Trabajo y Psicoanálisis

Work content as a moderator between work demands and fatigue

Fred R.H. Zijlstra,
University of Surrey
Guildford, UK


In the last two decades the face of work has changed considerably. Social economic changes, and technological innovations have had a great influence (Zijlstra et al, 1996). There are several changes to be noticed. Firstly, at present the majority (over 75 % in Holland) of the working population is employed in the service sector. This means that services have to be rendered to customers and clients, and people have to deal with patients. Client friendliness has become a keyword for many employees. This places specific demands on those people, they have to be friendly all the time, i.e. being patient and smile to the customers, in fact that means that they have to control their emotions. There is ample evidence that working with people puts specific demands on people in those professions, and that it is believed that they have a higher risk for the burn out syndrome (Schaufeli & Enzmann, 1998).

Secondly, due to technological innovations (information and communication technology) many people work with computers. Their main activity has become processing information, which implies that work places a strong demand on the cognitive system (information processing). According to Aronson (1989) this has lead to an increase in demands, in particular the demands on concentration and attention. For the workers this has resulted in intensified work.

This suggests that some different types of work have emerged: ‘working with people’, and ‘working with information’, as a contrast with the more traditional type of work in which people work primarily with material objects (things). The idea is that these types of work place a different pattern of demands on the workers. This presentation deals with the question whether this also leads to a different pattern of health risks, in terms of level of workload, work pressure, and fatigue.

Demographic, cultural and social changes have lead to a change in the workforce: increasing multi-ethnic variety, increasingly elder workers, higher number of women participating on the labour market. The latter means that the division of household duties (care) often has to be re-negotiated between husband and wife. Some people assume that this also contributes to higher levels of burnout in our society.


Generally it is assumed that when people have control in their work they will be able to mediate the potential stressors (work demands) (Karasek & Theorell, 1990). In particular when one has control over ones own task (Carayon & Zijlstra, 1999). Working with people may mean that one has to a lesser extent control over the object of work, because one will have to take into account that patients, clients and customers, pupils are individuals whom all may have their own specific wishes and/or demands which makes it always unexpected and difficult to serve them well. This may affect workers sense of control, i.e. that they feel that they have less control over their work, or that control has less influence in terms of reducing potential stressors (i.e. the stress-reduction hypothesis, Frese, 1989). This may become apparent in the contribution of control in a regression equation predicting work load or feelings of burn out.

The conceptual model for this study is presented in the figure (cf. Roe & Zijlstra, 2000).


Regarding the fact that the three groups represent three different types of work, and we assume that for each group there is a specific pattern of demands, we expected that for the groups that works primarily with people the emotional demands would be most prominent in predicting level of workload and fatigue, and that this group would have the highest level of burn out, as measured with the MBI- general version.

For the group working with data we expected that attentional demands (concentration) would be highest. We expected that this group would have the highest level of work pressure, since Aronsons (1989) research suggested that in particular the attentional demands which accompanies working with computers cause feelings of work pressure.

With respect to the third group, people who primarily work with things we expected that the physical demands would be highest.



A cross-sectional survey has been conducted among the working population of The Netherlands (N=1130). The survey appears to represent the working population in the Netherlands quite well, be it that the in sample the number of higher educated was over represented. However, that is standard with these methods.

Sample description

The sample consisted of 44 % women and 56 % men. Age distribution was from 17 to 70 years old, with a mean age of 39,5 (S.D. = 9.6). The average number of weekly working hours was 33.8 (S.D.= 10.1, and maximum is 84 hours) hours per week.


Apart form biographic information (age, sex, marital status, etc.) the questionnaire contained questions in which people were asked to indicated what kind of work they did (job label), and how many hours they worked per week (official, and actually), in which economic branch they worked, et cetera. Furthermore people were asked to describe their work, using the classification that has been made by Fine (1955) to establish whether they work primarily with things, people, or data.

And in addition established scales on work demands, workload, and work pressure (Roe & Zijlstra, 2000), and fatigue (CIS-20, Vercoulen et al., 1994), Burnout (the Dutch translation -Schaufeli & Dierendonck, 1994- of the MBI, Maslach & Jackson, 1981 ) were included.

Furthermore scales on rating effort investment (Zijlstra, 1993), job control including task control and organization control (Greenberger, et al., 1986; Carayon & Zijlstra, 1999) have been included.


Three groups have been created using the classification made by Fine (1955): a group of respondents who work primarily with things, a group that works primarily with people, and the third group works primarily with data.

With analyses of variance (ANOVA) it has been established whether the groups differ with respect to level of demands, level of workload, and work pressure, and fatigue and burnout.

After that a hierarchical analysis of regression has been done in various steps to establish to what extent the various demands, individual characteristics and subjective work characteristics (workload, work pressure) contribute to the level of fatigue and burnout for the three different groups.

