Technological innovation is important for industrial organizations trying to survive in competitive markets. However, innovation is never a simple nor a smooth process. Faced with major technological changes, people react differently; some seem to relish the challenge, but many show symptoms of stress. In the 1970s new computer technologies started to change nearly every workplace, and also influenced private life. People feared that stress and unemployment would be the future consequences of the technological revolution.
Scientists started empirical research that might support or contradict these expectations, and developed implementation strategies in order to reduce some of the potential negative consequences.
Nowadays computer systems have become standard, at least in modern industries and administration. The cycle of computer hardware and software innovation has accelerated. Every year or so new hardware and software systems are released, to which people have to adapt. The hardware components of new technologies are also changing more rapidly in our times. In the meantime, it has become evident that innovation and new technologies are not the source of unemployment or of low qualification requirements per se (Welsch, 1989). The question is how people react to such permanent innovation, whether it results in stress, and which practical psychological consequences should be drawn to help them cope with permanent technological innovation. One of the central psychological problems of all these different changes relates to highly complex interrelated systems and the resulting risk of errors. Complexity and error management, therefore, will be major problems that will have to be looked at in relation to stress and coping competencies.
This chapter gives a summary of research on new technologies and stress and discusses practical implications for the implementation process. It presents cross-sectional research and longitudinal studies. Following this, practical consequences concerning implementation strategies, job (re)design and software design are presented. The final section addresses the problem of permanent technological innovation and gives a vision of adaptive learning organizations. THE IMPACT OF NEW TECHNOLOGIES ON STRESS
Basic Stress Model
Faced with technological changes, not all people show stress reactions. According to the transactional stress model (Lazarus, 1976), the individual appraisal of the stressor, the resources, the reappraisal of the stressor and coping competences are important. Some people clearly seem to appreciate the challenge of technological novelty. Only subjects who expect a long-lasting aversive experience after an appraisal of the whole situation and its consequences will react with stress. Research into the impact of new technologies on stress therefore has to take into account many different factors and conditions that might affect the appraisal process and the stress reaction. ln order to develop a systematic research overview we have to consider types of stressors, different resources, short- and long-term consequences, and also the means of measuring these consequences.
A Definition of Stress in the Context of New Technologies
We define stress as a state of intense and aversive tension, which the subject strongly wants to avoid. The sensation of stress depends on the expected persistence, closeness and lack of control of the situation (Greif, 1991a). The application of this definition implies several seemingly trivial, but often neglected, consequences. Stress is not suffered by people who before or after the implementation of new technologies neither show nor expect persistent aversive tension, nor want to avoid the change. Most people, apparently, do not worry about technological changes which may or may not come in the long run. People who are able to control the technological problems, either by avoiding or by learning to manage them, will show no stress reactions.
Stressors can be defined simply as factors that are assumed to release stress reactions with high probability (Semmer, 1984). The technology itself, but also the expected direct or mediated consequences of its implementation, may be a source of stress. Therefore, we have to take into account the whole range of different possible factors and sources known from stress research that elicit stress reactions. Following the transactional stress model and action theory, Semmer (1984) distinguishes three major groups of hypothetical stressors at work, which can be applied here:
1. Additional demands on the action regulation process or control of task performance (for example, long and unpredictable system response time, cumbersome handling, and indirect consequences such as time pressure following from higher work demands),
2. Regulatory insecurity or insufficient control of actions resulting from overload or ambiguous performance feedback (for example, complex hardware and software systems or incomprehensible handbooks, unclear wording and feedback, the program’s error messages and, finally, higher complexity of the whole set of tasks).
3. Conflicting goals or lack of task and role clarity (for example, conflicts between time pressure and quality standards or unclear role changes resulting from the implementation of the technology).
In the scientific literature short-term and long-term consequences are differentiated. Examples of short-term consequences are biochemical and psychophysiological reactions such as increased blood pressure, pulse rate and catecholamine excretion. Moreover consequences like eye or musculoskeletal strain symptoms and lower performance efficiency (especially a higher rate of errors) are also short-term indicators that are found to be related to stress in some studies. The whole range of indicators mentioned above has also been applied in the research on stress induced by computer technology. Typical long-term consequences that have been examined in this field include reduced well-being, psychosomatic complaints and diseases. The following section gives a summary of research results on special factors or sources of stress possibly caused by new technologies.
