected to depend
not only on the quantity of labor but also on its “quality.” Education is
an important aspect of labor quality, because it can affect a range of
skills and attitudes.
9. Average wage: Wages may be related to worker quality, because more
capable workers will be able to command higher wages. In this case,
higher wages would be associated with higher productivity. However,
managers may take account of labor costs when they answer the question
on productivity. If they think of productivity in terms of output relative to
labor costs, higher wages will tend to go with lower reported productivity.
Because labor markets are structured by occupation, levels of edu-
cation and wages must be understood in relative terms. An employer
seeking to attract high-quality workers for a particular job will need to
consider the typical wage rate for that job and similar jobs. Hence, in
the analysis, education and wages are adjusted for occupational group.
Specifically, the relative wage (w*) in a workplace is w* = ∑pi(wi-mi),
in which pi is the proportion of the workers in occupational group i, wi
the workplace average wage of workers in that occupation, and mi is the
average wage of workers in that occupation across all workplaces.
Thus, a value of 0 for workplace wages indicates that wages are normal
given the occupational composition of the workplace, and positive or
negative values indicate that they are above ore below what would be
expected. A corresponding adjustment is used for education.
10. Organization status: The AWIRS includes all types of workplaces, not
just private for-profit firms. Because the standards of productivity may
differ depending on the type of organization, we include dummy vari-
ables for each category distinguished in the AWIRS (private, government
 Weakliem, Frenkel / Effects of Workplace Morale 347

business enterprise, noncommercial statutory, federal public service,

state public service, other noncommercial, other commercial).

Intervening Variables
Two additional variables are involved in the hypotheses about how
morale influences productivity—work effort and the difficulty of reaching
agreement. The AWIRS employee survey contains a straightforward mea-
sure of effort—a question asking workers whether they put “a lot of effort”
into their job. Unfortunately, they were offered only a choice of agree, dis-
agree, or neither agree nor disagree. The great majority of workers chose
agree, and in about 40% of workplaces, all the employees surveyed chose
that response. Hence, although the variable is conceptually appropriate,
the measurement is not very precise. A measure of the difficulty in reaching
agreement can be obtained from the management survey, which asks
whether there are any productivity-related changes that management would
like to make but cannot. About 40% of managers mention no such changes,
48% mention one, and 12% mention two or more. The total number of
desired changes can be taken as an indicator of difficulty in reaching agree-
ment. In principle, an agreement can always be devised so that both work-
ers and management benefit from changes that enhance productivity—some
of the additional revenue can be used to compensate workers for any dis-
ruption and job loss. Collom (2003) finds that greater worker influence over
production is associated with higher managerial salaries and suggests that
managers share in gains resulting from greater influence. Hence, failure to
implement a change would indicate that there was some barrier to develop-
ing an agreement that shared the benefits in a mutually satisfactory way.

Results

An analysis of variance shows statistically significant workplace differ-

ences in levels of morale.6 In principle, it is possible that some of these differ-
ences reflect differences in workforce composition—if some types of workers
tend to have higher morale than others, workplaces in which those types are
more numerous will have higher morale. However, the workplace rankings are
similar regardless of whether controls for occupation and other individual
characteristics such as gender, age, and education are included. Hence, the
estimates of workplace morale used in the analysis are not adjusted for the
composition of the workforce. Because the workplace averages are obtained
348 Work and Occupations

from samples of the employees, they are affected by sampling error, and the
magnitude of this error increases as the sample size decreases. The accuracy
of estimates of workplace morale can be increased by taking account of sam-
pling error. We use a procedure suggested by James and Stein (1961; see Efron
& Morris, 1975, for a simpler exposition), which uses a weighted average
of the overall mean and the workplace mean. Specifically, the weights are
k/(k + n) for the overall mean and 1-k/(k + n) for the observed workplace mean,
in which n is the number of workers surveyed in a workplace and k is an esti-
mate of the ratio of individual-level to workplace-level variation. In these data,
k is approximately 5.0. As mentioned above, the possible values of morale
range from 3 to 9. Using the James-Stein estimates, the workplace averages
cover most of the theoretically possible range. Figure 1 shows a histogram of
the levels of morale estimated by the James-Stein procedure. The figure makes
it clear that the differences among workplaces are not only statistically signif-
icant but also large enough to be of substantive interest.

