Job characteristics, motivators and stress among information technology consultants: A structural equation modeling approach

ARTICLE IN PRESS

International Journal of Industrial Ergonomics 37 (2007) 51–59 www.elsevier.com/locate/ergon

Job characteristics, motivators and stress among information technology consultants: A structural equation modeling approach Lars Go¨ran Wallgren, Jan Johansson Hanse Department of Psychology, Go¨teborg University, Box 500, SE 405 30 Go¨teborg, Sweden Received 4 November 2005; received in revised form 21 September 2006; accepted 3 October 2006 Available online 27 November 2006

Abstract The aim of the current study was to test a structural model of the relationship between job characteristics (job demand, job control) and perceived stress (e.g. stressed, pressed, tense) with ‘motivators’ (e.g. responsibility, recognition, achievement, possibility of growth) as the mediating variable. In this cross-sectional study a web-based questionnaire survey was conducted among information technology (IT) consultants in Sweden (N ¼ 167). The results show that job demand was positively related to perceived stress. The results further indicate that motivators in part mediate the relation between job control and perceived stress, i.e. high job control was significantly related to high appraisals of motivators, and motivators was negatively related to perceived stress. Relevance to industry The results point out the importance of motivators (e.g. responsibility, recognition, achievement, possibility of growth) among ITconsultants in the job stress and performance framework. r 2006 Elsevier B.V. All rights reserved. Keywords: IT consultants; Demand; Control; Motivation; Stress; Structural equation modeling

1. Introduction It is generally believed that knowledge firms stand for and symbolize the future of business and working life. According to a report from the European foundation (2000) the number of knowledge workers in the USA between 1990 and 1998 rose by 2.5 million, i.e. over 18% of the net employment gains recorded during that period, and according to a report from the European Foundation (2005) knowledge-intensive business services represent one of the fastest growing areas of the European economy. According to Alvesson (2003) the concept ‘knowledgeintensive’ can be used in at least three contexts: knowledgeintensive firms, knowledge-intensive work, and knowledge workers. Knowledge-intensive firms claim to produce qualified products and/or services, and also generate new Corresponding author. Tel.: +46 31 786 1000; fax: +46 31 786 4628.

E-mail address: [email protected] (J.J. Hanse). 0169-8141/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ergon.2006.10.005

and unique knowledge (Ka¨rreman et al., 2003). Furthermore, according to Ka¨rreman and colleagues, knowledgeintensive firms are typically engaged in complex and difficult tasks that cannot easily be converted into standardised work procedures. Example of knowledge-intensive firms are law and accounting firms, management, engineering, and IT consultancy companies. In knowledge-intensive firms the most work are of an intellectual nature and the major part of the workforce, the knowledge workers, are according Alvesson (2000) well-educated and qualified. According to Statistics Sweden’s forecast (2002) the number of people employed within the private service sector is predicted to increase by 13 per cent from 1999 to 2020, and the increase will mainly occur in consultancy services, chiefly in the business sector. According to Docherty and Huzzard (2003), the service is often produced in co-operation with the customer, who makes demands on the knowledge workers’ social and communication ability.

ARTICLE IN PRESS 52

L.G. Wallgren, J.J. Hanse / International Journal of Industrial Ergonomics 37 (2007) 51–59

The information technology (IT) sector has grown considerably during the last decade and will probably expand in the next decade (Iverga˚rd, 2000). IT consultants’ work exhibits many characteristics of knowledge work. IT consultants have to continuously deal with non-standard problems, originating from the customer, try to make sense out of them, and then try to provide satisfactory solutions to them. IT consultants work is surrounded by uncertainty and may therefore be stressful. In general, individual job control is a central concept in the understanding of relationships between stressful experience, behavior and health (Karasek and Theorell, 1990; Melin and Lundberg, 1997; Pousette and Johansson Hanse, 2002). The increased interest in knowledge-intensive firms can probably be explained by their increased growth rate and overall importance for the economy. When researchers study the phenomena knowledge of knowledge management they prefer to study knowledge-intensive firms and one of the reasons is the fast expansion of management and IT consultancy firms (Alvesson, 2003). Even if there is an increasing interest in knowledge-intensive firms, there are relatively few studies that describe the working conditions of IT consultants as regards the psychosocial work environment such as job demand, job control, workload and stress. Furthermore, if an IT consultancy firms will receive new tasks and survive, the consultants must perform a good job for the clients. As recent research shows that work motivation may predict job performance (e.g. Locke and Latham, 2002; Pritchard and Payne, 2003) it is interesting to include work motivation as a variable in the psychosocial work environment. 1.1. Psychosocial factors, work motivation and stress Job stress has become a major issue among employees in the advanced industrial societies. Stress often depends on high job demands in relation to the worker’s abilities, frustrated aspirations and dissatisfaction regarding valued goals (Kalimo and Mejman, 1987). A basic hypothesis in stress theory is that psychosocial stressors in the work environment, such as quantitative overload, qualitative underload, lack of control and lack of social support, and the interaction of such conditions, may have harmful effects on an individual’s health and well-being (Karasek and Theorell, 1990; Levi et al., 1986; Melin and Lundberg, 1997). A physiological approach conceptualises work stress as a negative psychological state which is a result from a dynamic interaction between the person and their work environment (Cox et al., 2000). With this physiological approach it is important to develop methods to describe this state and the most direct method for this is, according to Kjellberg and Wadman (2002), a self-report instrument. An important question is how many and in which dimensions this emotional state shall be measured with. Mackay et al. (1978) developed an adjective checklist to

