Nursing and Health Sciences (2007), 9, 205–211

Research Article

The psychosocial work environment and burnout among

Swedish registered and assistant nurses: The main,
mediating, and moderating role of empowerment
Jacek Hochwälder bsc, msc, phd
Stockholm Centre of Public Health, Stockholm and Mid Sweden University, Östersund, Sweden

Abstract The aim of the present study was to explore: the main effect of empowerment on burnout; empowerment as
a mediator between the work environment and burnout; and empowerment as a moderator of the association
between the work environment and burnout. In order to explore these effects, multiple regression analyses
were performed on questionnaire data from 838 registered nurses and 518 assistant nurses in Sweden. The
analyses showed that: empowerment has a negative association to burnout; empowerment has a mediating
effect between the work environment (especially for control and social support) and burnout; and the
moderating effect of empowerment on the association between the work environment and burnout was weak.
The results suggest that: empowerment explains variation with regard to burnout over and above what can be
explained by established work situation dimensions; the improvement of the work environment is associated
with a higher sense of empowerment which, in turn, is related to lower degrees of burnout; and individual and
group differences should be considered in workplace health promotion.

Key words burnout, empowerment, nurses, work environment.

INTRODUCTION association between empowerment and ill health (Waller-

stein, 1992). This assertion has been confirmed by a number
It is well-known that a poor work environment can result in
of studies that have observed a negative association between
burnout (e.g. McGrath et al., 2003). An important task for
empowerment and burnout (e.g. Laschinger et al., 2003;
researchers is to find the variables that might protect against
Hochwälder & Bergsten-Brucefors, 2005). A second theoreti-
burnout. Empowerment is a relatively new variable waiting
cal assertion is that empowerment mediates the effects of the
to be tested in the battle against burnout (Lee & Koh, 2001).
work environment on behavioral outcomes or, in other
One of the most established definitions of empowerment
words, that various work environment aspects affect various
has been proposed by Spreitzer (1997). Spreitzer focuses on
behavioral outcomes to some extent via empowerment
psychological empowerment at the workplace, which she
(Spreitzer, 1997). This assertion has received confirmation in
defines as a four-dimensional concept consisting of meaning
a study by Laschinger et al. (2003), who found that a good
(a fit between the requirements of the job tasks and the
work environment results in a higher sense of empowerment
subject’s own values), competence (the subject’s belief that
which, in turn, contributes to preventing burnout. A third
he or she possesses the skills and abilities necessary to
theoretical assertion is that empowerment can moderate the
perform a job or task well), self-determination (the subject’s
effects of various work environment variables on various
feeling of having control over his or her own work), and
behavioral outcomes or, in other words, that the effect of a
impact (the belief that the subject has a significant influence
given work environment variable on a given behavior
over strategic, administrative, or operational outcomes at
outcome might depend on the level of empowerment. No
work). The four dimensions can be used as four separate
study could be located that explored the role of empower-
indicators of empowerment or combined into a single overall
ment as a moderator on the association between the work
indicator of empowerment (Spreitzer, 1995a).
environment and burnout.
A general theoretical assertion is that empowerment has a
The aim of the present study was to explore, for registered
main effect on ill-health, which manifests itself in a negative
nurses and assistant nurses separately because of differences
with regard to a number of fundamental variables (e.g. edu-
cation, work tasks, and wages): (i) the main effect of empow-
Correspondence address: Jacek Hochwälder, Unit of Mental Health, Stockholm
erment on burnout; (ii) empowerment as a mediator between
Centre of Public Health, PO Box 175 33, SE-118 91 Stockholm, Sweden. Email:
jacek.hochwalder@miun.se the work environment and burnout; and (iii) empower-
Received 9 November 2006; accepted 9 March 2007. ment as a moderator of the association between the work