General Effects of Morale

Table 2 shows estimates of the general effects of morale on productivity.
Because the dependent variable has a small number of ordered categories, we
use the ordinal logistic regression proposed by McCullagh (1980), which can
be thought of as a regression on a continuous latent variable representing per-
ceived relative productivity. Analyses using conventional linear regression
produced similar results. The first column gives results from a baseline model
including only the control variables. The second and third columns add two
alternative measures of morale—the observed workplace average and the
James-Stein estimate. In general, measurement error in an independent vari-
able will reduce the size and t ratio of the estimated effect. Both estimates are
positive and significant, but those using the James-Stein measure are sub-
stantially stronger. Consequently, it appears that the procedure is successful
in reducing measurement error, and we will rely on the James-Stein estimates
in subsequent analyses. Because it is difficult to be sure that all appropriate
control variables have been included, we also estimated the models adding
dummy variables for the 46 industry categories distinguished in the AWIRS.
In this equation, the estimate for morale remains significant, and the estimate
and t statistic actually become somewhat larger.
The model in the fourth column allows morale to have a nonlinear effect
by including a squared term. The hypothesis of a diminishing effect implies
that this term will be negative. Although it has the expected sign, the esti-
mate is small and does not approach statistical significance. An alternative
 Weakliem, Frenkel / Effects of Workplace Morale 349

Figure 1
Distribution of Morale by Workplaces

160

140

120
Number of Cases

100
Std. Dev = .65
80 Mean = 5.37
N = 1,819.00
60

40

20

0
3.
3.
4.
4.
4.
4.
5.
5.
5.
5.
6.
6.
6.
6.
7.
7.
7.
63
88
13
38
63
88
13
38
63
88
13
38
63
88
13
38
63

Estimated Workplace Morale

way to model nonlinearity is to include a dummy variable for workplaces

with unusually low morale. Several different definitions of low morale were
attempted, and the dummy variable was never significant. Thus, the hypoth-
esis of a decreasing effect of morale is not supported.
We next consider the size of the morale effects. In the ordered logit
model, the estimates involve effects on the cumulative probabilities—that is,
the probability of being at or below a given level. More specifically, the odds
of being at or below Category j relative to being above Category j will
increase by a factor of eβ for every one-unit increase in morale, in which β is
the estimated effect of morale. Table 3 gives examples of how the estimated
350
Table 2
Estimated Effects of Morale on Productivity (N = 1,493)
Model 1 SE Model 2 SE Model 3 SE Model 4 SE

Union density –.065 .158 .006 .159 .017 .156 .017 .156
Workplace size (log) .090* .058 .114* .058 .117** .058 .118** .058
Average education –.049 .065 –.038 .066 –.037 .066 –.037 .066
Number of unions –.038 .036 –.034 .036 –.034 .036 –.033 .036
Number of delegates –.017 .011 –.017 .011 –.016 .011 –.016 .011
Firm size –.039 .029 –.043 .029 –.045 .029 –.044 .029
Time at location (log) –.102** .050 –.092* .050 –.090* .050 –.090* .050
Average wage –.002 .009 –.002 .009 –.002 .009 –.002 .009
Time producing –.129** .054 –.132** .054 –.133** .054 –.133** .054
Multiplant .002 .166 –.048 .167 –.051 .167 –.052 .167
New office equipment .218** .098 .213** .099 .214** .099 .215** .099
New machinery –.012 .109 –.021 .110 –.017 .109 –.018 .110
Morale (observed) .132*** .045
Morale (estimated) .271**** .081 .410** .183
Morale2 (estimated) –.013 .087

*p < .10. **p < .05. ***p < .01.

 Weakliem, Frenkel / Effects of Workplace Morale 351

Table 3
Predicted Distribution of Productivity,
by Level of Morale (in Percentages)
Productivity Low Morale Baseline High Morale

A lot higher 12.3 15.6 19.5

A little higher 30.3 33.7 36.5
About the same 41.8 38.3 33.7
A little lower 12.8 10.3 8.7
A lot lower 2.7 2.1 1.6
Total 100.0 100.0 100.0

Note: Baseline is the distribution of reported productivity in the entire sample. High morale
is 1.0 unit (about 1.5 standard deviations) above the mean level of morale, whereas low
morale is 1.0 unit below the mean.

effects from model 2 of Table 2 translate into the original probabilities. The
baseline is the distribution of productivity is the entire sample. The compar-
isons are the predicted distributions if morale were increased by one unit
(about 1.5 standard deviations) in one case and reduced by one unit in the
other, while holding all other independent variables constant. For example,
a one-unit reduction in morale would reduce the chance that productivity is
rated a lot higher than most competitors from .156 to .123, and a one-unit
increase would increase the chance to .195. That is, having high rather than
low morale would increase the probability of having the highest level of pro-
ductivity by about 50%. Whether an effect of this size should be considered
large or small is a matter of judgment, but the differences seem large enough
to be of practical interest.