Research on Different Factors and Sources of Stress
Several basic sources of organizational stress can be distinguished. Most of them have also been investigated in conjunction with work and new technologies: e.g. job demands, job control, job content as well as human factors constraints and career/future concerns (Briner & Hockey, 1988; Carayon, 1993; Carayon et al., 1995; Frese, 1991a). Human factor constraints are related to hardware and software properties of computing environments. Important hardware components that have been studied in experiments and field studies are workstation layout, input devices (keyboards, mice etc.) and visual display units (VDUs). In the case of VDUs, results vary. Whereas C¸ akir (1981) and Zeier et al. (1987) found no properties directly related to physiological reactions and musculoskeletal discomfort Sauter et al. (1991) reported an association between work with VDUs and musculoskeletal complaints.
Furthermore, empirical evidence has been found on the impact of delayed or unpredictable system response times on stress (Johansson & Aronsson, 1984). In a summary of experimental research applying psychophysiological measures, Boucsein (1988) infers that system-response times that are either too long or too short may induce stress. There is lack of evidence about the direct impact of software design on stress. Only particular problems such as system failures, especially crashes, definitely appear to cause stress. White-collar workers showed significantly different adrenalin excretion and diastolic blood pressure during a breakdown, in comparison to ordinary conditions, in an investigation by Johansson & Aronsson (1984). Zapf (1993) also identified the malfunctioning of computer systems as a typical “computer hassle” causing stress.
The impact that errors have on stress and emotional reactions has been investigated recently (Krone et al., 2000; Reason, 1990; Zapf, 1991, 1993). The practical relevance of the research on human errors is high. Observation shows that both novices and experts using standard software systems make mistakes or action slips every fewminutes. This may lead to negative consequences for the organization and for the employees as well. There is empirical evidence that shows that errors are related to work and system complexity. For example, Zapf (1993) found an increase in errors related to high work complexity. Corresponding experimental results show that errors related to planning processes were particulary made by inexperienced users confronted with high task and system complexity (Hamborg, 1996). Regarding the increasing complexity of software systems in office work, errors will become even more likely in the future. This effect is probably strengthened by faster hardware and software innovation. Today, most software systems and even hardware components are updated and replaced by new releases after one or two years. Very often the quality and stability of the first releases of new hardware and software systems remain insufficient. Shorter cycles of innovation require permanent learning and adaptation to new and potentially unstable changes.
How the resulting problems can be classified depends on the type of hardware and software malfunction and on individual coping competencies. If the problems are severe (for example, when tasks cannot be performed correctly, task achievement is impeded or information gets lost) such events may arouse high stress reactions and even states of panic. The frequent repetition of minor malfunctions, errors and action slips that users experience in their daily work with computers may be classified as “microstressors” or daily hassles. Zapf (1991, 1993) has studied the cumulative effects of such computer hassles. He found that mental as well as physical workload and stress reactions are related to an observable number of performance errors. An increased workload and stress result in a substantial decrease of performance efficiency and quality. Moreover unsuccessful troubleshooting affects the emotional and motivational state and thus leads to frustration and irritation in the long run. Carayon (1997, p. 330f ) points out that the cumulative effect of so-called acute stressors like slow computer performance and computer breakdowns can result in chronic symptoms of stress. Even though the technology may not cause any difficulties, the anticipation of errors due to lack of knowledge can, according to the above definition of stress, be enough to cause stress.
The design ofworkstations, hardware and software components, the furnishings and interior equipment are important factors related to stress. Scientific field research on ergonomic and usable design, arrangement of the hardware equipment (especially the screen, keyboards and furnishings) has been applied by industrial firms, and has become an essential aspect of their marketing strategy.
