Measurement of nurse job satisfaction using the McClosky/Mueller Satisfaction Scale

Ann Tourangeau1, Linda McGillis Hall2, Diane Doran3, Teresa Petch4

1.2,3,4
Lawrence S. Bloomberg Faculty of Nursing, University of Toronto, Toronto, Ontario, Canada

Published 2006 in Nursing Research, 55(2), 128-136

http://journals.lww.com/nursingresearchonline/pages/articleviewer.aspx?year=2006&issue=03000&article=00008&ty
pe=abstract

ABSTRACT
Background: Originally developed to rank rewards that nurses value and that encourage them to remain in their jobs, the
McCloskey/Mueller Satisfaction Scale (MMSS) is being used extensively in research and practice to measure nurse job
satisfaction. Since its original development in 1990, limited evidence of psychometric properties of the MMSS has been
reported.
Objective: To investigate and report the psychometric properties of the MMSS when used in 2003 to measure hospital nurse
job satisfaction.
Methods: Data from a survey of 8,456 nurses were used to establish psychometric properties of the MMSS. Dimensionality
was tested using confirmatory and exploratory factor analyses. Validity of new MMSS factors was tested by investigating
relationships of the new factors with theoretically related concepts and by testing ability of the new factors to predict nurses'
intentions to remain employed in their hospitals. Reliability coefficients of the new factors are reported.
Results: The original eight factors could not be replicated satisfactorily using confirmatory factor analysis. Exploratory factor
analysis found a seven-factor model rather than the original eight factors previously reported. Validity of this new model was
supported. However, similar to the original instrument, weak internal consistency reliability coefficients were found for three
of the new MMSS factors
Discussion: From a research perspective, using an instrument with 23 items that measure 7 aspects of nurse job satisfaction is
more desirable than an instrument with 31 items. However, MMSS items must be redeveloped to improve internal
consistency of factors.
Key Words: Factor analysis, job satisfaction, McClosky/Mueller Satisfaction Scale, psychometric testing