© 2007 The Author doi: 10.1111/j.1442-2018.2007.00323.x

Journal Compilation © 2007 Blackwell Publishing Asia Pty Ltd.
206
environment and burnout. The nurses were studied with
regard to burnout because burnout has been recognized as
one of the major work-related health problems among
health-care personnel (e.g. Duquette et al., 1994; McGrath
et al., 2003). The work environment was studied in terms of
job demands, control, and social support, which constitute
the fundamental components of the well-established job
demand-control-support model (Karasek & Theorell, 1990).
A number of variables that have similarities to empower-
ment have been studied in relation to the three work envi-
ronment dimensions and to various ill health indicators. For
example, sense of coherence has been studied by Söderfeldt
et al. (2000), locus of control by Rodríguez et al. (2001), and
self-efficacy by Schaubroeck and Merritt (1997). The most
essential difference between these variables and empower-
ment is that empowerment is more related to the work situ-
ation and is easier to affect.
METHOD
Participants
A total of 1356 nurses participated voluntarily in the present
study. The participants were recruited from three hospitals
and two primary health-care centres in the Stockholm area.
The overall response rate was 67%. Eight-hundred-and-
thirty-eight of the participants were registered nurses (94.5%
women and 5.5% men, mean age = 42.60 years, and standard
deviation [SD] = 10.35 years) and 518 were assistant nurses
(95.4% women and 4.6% men, mean age = 45.30 years, and
SD = 10.58 years).
Questionnaire
The participants completed a self-rating questionnaire com-
posed of 53 questions. The variables and how they were mea-
sured are now described.
Gender and age were noted. The subjects marked given
alternatives to indicate their gender and wrote down their age.
The psychosocial work environment was measured by a
scale developed by Karasek and Theorell (1990). The scale
measures three dimensions: demand, control, and social
support. The total scale consists of 17 items and demand is
measured by five items, control is measured by six items, and
social support is measured by six items. Examples of items for
each of the three dimensions are as follows: “Does your work
demand that you work very hard?” (demand), “Does your
work include you doing some tasks over and over again?”
(control), and “I feel comfortable with my workmates”
(social support). The responses were given on a four-point
Likert scale, ranging from 1 (“No, almost never”; alterna-
tively, “Not at all”) to 4 (“Yes, often”; alternatively, “Very
well”). An average value on each dimension was computed
for each respondent. For each of the three dimensions, the
theoretical minimum value was 1 and the theoretical
maximum value was 4. Higher values indicate higher levels of
demand, control, and social support.
Empowerment was measured with Spreitzer’s empower-
ment scale (Spreitzer, 1995b). The scale consists of four
© 2007 The Author
Journal Compilation © 2007 Blackwell Publishing Asia Pty Ltd.
J. Hochwälder
dimensions: meaning, competence, self-determination, and
impact. The total scale consists of 12 items and each of the
four dimensions is measured by three items. Examples of
items for each of the four dimensions are as follows: “The
work I do is meaningful to me” (meaning), “I am self-assured
about my capabilities to perform my activities” (compe-
tence), “I have considerable opportunity for independence
and freedom in how I do my job” (self-determination), and
“My impact on what happens in my department is large”
(impact). The response scale was a seven-point Likert scale,
ranging from 1 (strongly disagree) to 7 (strongly agree). In
order to make a concise presentation and analysis of the
mediating and moderating roles of empowerment, one mean
score was computed for each participant over the four
dimensions, as in some of Spreitzer’s (1995b) studies. This
means that the theoretical minimum value was 1 and the
maximum value was 7. Higher values indicate higher levels of
experienced empowerment.
Burnout was measured with the Maslach Burnout Inven-
tory, developed by Maslach and Jackson (1981; 1986). The
scale consists of three dimensions: emotional exhaustion,
depersonalization, and personal accomplishment. The total
scale consists of 22 items and emotional exhaustion is mea-
sured by nine items, depersonalization is measured by five
items, and personal accomplishment by eight items.
Examples of items for each of the three dimensions are as
follows: “I feel emotionally drained by my work” (emotional
exhaustion), “I feel I treat some recipients as if they were
impersonal objects” (depersonalization), and “In my work, I
deal with emotional problems very calmly” (personal accom-
plishment). The response scale was a seven-point Likert
scale, ranging from 0 (never) to 6 (every day). As recom-
mended by Maslach (Maslach & Jackson, 1986), a mean score
was computed for each participant on each of the three
dimensions. This means that on each dimension, the theoreti-
cal minimum was 0 and the theoretical maximum was 6. For
the dimensions, emotional exhaustion and depersonalization,
greater values indicate higher levels of burnout, whereas for
the dimension of personal accomplishment, lower values
indicate higher levels of burnout.
Procedure
The questionnaire, with written instructions accompanied by
a stamped addressed envelope, was sent to the home address
of the participants. The data were collected during 21 con-
secutive days. Where necessary, two reminders were sent
3 weeks apart. The study was approved by the ethical com-
mittee of the Karolinska Institute in Stockholm.
Analyses
All descriptions and analyses were performed using the SPSS
14.0 program (Norusis, 2006) and separately for registered
nurses and assistant nurses. Before performing the regression
analyses, all the variables (except for gender, which was
dummy-coded) were z-scored centered in order to avoid
problems with unessential multicollinearity (Cohen et al.,
Empowerment 207