measure the emotional state in two dimensions that they labelled ‘‘stress’’ and ‘‘arousal’’. A similar instrument, with the dimensions ‘‘stress’’ and ‘‘energy’’ were developed by Kjellberg and Wadman (2002). Their ambition was to develop a mood adjective checklist useful in research of work environment. The stress- and energydimensions are, according to Kjellberg and Wadman (2002), partly similar to Karasek’s job strain model dimensions (1979). De Jonge et al. (1999) found that a work situation considered as having both high job demands and job control was related to a high degree of work motivation and job satisfaction. Moreover, Sargent and Terry (2000) found that high levels of job control and job demands had a positive effect on work satisfaction. One main theory of motivation is Herzberg’s two-factor theory (‘‘Motivation-Hygiene Theory’’), which also deals with job satisfaction (Herzberg et al., 1959). Herzberg suggests that certain factors motivate (‘motivators’, ‘satisfiers’), whereas others are preventive and will serve to remove the hindrance regarding positive job attitudes (‘hygiene factors’, ‘hygienes’). According to this theory true motivators come from achievement, recognition, the work itself, responsibility, advancement and the possibility of personal growth. In general, motivation comes from the desire or need of workers to be and do the very best that is in them. Despite the fact that work motivation occupies a central role in research and in practical reality, this concept cannot, according to Schou (1991) be quantified or exactly defined. Schou states that work motivation is a hypothetic concept that, among other things serves as an overarching concept for a large part of the research about what drives and maintains job performance. Robbins (1997) argued that motivation is the willingness to do something and is conditioned by this action’s ability to satisfy needs for the individual. According to Sims et al. (1993) motivation refers to forces acting to initiate and direct behaviour. Motivation requires an act that leads to some degree of satisfaction, which can be described as a chain of casual relationships; motivation gives performance that in turn gives satisfaction (Alvesson, 1993). Generally speaking, the total effects of the work environment on health are underpinned by complex associations between psychosocial stressors by means of direct, indirect and moderation effects (Cox and Ferguson, 1994). According to Le Blanc et al. (2000) there are processes such as cognitive, evaluative and motivational that intervenes between stressful stimuli and the reaction. Our model is a mediational one, in that we propose that job demand and job control impact on perceived stress is indirect rather then direct. In the processes of mediation in the psychosocial work environment and health relation, Cox and Ferguson (1994) express that ‘‘A mediator variable is one that is responsible for the transmission of an effect, but does not alter the nature of that effect’’ (Cox and Ferguson, 1994, p. 101).

ARTICLE IN PRESS L.G. Wallgren, J.J. Hanse / International Journal of Industrial Ergonomics 37 (2007) 51–59

Given the lack of studies examining motivation as a mediator in the psychosocial work environment and health relation our model are somewhat exploratory.

p1 Perceived Stress

Demand p2 p3

1.2. Information technology The IT sector (e.g. consultants) represents the pioneers regard to new technology within the industrial and public sector (Caplen Jensen et al., 2003), where the whole branch relies on the latest technology and technologically qualified workers (e.g. programmers, systems architects, IT solutions advisors, etc.) (Davies and Mathieu, 2005). IT consultants’ work may be categorised as knowledge intensive work (Alvesson, 2000). The IT sector is interesting to investigate, not only the interrelationship of IT and the organisation, but more specifically in relation to ergonomics (e.g. psychosocial work environment). Previous research indicates some inappropriate working conditions in the psychosocial work environment within the IT sector (e.g. Brown, 2002; Lim and Teo, 1999). This type of work may increase intrinsic motivation and inspire IT consultants to learn and develop the skills they need to meet increasing job demands, but at the same time it raises levels of job strain and other negative health outcomes (e.g. sickness). To the best of our knowledge, very little research has been published pertaining to work conditions, work motivation and health of IT consultants (Lim and Teo, 1999). However, some previous studies are interesting in this context. In a study among a group of IT-consultants in the United States, Brown (2002) found that a job that provides autonomy, challenge, feedback and ability to use skills was important in promoting job satisfaction and work motivation. However, this study had some limitations due to a small sample size (N ¼ 21), and especially a low response rate (49 per cent). Moreover, an intervention study in a Swedish IT-company focused on strategies that the organisation and employer could use to increase health (So¨derstro¨m et al., 2003). The authors described the psychosocial work environment as a job with high rate of work, high job demands and a high level of customeroriented activity. Most of the workers reported high levels of stress due to short advance planning. The workers also experienced a negative balance between job demands, job control and social support at work. So¨derstro¨m concluded that job demands seem to be the most important risk indicator related to strain and burn out. In a qualitative study regarding possible causes of ill-health in an IT workplace (a division in a bank) in Sweden it was pointed out that ‘‘impossible deadlines’’ impaired the employees’ ability to perform a good job (Dahlman, 2004).