In the first step some person related variables (age, time for household, care, effort investment) have been entered in the regression equation, and subsequently the nine scales for work demands were entered as predictors. These analyses were repeated for each of the three groups.

Thus it was examined to what extent work demands had added value in predicting work pressure and emotional exhaustion.

Furthermore the role of two dimensions of control in predicting (subjective and physical) fatigue was examined, and the mediating role of the concept of work pressure.



First of all an analysis of variance was executed to check the differences between the three groups. Table 1 contains the results:

Table 1: results of analysis of variance; comparison of three job types with respect to job demands.

As can be seen form the table above, there are significant differences between the three job types concerning the various job demands and some of the outcome variables. The results indicate that the highest demands can be found in the ‘modern job types’. Furthermore there are some interesting points that need to be mentioned. First of all, the fact that the emotional demands are highest in the group working with people confirms our hypo-thesis. Those people also have the highest scores on responsibility, and indicate that their work is quite complex, and that multiplicity (i.e. having to do several things more or less at the same time) is highest in this job type. The group working with data appears to have the highest quantity of work, they are most hindered by interruptions during work, and are most often confronted with deadlines, resulting in time pressure. It is therefor not remarkable that they experience the highest amount of work pressure. The group working with people has to put in the most effort during the day, which clearly results in the highest workload ratings for that group. The traditional type of work (i.e. working with things) seems to be working the most hours, at least officially. When we look at the actual number of hours that people work per week, there appears to be no significant differences anymore. In all groups the number of actual working hours is higher than the official working hours, the average amount of hours is about 5 to 6 hours higher than people officially have to work. In the group working with people the highest number of part-time jobs can be found (< 32 hours per week), and people working on flexible contracts.

Table 2 shows the comparison of level of workload, pressure, effort and burnout.


Regression analysis:

Table 3 contains the results of the analysis of regression: workload predicted by the job demands.


In all three job types, quantity and complexity of work contribute to perceptions of workload.

In the ‘modern’ jobs some other dimension contribute as well. In particular demand 9, the hindrances/ and lack of support (obstacles at work), seem to be important, as is the lack of responsibility (notice negative Beta coefficient). This clearly suggests a different profile of work demands in modern job types versus the traditional job types (i.e. working with things)

The results in this table have been corrected for the number of weekly hours people work.


When a different variable is selected as dependent variable, respectively work pressure and the scores on the MBI-dimension ‘emotional exhaustion’, a remarkable decrease in R2 adj (the variance explained by the regression model) can be found. Table 4 shows those parameters for all the three groups. This clearly indicates that the relevance of the demands decrease in predicting the outcomes in terms of work load, work pressure, and burn out.

Table 4: the variance explained (R2 adj) by work demands for three different dependent variables

The work demands appear to have a more direct effect on work load, and have a rather remote effect on burn out. Other factors apparently have a greater influence in predicting whether people will experience work pressure, or suffer from burn out. This suggests a kind of process model: Work demands -----> work load ------> work pressure --------> burn out

As predicted in the conceptual model. This to some extent validates the conceptual model.


Contribution of work demands in predicting health outcomes for three groups

Tables 5 to 11 present regression models in which other factors than the work demand have been included as well for the three different job types. In particular some subjective evaluations of the job have been included in the model (i.e whether the job is perceived as being rewarding, people have been asked to indicate how much effort they have to invest in the morning when they start to work, and how much effort they have to invest at the end of their working day. This gives an indication of the energetic demands of the job. A rating on how much job control people have has been included in the model. Furthermore some time parameters have been included: i.e. the number of actual hours people work per week, and how much time they spend on household and care.

The results presented in Table 5 illustrate that in traditional type of jobs , the non-work demands do not contribute in predicting work pressure. Explained variance is only 6 %. Task control does have a negative contribution, as predicted by the ‘stressor-reduction hypothesis’. However, organization control has no contribution. Both dimensions of control do not contribute in predicting exhaustion. Only the amount of effort exerted during the day appears to be relevant.

Table 6 indicates that work demands add only a little bit of predicting power, as far as exhaustion is concerned (7 %). Work demands are much more relevant as far as work pressure is concerned.

Tables 7 and 8 present the results of the regression analysis for the group ‘working with people’. Effort investment during the day, again has an important contribution in predicting work pressure and exhaustion. Furthermore rewards seem to be important, and considering the negative Beta-weight, this might be interpreted as ‘lack of rewards’. Combined with the high effort investment, this finding seems to confirm the Effort-rewards imbalance (Siegrist, 1996). Table 8 indicates that the work demands again do explain very little variance in addition to the variables presented in Table 7, in particular as far as ‘exhaustion’ is concerned.

Interesting to note is the negative Beta-weight for the ‘time spend on household and care’. Probably the interpretation should be that people who feel exhausted do not spend much time on household activities (they may feel too tired).

A similar picture emerges in Table 9 and 10, although work demands appear to explain now a little additional variance again (9%).