2003). Furthermore, the data were checked for multicol-
linearity using two measures: the tolerance and the variance
inflation factor (Kleinbaum et al., 1988). One hierarchical
linear regressions analysis was performed for each of the
three burnout dimensions as the criterion. In Step 1 of the
analyses, the two control variables, age and gender, were used
as predictors because they were assumed to be related to
empowerment (Spreitzer, 1995b). In Step 2, the three psycho-
social work environment dimensions were added. In Step 3,
empowerment was added. A main effect is said to exist if,
after controlling for the effects of the demographic variables
in Step 1 and the work environment variables in Step 2,
empowerment has in the present step (Step 3) a statistically
significant relationship to a given burnout dimension (Feldt,
1997; Albertsen et al., 2001). A mediating effect is said to exist
if: (i) in the previous step of the analysis (Step 2), a given
work environment variable had a statistically significant rela-
tionship to a given burnout dimension; and (ii) in the present
step (Step 3), empowerment has a statistically significant
relationship to the given burnout dimension and the relation-
ship between the given work environment variable and the
burnout dimension diminishes or becomes statistically insig-
nificant (Albertsen et al., 2001; Seibert et al., 2004). Finally, in
Step 4, three bilinear interactions between each one of the
three psychosocial work environment dimensions and
empowerment were added. A moderating effect is said to
exist if: (i) any one of the coefficients for the interaction terms
is statistically significant; and (ii) any one of the interaction
terms significantly increases the amount of explained vari-
ance in the criterion (Albertsen et al., 2001). In order to inter-
pret the interactions, plots were made according to the
recommendations of Cohen et al. (2003). The values of the
predictor variables were chosen three SDs below (z = -3) and
three SDs above (z = 3) the mean and the regression lines
were produced by entering these values into the regression
equations.
RESULTS
Descriptive analyses
Table 1 shows the means, SDs, and Cronbach’s alpha coeffi-
cients. The only statistically significant differences between
the means of the registered nurses and the assistant nurses
were observed for the variable of age (t1354 = 2.24, P < 0.05
[two-tailed]) and control (t1354 = 3.99, P < 0.001 [two-tailed]).
For both groups, the alpha coefficients for all variables,
except for control, were satisfactory (ⱖ 0.70). Table 2 shows
the correlations between the variables. For both groups, it can
be noted that, except for the correlation between personal

Table 1. Means (M), standard deviations (SD), and Cronbach’s alphas (a) for all variables for registered nurses and for assistant nurses