53

p5 Motivators Control

p4

Fig. 1. Specification of the full structural equation model tested in this study, with the mediating effect of motivators. Only latent variables and their relations have been depicted, where p ¼ paths (regression weights). Manifest variables, error terms and disturbances in the endogenous variables have been omitted in the figure.

and perceived stress with motivators as the mediating variable. The hypothesized models were tested using full structural equation modeling (SEM). To the best of our knowledge, a process model has not been used in any previous empirical studies to analyse the associations between job characteristics, motivators and stress among IT-consultants. 1.4. Model specification (SEM) The model consists of the independent (exogenous) latent variables job demands and job control, and the dependent (endogenous) latent variable perceived stress. Support for the suggested covariance and path relationships in the structural model is justified by prior research findings. We hypothesise that: job demands is positively associated with perceived stress (p1) (e.g. Karasek and Theorell, 1990; Larsman et al., 2006) and job control (job autonomy) is negatively related to perceived stress (p2) (e.g. Karasek and Theorell, 1990; Melin and Lundberg, 1997). A full structural model used in the study is presented in Fig. 1. This process model consists of the independent latent variables job demands and job control, motivators which is treated as the mediating (intervening) variable and perceived stress which is considered as a dependent (endogenous) variable. Consistent with the first model, job demands is hypothesized to be positively associated with perceived stress (p1) and job control negatively associated with perceived stress (p2). Moreover, job demand is hypothesized to be negatively related to motivators (p3), job control positively related to motivators (p4), and finally, motivators are hypothesized to be negatively related to perceived stress (p5) (e.g. Levy, 2003; Moorhead and Griffin, 2004).

1.3. Aim

2. Methods

The lack of empirical data on IT-consultants’ working situation, and specifically the relationship between job characteristics and ‘work reactions’ gave rise to the following aim: to examine the relationship between job characteristics

2.1. Sample In this cross-sectional study, a questionnaire survey was conducted among non-management IT consultants at one

ARTICLE IN PRESS 54

L.G. Wallgren, J.J. Hanse / International Journal of Industrial Ergonomics 37 (2007) 51–59

company in Sweden. The study included 198 employed IT consultants and the overall response rate was approximately 84 per cent (N ¼ 167). Almost all consultants have a university degree of which approximately 80 per cent have a degree in computer science, engineering or business administration. The IT consultants worked full time, mainly with assignments connected to external customers within the industrial sector as well as the public sector. The IT consultants were hired by a customer for a specific purpose with a time limit and often worked within a projectoriented work organisation. As regards the period of employment at the company, approximately 9 per cent had more than 10 years of employment, 36 per cent between 4 to 9 years’ employment, and 55 per cent less than 4 years. Concerning work experience in the IT-sector, approximately 59 per cent had more than 10 years’ experience, 38 per cent between 4 to 9 years’ experience, and 3 per cent less than 4 years. Among the participants approximately 80 per cent were males and approximately 20 per cent were females. 2.2. Procedure of collecting data This study was web-based and accessible via Internet. Each participant received an e-mail that consisted of information about the study and also an Internet-link to the web page where the questionnaire was obtainable. In the e-mail, the purpose of the survey was explained and strict confidentiality was guaranteed. When the subjects were logged in at the web page an authorisation check was performed and data was stored on which subjects had answered the questionnaire. Thereby, these subjects were excluded from reminders. The questionnaire was constructed so that each question had to be answered to complete the questionnaire. The subjects’ answers were e1 e2 e3

D1 D2 D3

e5 e6

D4

All participants responded to a questionnaire containing questions on background variables, job demands, job control, motivators and perceived stress. Background variables: This part included questions concerning years of employment at the company (5-point response scale) and years of employment in the IT-sector (5-point response scale). Job demands: Workers’ attitudes to their job demands were operationalised by using a short version of the Swedish version of the ‘‘demand–decision latitude’’ questionnaire (Karasek, 1979), developed by Theorell and coworkers (Theorell et al., 1988; A˚kerstedt et al., 2002). This factor includes five questions: whether it was necessary to work fast (D1 in Fig. 2), hard (D2), with high effort (D3), enough time to do the job (reversed score) (D4), and whether demands were in conflict with each other (D5). The questionnaire uses a four-point response for each question and a frequency-based grading (never, seldom, sometimes and often). High values (scores) for job demands indicate high demands. Job control (Influence on and control over work): Workers’ attitudes to their job were assessed using part of a ‘‘Standardised’’ questionnaire’’ (Rubenowitz, 1997). The factor comprises five items and each item has five fixed response alternatives which are given points from 1 to 5, where 1 means very unsatisfactory and 5 very satisfactory. The items were: Influence on the rate of work (C1 in Fig. 2), influence on working methods (C2), influence on ,83

,70

ed2

,70

,64*** Demand

,47

D5 C1

2.3. Questionnaire

,62

,63

e4

stored in a database where the subjects were anonymous, i.e. it was impossible to track which subject had answered what questionnaire. It was only possible to check if the subject had answered the questionnaire. The questionnaire was accessible for 4 weeks and 3 reminders were sent out.

,04 -,30* -,14

,63

-,23

,85 ,77

Perceived ,83 Stress ,72

S1

e17

S2

e18

S3

e19

S4

e20

S5

e21

S6

e22

,80

,69

e7

C2

e8

C3

Control_1 ,65

,48**

e9

C4

e10

C5

,62 ,52

Control_2

Fig. 2. Structural equation model with no mediator among IT-consultants (N ¼ 167). Measurement and structural components with standardized estimates. The non-significant paths are marked as dotted lines. *po.05. **po.01. ***po.001.