Introduction individuals to seek to satisfy by acquiring incentives or

achieving goals. Further, the rewards desired by an individual
Measurement of job satisfaction has been undertaken for a are based on a hierarchy of needs. According to Maslow, this
variety of research and practice purposes. The hierarchy from lowest to highest consists of the following five
McCloskey/Mueller Satisfaction Scale (MMSS) is one categories of needs: physiologic, safety, social (belongingness),
instrument commonly used in nursing research and in healthcare ego (esteem), and self-actualization. The lowest level of
administrative practice to measure nurse job satisfaction. The unsatisfied need has the greatest motivating potential.
purpose of this paper is to describe the psychometric properties
When developing this original scale, McCloskey (1974)
of the MMSS as used with a large sample of Canadian nurses
grouped hospital nurse rewards and incentives into three distinct
working in Ontario acute care hospitals. This is accomplished by
categories or domains: safety, social, and psychological. The
reviewing the development of the MMSS, describing how the
safety dimension of job satisfaction was conceptualized to
MMSS has been used in nursing research and practice, reporting
include satisfaction with salary, benefits, and work scheduling.
on confirmatory and exploratory factor analyses and other tests
The social dimension of job satisfaction was conceptualized to
of validity, reporting reliabilities of the newly derived MMSS
include satisfaction with maternity leave support, childcare
subscales, and making recommendations about use and need for
facilities, immediate supervisor support, relationships with peers,
further redevelopment of the instrument.
and opportunities to socialize with colleagues. The psychological
dimension of job satisfaction included satisfaction with work
Development of the MMSS
responsibility, opportunities for further education, praise and
recognition, and control over work activities.
McCloskey (1974) laid the groundwork for this instrument
After 36 items were developed to reflect rewards and
by creating a scale to identify and rank rewards and incentives
incentives in the three categories, 94 nurses who had resigned
nurses valued and that encouraged them to remain in their jobs.
recently from their jobs in two urban hospitals were asked to rate
This initial scale was grounded in Maslow's theory of hierarchy
whether they might have stayed employed in their previous jobs
of needs (Maslow, 1954) and Burns' theory of motivation
had their employer offered more of what was identified in that
(Burns, 1969). According to Maslow and Burns, needs are
item. Though rewards and incentives related to each category of
physiological or psychological wants or desires that motivate
1
rewards and incentives were identified as important, McCloskey Brayfield Roth Global Satisfaction Scale and Hackman and
(1974) reported that the domain of psychological rewards had Oldhams's Job Diagnostic Survey. They reported positive
the greatest influence on nurse retention than were safety or correlations ranging between .53 and .70 when correlated with
social rewards. the MMSS subscales. They concluded that the 31-item MMSS
In 1990, Mueller and McCloskey revised McCloskey's instrument demonstrated adequate internal consistency, factor
original reward satisfaction scale with the intent of developing stability, and criterion-related and construct validity.
an easy to use, valid, and reliable instrument that measures nurse
job satisfaction. Thirty-three items were constructed based on Use of the MMSS in nursing research
McCloskey's previous work on the three dimensions of
incentives or rewards hypothesized to promote job satisfaction: The MMSS is one of the most commonly used tools to
safety, social, and psychological dimensions (Mueller & measure nurses' attitudes towards their jobs and has been used to
McCloskey, 1990). In this instrument, the safety dimension of identify job satisfaction in a variety of clinical and geographical
job satisfaction was conceptualized to include satisfaction with settings, including mental health (Brodell, 199; Flannery & Van
salary and benefits, balance of family and work, and Gaasbeek, 1998), long-term care (Robertson, Higgins, Rozmus,
opportunities to work straight days. The social dimension was & Robinson, 1999), public health (Cumbey & Alexander, 1998),
conceptualized to include satisfaction with supervisor support, home healthcare (Lynch, 1994), rehabilitation (Crose, 1999),
relationships with peers, and opportunities to socialize with rural settings (Anderko, Robertson, & Lewis, 1999), and
colleagues. The psychological dimension included satisfaction ambulatory care (Wilkinson & Hite, 2001). Healthcare
with praise and recognition, control over work activities, and administrators also have used the MMSS to measure levels of
professional opportunities (Mueller & McCloskey). nurse job satisfaction and to identify which aspects of job
Responses to items were rated using a 5-point Likert scale satisfaction should be strengthened strategically to promote
ranging from 1 (very dissatisfied) to 5 (very satisfied). This nurse recruitment and retention (Roberts, Jones, & Lynn, 2004;
instrument was administered to 190 nurses in 1983 and 1984 that Tang, 2003).
worked in one large urban hospital. Confirmatory factor analysis Even though the MMSS is based on factors thought to
was used to test the existence of the three hypothesized domains reflect American values and realities, this instrument has been
of job satisfaction. Mueller and McCloskey (1990) reported that used with international nursing populations. For example, the
a three-factor model did not fit the data. Exploratory factor MMSS was used to measure job satisfaction with nurses in the
analysis was completed and yielded eight distinct factors that West Bank. It was concluded that only four of the eight MMSS
were conceptualized as satisfaction with: extrinsic rewards, factors were considered important to nurses working in the West
scheduling, balance of family and work, co-workers, interaction Bank (Abu Ajamieh, Misener, Haddock, & Gleaton, 1996).
opportunities, professional opportunities, praise and recognition, Although the MMSS has been commonly used to measure
and work control and responsibility. Because two of the original nurse job satisfaction, some have suggested that the