Registered nurses Assistant nurses

Variable M SD a M SD a

Emotional exhaustion 1.95 1.11 0.90 1.82 1.14 0.90

Depersonalization 0.93 0.89 0.71 0.81 0.84 0.70
Personal accomplishment 4.82 0.76 0.78 4.60 0.99 0.78
Age 42.60 10.35 NA 45.30 10.58 NA
Gender NA NA NA NA NA NA
Demand 2.69 0.48 0.71 2.62 0.47 0.73
Control 3.02 0.37 0.52 2.88 0.40 0.53
Social support 3.22 0.45 0.83 3.19 0.48 0.85
Empowerment 5.29 0.82 0.86 5.24 0.90 0.87

NA, not applicable.

Table 2. Correlation coefficients for all variables for registered nurses below the diagonal and for assistant nurses above the diagonal

Variable 1 2 3 4 5 6 7 8 9

Emotional exhaustion 1.00 0.54*** -0.12** -0.01 0.00 0.53*** -0.16*** -0.48*** -0.41***
Depersonalization 0.54*** 1.00 -0.13** -0.14*** 0.01 0.19*** -0.20*** -0.22*** -0.31***
Personal accomplishment -0.22*** -0.23*** 1.00 -0.02 0.02 -0.01 0.18*** 0.08* 0.30***
Age -0.04 -0.18*** -0.03 1.00 -0.02 -0.01 0.00 -0.01 0.12**
Gender† -0.05 0.06 0.00 -0.08* 1.00 0.07 0.08 -0.04 0.04
Demand 0.54*** 0.24*** -0.04 -0.01 -0.04 1.00 0.01 -0.40*** -0.22***
Control -0.16*** -0.14*** 0.16*** -0.03 0.12*** -0.02 1.00 0.25*** 0.44***
Social support -0.33*** -0.18*** 0.17*** -0.03 -0.01 -0.30*** 0.24*** 1.00 0.40***
Empowerment -0.40** -0.35*** 0.36*** 0.20*** 0.02 -0.23*** 0.38*** 0.33*** 1.00

*P < 0.05 (two-tailed); **P < 0.01 (two-tailed); ***P < 0.001 (two-tailed). †Women were coded with 0 and men with 1.

© 2007 The Author

Journal Compilation © 2007 Blackwell Publishing Asia Pty Ltd.
208 J. Hochwälder

Table 3. Results from the hierarchical regression analyses, with emotional exhaustion as the criterion variable

Emotional exhaustion
Registered nurses Assistant nurses
Variable b†step 1 bstep 2 bstep 3 bstep 4 bstep 1 bstep 2 bstep 3 bstep 4

Age -0.05 -0.05 0.01 0.01 -0.01 -0.01 0.02 0.01

Gender‡ -0.05 -0.02 -0.03 -0.02 -0.01 -0.04 -0.03 -0.03
Demand – 0.49*** 0.45*** 0.46*** – 0.41*** 0.39*** 0.38***
Control – -0.11*** -0.02 -0.02 – -0.09** -0.01 0.00
Social support – -0.16*** -0.11*** -0.10*** – -0.29*** -0.23*** -0.23***
Empowerment – – -0.26*** -0.26*** – – -0.23*** -0.23***
Demand ¥ empowerment – – – -0.08** – – – -0.01
Control ¥ empowerment – – – -0.01 – – – -0.04
Social support ¥ empowerment – – – 0.01 – – – 0.02
Model F 1.90 85.53 87.75 60.20 0.06 60.94*** 58.92*** 39.38***
(d.f.) (2, 835) (5, 832)*** (6, 831)*** (9, 828)*** (2, 515) (5, 512) (6, 511) (9, 508)
R2 0.01 0.34 0.39 0.40 0.00 0.37 0.41 0.41
DR2 0.01 0.34*** 0.05*** 0.01* 0.00 0.37*** 0.04*** 0.00

*P < 0.05; **P < 0.01; ***P < 0.001. †b is the standardized regression coefficient; ‡women were coded with 0 and men with 1.