ARTICLE IN PRESS L.G. Wallgren, J.J. Hanse / International Journal of Industrial Ergonomics 37 (2007) 51–59

stress dimension was considered. It was measured using the items ‘‘rested’’ (reversed score) (S1 in Fig. 2), ‘‘relaxed’’ (reversed score) (S4), ‘‘calm’’ (reversed score) (S6), ‘‘tense’’ (S2), ‘‘stressed’’ (S3), and ‘‘pressured’’ (S5). The respondents were instructed to think about how they usually feel at the end of a normal workday. There were six fixed response alternatives ranging from ‘‘not at all’’ to ‘‘to a very great degree’’.

the allocation of tasks (C3), technical control (C4), and influence of rules and regulations (C5). Previous studies indicate that the job control factor, as measured in this study, may be divided into two sub-factors (e.g. Johansson et al., 1993) and in some studies only the manifest variables C1–C3 have been used to measure job control (autonomy) (Pousette and Johansson Hanse, 2002). ‘‘Motivators’’ was operationalized by using a short form of the Minnesota Satisfaction Questionnaire (MSQ) (Weiss et al., 1967). The questionnaire began with an all-purpose question: ‘‘How satisfied am I with this aspect of my job?’’. In the current study, measures were across five different dimensions, with one question on each dimension: Responsibility (‘‘The freedom to use my own judgement’’, M1 in Fig. 3); Work itself/Variety (‘‘The chance to do different things from time to time’’, M2); Advancement (‘‘The chances for advancement on this job’’, M4); Recognition (‘‘The praise I get for doing a good job’’, M5); Achievement (‘‘The feeling of accomplishment I get from the job’’, M6). In the current study a question for the motivator ‘‘Possibility of growth’’ (Herzberg et al., 1959) was created (‘‘The possibility of learning new skills so I develop and grow as a person’’, M3). The questionnaire uses the following five response alternatives: Very Satisfied, Satisfied, Neither Satisfied nor Dissatisfied (‘‘N’’), Dissatisfied, Very Dissatisfied. Perceived stress: Perceived stress was assessed using a mood adjective checklist (Kjellberg and Wadman, 2002). This checklist, which was constructed to be used to describe mood during work, contains two mood dimensions—stress and energy. In the present study only the

e1

D1

e2

D2

e3

D3

e4 e5 e6

D4

,83

ed2

,70

,25 ,47

,72 -,10

-,31**

-,38*

-,14

C3 C4

S2

e18

S3

e19

S4

e20

S5

e21

S6

e22

,55*** ,35**

,61

e17

,11

Control_1

C2

S1

,80

,64

,49**

e9

,77

Perceived ,83 Stress

Demand

,65

e8

,85

,68***

,68

e7

Input data to SEM consisted of the raw data that was stored in SPSS version 11.5. The current study had no internal missing values. Structural equation modelling was performed using the maximum likelihood methods of AMOS version 5.0 (Arbuckle, 2003). The focal point when analysing structural equation models is the extent to which the hypothesised model fits or not, in other words, adequately describes the sample data. In accordance with the classification of recommended fit indices, a number of fit indices were considered (Schumacker and Lomax, 1996). The Chi-square value statistic is a goodness-of-fit measure that assesses the magnitude of the discrepancy between the sample (the observed) covariance matrix and the estimated (fitted) covariance matrix (Hu and Bentler, 1995). Jo¨reskog and So¨rbom (1993) propose that the Chi-square value statistic be regarded as a measure of fit rather than as a test statistic, i.e., a measure of overall fit of the model to the data. A large, statistically significant value relative to the degrees of freedom indicates poor model fit. The Chi-square value

,70

D5 C1

2.4. Statistical analysis

,61

,64

55

,73

,70

Motivators ,43

,51

ed1

,66

,53

M1 M2

M3 M4

M5 M6

e11

e13

e15

,52

e10

C5

Control_2 e12

e14

e16

Fig. 3. Full structural equation model with motivators as mediator among IT-consultants (N ¼ 167). Measurement and structural components with standardized estimates. The non-significant paths are marked as dotted lines. *po.05. **po.01. ***po.001.

ARTICLE IN PRESS 56

L.G. Wallgren, J.J. Hanse / International Journal of Industrial Ergonomics 37 (2007) 51–59

statistic is sensitive to sample size and with a large sample size even trivial differences may result in the rejection of the specified model (Hu and Bentler, 1995). The normed Chisquare measure is the ratio of the Chi-square value to its degrees of freedom, where ratios in the range of 2 to 1 are indicative of an acceptable fit between the hypothetical model and the sample data. Values below 1.0 indicate an ‘‘overfitted’’ model (Schumacker and Lomax, 1996), and values larger than 2.0, or the more liberal limit of 5.0, indicate that the model does not fit the observed data and needs improvement. However, there is no consensus on what precisely represents a good fit (Bollen, 1989). The root mean square error of approximation (RMSEA) is a measure of the discrepancy per degree of freedom for the model (Browne and Cudeck, 1993). Values of about 0.05 or less indicate a close fit of the model to the data and values of about 0.08 or less indicate a reasonable error of approximation (Browne and Cudeck, 1993). Finally, PCLOSE is a ‘p-value’ for testing the null hypothesis that the population RMSEA is no greater than .05 (i.e. H0: RMSEA ¼ o.05) (Arbuckle, 2003). To measure mediation the Sobel test was used (Sobel, 1982), which test the significance of the intervening variable effect, also called indirect effect. 3. Results 3.1. A model of job characteristics and perceived stress Structures for relationships between the latent variables was specified and analysed. The standardised path coefficients for the model are reported in Fig. 2. The structural model with no mediator (See Fig. 2) showed good fit to the data (Chi-square ¼ 143.17; df ¼ 98; po.01; Normed Chi-square ¼ 1.46; RMSEA ¼ .053; PCLOSE ¼ .39). All paths (factor loadings) in the measurement models were significant. In the structural part, only the path between job demand and perceived stress was significant in the expected direction. High job demand (statistically) predicted high perceived stress among the ITconsultants. The non-significant paths were from the job control variables to perceived stress (when controlling for other independent latent variables). 3.2. A process model of job characteristics, motivators and perceived stress The second proposed model, i.e. the full structural model with a mediator (See Fig. 3), showed good model fit (Chi-square ¼ 299.27; df ¼ 199; po.001; Normed Chisquare ¼ 1.50; RMSEA ¼ .055; PCLOSE ¼ .25). All paths in the measurement models were significant, including the measurement model of the mediator. In the structural part, four out of seven paths were significant and also in the expected direction (when controlling for other latent variables). In accordance with the first structural model, the path between job demand and perceived stress was