Interactions
The analysis of interactions is largely exploratory, because even with a
sample of this size, the estimates of interaction terms have relatively large
standard errors. To give a sense of magnitude, we show the minimum and
maximum effects of morale implied by the estimated interaction term. For
example, the smallest possible value of autonomy is 1, when all respondents
in a workplace say that they have no influence, and the largest is 4, when all
say that they have a lot of influence. The minimum and maximum column
shows the implied effect of morale on productivity at the minimum and max-
imum possible values of autonomy. Because a large number of possible
interactions are considered, it is not practical to include them all in a single
352 Work and Occupations

model. Rather, the figures are from models in which a single interaction
effect added to the independent variables included in Model 2 in Table 2.
The estimated interactions involving autonomy, occupation, and firm
size are all well short of statistical significance and small in magnitude.7 For
example, the estimated effect of morale is only about 25% larger in fully
unionized workplaces than in workplaces with no union members (.306 vs.
.245), and the t ratio of the interaction effect is less than 0.5. However, the
estimated interactions involving management strategy and the importance
of quality are substantial and approach statistical significance. The interac-
tion involving stated emphasis on human resources is significant at the 10%
level, and the estimate suggests that the effect of morale disappears entirely
in organizations that do not claim to emphasize human resources. The inter-
action involving emphasis on corporate culture is also substantial and has a
t ratio of greater than 1.0. The estimated interaction involving quality as the
most important factor in success is close to the 10% significance level,
whereas the interaction involving preference for quality improvement is
also in the expected direction and has a t ratio of approximately 1.0. There
is little or no evidence of interactions involving the other variables.
One response to these results would be simply to conclude that there is no
compelling evidence of any interactions. That is, we fail to reject the hypoth-
esis that the effects of morale are uniform with respect to all of the variables
considered here. However, given the large standard errors, there is also a sub-
stantial possibility of a Type II error: failure to reject a false null hypothesis.
The pattern of results in Table 4 suggests that the possibilities of interactions
with management strategy and emphasis on quality deserve further consider-
ation. The items can be combined into two indexes: one composed of human
relations and corporate culture; the other of preference for quality improve-
ment and quality as the most important factor in success.8 Also, accounts of
high-performance workplaces suggest there will be some overlap between
these two features—when management regards quality as important, it will
attempt to foster organizational commitment. Hence, it is reasonable to com-
bine all items into a single index, which could be interpreted as the degree to
which the workplace follows the high-performance model.
Table 5 shows fit statistics and relevant coefficients from models includ-
ing interactions of morale and the indexes. The interactions with each index
are significant at the 10% level, and the interaction with the combined index
is significant at the 5% level. The magnitude of the estimated effects is
substantial—the estimates from Model 5 imply an average morale effect of
about .23 with a standard deviation of about .16. That is, the estimates sug-
gest that in workplaces that are farthest from the high-performance model,
 Weakliem, Frenkel / Effects of Workplace Morale 353

Table 4
Estimated Interactions Between Morale and Other Factors
Minimum Maximum
Estimate SE Morale Effect Morale Effect N

Autonomy-1 –.039 .158 .226 .343 1,493

Autonomy-2 –.174 .166 .096 .618 1,493
Autonomy-3 –.085 .150 .173 .428 1,493
Unionization .061 .211 .245 .306 1,493
Delegates –1.86 2.99 –.115 .313 1,488
Managers or professionals .350 .398 .208 .558 1,493
Operatives or laborers .114 .256 .294 .398 1,493
Choose quality improvement .082 .083 –.008 .320 1,489
Quality most important .279 .172 .201 .480 1,493
Corporate ethic .104 .082 –.056 .360 1,489
Human resources .144* .081 –.169 .407 1,492
Workplace size .009 .070 .263 .315 1,493
Firm size .000 .033 .275 .280 1,492
Multiplant –.173 .210 .250 .423 1,489
Average education .045 .102 .143 .480 1,493

*p < .10.

morale has little or no effect on productivity. The estimates of statistical

significance in Table 5 should be regarded with some caution, because the
hypotheses were obtained after examination of a larger number of non-
significant results. However, the results are suggestive, and the construction
of the index is consistent with the account of high-performance workplaces
offered in Appebaum et al. (2000).