accomplishment and demand, all correlations between the

three burnout dimensions and the three work environment
dimensions, as well as between the three burnout dimensions
and empowerment, were statistically significant and went in 2
the expected directions.
Emotional exhaustion

Regression analyses with emotional exhaustion

as the criterion variable
As can be seen in Table 3, neither age nor gender were sig-
nificantly related to emotional exhaustion in the two groups
of nurses. For both groups, demand, control, and social
support were related as expected to emotional exhaustion,
explaining 34%/37%, respectively, of the variance in the cri-
terion variable. It may be noted that for assistant nurses, as
compared to registered nurses, social support had a higher
negative association with emotional exhaustion (b = -0.29/
b = -0.16, respectively). For both groups, the main effect of
empowerment on emotional exhaustion was negative and
statistically significant as expected, accounting for an addi-
tional 5%/4%, respectively, of the variance. For both groups,
empowerment had a mediating effect between the three
environment variables and emotional exhaustion. As can be
seen in Table 3, the regression coefficients for the three envi-
ronment variables diminished and, for control, were even
insignificant when empowerment was added. Only for regis-
tered nurses was one statistically significant moderator effect
of empowerment on the association between the environ-
ment variables and emotional exhaustion observed, account-
ing for only an additional 1% of the variance. As can be seen
from Figure 1, registered nurses with high empowerment
were less emotionally exhausted than those with low empow-
erment and an increase in demands was related to a some-
what lower increase in emotional exhaustion in registered
nurses with high empowerment as compared to those with
low empowerment.
0
–2
Low High
Demands
Figure 1. The interaction between demands and empowerment
with regard to emotional exhaustion for registered nurses. ---, low
empowerment; ––, high empowerment.
Regression analyses with depersonalization as the
criterion variable
As can be seen in Table 4, in the first step, only age was
significantly related to depersonalization in the two groups,
explaining 3%/2%, respectively, of the variance in the crite-
rion. For both groups, demand, control, and social support
were related as expected to depersonalization, explaining an
additional 9%/9%, respectively, of the variance. It may be
noted that for registered nurses, as compared to assistant
nurses, demand had a somewhat higher positive association
with depersonalization (b = 0.22/b = 0.14, respectively). For

© 2007 The Author

Journal Compilation © 2007 Blackwell Publishing Asia Pty Ltd.
Empowerment 209

Table 4. Results from the hierarchical regression analyses, with depersonalization as the criterion variable

Depersonalization
Registered nurses Assistant nurses
Variable b†step 1 bstep 2 bstep 3 bstep 4 bstep 1 bstep 2 bstep 3 bstep 4

Age -0.17*** -0.18*** -0.12*** -0.12*** -0.14*** -0.14*** -0.12** -0.12**

Gender‡ 0.04 0.06* 0.06* 0.06* 0.01 0.01 0.01 0.01
Demand – 0.22*** 0.18*** 0.18*** – 0.14** 0.11* 0.10*
Control – -0.13*** -0.04 -0.04 – -0.17*** -0.09 -0.08
Social support – -0.08* -0.03 -0.03 – -0.12** -0.07 -0.08
Empowerment – – -0.26*** -0.26*** – – -0.21*** -0.21***
Demand ¥ empowerment – – – -0.06 – – – -0.01
Control ¥ empowerment – – – 0.02 – – – -0.10*
Social support ¥ empowerment – – – 0.01 – – – 0.06
Model F 14.30*** 22.30*** 28.46*** 19.54*** 5.56 12.07 13.42 9.58
(d.f.) (2, 835) (5, 832) (6, 831) (9, 828) (2, 515) (5, 512) (6, 511) (9, 508)
R2 0.03 0.12 0.17 0.18 0.02 0.11 0.14 0.15
DR2 0.03*** 0.09*** 0.05*** 0.01 0.02** 0.09*** 0.03*** 0.01