33 items did not load onto any factor, these two items were psychometric properties no longer are considered adequate
recommended for omission in the instrument. This instrument (Roberts, Jones, & Lynn, 2004). Roberts et al. (2004) used the
was named the MMSS. MMSS to examine job satisfaction of recent baccalaureate
Mueller and McCloskey (1990) suggested that these eight graduates employed in various settings. They questioned the
factors could be grouped into the three originally hypothesized current relevance of the eight-factor structure and inadequate
job satisfaction domains. Safety rewards included three job reliabilities of four of the eight MMSS subscales. They
satisfaction factors: extrinsic rewards, scheduling, and balance of concluded that further exploration of the psychometric properties
family and work. Social rewards included two job satisfaction of the MMSS was required.
factors: co-workers and interaction opportunities. Psychological In this study, the theoretical perspective of three categories
rewards included the remaining three job satisfaction factors: of rewards and incentives (safety, social, and psychological)
professional opportunities, praise and recognition, and work leading to nurse job satisfaction proposed by Mueller and
control and responsibility (Mueller & McCloskey). Reliabilities McCloskey (1990) was adopted, recognizing that these three
for the eight subscales ranged between .52 and .84. Only four of categories of rewards and incentives could be measured using
the subscales had alpha reliabilities of .70 or higher and four had the eight factors comprising the MMSS. This paper contains a
reliabilities less than the desirable minimum value of .70 description of findings of the psychometric properties of the
(Nunnally & Bernstein, 1994). MMSS when used with a sample of Canadian hospital nurses.
Mueller and McCloskey (1990) also suggested that a global
job satisfaction scale could be calculated by summing all 31 Methods
items. The global job satisfaction indicator was reported to have
a Cronbach alpha coefficient of .89. They acknowledged the The study context
limitations of combining all items to form a global job
satisfaction scale as factor analysis demonstrated that the 31 Over 13,000 nurses working in Ontario, Canada, acute care
items yielded eight factors measuring distinct aspects of job hospitals were surveyed between February and May 2003 as part
satisfaction. of two studies examining nursing-related determinants of 30-day
Mueller and McCloskey (1990) also reported criterion- mortality and unplanned hospital readmission. The sampling
related validity of the MMSS instrument compared to several plan was developed from the 2003 College of Nurses of Ontario
other established job satisfaction instruments, including the registration database. All registered nurse and registered
2
(licensed) practical nurse registrants who reported on their 2003 a nurse, percentage of males, percentage working full-time,
registration renewal form that they worked in a medical, percentage routinely working 12-hr shifts, and percentage of
surgical, or critical care area in an Ontario teaching or baccalaureate educated nurses for the whole sample and for each
community acute care hospital were included in the study. of the registered nurse and registered practical nurse categories.
Surveys were mailed to nurses' homes. A reminder card was
mailed 10 days after the initial survey was mailed. A second Analysis
complete survey was mailed to non-responders 4 weeks after the
initial survey was mailed and was followed 10 days later by a Analyses were completed using the SPSS version 11.5
reminder card. Ethical approval for these studies was obtained (Chicago, IL). Analyses focused on testing MMSS instrument
from the university ethics review board. dimensionality, validity testing, and factor internal consistency
reliability.
Study instruments Using the theoretical framework adopted in this study, there
are three domains of rewards and incentives that reflect nurse job
The Ontario Nurse Survey 2003 was nine pages in length satisfaction and these three domains can be measured with the
and included sections inviting respondents to describe where eight-factor MMSS; thus, dimensionality of the MMSS was first
they worked within their hospitals, their evaluation of quality of examined by means of a confirmatory factor analysis with
patient care, their career intentions, their history of injury from a principal components analysis (PCA) using orthogonal rotation
patient-contaminated sharp, their job-related feelings (burnout), (varimax) to obtain a distinct and interpretable solution. Because
the condition of their professional practice environments, their this confirmatory factor analysis was unable to acceptably
job satisfaction, selected patient discharge processes of care in replicate the eight-factor solution found by Mueller and
their units, and demographic information. Three commonly used McCloskey (1990), an exploratory PCA with varimax was
instruments were included in the Ontario Nurse Survey 2003. implemented to explore instrument dimensionality.
The Maslach Burnout Inventory (MBI), a 22-item Three criteria are used to validate inclusion of items loading
instrument with 7-point response options, was used to assess on a factor. First, items must have a factor loading of at least .50
nurse burnout. This instrument is well validated and remarkably to be included in a factor. In orthogonal rotated factors, a factor
stable and was used to measure three components of burnout: loading expresses the correlation between the item and the
emotional exhaustion, depersonalization, and personal underlying dimension or factor. It is through examination of
accomplishment. Evidence of both reliability and validity of the factor loadings that one is able to determine which items belong
MBI as well as scoring procedures have been documented to that factor (Polit & Beck, 2004; Streiner & Norman, 2003).
(Maslach, Jackson, & Leitner, 1996). The choice of a minimum factor loading is usually set between
The Revised Nursing Work Index (NWI-R), a 49-item .40 and .60. In this study, the minimum factor loading was set as
instrument with 4-point response options, was used to measure .50. Second, to find a distinct and parsimonious factor structure,
various components of the condition of the professional nursing if an item cross-loads on two factors with a loading greater than