Research shows that it is insufficient to consider human factors constraints and the ergonomic design of the equipment as the only source of stress. Several authors emphasize that the type ofwork carried out at VDUs and the embedding organizational conditions seem to be the main causes of health complaints and stress reactions (Agervold, 1987; Briner & Hockey, 1988; C¸ akir, 1981; Frese, 1991a). New technology may lead to changes in work structure and human–machine redivision of labour (Buchanan & Boddy, 1982; Levi, 1994; Turner & Karasek, 1984). This may result in changes of work demands, e.g. work overload or underload, time pressure (Saupe & Frese, 1981; Schulz & H¨ofert, 1981) and interruptions (Johansson & Aronsson, 1984; Leitner et al., 1993), as well as anxiety (Mohr, 1991), uncertainty (Saupe & Frese, 1981; Turner & Karasek, 1984), lack of job control (Buchanan & Boddy, 1982, Sauter et al., 1983) and career concerns like uncertainty about job future and career advancement (Carayon, 1993). The following section briefly describes the role of intervening or moderating variables. Field studies on the overall consequences of new technologies and their implementation are addressed later.
Resources
It is well known from stress research that stress reactions and the resulting long- and shortterm consequences are mediated by so-called intervening variables or resources, serving as buffers protecting the person from stress effects. Control, technological knowledge and competence, as well as support are the major resources that have been studied in this field.
Control
Control, defined as the subjective probability of reducing stress reactions, is often mentioned as an important buffering resource (cf. Frese & Brodbeck, 1989; Greif, 1991a; Johansson, 1989). This kind of control at work implies the possibility of successfully changing environmental conditions or one’s own activities (cf. Frese, 1989). Low decision latitude and little control over scheduling of tasks may follow from the inadequate (re)design of jobs after the implementation of new technologies (Buchanan & Boddy, 1982; Frese & Zapf, 1987). Carayon (1993), however, has found that while job demands and career/future concerns are related to stress outcomes in office work, job control is not. According to Jones and Fletcher (this volume) the latter point seems not to be a clear cut issue.
Technological Knowledge and Competences
Technological knowledge and coping competences are very powerful resources meeting the challenge of technological change. Such changes require intensive, adaptive effort. Briner & Hockey (1988) supposed that in the short-term, differences between old and new work demands and the lack of competences are likely to be the main source of stress. Training and development of the necessary technological knowledge and skills provide security and selfconfidence. Expert knowledge and competences reduce the probability of complex errors. Furthermore, the development of special error management competences appears to be an important resource preventing stress (Frese et al., 1991b). Therefore, the implementation of new technologies should be accompanied by the extensive training of novices.
Social Support and Help
Social support and help by colleagues and experts is a well-known resource moderating stress reactions (Frese&Semmer, 1991; Udris, 1989). In the implementation process of new technologies the hardware and software retailers are normally expected to provide a guaranteed hotline to help the users. The development of an internal social support network within the work environment seems to facilitate individual learning, problem solving and error management beyond the initial period of formal training (Briner & Hockey, 1988; Dutke & Sch¨onpflug, 1987; Greif, 1986). The results of some studies show that the implementation of new technologies may reduce social interaction (Buchanan & Boddy, 1982; Turner & Karasek, 1984; Stellman et al., 1987). Therefore, special investments in the development of personal help networks, and a positive social team and organizational climate may be necessary to compensate for an impairment of such resources. For example in training seminars, team development, reinforcement of supporting interactions between the participants and agreement to help each other when faced with problems in the workplace may be at least as important as filling any technological knowledge gaps that the participants may have. RESEARCH ON THE IMPLEMENTATIO OF NEW TECHNOLOGIES
Cross-Sectional Studies
Studies on the overall influence of the implementation of new technologies show controversial results on strain and stress reactions. Some studies report an increase, others a decrease and several no differences in stress-related reactions linked to the introduction of new technologies.
The implementation strategy is an important variable that has to be considered. Often the introduction follows a strategy of simply implementing the latest or the “best” hardware and software system and “muddling through” the resulting problems of organizational change (von Benda, 1990; Greif, 1991b). Here it is not predictable whether the technological changes result in repetitious tasks and a loss of skill or in job enrichment. By following this strategy it remains uncertain how the job demands and decision latitude will be affected. A typical unsolved problem linked to the preference for “muddling through” seems to be the lack of advanced knowledge about what computers can or cannot do and how the future working situation may be affected by the new technology (Briner & Hockey, 1988). As a consequence the fear of unemployment or negative job changes such as a lower position, social isolation, role change and increasing supervision may arise (von Benda, 1990; Frese, 1991a; Frese & Brodbeck, 1989).
