My goal in pursuing research in Industrial/Organisational psychology is to generate knowledge that is recognised for its contribution to theory and application internationally and (currently) within Australia. I seek to fulfil this goal through the publication of research findings in refereed research journals, through the dissemination of knowledge at international conferences, through the supervision of research students and by teaching courses that reflect current developments in my field. The best way to illustrate my approach to applied research is to provide brief summaries of some of my currently funded research projects.

A major research theme that I am currently pursuing is how technology is introduced to work environments and how different forms of work design might be used to encourage knowledge and skill development. The industries in which we are working currently are that of food processing and water effluent treatment. Both are industries that have undergone dramatic change in recent years due to the introduction of new technology in the case of the latter, and a drastic increase in global competition in the former necessitating a large reduction in the workforce.

The impacts of skill development on employee well-being, satisfaction and general productivity are the broad outcome variables of interest both for the companies involved and us as researchers. At one level our research projects are concerned with bringing change to the industries in which we work, at another we seek to test and develop theories of job design in the context of modern manufacturing and management philosophies.

At a general theoretical level we have been interested in how different forms of job design can foster both enhanced feelings of well-being and productivity especially in organisations that are now placing an emphasis on ‘smart work’ and multi-skilling. Such organisations are of interest from a research point of view as it seems likely that the traditional theories of job design that have sought to energise and motivate employees to work harder may not apply when the task demands require mental agility rather than physical dexterity. More specifically we are interested in the psychological processes that mediate between job characteristics and work outcomes. Building on the work of Toby Wall and his colleagues in Sheffield, we have hypothesised that one of the mediating mechanisms that facilitates learning and a sense of mastery is related to, or is reflected in, opportunities for skill use and skill development. In hypothesising a mediating role for such a variable, which we refer to as skill utilisation, it was important to establish first that it might exist as a separate psychological construct. Separate, that is, from other familiar job design constructs such as autonomy and skill variety, which we believe have been confounded with skill utilisation. Having defined skill utilisation as an independent construct we are currently in the process of attempting to show how work practices, which give employees more control over their work methods and pacing, encourages skill development as well as fostering higher levels of employee well-being. Thus, skill utilisation and development is a consequence of job design not simply another exogenous job design variable.

The research program described above has been both time consuming and painstaking as in order to get good longitudinal data, which is necessary to make a strong case for inferring causal relations between various constructs from correlational data, it has been important to establish long term relationships with various companies. As a consequence we have tried to be economical with our efforts and to address several other issues of theoretical interest to us. One of these is the nature of the relationship between different job characteristics and work outcomes. Foe example, which job characteristics combine, how do they combine and when do they combine to achieve higher levels of productivity and employee well-being? Here we draw together research from both the job stress (eg., Karasek) and job design (eg., Hackman and Oldham) literatures. Among other things, we seek to understand why it is that Karasek’s much hypothesised demand-control interaction is reported only intermittently in the literature. In the context of our research program, we have the opportunity to address this topic as we have not only access to companies in ways that allow us to collect panel data but, in addition, we have been able to use the same measures across companies and industries over time. There are few of research programs that we know of that have achieved this. In terms of outcomes, its seems clear that the nature of the relationship between broad categories of job characteristics such as demand and control varies across the range of the job demand continuum. Furthermore, job control, far from always being a help in managing a difficult job, can also contribute to the stress experienced on the job.

Finally, at another level in our research on job design, we have also become interested with the organisational processes and management practices that facilitate or hinder the process of change. Again, using our privileged access to the companies with which we have been involved in trying to facilitate change, he nature of the relationship between supervisors and supervisees has been of some interest. Especially so in the context of the development of trust and the psychological climate that fosters innovation and continuous learning. This is an important topic since the role of management in supporting change projects has largely been ignored in the literature and yet it is estimated that up to 80% of job redesign projects fail because of lack of support from management. Unlike much previous and contemporary research, our approach has a quantitative bias and makes use of structural equation modelling and multi-level analytical techniques. Such techniques allow us to test statistical models of the process of change which allows the multivariate nature of the change process to be studied in ways hitherto precluded from more straightforward univariate statistical techniques and qualitative approaches to research.

A second major area in which I have been working in recent years is that of occupational health and safety, rehabilitation and compensation. The major sponsor of this work has been a Government Commission (WorkCover). Initially this project began with a simple actuarial analysis of the workers compensation system using data supplied by insurance companies regarding the nature of the injuries sustained, their severity and other demographic details regarding the employee and employer. To date we have analysed over 100,000 workers compensation claims and generated statistical models capable of predicting claims duration and costs at levels superior to professional insurance claims assessors. The practical value of this is that potentially protracted claims can be identified early and appropriate rehabilitation commenced. In recent years we have written extensively with regard to how claims are settled and what factors influenced return to work rates. We have also conducted a cost benefit analysis of rehabilitation programs and recently researched the organisational characteristics that facilitate rehabilitation.