Paths Connecting Morale and Productivity

As discussed above, the most straightforward possibility is that workers
with high morale are willing to put more effort into their jobs and that more
effort leads to more productivity. A second possibility is that high morale
increases the effectiveness of work effort, implying that the effects of
morale and work effort on productivity are interactive. A third possibility is
that morale may facilitate agreement between workers and management,
making it easier to introduce productivity-enhancing changes.
Table 6 shows results from regressions of average reported effort and the
number of desired changes desired by management on workplace morale, with
the same controls used in the analysis of productivity. The effort equation is a
354
Table 5
Interactions of Morale With Management Strategy and Importance of Quality (N = 1,483)
Model 1 SE Model 2 SE Model 3 SE Model 4 SE Model 5 SE

Morale .212*** .082 –.302 .290 –.190 .235 –.530 .333 –.533 .333
Strategy .094*** .033 –.225 .176 .094*** .033 –.183 .264
Quality .168*** .045 .167 .045 –.265 .241 –.166 .364
High-performance model –.173 .130
Morale × Strategy .088* .048 .071 .049
Morale × Quality .120* .065 .097 .067
Morale × High-Performance .081** .035
Model
–2 log-likelihood 3,907.7 3,904.5 3,904.5 3,902.5 3,904.0

Note: Models also include the control variables in Table 2. Estimates for the control variables are omitted to save space.
*p < .10. **p < .05. ***p < .01.
 Weakliem, Frenkel / Effects of Workplace Morale 355

Table 6
Effect of Morale on Effort and Number
of Desired Productivity Changes
Effort SE Number of Changes SE

Union density –3.972*** 1.295 0.567*** 0.168

Workplace size (log) 0.659 0.434 0.066 0.056
Average education –0.395 0.514 0.074 0.067
Number of unions –0.002 0.274 0.051 0.035
Number of delegates 0.692 18.67 3.167 2.403
Firm size –0.010 0.223 –0.005 0.029
Time producing 0.552 0.383 –0.023 0.050
Average wage –0.106 0.070 –0.018 0.080
Time at address (log) –0.093 0.418 0.122* 0.055
Multiplant –0.568 1.302 0.083 0.170
New office equipment –0.299 0.767 0.229** 0.100
New machinery 0.032 0.850 0.072 0.110
Morale (estimated) 3.287*** 0.621 –0.018 0.080
N 1,670 1,666

*p < .10. **p < .05. ***p < .01.

standard linear regression, whereas the number of changes equation is an

ordered logit. Morale is clearly associated with more effort, with a t ratio of
about 5.0, but there is no evidence that morale affects the number of changes
that management would like to make.
Table 7 gives a summary of the effect of morale and work effort on pro-
ductivity. The control variables used in the previous productivity equations
are also included but are omitted to save space. The model in the first col-
umn includes separate effects of morale and work effort, with no interac-
tion. The estimate for morale is positive and statistically significant,
whereas the estimate for work effort is positive but not statistically signifi-
cant. Because of the weaknesses in the measurement of effort, the results
should not be taken to imply that work effort has no effect on productivity
or even that morale has more effect than work effort. However, they suggest
that morale has some effect on productivity apart from work effort.
The second column adds an interaction between morale and work effort,
which is positive and significant at the .10 level. The model in the third
column provides an alternative way of representing interactions, dividing
workplaces into groups of low, medium, and high morale and estimating
the effects of work effort separately in each.9 The results from Model 3 are
notably stronger than those of Model 2, even though the groups were not
356 Work and Occupations

Table 7
Effects of Morale and Work Effort on Productivity (N = 1,491)
Model 1 SE Model 2 SE Model 3 SE

Morale .265*** .081 –1.944 1.209 .220 .177

Effort .235 .293 –2.498 1.538
Morale × Effort .774* .423
Effort × Low Morale –.172 .539
Effort × Medium Morale –.102 .417
Effort × High Morale 1.774*** .628
–2 log-likelihood 3,954.3 3,951.1 3,943.9

Note: All models also include control variables from Table 2. Estimated effects of the
controls are omitted to save space.
*p < .10. **p < .01.

chosen to maximize the relationship. The estimates from this model imply
that work effort has essentially no relationship to productivity in the two
lower groups but a positive effect in the highest group. The results of
Models 2 and 3 suggest that the effect of work effort on productivity
increases with morale. As suggested above, this may mean that workers
direct their efforts in more useful ways: Rather than considering only what
is best for themselves, they also consider what is best for their coworkers
and the general goals of the organization. Unfortunately, the AWIRS does
not contain any variables that could be used to evaluate this interpretation.
However, the results suggest that the usual model of morale affecting pro-
ductivity through a simple increase in work effort may be incomplete.