An increase in strain due to VDU usage, especially in connection with high workload and repetitious tasks, is reported in a comprehensive study by Stellman et al. (1987). The authors investigated more than 1000 female office clerical workers. They studied the differences between five groups: part-day typists, all-day typists, clerical workers, part-day VDU operators and full-day VDU operators (as they call them). All-day terminal users reported higher levels of job and physical environment stressors than part-day VDU users, typists and other clerical workers. Musculoskeletal strain, symptoms such as eye strain, and dissatisfaction were also highest among all-day users of terminals. However, no consistent or significant differences in the levels of psychological symptoms (depression, anxiety, hopelessness, irritation) were observed between the groups. Typists and clerical workers who also held supervisory positions reported fewer stressors and greater job satisfaction than workers with no supervisory tasks. However, there were no such differences between supervisors and non-supervisors engaged in all-day VDU work with terminals.
Concerning full term VDU operators, the authors suppose that the potential advantage of increased supervisory responsibilities may be annihilated when a worker is involved in highly demanding, repetitivework. This group reported the highest mean levels ofworkload demands and repetitiousness, and also the lowest mean levels of decision-making latitude, ability to learn new things on the job, understanding of the overall work process, and the meaning of the material dealt with. Furthermore, full term users of terminals reported the highest level of ergonomic sources of stress, although they had a greater ability to adjust the height and back of their chairs in comparison to other groups.
The results of Stellmann et al. (1987) correspond with a study on insurance staff by Johansson & Aronsson (1984), which showed that the highest level of stress was found among those doing repetitive tasks and constant work with VDUs. Differing from these results, Agervold (1987) found, in an investigation of a representative sample of 907 white-collar workers, that the incidence of mental fatigue, stress and psychosomatic complaints was the same in the sub-groups working with or without new technology. A close comparison was carried out. The results indicated there was no correlation between the impairment of the psychological work enviroment and new technologies. On the contrary, there seemed to be some improvement in the quality of work, although this was combined with an increase in workload (pressure of work and mental strain). Concerning psychological strain, the only effect of working with new technology was a tendency of higher levels of mental fatigue. Stress and psychosomatic reactions, however, seemed to remain unaffected.
Only where technological changes resulted in a deterioration of working conditions (e.g. less personal influence, fewer cognitive demands, greater isolation, more pressure at work, higher mental and physical workload) did the study of Agervold (1987) show a decrease in the quality of work related to the level of stress. Agervold concludes that “new technology seems only to have negative consequences in terms of stress if it is combined with changes in important psychological aspects of work” (Agervold, 1987, p. 149). The results of this study indicate that the consequences of the introduction of new technology concerning job content, quality of work, influence, satisfaction and stress are determined by the kind of job and the degree of change in the organization of work, combined with changes in work pressure.
In a study investigating the impact of computer technology onwork content, feedback, job control and mental strain in text preparation in printing shops, Kalimo & Lepp¨anen (1985) found that subjects working with the most advanced technology assessed their mental activity and self-determination at work more positively than subjects whose tasks involved less advanced technology. The former subjects were more satisfied with their work than the others. Their tasks demanded more decision making and were more complex. Even for subjects with minimal initial task variety and challenge, however, Kalimo & Lepp¨anen found that computer technology and the application of VDUs may increase performance feedback and quality control. Therefore, this kind of work was associated with a positive impact on the whole work setting and positive changes in the daily workload as well. The results of this investigation show that new technologies may diminish stress even when combined with simple tasks.
The summarized studies demonstrate that the impact of new technologies on stress and its consequences can be extremely varied. We should be careful not to attribute observed increases of stress reactions to new technologies per se. The results of studies taking into account the design of task demands, performance feedback and other mediating organizational factors show that these are more critical factors, which have to be designed carefully in the implementation process (Briner & Hockey, 1988; Frese & Brodbeck, 1989; Greif, 1986). Moreover, new technology provides the “option” for an improvement in working conditions (Ulich et al., 1989). Stress is expected to decline as a consequence.
Cross-sectional field research has often been criticized in that it can be misleading when it comes to causal inferences. Correlations between hypothetical stressors and stress reactions may result from hidden factors or even from a reversed causal relation (for example, a stress reaction may increase time pressure). The same is true for observed mean differences between groups. It is impossible to control all relevant conditions and factors which may influence the results of field studies. Therefore, controversial results should not be surprising. Researchers prefer longitudinal studies because here, at least, the chronological order of hypothetical causes and effects can be partly controlled. Although longitudinal studies demand higher and longer research investments, several studies of this kind have been conducted especially in the past decade. The following section summarizes the results of several longitudinal studies on the consequences of implementing new technologies.