practice environment including nurse manager ability and .30 on the second factor, it must be eliminated from both factors.
support, nurse participation in hospital affairs, nursing Third, because items in factors must make sense conceptually to
foundations for quality care, adequacy of staffing and resources, enable interpretation of the essence of that factor, any item that
and collegial relationships among nurses and physicians (Lake, is conceptually unacceptable to a factor should also be deleted
2002). Evidence of there liability and validity of the NWI-R and from the factor (Streiner & Norman, 2003; Tourangeau &
its subscales as well as scoring procedures have been McGilton, 2004).
documented (Estabrooks et al., 2002; Lake).
The MMSS was used to measure eight components of nurse Table 1 Sample description
job satisfaction, including satisfaction with extrinsic rewards, Variable All Registered Registered
scheduling, balance of family and work, co-workers, interaction Sample Nurses Practical Nurses
opportunities, professional opportunities, praise and recognition, Age* 43.1 (9.4) 42.6 (9.5) 45.8 (8.7)
and control and responsibility (McCloskey & McCain, 1987; Years experience as 17.8 (9.9) 17.6 (9.9) 19.6 (9.9)
a nurse*
Mueller & McCloskey, 1990). Respondents rated their Percentage male 3.6 3.8 3.0
satisfaction for each item on a 5-point Likert scale ranging from Percentage full 59.0 60.7 50.4
very dissatisfied (1) to very satisfied (5). time*
Percentage working 69.8 76.5 34.7
12-hr shift
Sample Percentage with 18.7 21.8 2.3
baccalaureate*
Of the 13,093 nurses surveyed, 8,456 nurses (65% of
surveyed nurses) completed the survey. An additional 3% of Note. * Statistically significant on the variable between
nurses returned incomplete surveys. Respondents worked in 105 registered nurses and registered practical nurses (p < .05).
different hospital sites. Of these, 14 were teaching hospitals, 40
were large community hospitals, and 51 were medium-sized
community hospitals. The sample of nurses included 6,856 Validity of the newly derived MMSS factors is explored by
registered nurses and 1,325 registered practical nurses. Table 1 testing hypothesized correlations between the new subscales
contains descriptions of the mean age, mean years experience as with measures of theoretically related constructs and by testing
3
the ability of new MMSS factors to predict nurses' intentions to (items 7, 20, and 26). The remaining 23 items each had a
remain employed. To examine internal consistency of each new loading greater than .50 on one factor and fit conceptually with
MMSS factor, reliability coefficients of the new factors are other items in each factor.
measured. Table 4 contains a summary of the seven new factors
including eigenvalues and the amount of variance explained by
Results and Interpretation each factor. Two newly derived factors reflected nurses'
satisfaction with psychological rewards. The new Factor 1,
"satisfaction with work conditions and supervisor support,"
Dimensionality accounted for 13.34% of variance. All six items constituting this
new factor were also on the two original MMSS subscales of
A confirmatory factor analysis was implemented using PCA "satisfaction with praise and recognition" and "satisfaction with
with varimax to force an eight-factor model solution for the 31- work control and responsibility." In these analyses, those two
item MMSS instrument. The eight factors explained 61.1% of original factors collapsed into one factor. The new Factor 5,
variance. Once the three previously described criteria were "satisfaction with scholarly opportunities," also reflected
implemented to validate inclusion of items loading on a factor, 6 satisfaction with psychological rewards and explained 7.40% of
factors with 16 items remained. Four of the 31 items did not load variance. This new factor consisted of two items from the
on any factor (items 4, 10, 31, and 23), 9 of the 31 items double- original MMSS factor "satisfaction with professional
loaded on two factors (items 7, 14, 16, 17, 20, 24, 25, 26, and opportunities." The other two items from the original
29), and two items loaded on factor eight that had an eigenvalue "satisfaction with professional opportunities" subscale did not
less than 1.0 (items 11 and 12). Further, the seventh factor in the load satisfactorily on any factor.
model did not have any items that loaded satisfactorily making it Three newly derived factors reflected nurse satisfaction with
a non-interpretable factor. Table 2 contains a comparison of the safety rewards. New Factor 2 "satisfaction with scheduling"
confirmatory factor analysis findings with the original MMSS explained 10.69% of variance and consisted of four of six items
factor structure, including eigenvalues and the percentage of in the original MMSS subscale "satisfaction with scheduling,"
explained variance for each factor. Eigenvalues and percentage with the other two items of the original subscale not loading on
of explained variance were not reported for each of the original any factor. Newly derived Factor 6 "satisfaction with salary and
eight factors and could not be compared with those found in this benefits" consisted of the identical items as in the original
study. MMSS subscale "satisfaction with extrinsic rewards." The third
Because using confirmatory factor analysis could not factor reflecting satisfaction with safety rewards was new Factor
replicate a satisfactory eight-factor structure, an exploratory 7 "satisfaction with support for family responsibilities" and
PCA with varimax was implemented and yielded seven factors consisted of two of three items of the original MMSS subscale
with eigenvalues greater than 1.0 (Table 3). These seven factors "satisfaction with balance of family and work." This new factor
explained 57.9% of variance. Twenty-six of the original 31 items explained the least amount of unique variance of all seven
had loadings greater than .50 on at least one of seven factors. factors (4.13%). The structure of factor seven is noteworthy as it
Five of the 31 items did not load on any factor and, therefore, contained only two items (numbers 11 and 12) that had different
should be eliminated from use (items 4, 10, 16, 21, and 23). factor loading directions (.711 and -.706) suggesting that
Examination of all item loadings showed that three additional although these two items belong together in a factor, respondents
items loaded on two factors and also should be eliminated from