Summary

This study considered three major questions: whether morale affects

productivity, whether the effects are uniform or differ depending on some
other factors, and the paths by which any effects take place. On the first
point, there is clear evidence that higher morale is associated with higher
productivity and no sign that increases in morale have diminishing returns.
Admittedly, because productivity is measured by managers’ estimates, it is
necessarily subject to error. Future research could consider the validity and
reliability of managerial estimates of productivity. However, because the
estimates of morale are obtained from workers rather than management, we
can be confident that the estimated relationship is real, not an artifact of
measurement.
 Weakliem, Frenkel / Effects of Workplace Morale 357

Conclusions on the second question are more tentative, because estimates

of interaction effects have a good deal of error, even with a relatively large
sample. There is no evidence of interactions with skill or direct measures of
autonomy, although these factors have been prominent in many previous
accounts. However, there is some evidence that morale is more important
when quality is more important to success and management emphasizes
human relations and a corporate culture. Taken together, these two features
can be interpreted as a high-performance model. The results suggest that
morale is important in workplaces that follow this model but has little or no
effect in those workplaces that do not. Finally, we find that morale is asso-
ciated with greater work effort and that work effort has a larger positive
impact on productivity in which morale is higher. The existence of this inter-
action effect suggests that higher morale helps to redirect worker efforts in
ways that promote productivity. The models suggest that there is also some
effect of morale apart from work effort, although no evidence was found for
the hypothesis that morale makes it easier for workers and management to
reach agreement on measures that enhance productivity.

Conclusions

These results support an assumption that is implicit in many accounts of

workplace organization and change—that positive feelings about the orga-
nization and management lead to higher productivity. Moreover, this effect
is not limited to a specific type of workplace—none of the interactions with
workforce composition, unionization, or size even approached statistical
significance, and most were small in magnitude. Thus, the results of this
analysis generally support arguments in favor of the high-performance
model of workplace organization: Enhancing morale not only makes work
more pleasant but also produces productivity gains.
However, there is some evidence that the influence of morale depends on
management strategy. Fifty years ago, Shister (1950) criticized advocates
of workplace reform for failing to clearly address the question of managerial
interests, and his point is applicable to many contemporary discussions as
well. If the relationship between morale and productivity is universal, man-
agers always have an interest in cultivating morale, even if they have no intrin-
sic concern for their employees. In this case, managerial neglect of morale
must be the result of factors such as incompetence or ignorance. On the other
hand, if there are circumstances under which morale does not affect produc-
tivity, ignoring morale may be the optimal strategy. The results in Table 5 sug-
gest that this may be the case in some workplaces. For example, in Model 4,
358 Work and Occupations

the estimated effect of a human relations approach is –.183 + .071 × Morale,

which implies a “break even” value of about 2.6 for morale. If morale is below
that level, as it is in about 7% of workplaces, paying attention to a human
resources and attempting to develop a corporate ethic are associated with lower
productivity. Particularly if management is not confident of its ability to raise
morale, de-emphasizing human relations may be a more appealing strategy
when the pre-existing level of morale is low. If this is the case, changes that
enhance morale would be likely to require outside intervention.
Consequently, further examination of possible variations in the influence
of morale is needed. Additional analysis of the paths through which morale
affects productivity would also be fruitful. Our analysis suggested that the
effect is not simply a result of morale producing increased effort but also of
increasing the effectiveness of efforts. More detailed studies of particular
workplaces could shed light on the ways in which this enhancement might
take place.
The magnitude of the effect of morale on productivity is also an impor-
tant question, especially for practice. If measures that would enhance morale
have costs, as many do, decisions about whether to implement them require
information about the size of the productivity gains. The measure of pro-
ductivity used here is subjective and does not have a well-defined scale, so
it does not permit strong conclusions on the size of effects. Objective mea-
sures of productivity such as those used by Tomaskovic-Devey and Skaggs
(1999) are necessary to obtain more accurate estimates.
Ultimately, the relationship between morale and performance is a dynamic
one. Authors such as Fox (1974) and Hodson (2002) have noted the possibil-
ity of vicious or virtuous cycles of relations between workers and manage-
ment. This analysis touched on the issue of long-term effects by considering
the hypothesis that morale facilitates agreement between workers and man-
agement. Although no evidence for this hypothesis was found, agreement
is measured by managers’ statements about changes that they would like to
make, not by information on changes that are actually attempted. Hence,
introducing a dynamic element, whether by traditional case studies or by the
analysis of longitudinal data, is the crucial next step in research.