Longitudinal Studies
Frese & Zapf (1987) investigated the impact of new technologies on qualification demands, decision latitude and stress within a longitudinal design. The study used two measurement phases, before and after the technological changes. In 1979, before the changes, 218 bluecollar workers in the German car and steel industry participated in a comprehensive survey. At that time hardly any computer-aided machines or robots were used in the workplaces. Six years later, in 1985, 166 subjects of the first sample were studied again. In the study the changes between the following five groups of technological demands were compared:
1. Traditional jobs which had not experienced technological changes.
2. Computer-supported work without programming tasks.
3. Computer-supported work with some influence on the programming (but not practised by the employees themselves).
4. Computer-suported work with programming tasks.
5. Operators of industrial robots.
In their statistical analysis of standard scales of job demands, stressors and hypothetical resources mediating possible stress reactions, the authors found only some minor changes and differences. One result was a low but significant increase of job decision latitude for all groups, from 1979 to 1985, and also a minor but significant general decrease in time pressure and concentration demands, with the exception of group 4. Job satisfaction increased significantly for groups 3 and 4 after the changes.
The results of the study show that job demands, stressors and their relations to resources have remained stable over time for the different groups. The authors conclude that new technology has only a very low—if any—impact on changes of stress and resources. Workers who had a high level of time pressure or job discretion and complexity before the implementation of the new technology were selected for jobs with similar attributes after the change.
The study of Frese & Zapf (1987) raises the problem of interactions between the implementation of newtechnologies and personnel selection by the organization or by the process of self-selection. When introducing new and expensive computer-based technologies, the organization typically selects experts or volunteers with high ability. Therefore also in longitudinal studies a comparison of mean values of selected sample groups may be completely misleading. Comparing individual changes in their longitudinal study, Majchrzak & Cotton (1988) have shown that stress reactions can be found only for subjects with unfavourable starting conditions who face strong technological changes. K¨uhlmann (1988), in his longitudinal study, assessed individual attitudes of employees facing the implementation of new technologies and their expectations of negative changes. Immediately before the change most employees developed an optimistic attitude and seemed to underestimate the possible negative job changes in comparison to their later observations. This tendency to simplify and belittle unpredictable future difficulties could be interpreted as an important type of cognitive coping with future uncertainty. Such a way of coping seems to be an important ability facing new and unpredictable situations. But as K¨uhlmann (1988) has found, a critical group of employees who, before the changes, are very concerned and worried about whether they will be able to cope with the changes should get special psychological support. While Frese & Zapf (1987), Majchrzak & Cotton (1988) and K¨uhlmann (1988) investigated the impact of the change before and after implementation of new technologies, Korunka et al. (1993) also included a comparison of different types of implementation processes in their longitudinal study. In addition to the possible stress reactions related to new technologies, the authors suppose that the process of implementation can be an important source of strain. The authors applied a longitudinal design with three measurement phases: prior to implementation, during implementation, and one year after implementation. Strain reactions and satisfaction were assessed for all three times. In particular, stress induced by job content, organizational aspects and physical conditions of the environment were considered. The sample consisted of 279 employees either using computer-aided design (CAD) software, doing clerical work or carrying out telephone information tasks. To assess the style of the implementation process, the authors distinguished three property classes: organization of the project, (ii) participation of employees, and (iii) training and supervision.
The results of the study by Korunka et al. (1993) show a significant increase of subjectively experienced stress over the three phases. Paired comparisons of the phases showed a significant increase of subjectively experienced stress, as well as a significant increase of dissatisfaction in the interval between phase (“prior to implementation”) and phase (“during the implementation”). Furthermore, with the exception of eye problems, a significant increase in physical complaints (neck pains, shoulder pains, back pains, pains in the arms and arrhythmic heart rate) was registered from phase (“prior to the implementation”) to phase (ii) (“during the implementation”) and from phase to phase (“one year after the implementation”). The authors suppose that the major cause for this increase is likely to be found in an insufficient provision of ergonomic furnishings. To analyse the effect of project organization and inclusion—participation of employees— Korunka et al. (1993) classified the sample into different dichotomized subgroups. The results showed an interaction between stress and participation. Stress following implementation decreased significantly in those companies that practised more active employee participation in the change process. On the other hand, no differences over the same period were observed in companies with lower participation. Moreover, inclusion generally resulted in higher ratings of satisfaction with the technological changes. Across the three measurement phases, satisfaction showed an interaction effect too. High participation led to a decrease while low inclusion led to an increase in dissatisfaction. Furthermore, the overall extent of physical complaints is attenuated by participation. The authors conclude: “It seems to be that a participatory managerial style may counteract any negative effects of the new technologies” (Korunka et al., 1995, p. 138).