Nevertheless, a substantial part of the variance is explained by other factors than work demands.

The last two figures focus on the role of control in predicting fatigue (physical and subjective fatigue).

As can be seen in Table 11, task control does not contribute in predicting physical fatigue, while organization control does, in the total sample. When work pressure is included in the regression equation this does not change. This means that work pressure does not have a mediating role (Carayon & Zijlstra, 1999).

This, however, is different for traditional type of jobs (working with things). Task control seems to be important instead of organization control, but again no mediating role for work pressure.

Table 12 shows a slightly different picture: task control is important in predicting subjective feelings of fatigue, it evidently helps to counteract feelings of fatigue. And work pressure indeed seems to have a mediating role, considering the fact that when work pressure is added in the equation organization control now has a significant contribution. Although it doesn’t appear to have this role in traditional type of jobs. This suggests that when work pressure comes into play, people in modern type of jobs have opportunities to counteract potential negative effects. Having organization control means that people have some influences over organization’s policy, delegation of work, et cetera, and this is helpful in modern type of jobs.


Results have indicated that there are significant differences in levels of demands between the various groups, suggesting that he ‘modern’ jobs generally have higher level of demands.

In particular the job demands ‘responsibility’, ‘concentration’ and ‘experiencing hindrances and/or lack of support’ seem to make modern jobs quite demanding.

Other factors than job demands seem to explain quite a lot of variance in predicting health outcomes, in particular in ‘modern’ type of jobs.

The data seem to support the conceptual model.

There are differential effects of ‘job control’ in various job types. Distinction is necessary between ‘physical fatigue and ‘subjective fatigue’.

Practical recommendations:

Non-work demands seem to rather important in predicting exhaustion in "modern jobs". Means that this problem can not only be solved within organisations, and / or by employers.

Societal changes call for societal solutions, such as: Extended child care facilities, Opening times for shops, etcetera.



Aronson, G. (1989). Changed Qualification Demands in Computer Mediated Work.

In: Applied Psychology, 38(1). p. 57-71.

Carayon, P., & Zijlstra, F.R.H., (1999) Relationship between job control, work pressure and strain. Studies in the USA and The Netherlands. Work & Stress, 13. No 1, 32-48.

Frese, M. (1989). Theoretical Models of Control and Health. In: S.L. Sauter, J.J. Hurrell Jr. & C.L. Cooper (Eds.), Job Control and Worker Health. (107-129). Chichester: Wiley & Sons.

Fine, S.A., (1955). Functional job analysis. Journal of Personal Administration and Industrial Relations, 2. 1-16

Greenberger, D.B., Strasser, S., Cummings, L.L., and Dunham, R.B., (1989). The impact of personal control on performance and satisfaction. Organizational Behavior and Human Decision Processes, 43, 29-51.

Karasek, R.A. & Theorell, T., (1990). Healthy Work. Stress, Productivity, and teh Reconstruction of Working Life. New York: Basic books

Maslach, C., & Jackson, S.E. (1981). The measurement of experienced burnout. Journal of Occupational Behaviour, 2, 99-113.

Roe, R.A. and Zijlstra, F.R.H., (2000). Work pressure. Results of a conceptual and empirical analysis. In: M. Vartiainen, F. Avallone, and N. Anderson (Eds.), Innovative Theories, Tools and Practices in W&O Psychology. Göttingen: Hogrefe. Pp. 29-46.

Schaufeli, W.B., & van Dierendonck, D., (1994). Brunout, een concept gemeten: De Nederlandse versie van de Maslach Burnout Inventory (MBI-NL). Gedrag en Gezondheid, 22, 153-172.

Schaufeli, W. & Enzmann, D. (1998). The Burnout Companion to Study & Practice: A critical analysis. Washington, DC: Taylor & Francis.

Siegrist, J., (1996). Adverse health effects of high-effort/low-reward conditions. Journal of Occupational Health Psychology, 1, 27-41.

Vercoulen, J.H.M.M., Swanink, C.M.A., Galama, J.M.D., Fennis, J.F.M., van der Meer, J.W.M., & Bleijenberg, G. (1994). Dimensional assessment in chronic fatigue syndrome. Journal of Psychosomatic Research, 38, pp. 383-392.

Zijlstra, F.R.H., (1993). Efficiency in Work Behaviour; A Design Approach for Modern Tools. Delft: Delft University Press. ISBN 90-6275-918-1 (Ph.D. thesis).

Zijlstra, F.R.H., Schalk, M.J.D., & Roe, R.A., (1996). Veranderin-gen in de Arbeid. Conse-quen-ties voor Werkenden. Tijd-schrift voor Arbeids-vraagstukken, vol.12 (3). 251-263.

Volver al menu general de capítulos del foro

Colabore con una Membresía Solidaria y obtenga sus contraseñas únicas de acceso a todas las áreas de PsicoMundo (revistas, seminarios, etc.)

PsicoMundo - La red psi en Internet