Table 2 Comparison of confirmatory factor analysis findings with original McClosky/Mueller Satisfaction Scale
factor structure
Original Factor Title Original Newly Derived Factor Eigenvalue Percentage of Explained
Factor Items Items (New) Variance (New)
Control and responsibility 22,23,29,30,31 13,22,30,31 8.84 28.53
(24 and 29 double loaded)
Scheduling 4,5,6,8,9,10 5,6,8,9 2.35 7.59
Interaction opportunities 16,17,18,19 18,19 1.77 5.72
(17 and 20 double loaded)
Professional opportunities 20,21,27,28 27,28 1.39 4.49
Extrinsic 1,2,3 1,2,3 1.38 4.45
Coworkers 14,15 15 1.18 3.81
(14 and 16 double loaded)
Praise and recognition 13,24,25,26 No items 1.03 3.30
(25 and 26 double loaded)
Balance of family and work 7,11,12 11,12 0.99 3.21

Note. Total explained variance is 61.1%. Reported eigenvalues and percentages of explained variance for each factor
reflect values when all items are retained in the factor. For those factors where items are not included, these values would
be less than the stated values. Refer to Table 3 for brief description of content for each item

4
 Table 3 Factor structure for seven new factors

Item Number and Descriptor Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6 Factor 7
1.Salary .054 .057 .019 .126 .015 .725 .099
2.Vacation .148 .258 .068 .080 .089 .697 .021
3. Benefits .129 .098 .080 .031 .082 .729 -.047
4. Work hours .168 .488 .107 .142 .069 .335 .175
5. Schedule flexibility .253 .702 .113 .087 .090 .113 .080
6. Work days .182 .692 .056 .035 .114 -.016 -.119
7. Part-time -.014 .504* .096 .113 .164 .003 .362
8. Weekends off .106 .725 .062 .066 .010 .222 -.031
9. Flexibility weekends .176 .816 .089 .043 .040 .079 -.045
10. Weekend pay .140 .472 .103 .011 .014 .421 -.083
11.Maternity leave .040 .124 .125 .033 .055 .173 .711
12.Child care .039 .152 .088 .062 .201 .116 -.706
13.Supervisor .613 .142 .088 .132 .000 .022 -.046
14.Peers .112 .048 .179 .776 .006 .034 .050
15.Physicians .207 .064 .201 .520 .032 .123 -.086
16.Care delivery .355 .104 .348 .454 -.069 .029 .079
17.Work social contact .199 .156 .723 .293 -.003 .044 .079
18.Social contact outside .104 .111 .762 .241 .071 .039 .052
19.Interact disciplines .250 .109 .696 .152 .178 .098 -.024
20.Interact faculty .204 .080 .515* -.040 .475 .104 -.107
21.Committee .359 .096 .420 .008 .373 .127 .071
22.Control setting .751 .161 .222 .107 .154 .102 -.008
23.Career advance .479 .205 .145 .067 .456 .184 .055
24.Superior recognition .680 .201 .063 .255 .241 .110 -.063
25.Peer recognition .188 .090 .084 .727 .213 .102 .009
26.Feedback .541* .173 .062 .486 .232 .092 -.047
27.Research .257 .100 .088 .152 .786 .049 -.012
28.Publish .131 .088 .087 .082 .824 .036 -.049
29.Responsibility .532 .120 .192 .275 .053 .120 .120
30.Control conditions .777 .168 .153 .094 .084 .116 .029
31.Decision making .714 .105 .148 .076 .276 .114 .021

Note. *Items that loaded > .5 on one factor but also had a loading > .3 on another factor. These items are
deleted from any factor.

Validity Testing
who reported being satisfied with maternity benefits also
reported being unsatisfied with hospital child care services. In Hypothesis testing was completed to examine validity of the
Canada, residents have access to up to 1 year of paid parental seven new MMSS factors. Relationships were tested between
leave but much less access to child care services associated with newly derived MMSS factors and subscales of the MBI and
their places of employment. NWI-R. First, it was hypothesized that nurses who experienced
The two other newly derived factors reflected satisfaction with higher burnout were more likely to report lower job satisfaction.
social rewards. New Factor 3 "satisfaction with social and It was expected that all newly found factors of the MMSS would
interaction opportunities" consisted of three of four items in the be mildly to moderately inversely correlated with the emotional
original MMSS subscale "satisfaction with interaction exhaustion and depersonalization subscales of the MBI and
opportunities." New Factor 4, "satisfaction with collegial positively related with the personal accomplishment subscale of
relationships and recognition," consisted of the same two items the MBI. Each of these three MBI subscales is known to
as in the original "satisfaction with co-workers" subscale plus measure aspects of burnout. Job-related burnout is
one additional item about satisfaction with recognition from conceptualized as high levels of emotional exhaustion and
peers. In these analyses, recognition from peers did not group feelings of depersonalization and low levels of feelings of
with recognition from supervisors. personal accomplishment (Maslach, Jackson, & Leitner, 1996).
Correlations among the original eight subscales as used in These relationships were hypothesized because of previously
this study ranged from .21 to .66. Correlations among the new found evidence of associations between job satisfaction and both
factors ranged from -.01 to .50 (Table 5). As expected, Factors 1 nurse burnout and condition of the nursing practice environment
through 6 are moderately and positively correlated with each (Table 6; Aiken, Clarke, & Sloane, 2002; Fletcher, 2001;
other ranging between .20 and .50. Correlations of factors one Greenglass, Burke, & Fiksenbaum, 2001). Evidence to support
through six with factor seven are much smaller and range from - the hypothesis was found with one consistent exception.
.11 to .04. These findings likely reflect the different directions of Moderate inverse relationships were found ranging from -.11 to -
the factor loadings of the two items, resulting in weak and .46 between the first six new MMSS factors with the three MBI
sometimes small negative correlations with other factors. scales (p < .0001). There was no correlation between any of the
5
 Table 4 Summary of new factors

Factor Number and Description Items in Factor Factor Explained Reliability

Eigenvalue Variance Coefficient
1. Satisfaction with work conditions 6 items (numbers 13, 22, 24, 29, 8.84 13.34% .85
and supervisor support 30, and 31)
2. Satisfaction with scheduling 4 items (numbers 5, 6, 8, and 9) 2.35 10.69% .80
3. Satisfaction with social and 3 items (numbers 17, 18, and 19) 1.77 8.1% .78
interaction opportunities
4. Satisfaction with collegial 3 items (numbers 14, 15, and 25) 1.39 7.44% .64
relationships and support
5. Satisfaction with scholarly 2 items (numbers 27 and 28) 1.38 7.40% .80
opportunities
6. Satisfaction with salary and 3 items (numbers 1, 2, and 3) 1.18 6.84% .67
benefits
7. Satisfaction with support for family 2 items (numbers 11 and 12) 1.03 4.13% .31
responsibilities

Note. Total variance explained by these seven factors was 57.93%. Refer to Table 3 for brief description of content for each item.

