Soc Psychiatry Psychiatr Epidemiol (2004) 39 : 576580 DOI 10.1007/s00127-004-0791-z
Abstract Background The working environment inmental health wards has been shown to have profound ef-fects on the health and work stability of mental healthworkers. Despite an apparent need for regular measure-ment of work environment,development of short and re-liable instruments for such measurements has beenlargely neglected. The aim of the present study was toevaluate the psychometric properties of the Working En-vironment Scale-10 (WES-10). Methods During the pe-riod 1990 through 2000, a total of 640 staff members on42 wards for psychotic patients completed the WES-10.To establish the number of subscales, a factor analysiswas carried out.The internal consistency of the subscaleswas calculated as Cronbachs alpha.We also collected dataconcerning satisfaction with the ward, its patients andstaff, and for how long the respondents had worked andexpected to continue to work at the ward. Results Weidentified four subscales named: Self Realization,Work-load,Conflict and Nervousness.The psychometric prop-erties of the subscales proved to be acceptable. All thesubscales were significantly correlated with at least onesatisfaction item, and/or the time the staff expected tocontinue at the ward. Most notably, the Self Realizationsubscale was strongly correlated to general satisfactionwith the ward, and to the time the staff expected to workon the ward in the future,while Conflict was strongly neg-atively correlated with liking for staff. Conclusion TheWES-10 appears to measure four clinically meaningfulsubscales. It seems well suited for use in further researchand for evaluation of clinical milieus.
Key words work environment psychometrics staffmembers satisfaction questionnaires
J. I. Rssberg . Eiring S. Friis
Work environment and job satisfactionA psychometric evaluation of the Working Environment Scale-10
Accepted: 18 February 2004
An extensive literature has been generated about the mi-lieu in which health professionals work and the impactit has on both mental and physical health . The po-tentially stressful nature of mental health work has beenamply demonstrated. Several studies have reported ahigh level of burnout and poor mental health amongpsychiatric staff members .A poor work environ-ment has proved to be associated with reduced job sat-isfaction, absenteeism, somatic complaints, burnout anddepression . It has been reported how a poorwork environment might influence the work perfor-mance negatively , and promote negative and cyni-cal attitudes towards patients and colleagues . Apoor work environment is probably one of the main rea-sons for the high staff turnover rate [1, 20, 21] and poorinpatient satisfaction and outcome [22, 23].
Against this background, there is an obvious need forregular studies of the work environment on psychiatricwards.As pointed out by Burnard et al. , the workingenvironment should be regularly measured, as one wayto measure service quality. Measures of the working en-vironment and job satisfaction may also be usefulbenchmarks for evaluating future changes and develop-ments in the psychiatric wards, and to monitor and im-prove the clinical working environment.
To regularly study the working environment,we needan instrument that is clinically meaningful, easy to useand with acceptable psychometric properties.To be clin-ically meaningful, the instrument has to measure thecentral dimensions. Reviewing the literature, it seemsthat previous studies have identified three core dimen-sions named: Workload (Work pressure, Task require-ments) [2, 21, 2426], Personal growth (Professionalgrowth, Support, Achievement value and growth) [14,21, 26] and Conflict [2, 25, 27]. Some studies come upwith additional dimensions like safety, role clarity,salary, work hazards, home-work conflict, professionalstatus and organizational issues [7, 14, 21, 24, 25].
J. I. Rssberg, MD () . Eiring, MD S. Friis, MD, PhDDept. of PsychiatryUllevaal University Hospital0407 Oslo, NorwayTel.: +47-22/118370Fax: +47-22/117848E-Mail: email@example.com
Many of the original instruments seem to have beenso large and complex that they have been difficult to use.Consequently, many studies of the working environ-ment tend to have small and unrepresentative samplesof psychiatric staff members . Admittedly, severalworking environment scales have been used in shortforms to facilitate ease of use or they have been changedto become more suitable for mental health workers ,but, as indicated in a recent review, analysis of the psy-chometric properties of different instruments seemsmostly to be lacking .