Furthermore four job clusters linked to qualification demands were distinguished and analysed with regard to the degree of participation and the impact of the introduction of the new technologies. Employees in the more highly ranked job clusters (using “computer aided drawing” as they call it, and also “Clerical work”) had more opportunities to participate in the implementation process than personnel of the less qualified clusters (Korunka et al., 1993, 1995). In the cluster with the lowest qualification (“Extremely monotonous work”) most psychosomatic complaints and decreased job satisfaction were observed. This result corresponds to a similar outcome of the study of Agervold (1987), cited above. On the other hand the employees in the highly technological qualified cluster (“Computer aided drawing”) showed an increase in job satisfaction and only slightly increased psychosomatic complaints after the introduction of new technologies (Korunka et al., 1995).
Korunka et al. underline that their results fit the model of Karasek (Karasek & Theorell, 1990), who postulate that the most strain is associated with high demands and low decision latitude. High employee participation in the implementation process appears to generate a higher level of acceptance and leads to a high degree of attenuation of subjectively experienced stress and dissatisfaction. On the other hand, low participation results in an increase of these variables.
Furthermore, Korunka et al. (1997) report preliminary results of a follow-up study about the effects of continuous implementation (implementation in workplaces already equipped with computers) of newtechnologies on employees. In a longitudinal research design, strain and dissatisfaction of employees were investigated in 5 of 10 different companies (only in 5 organizations had new information technologies already been installed). Compared to a control sample, employees in the implementation sample showed (two weeks after the implementation) an increase in subjectively experienced stress but not in dissatisfaction. Regarding the implementation subsamples at the individual level of different organizations, all samples except one followed the general pattern of an increase in subjectively experienced stress. In only one organization, however, was this difference statistically significant.
Dissatisfaction increased in two organizations significantly. The authors emphasize that the reported effects of implementation are preliminary as no data on organizational contextual factors (e.g. implementation style, participation, type of training) were available at this stage of the study. To analyse the effects of job profiles and external workload, the employees were grouped into persons performing jobs characterized by low versus high decision latitude and with regard to high versus low external workload. Subjectively experienced stress caused by implementation only seemed to increase in workplaces with low decision latitude and high external workload. The marked statistical trend corresponds with the results of the first longitudinal study (Korunka et al., 1993, 1995). It seems worth mentioning that the subjectively experienced stress in general was higher for participants with high decision latitude in contrast to low decision latitude in the pre-implementation period whereas dissatisfaction was lower in this group. The preliminary results of the study of Korunka et al. (1997) suggest that not only the initial introduction of new technologies but also the continuous changes of information technologies affect employees’ strain and satisfaction.
In an advanced analysis of the above mentioned follow-up study, Korunka & Vitouch (1999) investigated the impact of personal factors (individual differences, external workload), situational factors like job design (e.g. job complexity, decision latitude), implementation content factors on strain (psychosomatic complaints, subjectively experienced stress) and satisfaction using structural equation modelling. Implementation content factors were (i) adaptational demands of the employees due to the newtechnology (qualification, duration of training and changes in program functionality), (ii) software-ergonomic changes and (iii) participation. Because in some of the 10 organizations implementation was delayed or was not yet in operation at the second measurement point, a “control sample”was available. Data were collected during a period of one and a half to five and a half months before implementation and three to six months after implementation of information technology. It was assumed that personal and situational factors at the job design level have general effects on users’ strain and satisfaction. Implementation content factors were supposed to have causal effects in users’ strain and satisfaction. In a two-step model testing procedure with a base model, (i) the factorial structure of the dimensions of strain and satisfaction and the measurements before and after the implementation were tested. Following the authors, the base model confirms the scales used to measure stress reactions and satisfaction to be solid markers for their latent factor and proves their factorial structure as unchanged over time. A general implementation factor showed no significant effect. However, as expected, a significant influence of personal dimensions on strain and satisfaction was found: internal locus of control, higher self-esteem and positive attitudes towards information technologies were associated with job satisfaction and less strain. In a second model the implementation content factors were included and revealed significant effects on the changes in users’ strain and satisfaction at the second measurement point. The second model shows that participation, adaptational demands and ergonomic software design are relevant factors explaining employees’ reactions to technological change. Ergonomic software design is related to the change of the user interface: software implementations retaining a character-based interface showed strong negative effects on strain and satisfaction compared to implementations changing from character-based to graphical user interfaces.