three MBI subscales and the seventh new MMSS factor less than .05: Factors 1, 4, 5, 6, and 7. Each of these job
"satisfaction with support for family responsibilities." This satisfaction factors was positively related with intention to
finding is likely a reflection of the different directions of factor remain employed. This finding supports predictive ability of the
loadings for the two items constituting factor seven (items 11 new MMSS factors.
and 12).
Second, it was hypothesized that nurses who worked in Reliability Testing
stronger professional nursing practice environments
characterized by supportive and able managers, adequate staffing In this study, internal reliability coefficients for the original
and resources, and collegial work relationships would experience eight MMSS subscales ranged from .29 to .84. Three of the
higher job satisfaction. Therefore, it was expected that the original MMSS subscales had unacceptably low alpha reliability
following three subscales of the NWI-R measuring the coefficients: satisfaction with extrinsic rewards (.67), satisfaction
professional nursing practice environment would be mildly to with balance of family and work (.29), and satisfaction with
moderately correlated with all new MMSS factors: nurse coworkers (.56). These low reliability coefficients are consistent
manager ability and support, adequacy of staffing and resources, with low coefficients originally found by Mueller and
and collegial relationships among nurses and physicians (Table McCloskey (1990). Internal reliability coefficients for the seven
6). Evidence to support the hypothesis was found with the same new factors ranged from .31 to .85 (Table 4). Three of the new
consistent exception found previously. Moderate positive factors have reliability coefficients less than the acceptable
relationships ranging from .19 to .73 were found between the minimum criterion of .70: satisfaction with collegial
first six new MMSS factors with the three NWI-R subscales relationships and support, satisfaction with salary and benefits,
measuring the condition of the practice environment (p < .0001). and satisfaction with support for family responsibilities. These
There was no correlation between any of the three NWI-R three factors have similar structures as the original three MMSS
subscales and the seventh new MMSS factor "satisfaction with factors that demonstrated weak internal consistency coefficients.
support for family responsibilities."
Validity of the new MMSS factors was further evaluated by Discussion
exploring whether the new MMSS factors predicted nurse
intention to remain employed. Others have found that job A seven-factor job satisfaction factor structure consisting of
satisfaction is the most important predictor of nurse intention to 23 items was found instead of the original eight-factor model
remain employed (Lu, Lin, Wu, Hsieh, & Chang, 2002; Shader, that included the 31 items developed and validated by Mueller
Broome, Broome, West, & Nash, 2001; Sourdif, 2004). Nurse and McCloskey (1990). The seven-factor model found in this
respondents were asked to rate how likely they were to continue study continues to be consistent conceptually with the theoretical
working at their current hospital until retirement on a 4-point foundation of the original MMSS. The three conceptual
scale ranging from 1 (very unlikely) to 4 (very likely). Multiple dimensions of rewards including safety, social, and
regression analysis was used to regress nurse intention to remain psychological dimensions that reflect nurse job satisfaction
employed scores on the seven new MMSS subscales as well as remain distinguishable in the new factor structure. Safety
nurse age. This model explained 25% of variance in intention to rewards include nurse satisfaction with scheduling, satisfaction
remain employed, F = 238.88, p < .0001. However, only five of with salary and benefits, and satisfaction with support for family
seven job satisfaction factors were found to be statistically responsibilities. Together, these three factors explain 21.66% of
significant predictors of intent to remain employed with p values variation in nurse job satisfaction. Social rewards include nurse
6
 Table 5 Correlations among seven new factors

New Factors Factor 1: Work Factor 2: Factor 3: Social Factor 4: Collegial Factor 5: Scholarly Factor 6: Salary
Conditions Scheduling Opportunities Relationships Opportunities and Benefits
Factor 2: Scheduling 0.45
p < .001
Factor 3: Social 0.42 0.27
opportunities p < .001 p < .001
Factor 4: Collegial 0.50 0.26 0.43
relationships p < .001 p < .001 p < .001
Factor 5: Scholarly 0.43 0.25 0.24 0.27
opportunities p < .001 p < .001 p < .001 p < .001
Factor 6: Salary 0.33 0.34 0.21 0.24 0.20
and benefits p < .001 p < .001 p < .001 p < .001 p < .001
Factor 7: Support -0.03 -0.02 0.04 -0.01 -0.11 0.02
for family p = .001 NS p = .0002 NS p < .001 NS

Note. NS refers to correlation not statistically significant at the .05 probability level.