*P < 0.05; **P < 0.01; ***P < 0.001. †b is the standardized regression coefficient; ‡women were coded with 0 and men with 1.

both groups, the main effect of empowerment on deperson- 2

alization was negative and statistically significant as
expected, accounting for an additional 5%/3%, respectively,
of the variance. For both groups, empowerment had a medi-
ating effect between the three environment variables and 1
depersonalization. As can be seen in Table 4, the regression
Depersonalization

coefficients for the three environment variables diminished

and, for control and social support, even became insignificant
when empowerment was added. Only for assistant nurses was 0
one moderator effect of empowerment on the association
between the environment variables and depersonalization
observed. As can be seen from Figure 2, for assistant nurses
–1
with high empowerment, an increase in control resulted
in less depersonalization while for assistant nurses with
low empowerment, an increase in control resulted in more
depersonalization. –2

Regression analyses with personal accomplishment as
the criterion variable
As can be seen in Table 5, in the first step, neither age nor
gender was significantly related to depersonalization in the
two groups. For registered nurses, demand, control, and
social support were related as expected to personal accom-
plishment, explaining an additional 4% of the variance but,
for assistant nurses, demand and control, but not social
support, were related as expected, explaining an additional
3% of the variance. It may be noted that the association
between social support and personal accomplishment was
stronger for registered nurses as compared to assistant
nurses (b = 0.14/b = 0.04, respectively). For both groups, the
main effect of empowerment on personal accomplishment
was positive and statistically significant as expected,
accounting for an additional 10%/7%, respectively, of the
variance. For both groups, empowerment had a mediating
effect between the environment variables and personal
Low High
Control
Figure 2. The interaction between control and empowerment with
regard to depersonalization for assistant nurses. ---, low empower-
ment; ––, high empowerment.
accomplishment. As can be seen in Table 5, the regression
coefficient for the environment variables that were statisti-
cally significant in the second step diminished and became
insignificant when empowerment was added in the third
step. Only for assistant nurses was one statistically signifi-
cant moderator effect of empowerment on the association
between the environment variables and personal accom-
plishment observed. As can be seen from Figure 3, an
increase in social support is related to higher degrees of
personal accomplishment in assistant nurses with low
empowerment and to lower degrees of personal accom-
plishment in assistant nurses with high empowerment.

© 2007 The Author

Journal Compilation © 2007 Blackwell Publishing Asia Pty Ltd.
210 J. Hochwälder

Table 5. Results from the hierarchical regression analyses, with personal accomplishment as the criterion variable

Personal accomplishment
Registered nurses Assistant nurses
b†step 1 bstep 2 bstep 3 bstep 4 bstep 1 bstep 2 bstep 3 bstep 4

Age -0.03 -0.03 -0.11*** -0.11** -0.02 -0.02 -0.06 -0.05

Gender‡ -0.00 -0.02 -0.01 -0.01 0.02 0.01 0.00 0.00
Demand – 0.01 0.06 0.07 – 0.01 0.04 0.04
Control – 0.13*** 0.00 0.00 – 0.17*** 0.05 0.04
Social support – 0.14*** 0.07 0.07* – 0.04 -0.04 -0.05
Empowerment – – 0.37*** 0.37*** – – 0.31*** 0.29***
Demand ¥ empowerment – – – -0.05 – – – 0.00
Control ¥ empowerment – – – -0.04 – – – -0.04
Social support ¥ empowerment – – – 0.02 – – – -0.11*
Model F 0.47 7.90*** 23.36*** 16.10*** 0.25 3.61 9.43 7.51
(d.f.) (2, 835) (5, 832) (6, 831) (9, 828) (2, 515) (5, 512) (6, 511) (9, 508)
R2 0.00 0.04 0.14 0.15 0.00 0.03 0.10 0.12
DR2 0.00 0.04*** 0.10*** 0.01 0.00 0.03*** 0.07*** 0.02*