Adaptational demands like changes in the users’ qualification, duration of training, and changes in the softwares’ functionality are inversely related to strain measures. Concerning employee participation, the authors state that the study provides additional support for the positive effects of user participation in the case of the continous implementation of new technologies. They conclude that in the case of IT implementations, increased attention should be paid to current developments in user interface design and to active participation of employees. Furthermore, adaptational demands should accompany the enhancement of employee qualifications.
Slightly positive effects of the implementation of new technologies are reported in some longitudinal Scandinavian studies investigating the implementation process of new technologies. In his review of several longitudinal studies carried out in the field of public service (state institutions), a bank, a library and an insurance company, Huuhtanen (1997) summarizes a positive overall impact of new technology on work content experienced by employees in all occupation and age groups. He states that new technologies seem to have more often increased than decreased those characteristics that are important for mental well-being at work: “Compared with the expectations before the change, the work has become more interesting and the employees felt that they could use their abilities better” although office tasks have become somewhat more difficult, and the work pace has increased.
Mixed results are presented by J¨arvenp¨a¨a (1997), who conducted a longitudinal study over a 4-year period on the implementation of office automation at a district court. After the implementation, office workers perceived their jobs as slightly more interesting than before the implementation. Short-term mental strain was also slightly lower after the implementation, but this positive effect seemed to decrease over time. On the other hand, a slight increase in office workers’ long-term strain (e.g. stomach symptoms, chest pain, restlessness, fatigue and eye symptoms) was observed but no effect relating to job satisfaction was found.
An important aspect regarding results of longitudinal studies on job design and stress was given by Carayon et al. (1995). The results of their 3-year longitudinal study investigating the relationship between job design variables and strain of office workers from a public service organization indicate that the relationship between job design and strain changes over time. At the first point of measurement, quantitative workload, work pressure and supervisor support were related to most measures of worker strain. At the second point supervisor support was related to all of the strain measures as well as task clarity except for one. Job future ambiguity, quantitative workload and job control were respectively related to six and four of the eight measures of worker strain. Eventually, at the last point of measurement, task clarity was related to seven, attention and job future ambiguity to five and job control to four measures of strain. The authors suggest that the lack of stability of the correlations between work stressors and worker strain may be due to changes in management or, more generally, that work environments and people may change over time. Due to this assumption they conclude that in theories of job design and worker strain we have to take into account the flexibility of working environments.
Summarizing the results of the cross-sectional and longitudinal field studies on the implementation of new technologies, there is, as in the results of the reported cross-sectional studies, no definite support for the simple assumption that the computer technologies per se might excert stress. Malfunctions of systems that have not been tested sufficiently by the manufacturer, and poor ergonomic design of hardware, software and ergonomic furnishings are problems, which in most cases can be reduced to a tolerable level by professional experts by means of usability engineering. Where stress reactions are found, they seem to result from negative interactions between technology and job demands (especially monotony and time pressure; Carayon, 1997, p. 325) or sources of stress resulting from demands parallel to the technological changes or insufficient training.
However, the most difficult practical problem is an increase in stress that can be observed as a consequence of an indaquate implementation process of newtechnologies. Most experts recommend as remedies user participation during the implementation process (Briner & Hockey, 1988; Huuhtanen, 1997; Karasek & Theorell, 1990; Korunka & Vitouch, 1999), concrete information in advance, redesign ofwork, and training of the employees to manage the complex and cyclic technological changes (Korunka&Vitouch, 1999).