To overcome many of the previous difficulties, wehave developed a short and user-friendly ten-itemworking environment scale. We collected data from alarge number of psychiatric staff members on 42 differ-ent psychiatric wards. The main aims of this study wereto evaluate the psychometric properties of the WorkingEnvironment Scale-10 (WES-10) on a ward level.We de-cided to examine whether the WES-10 measures clini-cally meaningful subscales worth using in future re-search and in a clinical context.Another objective was toexamine how these subscales were related to general jobsatisfaction and the staff members intention to leavetheir job. More specifically, we wanted to answer the fol-lowing questions:1. Does the factor structure of the WES-10 point to spe-
cific subscales?2. What are the internal consistencies, mean scores,
standard deviations,and intercorrelations of the sub-scales?
3. What is the relationship between the WES-10 sub-scales and the staff members general satisfactionwith the psychiatric ward, liking for patients andstaff?
4. What is the relationship between the WES-10 sub-scale scores and how long the staff members haveworked at the psychiatric wards and how long theyintend to work at the wards?
Subjects and methods
During the period 1990 through 2000, a total of 640 staff members on42 wards for psychotic patients completed the WES-10. The WES-10is a self-report questionnaire and comprises the ten items displayedin Appendix 1. Three items (items 1, 2 and 3) were taken from the Per-sonal Development Scale developed by Moos . The other sevenitems were developed by a clinical research group at Ullevaal Univer-sity Hospital (1981) with the intention to evaluate the restructuring ofan acute ward . The items comprising the WES-10 were rated ona five-point scale ranging from 1, not at all or never, to 5, very often orto a large extent.
This study included wards where more than two-thirds of the pa-tients were diagnosed in the psychotic spectrum of mental diseases.Most of the wards were short-term (n = 36), but some intermediateand long-term wards (n = 6) were also included.
Three questions were applied to capture general satisfaction withthe milieu. The questions were rated on a five-point scale rangingfrom 1, not at all, to 5, very much, and were: (1) How satisfied are youwith this ward?, (2) How much do you like the patients on thisward? and (3) How much do you like the staff on this ward?. Themean and SD concerning these three questions were respectively: 4.03(0.33), 4.20 (0.20) and 4.43 (0.21).
To establish the number of subscales of the WES-10, we carried out afactor analysis (PCA) with varimax rotation. For each factor, we madea corresponding subscale. Items were included into the subscales ifthey had a loading on the corresponding factor 0.50 and if the dif-ference between this loading and the highest loading on non-corre-sponding factors was 0.05.
The internal consistency of the subscales was calculated as Cron-bachs alpha. We calculated the Corrected Item Total subscale Corre-lation (CITC) for the subscale, which comprised four items. Pearsoncorrelations were used to assess the relationships between the sub-scales and between the subscales and the other variables.
A total of 105 individual therapists and 529 nursing staffmembers completed the WES-10. A total of six staffmembers did not report whether they were individualtherapists or nursing staff. All staff members were in-cluded in the study except for night staff. The limitationwas due to difficulties in obtaining an acceptable num-ber of completed questionnaires from the latter group ofemployees.
All staff members were asked for how long they hadworked at the psychiatric ward and how long they ex-pected to work on the ward. A total of 15 % of the staffmembers had been employed less than 6 months, 22 %between 6 months and 18 months and 20 % between 18months and 3 years.Of the staff members,43 % had beenemployed more than 3 years. A total of 15 % were plan-ning to quit their jobs during the following 6 months.Anadditional 15 % were planning to quit before 1 year hadpassed.The largest group,47 %,thought they would con-tinue to work at the same psychiatric ward somewherebetween 1 and 3 years. A total of 23 % were planning tocontinue for more than 3 years. For each ward, we calcu-lated the percentage of staff members who had beenworking: (a) for less than 18 months, and (b) for morethan 3 years. We also calculated the percentage that in-tended to continue: (a) for less than 1 year, and (b) formore than 3 years.
The factor analysis revealed four factors with anEigenvalue > 1. These four factors accounted for 79 % ofthe variance. The factor loadings of the items are listedin Table 1. The four subscales were named: Self Realiza-tion, Workload, Conflict and Nervousness.
The Self Realization subscale explained 37 % of thevariance and comprised four items (items 1, 2, 5 and 6).This subscale measures to what extent the staff membersfeel supported, whether they achieve more confidenceand whether they experience being able to use theirknowledge working on the ward.The Workload subscaleexplained 19 % of the variance and comprised two items(items 9 and 10). This subscale measures the number oftasks imposed on the staff members and also to what ex-tent they feel they should have been on several places atthe same time. The last two subscales, Conflict and Ner-vousness, each comprised two items and explained re-spectively 13 % and 12 % of the variance. The Conflict
subscale measures to what extent the staff members ex-perience conflicts and loyalty problems among them(items 7 and 8). The Nervousness subscale measures towhat extent the staff members are worried about goingto work and to what extent they feel nervous or tense onthe ward (items 3 and 4). All the items were included inthe subscales.