satisfaction with social and interaction opportunities and or a subscale of an instrument, is measuring the same trait and is
satisfaction with collegial relationships and support. Together a common approach to assessing measurement error that arises
these two factors explain 15.44% of variation in nurse job from sampling of items (Polit & Beck, 2004). Similar low
satisfaction. Psychological rewards include nurse satisfaction reliabilities were reported in this study as were reported with the
with work conditions and supervisor support as well as original MMSS instrument development and testing (Mueller &
satisfaction with scholarly opportunities. Together these two McCloskey, 1990). Low reliability coefficients pose a particular
factors explain 20.74% of variation in job satisfaction. Mueller challenge when an instrument is used for research purposes
and McCloskey (1990) had reported that psychological rewards because the minimum criterion for subscale reliability is usually
were considered by hospital nurses to be more important than set at .70. Generally, further analyses are not completed on
were safety or social rewards for nurse job satisfaction that led to subscales with reliability coefficients that fall below that
nurse retention. In this study, safety rewards and incentives criterion. For this reason, it is essential that the MMSS be
explained the most variation in job satisfaction, followed by redeveloped to promote improved internal reliability within
psychological rewards or incentives. Social rewards explained subscales.
the least amount of variance in job satisfaction. According to Streiner and Norman (2003), a variety of strategies
Validity of the seven new factors of the MMSS was can be implemented to promote higher internal consistency
supported through hypothesis testing and tests of predictive among items in instrument subscales. These strategies include
validity. With the exception of factor seven, hypothesized revalidating and exploring additional constructs and related
relationships were found between MMSS factors and subscales items with key nurse informants, rewording items to ensure
of the MBI and the NWI-R. Using multiple regression analysis, terms used are current (e.g., remove term "maternity leave" and
the seven new MMSS factors and nurse age predicted 25% of substitute with "parental leave"), using minimal but sufficient
variation in nurses' intentions to remain employed in their number of words in each item (e.g., instead of item "your
current hospitals. Although these seven new factors immediate supervisor," modify term to "relationship with your
demonstrated some adequate psychometric properties related to immediate supervisor"), and omitting value-laden or biased
validity and reliability, the instrument continued to exhibit terms (e.g., replace current item "opportunity to work straight
weaknesses, particularly with respect to low internal consistency days" with revised item "opportunity to work preferred shifts").
coefficients for three factors. There are a variety of explanations for differences found in
Internal consistency refers to the extent that an instrument, factor analysis findings between this study and the original

Table 6 Summary of correlations among new MMSS factors with MBI and NWI-R subscales
New MMSS Factors MBI-EE MBI-DP MBI-PA NWI-R Manager NWI-R Staffing NWI-R Collegial
Factor 1: Work conditions -.46 -.32 .26 .73 .56 .44
Factor 2: scheduling -.29 -.21 -.17 .32 .31 .23
Factor 3: Social opportunities -.29 -.17 .16 .29 .34 .29
Factor 4: Collegial relationships -.30 -.24 .24 .33 .31 .45
Factor 5: Scholarly opportunities -.20 -.15 .11 .29 .25 .24
Factor 6: Salary and benefits -.22 -.14 .12 .21 .23 .19
Factor 7: Support for family -.02 (NS) .02 (NS) .00 (NS) -.01 (NS) -.02 (NS) -.01 (NS)
Note. All correlations have probabilities G .0001 except where noted with NS (not significant correlation). MMSS = McCloskey/Mueller Satisfaction
Scale; MBI-EE = the emotional exhaustion subscale of Maslach Burnout Inventory; MBI-DP = the depersonalization subscale of Maslach Burnout
Inventory; MBI-PA = the personal accomplishment subscale of Maslach Burnout Inventory; NWI-R manager = the nurse manager ability and support
subscale of the Revised Nursing Work Index (NWI-R); NWI-R staffing = the adequacy of staffing and resources subscale of the NWI-R; NWI-R collegial
= the collegial relationships among nurses and physicians subscale of the NWI-R.