significant. The two job control latent variables were both significantly related to motivators, which mean that high job control (statistically) predicted high scores in the latent variable (‘‘motivators’’). Furthermore, motivators were significantly, but negatively (i.e. expected direction), related to perceived stress. In sum, job control was significantly related to changes in motivators, which in turn affects perceived stress. Test of the indirect effect show that the effect of control_1 on perceived stress through motivators was almost statistically significant (‘‘tendency’’) (Sobel test ¼ 1.90; p ¼ :06). The indirect effects of job demand, respectively, job control_2 on perceived stress through motivators was not significant (p ¼ :30, respectively, p ¼ :12). 4. Discussion This study has examined the impact of specific job characteristics and motivators among IT-consultants on perceived stress. This particular group of workers constitutes an interesting population to study within the knowledge intensive sector. The focus in the current study was on IT-consultants who had knowledge-intensive work within a project-oriented work organisation. The results show that job demand was positively related to perceived stress, which confirms earlier studies (e.g. Karasek and Theorell, 1990; Melin and Lundberg, 1997). The results in the current study also indicate the importance of motivators (e.g. responsibility, recognition, achievement, possibility of growth) among IT-consultants in the job stress and performance framework. 4.1. Job demand, job control and stress In general, in a work situation characterised by prolonged exposure to high job demands and low job control the IT-consultant may not be able to reduce his or her state of stress (high activation is sustained) which can, for example, give rise to tension (Melin and Lundberg, 1997), difficulties in decision making (cognitive), headaches and musculoskeletal symptoms (Karasek and Theorell, 1990). On the organisational level this can lead to poor work performance and declined productivity (Le Blanc, de Jonge and Schaufeli, 2000). How IT-consultants respond to the job demands will partly depend on their role in the company and their mental and physical capacities. The result from the present study show that job demands have a significant impact on perceived stress, which confirms earlier research result (e.g. Jacobsson et al., 2001). These studies also showed that job control did not have any significant impact on stress, which the results from the present study also demonstrate. An explanation for the observation that job control did not have an equally strong direct impact on stress as job demand could be compared to what Frankenhaeuser and O¨dman (1983) call ‘‘decision stress’’, which can affect

ARTICLE IN PRESS L.G. Wallgren, J.J. Hanse / International Journal of Industrial Ergonomics 37 (2007) 51–59

certain persons with high job control. This decision stress means a permanent readiness to review decisions and to take new decisions and it is not possible to specify when this type of job control changes from being something positive to something negative. Another explanation can be that opportunity for job control is, according to Warr’s (1994) framework, an ‘‘AD-factor’’. In his ‘‘vitamin model’’ he draws an analogy between the way in which vitamins act on the human body and the effects of job characteristics on people’s affective well-being (mental health). According to Warr certain (AD) factors exist that have a positive effect, but only up to a certain level. If this level is exceeded, affective well-being is influenced negatively (i.e. inverted U-shaped curvilinear effects). 4.2. Motivators as mediator To motivate an individual (e.g. an IT consultant), a job itself must be challenging, have scope for enrichment and be of interest. Motivation is a major component of any theory that attempts to predict and explain organisational behaviour and performance and this central role has been reflected in the literature. However, for those who have tried to define motivation it has turned out to be a complicated process. Motivation is, according to Jewell (1998) a hypothetical construct and it is something that cannot be seen. Furthermore Mitchell’s (1997) opinion is that motivation is not directly observable and therefore must be tested as a part of a larger theory in which antecedent conditions are linked to behavioural actions. The motivation theory which maybe mostly deals with the individual’s behaviour in the organisation is the ‘‘twofactor theory of motivation’’ evolved by Herzberg et al. (1959). Factors, which increase the individual’s job satisfaction, were labelled ‘‘motivators’’ and contrariwise factors that lead to an unhappy employee were called ‘‘hygiene factors’’. Instead of trying to define motivation itself Herzberg strived to define antecedent conditions which presumably lead to motivation. This distinction between motivators and hygiene factors in the Herzberg theory provides, according to Hackman and Oldham (1980), a clear and straightforward way of thinking about employee motivation and of predicting the likely impact of various planned changes on motivation. In the present study ‘‘motivators’’ was measured using selected questions in the Minnesota Satisfaction Questionnaire (MSQ) (Weiss et al., 1967) which were compatible with the factors defined as motivators by Herzberg et al. (1959). The results from the full structural equation model with motivators as the mediator (Fig. 3) show that job demand did not have any significant effect on the motivators (when controlling for job control). Generally speaking, similar results were established by Van Yperen and Hagedoorn (2003) in their study of nurses, and in a longitudinal study of bank officials and teachers by Houkes et al. (2003). Moreover, the present study shows that job control (both control_1 and control_2 in Fig. 3) has a significant impact