The psychometric properties of the subscales provedto be satisfactory. Cronbachs alpha for the four sub-scales were: Self Realization: 0.85, Workload: 0.84, Con-flict: 0.69 and Nervousness: 0.66. The four items in theSelf Realization subscale had a mean CITC of 0.52. Theintercorrelations between the two items comprising thelast three subscales were respectively: Workload: 0.72,Conflict: 0.53 and Nervousness: 0.49.
The mean scores and standard deviations for the foursubscales were respectively: Self Realization: 3.73 (0.23),Workload: 3.41 (0.36), Conflict: 2.06 (0.28) and Nervous-ness: 1.98 (0.21).
As seen from Table 2, the subscales were only moder-ately intercorrelated (range: 0.130.37, median: 0.30).
As seen from Table 3, the Self Realization subscalewas most strongly correlated with general satisfactionand liking for the patients. The Conflict subscale wasmost strongly correlated (negatively) with liking for theother staff members.
Table 4 shows that only the Workload subscale wassignificantly correlated with the time staff members hadworked at the ward. This subscale showed a moderatenegative correlation with the percentage that hadworked less than 18 months at the ward. Two subscales
(Self Realization and Nervousness) were significantlyrelated to the time the staff expected to work on the wardin the future. The relationship was most pronounced forSelf Realization, which was moderately negatively corre-lated with the percentage who intended to continue lessthan 1 year, and positively with the percentage who in-tended to continue for more than 3 years.
The main finding of this study is that the WES-10 fea-tures acceptable psychometric properties on a wardlevel.Although not directly compared to other scales, theresults of this study clearly indicate that the core aspectsof the working environment can be captured with aneasy-to-use instrument that only comprises ten items.The instrument seems well suited for use in research andevaluation as it reliably measures four clinically mean-ingful subscales.
The validity is demonstrated by the fact that Conflictwas strongly negatively correlated with liking for staff,while Self Realization was fairly strongly positively cor-related with liking for patients. Furthermore, three ofthe four WES subscales were significantly related to gen-eral satisfaction with the ward. It also makes sense thatNervousness was significantly related to a high percent-age who intended to continue at the unit for less than 1year.
It seems that the WES-10 is able to capture the coredimensions of the working environment. The Workloadand Conflict subscales frequently appear in other stud-
Table 1 The factor loadings for each item. Only factor loadings 0.30 are included
Items from the I II III IVWES-10
4 0.45 0.32 0.63
5 0.63 0.43
Table 2 The intercorrelations between the four subscales
Self Realization Conflict Workload Nervousness
Self Realization X 0.37* 0.13 0.35*
Conflict X 0.20 0.34*
Workload X 0.25
* Correlation is significant at the 0.05 level (2-tailed)
Table 3 The intercorrelations between the four subscales and general satisfactionwith the psychiatric wards, patients and staff members
How satisfied How much do you How much do youare you with like the patients like the staff onthis ward? on this ward? this ward
Self Realization 0.56** 0.49** 0.40**
Conflict 0.41** 0.16 0.61**
Workload 0.20 0.15 0.17
Nervousness 0.30 0.27 0.15
** Correlation is significant at the 0.01 level (2-tailed)* Correlation is significant at the 0.05 level (2-tailed)
Table 4 The correlations between the four subscales and how long the staff mem-bers have been employed and how long they expect to work at the psychiatric ward
% worked % worked % intend % intend< 18 months > 3 years to continue to continue
< 1 year > 3 years
Self Realization 0.16 0.07 0.38* 0.32*
Conflict 0.05 0.15 0.25 0.26
Workload 0.37* 0.26 0.28 0.29
Nervousness 0.17 0.05 0.38* 0.25
* Correlation is significant at the 0.05 level (2-tailed)
ies and the importance of measuring these two dimen-sions of the work environment is widely described [2,25,27]. The Self Realization subscale measures the same as-pect of the work environment as the Professionalgrowth, Personal growth, Support and Achievementvalue and growth dimensions of other studi...