At first blush, it might appear that this research area is simply an actuarial analysis of existing data, the results of which may have a large practical consequence but in and of itself is not theoretically interesting. However, as with the research in the meat processing industry, our approach has been to develop a working relationship with those in industry in the hope that some more theoretically oriented research can be undertaken. Indeed, in one funded project we have been able to develop a program of research that is aimed at testing and extending a model of decision making (Ajzen and Fishbein’s theory of planned behaviour) with respect to unsafe behaviours. Initially we hypothesised that contextual variables, largely ignored by the model, such as the design of the job would account for some additional unique variance in safety violations. The results did not support this and perceived job demands and management behaviour had no direct influence on propensity to commit safety violations. Instead we have found that the organisational climate, transmitted through different management styles, had a strong indirect effect on safety behaviour. The practical utility of this finding is that management has a large part to play in the safety records of their employees.

In another project we have linked up several different sources of data with respect to work injuries. In addition to the insurance company data, we have captured information regarding organisational practices, employee perceptions of those practices and the behaviour of rehabilitation support workers (GPs, rehabilitation providers, and human resources among others) for specific injured workers. In particular we have examined how the motivation of injured employees to return to work is influenced by the behaviour of various stakeholders involved the recovery process of injured employees. The practical utility of this project is that we have sought to discover the relative weightings and influence of variables that affect the decision to return to work. In this project, once more, we have attempted to marry theoretical interests of our own with the needs and interests of participant organisations.

Over the past 10 years I have been concerned with how individuals solve complex problems such as the fault diagnosis tasks required of process control operators who work in complex industrial systems. In the early part of my career I was interested in how fault diagnosticians could be enabled to be more effective for novel failures in particular. Problems of this type are interesting since they have not been predicted in advance by system designers and yet being able to deal with such failures is often the justification for retaining human operators in complex systems. It is ironic that humans are largely retained to save novel problems since the very thing that the operators are supposedly able to do is the one thing that they have least practice at and that we understand least about. This is especially so in highly automated complex systems such as industrial process control. Indeed, it has proved extremely difficult to predict who will do well at solving novel failures and, it is a skill that has proved to be remarkably resistant to traditional methods of training. This is perhaps, hardly surprising since, by their very nature, novel problems are difficult to specify in advance and therefore it is difficult to specify a training curriculum.

I have pursued three avenues for improving general fault finding skill: (i) problem presentation and interface design; (ii) training and (iii) selection. In each case I have utilised cognitive theory to make predictions about what type of interface design, training and selection method should improve overall skill for dealing with novel problems. In more recent times I have attempted to understand the cognitive mechanisms that might be responsible for how complex cognitive skills such as fault diagnosis skills might be learned and the consequences of the way in which the skill was developed for the solution of novel failures. With these studies, supported by the Australian Research Council (ARC), the mechanisms with regard to how skills are learned and what changes in mental representation take place as expertise develops has been investigated. In particular I am interested in how different forms of knowledge are developed and what the consequences of the method of learning and type of knowledge are for future task performance. Specifically we have been interested in the distinction between implicit and explicit knowledge and their consequences. Following the work of Broadbent (see for example Hayes and Broadbent, 1988) and Anderson (1993) we have been interested in the notion that different learning mechanisms exist and may be involved in how complex skills are represented in memory. The idea that expert subjects may learn to control a complex dynamic task but not describe how they do it, and vice versa, has implications for training and the on-going maintenance of cognitive skills in low event environments (eg., automated process control).

At present we are conducting research aimed at training subjects to diagnose fairly complex logic network problems. Subjects learn diagnostic skills in one of two ways. Some subjects develop their expertise through an explicit verbal declarative route and continue developing their skill level until the task has been proceduralised (or in the parlance of Shiffrin and Schneider (1977), the skills have become automatic and do not require conscious cognitive control). Others, the implicit group, learn the task in a way that suppresses active rehearsal and articulation of the problem solving processes and other cognitive strategies which may be involved during different stages of practice. The hypothesis is that for diagnosis of novel failures, subjects must be able to uncouple mental routines that have been ‘chunked’ in order that they can be re-combined in new ways to solve the novel problem. The implicit group should not be able to do this since the separate mental routines never consciously existed for them. If we are right, the implications for the training of process control operators will be profound. First the reliance in industry of ‘on-the-job’ training will be questioned since the unstructured development of skill with complex systems will proceed in a manner that may ultimately be counter-productive with respect to the safe and effective operation of such systems. Second, ‘on-the-job’ experiences may have to be carefully managed in a way that ensures that operators who have learned their skills in a way that allows them to ‘uncouple’ mental routines appropriately, continue to be able to do so after long periods of inactivity and/or infrequent experience in dealing with novelty.