*P < 0.05; **P < 0.01; ***P < 0.001. †b is the standardized regression coefficient; ‡women were coded with 0 and men with 1.

ating effect between the three work environment dimensions

and the three burnout dimensions. The mediating effect of
2 empowerment was especially apparent between the control
dimension and the three burnout dimensions. This result
implies that promoting aspects of control (and also social
Personal accomplishment

support) in the work environment should result in a higher

1
0 moderating effects (Albertsen et al., 2001), in that few (three
–1
–2
Low
Social support
Figure 3. The interaction between social support and empower-
ment with regard to personal accomplishment for assistant nurses.
---, low empowerment; ––, high empowerment.
DISCUSSION
In the present study, and in line with both theoretical reason-
ing (Wallerstein, 1992) and empirical results (Hochwälder &
Bergsten-Brucefors, 2005), empowerment had a significant
main effect on burnout. After controlling for demographic
and work environment variables, empowerment accounted
for an additional 3–10% of the variation in the three burnout
dimensions. This result legitimates the use of empowerment
as a complement to the job demand-control-support model.
As suggested by Spreitzer (1997) and observed by Lasch-
inger et al. (2003), we found empowerment to have a medi-
sense of empowerment which, in turn, should lead to lower
degrees of burnout.
The present study is similar to other studies exploring
significant effects out of the 18 possible effects) moderating
effects were observed. With regard to the first significant
interaction, it is plausible that an increase in demands should
be related to a lower increase in emotional exhaustion in
subjects (registered nurses) with high empowerment as com-
pared to subjects with low empowerment. The second inter-
action showed that for subjects (assistant nurses) with high
High empowerment, an increase in control resulted in less de-
personalization, while for assistant nurses with low em-
powerment, an increase in control resulted in more
depersonalization. Thus, having the opportunity to have
control over one’s work (in terms of, for example, how and
what should be done) without having any power (in terms of,
for example, competence and meaning) can lead to deper-
sonalization. This result is in line with Wallerstein’s (1992)
assertion that trying to change the individual’s sense of
control without changing the environmental conditions
might result in even more ill health. This assertion has been
confirmed by Schaubroeck and Merritt (1997), who found
that control can have negative health consequences among
individuals with low self-efficacy. The third interaction
showed that an increase in social support was related to
higher degrees of personal accomplishment in subjects (assis-
tant nurses) with low empowerment and to lower degrees of
personal accomplishment in subjects with high empower-
ment. Thus, an increase in social support might facilitate per-
sonal accomplishment for those with low empowerment but