57

on motivators. In an exploratory study of job satisfaction and work motivation among IT consultants Brown (2002) got a similar result. She found the strongest correlation between motivation, measured with Hackman and Oldham’s (1980) Motivating Potential Score (MPS), and ‘‘autonomy’’. According to Hackman and Oldham (1980), autonomy is the degree to which the job provides substantial freedom and independence and can be compared with job control in the present study. But, perhaps Brown’s result is not reliable, partly for the reason that autonomy is a part of the calculation of MPS which gives a high inter-correlation, and partly due to a low response rate. In addition, the result in the current study agrees only partially with Karasek and Theorells (1990) opinion as regards the degree of motivation as a consequence of level of job demands and the degree of job control. However, their opinion is based on a general model. The current result indicates that it is important to examine specific occupational context to better understand how certain job characteristics influence motivators and stress. In general, Pousette and Johansson Hanse (2002) have the same views in a study comparing a generic model with occupationspecific (i.e. multiple-sample) models across occupations (e.g. industrial white-collar workers, blue-collar workers). Pousette and Johansson Hanse (2002) found that values of model parameters vary across occupations, which have important implications for the differentiation of prevention and intervention in different occupational contexts. It is important to address the relation between the individual IT consultant and the work organisation. The current study showed that job characteristics have an impact on perceived stress and, in addition, the importance of motivators in the job stress framework. Comprehensive stress interventions are suggested, which address both the stress exhibited by IT consultants as well as the organisational origins of stress at work. Since the managers have a substantial influence on the work organisation, it is important to involve and educate managers on the subject. 4.3. Study methods and limitations Our hypotheses concerning job demand, job control, motivators and perceived stress are theoretically based on earlier research and, as such, considered plausible. However, the proposed model can only be seen as an ‘‘as if’’ model of causality (Kline, 1998). The data are crosssectional, which implies that the relationships observed cannot be interpreted causally and will need to be replicated longitudinally within the IT-sector. Moreover, a deeper understanding of what motivates IT-consultants should be advantageous to most industrial organisations— it is important to have the stories behind the statistics. The measures were limited to self-reports. However, the use of the self-reported work environment is based on the fundamental principle that people know their own problems best and they should be encouraged to speak for themselves (Levi, 1987; Rubenowitz, 1997). The

ARTICLE IN PRESS 58

L.G. Wallgren, J.J. Hanse / International Journal of Industrial Ergonomics 37 (2007) 51–59

individual’s own perception of the working conditions must be considered if one is to obtain a good picture of how work influences his or her perceived stress. Jewell (1998) also presents the same argument in relation to measuring experienced job satisfaction. To use multiple measures (questions) of each construct (latent variable) tends to reduce the effect of measurement error (Kline, 1998). The constructs used in the present study are based on earlier theoretical and empirical research by Herzberg et al. (1959), Karasek and Theorell (1990), Kjellberg and Wadman (2002), and Weiss et al. (1967). Generally speaking, it is possible to assess construct validity (Shadish et al., 2002) using confirmatory factor analysis when the underlying theory is accessible. The present study shows good construct validity, thus confirming previous theoretical and empirical research. However, it was not possible to verify the way Rubenowitz (1997) operationalised ‘job control’. Instead the results show a better model fit and measurement model (i.e. higher factor loadings) if the construct was divided into two parts, which has been indicated in earlier studies (Johansson et al., 1993). 5. Conclusion It seems reasonable to conclude that the job characteristics demand and control constitute important factors to illuminate the genesis of perceived stress among ITconsultants. The results also point out the importance of motivators (e.g. responsibility, recognition, achievement, possibility of growth) among IT-consultants in the job stress and performance framework. The results indicate that motivators in part mediate the relation between job control and perceived stress. To strengthen the results in the current study, it is important to replicate the model with a longitudinal design in order to substantiate causality. Acknowledgements The authors gratefully acknowledge Ph.D. Pernilla Larsman at Go¨teborg University for valuable comments on using the SEM technique. In addition, the authors would like to express their gratitude to the anonymous reviewers for their helpful comments and suggestions on an earlier version of this paper. References A˚kerstedt, T., Knutsson, A., Westerholm, P., Theorell, T., Alfredsson, L., Kecklund, G., 2002. Sleep disturbances, work stress and work hours. A cross-sectional study. Journal of Psychosomatic Research 53, 741–748. Alvesson, M., 1993. Organisationsteori och teknokratiskt medvetande. Nerenius & Sante´rus fo¨rlag, Stockholm (in Swedish). Alvesson, M., 2000. Social identity and the problem of loyalty in knowledge-intensive companies. Journal of Management Studies 37, 1101–1123.