7
development of the MMSS. The original instrument was in rehabilitation nursing. Rehabilitation Nursing, 24,
developed more than 20 years ago and tested on a relatively 95Y102.
small group of nurses (Mueller & McCloskey, 1990). The Cumbey, D. A., & Alexander, J. W. (1998). The relationship of
current study data were collected in 2003. With passage of time, job satisfaction with organizational variables in public health
work conditions and employment agreements have changed. For nursing. Journal of Nursing Administration, 28(5), 39Y46.
example, nurses' pay and benefits have increased and improved Estabrooks, C. A., Tourangeau, A. E., Humphrey, C K., Hesketh,
supports for balancing family and work have been put in place K. L., Giovannetti, P., Thomson, D., et al. (2002). Measuring
by many employers. In Canada, parental leave and financial the hospital practice environment: A Canadian context.
support is available for all employees for extended periods of Research Nursing & Health, 25, 256Y268.
time to support family life. This may explain why new factor Flannery, J., & Van Gaasbeek, D. E. (1998). Factors of job
seven "satisfaction with support for family responsibilities" has satisfaction of the psychiatric clinical nurse specialist.
such poor reliabilities when used with Canadian nurses. One of Nursing- Connections, 11(4), 29Y36.
the two items refers to satisfaction with maternity leave and the Fletcher, C. E. (2001). Hospital RNs’ job satisfactions and
other refers to satisfaction with child care. Because Canadian dissatisfactions. Journal of Nursing Administration, 31(6),
legislation supports extended paid parental leave for up to 1 year, 324Y331.
Canadian respondents likely would rate their satisfaction with Greenglass, E. R., Burke, R. J., & Fiksenbaum, L. (2001).
this item as relatively high. However, there are not yet consistent Workload and burnout in nurses. Journal of Community &
supports for child care across Canada and, therefore, Canadian Applied Social Psychology, 11, 211Y215.
nurses are more likely to rate their satisfaction with this item as Lake, E. T. (2002). Development of the practice environment
much lower than their satisfaction with maternity leave. scale of the Nursing Work Index. Research in Nursing &
As well as having strong validity and reliability properties, Health, 25, 176Y188.
research instruments should be as succinct as possible to Lu, K. Y., Lin, P. L., Wu, C. M., Hsieh, Y. L., & Chang, Y. Y.
minimize respondent burden and research costs related to data (2002). The relationships among turnover intentions,
collection, data coding and cleaning, and data analyses professional commitment, and job satisfaction of hospital
(Tourangeau & McGilton, 2004). In research, including fewer nurses. Journal of Professional Nursing, 18(4), 214Y219.
measures of concepts usually leads to higher statistical power. If Lynch, S. A. (1994). Job satisfaction of home health nurses.
nurse job satisfaction can be effectively measured using 23 items Home Healthcare Nurse, 12, 21Y28.
collapsing into seven subscales as found in this study rather than Maslach, C., Jackson, S., & Leitner, M. (1996). Maslach
eight subscales, analytical models using these subscales will Burnout Inventory Manual (3rd ed.). Palo Alto, CA:
have more power. In the Canadian context and in other Consulting Psychologists Press.
jurisdictions with similar support for parental leave and lack of Maslow, A. (1954). Motivation and personality. New York:
support for childcare, it is advisable to consider use of only six Harper & Row.
scales as new Factor 7 "satisfaction with support for family McCloskey, J. (1974) Influence of rewards and incentives on
responsibilities" has very poor internal consistency. staff nurse turnover rate. Nursing Research, 23(3), 239Y247.
Further redevelopment and testing of the MMSS is required McCloskey, J. C., & McCain, B. E. (1987). Satisfaction,
to minimize potential sources of error related to adequacy of commitment and professionalism of newly employed nurses.
sampling of items. Improving internal consistency of the Image: The Journal of Nursing Scholarship, 19(1), 20Y24.
instrument will lead to increased utility and credibility of the Mueller, C. W., & McCloskey, J. C. (1990). Nurses’ job
MMSS as a valid and reliable measure of nurse job satisfaction satisfaction: A proposed measure. Nursing Research, 39,
113Y117.
References Nunnally, J., & Bernstein, I. H. (1994). Psychometric theory (3rd
ed.). New York: McGraw-Hill.
Abu Ajamieh, A. R., Misener, T., Haddock, K. S., & Gleaton, J. Polit, D. F., & Beck, C. T. (2004). Nursing research: Principles
U. (1996). Job satisfaction correlates among Palestinian and methods (7th ed.). Philadelphia, PA: Lippincott Williams
nurses in the West Bank. International Journal of Nursing & Wilkins.
Studies, 33(4), 422Y432. Roberts, B. J., Jones, C., & Lynn, M. (2004). Job satisfaction of
Aiken, L. H., Clarke, S. P., & Sloane, D. M. (2002). Hospital new baccalaureate nurses. Journal of Nursing
staffing, organization, and quality of care: Cross-national Administration, 34(9), 428Y435.
findings. International Journal for Quality in Health Robertson, E. M., Higgins, L., Rozmus, C., & Robinson, J. P.
Care,14(1), 5Y13. (1999). Association between continuing education and job
Anderko, L., Robertson, J., & Lewis, P. (1999). Job satisfaction satisfaction of nurses employed in long-term care facilities.
in a rural differentiated setting. NursingConnections, Journal of Continuing Education in Nursing, 30(3),
12(1),49Y58. 108Y113.
Burns, T. (1969). Human relations in management: Motivation Shader, K., Broome, M. E., Broome, C. D., West, M. E., &
and human relations. American Journal of Medical Nash, M. (2001). Factors influencing satisfaction and
Technology, 35, 12Y23. anticipated turnover for nurses in an academic medical
Brodell, E. (1996). Nursing career satisfaction: The effects of center. Journal of Nursing Administration, 31(4), 210Y216.
autonomy, social integration, and flexible scheduling. Prairie Sourdif, J. (2004). Predictors of nurses’ intent to stay at work in
Rose, 65(3), 4Y6. a university health center. Nursing & Health Sciences, 6,
Crose, P. S. (1999). Job characteristics related to job satisfaction 59Y68.
8
 Streiner, D. L., & Norman, G. R. (2003). Health
measurement scales: A practical guide to their development
and use (3rd ed.). New York: Oxford University Press.
Tang, J. H. (2003). Evidence-based protocol: Nurse retention.
Journal of Gerontological Nursing, 29(3), 5Y14.
Tourangeau, A. E., & McGilton, K. (2004). Measuring
leadership practices of nurses using the Leadership Practices
Inventory. Nursing Research, 53(3), 182Y189.
Wilkinson, C. S., & Hite, K. J. (2001). NurseYphysician
collaborative relationship on nurses’ self-perceived job
satisfaction in ambulatory care. Lippincott’s Case
Management. 6(2), 68Y78.