© 2007 The Author

Journal Compilation © 2007 Blackwell Publishing Asia Pty Ltd.
Empowerment
might hinder personal accomplishment for those with high
empowerment. In sum, the implications of these findings are
that individual and group differences should be considered in
workplace health promotion.
To conclude, on the theoretical level, the obtained results
contribute by both confirming (the main and mediating) and
expanding (the moderating) effects in the nomological
network of empowerment. On the practical level, the obtained
results contribute by indicating that: (i) improving aspects of
the work environment should result in a higher sense of
empowerment which, in turn, should be a protective factor
against burnout (and consequently against suffering); and (ii)
individual differences (e.g. with regard to empowerment) and
group differences (e.g. with regard to profession) should not
be overlooked in workplace health promotion. Concretely
and simplified, these findings imply that burnout among
nurses might be reduced by enhancing their empowerment
(defined in terms of meaning, competence, impact, and self-
determination), which can be done either by objectively
changing the work environment (e.g. reduce role ambiguity,
enhance access to information and resources, enhance socio-
political support, and review organizational culture: cf. Spre-
itzer, 1995b) and/or by changing the nurses’ interpretation of
the work environment (e.g. by changing their attribution,
evaluation, and envisioning: cf. Thomas & Velthouse, 1990).
Furthermore, the three significant interaction effects imply
that both empowerment and work environment factors must
be considered in the struggle against burnout among nurses.
REFERENCES
Albertsen K, Nielsen ML, Borg V. The Danish psychosocial work
environment and symptoms of stress: the main, mediating, and
moderating role of sense of coherence. Work Stress 2001; 15: 241–
253.
Cohen J, Cohen P, West SG, Aiken LS. Applied Multiple Regression/
Correlation Analysis for the Behavioral Sciences, 3rd edn. Hills-
dale, NJ: Erlbaum, 2003.
Duquette A, Kérouac S, Sandhu BK, Beaudet L. Factors related to
nursing burnout: a review of empirical knowledge. Issues Mental
Health Nurs. 1994; 15: 337–358.
Feldt T. The role of sense of coherence in well-being at work: analysis
of main and moderator effects. Work Stress 1997; 11: 134–147.
Journal Compilation © 2007 Blackwell Publishing Asia Pty Ltd.
211
Hochwälder J, Bergsten-Brucefors A. Psychological empowerment
at the workplace as a predictor of ill health. Personal. Individ. Diff.
2005; 39: 1237–1248.
Karasek RA, Theorell T. Healthy Work: Stress, Productivity and the
Reconstruction of Working Life. New York: Basic Books, 1990.
Kleinbaum DG, Kupper LL, Mueller KE. Applied Regression Analy-
sis and other Multivariable Methods. Boston: PWS-KENT, 1988.
Laschinger HKS, Finegan J, Shamian J, Wilk P. Workplace empow-
erment as a predictor of nurse burnout in restructured healthcare
settings. Longwoods Rev. 2003; 1: 2–11.
Lee M, Koh J. Is empowerment really a new concept? Int. J. Hum.
Res. Manag. 2001; 12: 684–695.
McGrath A, Reid N, Boore J. Occupational stress in nursing. Int. J.
Nurs. Stud. 2003; 40: 555–565.
Maslach C, Jackson SE. The measurement of experienced burnout.
J. Occup. Behav. 1981; 2: 99–113.
Maslach C, Jackson SE. Maslach Burnout Inventory, 2nd edn. Palo
Alto, CA: Consulting Psychologists Press, 1986.
Norusis MJ. SPSS 14.0 Guide to Data Analysis. Upper Saddle River,
NJ: Prentice Hall, 2006.
Rodríguez I, Bravo MJ, Peiró JM, Schafeli W. The Demands-
Control-Support model, locus of control and job dissatisfaction: a
longitudinal study. Work Stress 2001; 15: 97–114.
Schaubroeck J, Merritt DE. Divergent effects of job control on
coping with work stressors: the key role of self-efficacy. Acad.
Manage. J. 1997; 40: 738–754.
Seibert SE, Silver SR, Randolph WA. Taking empowerment to the
next level: a multiple-level model of empowerment, performance,
and satisfaction. Acad. Manage. J. 2004; 47: 332–348.
Söderfeldt M, Söderfeldt B, Ohlson C-G, Theorell T, Jones I. The
impact of sense of coherence and high-demand/low-control job
environment on self-reported health, burnout and psychophysical
stress indicators. Work Stress 2000; 14: 1–15.
Spreitzer GH. An empirical test of a comprehensive model of intrap-
ersonal empowerment in the workplace. Am. J. Community
Psychol. 1995a; 23: 601–629.
Spreitzer GH. Psychological empowerment in the workplace: dimen-
sions, measurement, and validation. Acad. Manage. J. 1995b; 38:
1442–1465.
Spreitzer GH. Toward a common ground in defining empowerment.
Res. Organ. Change Dev. 1997; 10: 31–62.
Thomas KW, Velthouse BA. Cognitive elements of empowerment.
An “interpretive” model of intrinsic task motivation. Acad.
Manage. Rev. 1990; 15: 666–681.
Wallerstein N. Powerlessness, empowerment, and health: implica-
tions for health promotion programs. Am. J. Health Promot. 1992;
6: 197–205.
© 2007 The Author