Alvesson, M., 2003. Knowledge Work and Knowledge-Intensive firms. Oxford University Press, Oxford. Arbuckle, J.L., 2003. Amos 5.0 update to the Amos user’s guide. SmallWaters Corporation, Chicago. Bollen, K., 1989. Structural Equations with Latent Variables. Wiley, New York. Brown, M.M., 2002. An exploratory study of job satisfaction and work motivation of a selected group of information technology consultant in the Delaware valley. Doctoral Dissertation. Wilmington College, United States. Browne, M.W., Cudeck, R., 1993. Alternative ways of assessing model fit. In: Bollen, K.A., Long, J.S. (Eds.), Testing Structural Equation Models. Sage Publications Inc, Newbury Park, pp. 136–162. Caplen Jensen, B., Netterstro¨m, B., Borg, V., 2003. ARIT—Arbetsmiljo¨et indenfor IT-branchen. Arbejdsmiljo¨instituttet, Copenhagen (in Danish). Cox, T., Ferguson, E., 1994. Measurment of subjective work environment. Work & Stress 8, 98–109. Cox, T., Griffiths, A., Rial-Gonzales, E., 2000. Research on Work-related Stress. Office for official publications of the European communities, Luxembourg. Dahlman, I., 2004. En studie av mo¨jliga orsaker till oha¨lsa pa˚ en ITarbetsplats. Examensarbete vid Arbetslivsinstitutets fo¨retagsla¨karutbildning. Arbetslivsinstitutet (National Institute for Working Life, NIWL), Stockholm (in Swedish). Davies, K., Mathieu, C., 2005. Gender inequality in the IT sector in Sweden and Ireland. Work Life in Transition 2005. National Institute for Working Life (NIWL), Stockholm, p. 3. De Jonge, J., van Breukelen, J.P., Landeweerd, J.A., Nijhuis, F.J.N., 1999. Comparing group and individual level assessments of job characteristics in testing the job demand-control model: a multilevel approach. Human Relations 52, 95–122. Docherty, P., Huzzard, T., 2003. Marknads-, management- och medarbetartrender 1985–2005. In: von Otter, C. (Ed.), Ute och inne i svenskt arbetsliv. Arbetsliv i omvandling 2003, Vol. 8. Arbetslivsinstitutet (National Institute for Working Life, NIWL), Stockholm, pp. 135–157 (in Swedish). European Foundation, 2000. Living conditions, working conditions and industrial relations in the knowledge society. European Foundation for the Improvement of Living and Working Conditions, Dublin. European Foundation, 2005. The knowledge-intensive business services sector—what future? European Foundation for the Improvement of Living and Working Conditions, Dublin. Frankenhaeuser, M., O¨dman, M., 1983. Stress en del av livet. Brombergs Bokfo¨rlag AB, Va¨rnamo, Sweden (in Swedish). Hackman, J.R., Oldham, G.R., 1980. Work Redesign. Addison-Wesley Publishing Company, Reading, MA. Herzberg, F., Mausner, B., Snyderman, B.B., 1959. The Motivation to Work. Wiley, New York. Houkes, I., Janssen, P.P.M., De Jonge, J., Bakker, A.B., 2003. Specific determinants of intrinsic work motivation, emotional exhaustion and turnover intention: a multisample longitudinal study. Journal of Occupational and Organizational Psychology 76, 427–450. Hu, L.T., Bentler, P.M., 1995. Evaluating model fit. In: Hoyle, R.H. (Ed.), Structural Equation Modeling. Concepts, Issues, and Applications. Sage, London, pp. 76–99. Iverga˚rd, T., 2000. An ergonomics approach for work in the next millennium in an IT world. Behaviour & Information Technology 19, 139–148. Jacobsson, C., Pousette, A., Thylefors, I., 2001. Managing stress and feelings of mastery among Swedish comprehensive school teachers. Scandinavian Journal of Educational Research 45, 37–53. Johansson, J.A˚., Kadefors, R., Rubenowitz, S., Klingenstierna, U., Lindstro¨m, I., Engstro¨m, T., Johansson, M., 1993. Musculoskeletal Symptoms, Ergonomic Aspects and Psychosocial Factors in Two Different Truck Assembly Concepts. International Journal of Industrial Ergonomics 12, 35–48. Jewell, L.N., 1998. Contemporary Industrial/Organizational Psychology, third ed. Brooks/Cole Publishing Company, Pacific Grove.