Notes
1. It should be noted that this sense of autonomy is somewhat narrower than the usual
sense. A worker who has control over how to do the work but whose output is easily moni-
tored is not autonomous in the sense used here.
2. The World Values Surveys (Inglehart, Basanez, Diaz-Medrano, Halman, & Luijkx,
2004) include a question on whether “one should follow one’s superiors instructions even
 Weakliem, Frenkel / Effects of Workplace Morale 359

when one does not fully agree” or follow instructions “only when one is convinced that they
are right.” Educated people are substantially more likely to choose the second answer in nearly
all nations, including Australia.
3. Another component of morale in this sense might be views about the value of the prod-
uct or service produced, but the Australian Workplace Industrial Relations Survey does not
contain any information on this topic.
4. Although all categories of workers are surveyed, we exclude managerial employees
from these calculations, because their views on these topics may diverge from those of ordi-
nary workers.
5. The survey question refers to “union delegates,” which are equivalent to American shop
stewards.
6. The F statistic is 2.683 with 1,824 degrees of freedom in the numerator and about
15,000 in the denominator, far beyond the 0.1% critical value of 1.13.
7. In addition to the variables shown in Table 4, the Australian Workplace Industrial
Relations Survey includes three others that might be considered forms of autonomy, control
over “when you start and finish work,” over “the way that this workplace is managed or organ-
ised,” and “decisions which affect you at this workplace.” The estimated interactions for these
variables also do not approach significance.
8. Because both the management strategy variables are measured on the same scale (1 to 5),
the index is the unweighted sum of the two items. The two quality variables are measured on
different scales, so they are weighted to give them approximately equal influence in the index.
9. The groups are defined as ranking in the lowest 25%, 25% to 75%, and top 25% of morale.

References
Addison, J. T., & Belfield, C. R. (2001). Updating the determinants of firm performance:
Estimation using the 1998 UK Workplace Employee Relations Survey. British Journal of
Industrial Relations, 39, 341-366.
Appelbaum, E., Bailey, T., Berg, P., & Kalleberg, A. L. (2000). Manufacturing advantage:
Why high-performance work systems pay off. Ithaca, NY: ILR Press.
Bakker, A. B., & Guerts, S. E. B. (2004). Toward a dual-process model of work-home inter-
ference. Work and Occupations, 31, 345-366.
Bewley, T. F. (1999). Why wages don’t fall in a recession. Cambridge, MA: Harvard
University Press.
Blinder, A. S., & Choi, D. H. (1990). A shred of evidence on theories of wage stickiness.
Quarterly Journal of Economics, 105, 89-96.
Campbell, C. M., & Kamlani, K. S. (1997). The reasons for wage rigidity: Evidence from a
survey of firms. Quarterly Journal of Economics, 112, 760-789.
Collom, E. (2003). Two classes and one vision: Managers’ and workers’ attitudes toward
workplace democracy. Work and Occupations, 30, 62-96.
Efron, B., & Morris, C. (1975). Data analysis using Stein’s estimator and its generalizations.
Journal of the American Statistical Association, 70, 311-319.
Fox, A. (1974). Beyond contract: Work, power, and trust relations. London: Faber & Faber.
Freeman, R. B., & Medoff, J. L. (1984). What do unions do? Cambridge, MA: Harvard
University Press.
Frenkel, S. J. (2003). The embedded character of workplace relations. Work and Occupations,
30, 135-153.
360 Work and Occupations