ARTICLE IN PRESS L.G. Wallgren, J.J. Hanse / International Journal of Industrial Ergonomics 37 (2007) 51–59 Jo¨reskog, K., So¨rbom, D., 1993. LISREL 8: Structural Equation Modeling with the SIMPLIS command language. Lawrence Erlbaum Associates Publisher, London. Kalimo, R., Mejman, T., 1987. Psychological and behavioural responses to stress at work. In: Kalimo, R., El-Batawi, M.A., Cooper, C.L. (Eds.), Psychosocial Factors at Work and their Relation to Health. World Health Organization (WHO), Geneva, pp. 23–36. Karasek, R., 1979. Job demands, job decision latitude, and mental strain: implications for job redesign. Administrative Science Quarterly 24, 285–308. Karasek, R., Theorell, T., 1990. Healthy work. Stress, Productivity, and the Reconstruction of Working Life. Basic Books Inc, New York. Kjellberg, A., Wadman, C., 2002. Subjektiv stress och dess samband med psykosociala fo¨rha˚llanden och besva¨r: En pro¨vning av Stress-Energimodellen (Arbete och Ha¨lsa 2002), Vol. 12. Arbetslivsinstitutet (National institute for working life, NIWL), Stockholm (in swedish). Kline, R.B., 1998. Principles and Practice of Structural Equation Modelling. The Guilford Press, New York. Ka¨rreman, D., Sveningsson, S., Alvesson, M., 2003. The return of Machine Bureaucracy? Management control in the work settings of professionals. International Studies of Management & Organization 32, 70–92. Larsman, P., Sandsjo¨, L., Klipstein, A., Vollenbroek-Hutten, M., Christensen, H., 2006. Perceived work demands, felt stress, and musculoskeletal neck/shoulder symptoms among elderly female computer uses. The NEW study. European Journal of Applied Physiology 96, 127–135. Le Blanc, P., De Jonge, J., Schaufeli, W., 2000. Job stress and health. In: Chmiel, N. (Ed.), Introduction to Work and Organizational Psychology: A European Perspective. Blackwell Publishers Ltd, Oxford, UK, pp. 148–177. Levi, L., 1987. Fitting work to human capacities and needs: improvements in the content and organization of work. In: Kalimo, R., El-Batawi, M.A., Cooper, C.L. (Eds.), Psychosocial Factors at Work and their Relation to Health. World Health Organization (WHO), Geneva, pp. 168–184. Levi, L., Frankenhaeuser, M., Gardell, B., 1986. The characteristics of the workplace and the nature of its social demands. In: Wolf, Jr., S.G., Finestone, A.J. (Eds.), Occupational Stress. Health and Performance at Work. PSG Publishing Company, Inc., Littleton, Massachusetts, pp. 54–67. Levy, P.E., 2003. Industrial/organizational Psychology. Understanding the Workplace. Houghton Mifflin Company, New York. Lim, V.K.G., Teo, T.S.H., 1999. Occupational stress among information technology personnel in Singapore. International Journal of Information Management 19, 277–291. Locke, E.A., Latham, G.P., 2002. Building a practically useful theory of goal setting and task motivation: a 35-year odyssey. American Psychologist 57, 705–717. Mackay, C.J., Cox, T., Burrowa, G., LA, J., 1978. An inventory for the measurement of self reported stress and arousal. British Journal of Clinical Psychology 17, 283–284. Melin, B., Lundberg, U., 1997. A biopsychosocial approach to work-stress and musculoskeletal disorders. Journal of Psychophysiology 11, 238–247.

59

Mitchell, T.R., 1997. Matching motivational strategies with organizational contexts. In: Cummings, L.L., Staw, B.M. (Eds.), Research in Organizational Behavior. JAI Press Inc., Greenwich, pp. 57–149. Moorhead, G., Griffin, R.W., 2004. Organizational Behavior. Managing People and Organizations, seventh ed. Houghton Mifflin Company, New York. Pousette, A., Johansson Hanse, J., 2002. Job characteristics as predictors of ill-health and sickness absenteeism in different occupational types— a multigroup structural equation modelling approach. Work & Stress 16, 229–250. Pritchard, R.D., Payne, S.C., 2003. Performance management practices and motivation. In: Holman, D., Wall, T.D., Clegg, C.W., Sparrow, P., Howard, A. (Eds.), The New Workplace: A Guide to the Human Impact of Modern Working Practices. Wiley, West Sussex, pp. 219–244. Robbins, S.P., 1997. Essentials of Organizational Behavior, fifth ed. Prentice-Hall, Englewood Cliffs, NJ. Rubenowitz, S., 1997. Survey and intervention of ergonomic problems at the workplace. International Journal of Industrial Ergonomics 19, 271–275. Sargent, L.D., Terry, D.J., 2000. The moderating role of social support in Karasek0 s job strain model. Work & Stress 14, 245–261. Schou, P., 1991. Arbetsmotivation—En studie av ingenjo¨rer (Work motivation—a study of engineers). Doctoral Dissertation. Stockholm School of Economics, Sweden (in Swedish). Schumacker, R.E., Lomax, R.G., 1996. A Beginner’s Guide to Structural Equation Modeling. Lawrence Erlbaum, Mahwah, NJ. Shadish, W.R., Cook, T.D., Campbell, D.T., 2002. Experimental and Quasi-experimental Designs for Generalized Causal Inference. Houghton Mifflin Company, New York. Sims, D., Fineman, S., Gabriel, Y., 1993. Organizing & Organizations. SAGE Publications Ltd, London. Sobel, M.E., 1982. Asymptotic confidence intervals for indirect effects in structural equation models. In: Leinhardt, S. (Ed.), Sociological Methodology 1982. American Sociological Association, Washington DC, pp. 290–312. Statistics Sweden, 2002. Trender och Prognoser 2002. Statistiska Centralbyra˚n (SCB), Stockholm (in Swedish, English summary). So¨derstro¨m, M., Jeding, K., Ekstedt, M., Kecklund, G., A˚kerstedt, T., 2003. Arbetsmiljo¨, stress och utbra¨ndhet inom ett fo¨retag i ITbranschen (The work environment, stress and burn out in a Swedish IT company). Stressforsknings-rapport, nr 312. Institutet fo¨r psykosocial medicin (IPM), Karolinska institutet, Stockholm (in Swedish, English summary). Theorell, T., Perski, A., A˚kerstedt, T., Sigala, F., Ahlberg-Hulte´n, G., Svensson, J., Eneroth, P., 1988. Changes in job strain in relation to changes in physiological state. A longitudinal study. Scandinavian Journal of Work, Environment & Health 14, 189–196. Van Yperen, N.W., Hagedoorn, M., 2003. Do high job demands increase intrinsic motivation or fatigue or both? The role of job control and job social support. Academy of Management Journal 46, 339–348. Warr, P., 1994. A conceptual framework for the study of work and mental health. Work & Stress 8, 84–97. Weiss, D.J., Dawis, R.V., England, G.W., Lofquist, L.H., 1967. Manual for the Minnesota Satisfaction Questionnaire. Industrial Relations Center, University of Minnesota, Minneapolis.