Grunberg, L., Anderson-Connolly, R., & Greenberg, E. S. (2000). Surviving layoffs: The effects
on organizational commitment and job performance. Work and Occupations, 27, 7-31.
Haire, M. (1954). Industrial social psychology. In G. Lindzey (Ed.), Handbook of social
psychology (pp. 1104-1123). Cambridge, MA: Addison-Wesley.
Harter, J. K., Schmidt, F. L., & Hayes, T. L. (2002). Business-unit-level relationship between
employee satisfaction, employee engagement, and business outcomes: A meta-analysis.
Journal of Applied Psychology, 87, 268-279.
Hodson, R. (2001). Dignity at work. Cambridge, UK: Cambridge University Press.
Hodson, R. (2002). Management citizenship behavior and its consequences. Work and
Occupations, 29, 64-96.
Inglehart, R., Basanez, M., Diaz-Medrano, J., Halman, L., & Luijkx, R. (2004). Human beliefs
and values. Mexico City: Siglo XXI.
James, W., & Stein, C. (1961). Estimation with quadratic loss. In J. Neyman (Ed.),
Proceedings of the fourth Berkeley symposium on mathematical statistics and probability,
(Vol. 1, pp. 361-379). Berkeley: University of California Press.
Judge, T. A., Thoresen, C. J., Bono, J. E., & Patton, G. K. (2001). The job satisfaction–job per-
formance relationship: A qualitative and quantitative review. Psychological Bulletin, 127,
376-407.
Kalleberg, A. (2003). Flexible firms and labor market segmentation. Work and Occupations,
30, 154-175.
Kohn, M. L., & Schooler, C. (1969). Class, occupation, and orientation. American
Sociological Review, 34, 659-678.
Leibenstein, H. (1975). Aspects of the X-efficiency theory of the firm. Bell Journal of
Economics, 6, 580-606.
Lincoln, J. R., & Kalleberg, A. L. (1990). Culture, control, and commitment. New York:
Cambridge University Press.
Machin, S., & Stewart, M. (1990). Unions and the financial performance of British private sec-
tor establishments. Journal of Applied Econometrics, 5, 327-350.
Manning, F. J. (1991). Morale, cohesion, and esprit de corps. In R. Gal & A. D. Mangelsdorff
(Eds.), Handbook of military psychology (pp. 453-470). Chichester, UK: Wiley.
McCullagh, P. (1980). Regression models for ordinal data. Journal of the Royal Statistical
Society (series B), 42, 109-142.
Morehead, A., Steele, M., Alexander, M., Stephen, K., & Duffin, L. (1997). Changes at work:
The 1995 Australian Workplace Industrial Relations Survey. Melbourne. Australia:
Addison Wesley Longman.
Moreton, D. (1999). A model of labour productivity and union density in British private sec-
tor unionised establishments. Oxford Economic Papers, 51, 322-344.
Organ, D. W. (1988). Organizational citizenship behavior. Lexington, MA: D. C. Heath.
Organ, D. W. (1997). Toward an explication of ‘morale’: In search of the M factor. In C. L.
Cooper & S. E. Jackson (Eds.), Creating tomorrow’s organizations (pp. 493-504).
Chichester, UK: Wiley.
Perrow, C. (1986). Complex organizations: A critical essay (3rd ed.). New York: Random House.
Porter, L. W., Steers, R. M., Mowday, R. T., & Boulian, P. V. (1974). Organizational commit-
ment, job satisfaction, and turnover among psychiatric technicians. Journal of Applied
Psychology, 59, 603-609.
Roethlisberger, F. J. (1941). Management and morale. Cambridge, MA: Harvard University Press.
Shister, J. (1950). Trade union policies and nonmarket values. American Economic Review, 40,
292-305.
 Weakliem, Frenkel / Effects of Workplace Morale 361

Slichter, S. (1961). Industrial morale. In S. Slichter (Ed.), Potentials of the American economy
(pp. 167-183). Cambridge, MA: Harvard University Press. (Original work published 1920)
Tomaskovic-Devey, D., & Skaggs, S. (1999). An establishment-level test of the statistical dis-
crimination hypothesis. Work and Occupations, 26, 422-445.
Whyte, W. F. (1955). Money and motivation: An analysis of incentives in industry. New York:
Harper.

David L. Weakliem is a professor of sociology at the University of Connecticut. His main

research interests are in political sociology, the sociology of work, and quantitative methods.
He current research projects include a study of public opinion on income inequality and a his-
torical analysis of changes in class politics.

Stephen J. Frenkel is a professor specializing in the sociology of work and organization in

the Australian Graduate School of Management. His current research focuses on globalization
and work, particularly the experience of human resource managers, service workers and
knowledge workers in the advanced societies, and labor standards in Asia.

View publication stats