Sociotechnical systems theory represents a very different paradigm compared with the technology-centric safety and risk management approach currently applied in rail level crossing design (see Chapter 6). Two fundamental issues that limit the application of sociotechnical systems theory are the extensive resources required such as time, cost, expertise (Clegg 2000) and stakeholder acceptance. Within the current project, we introduced sociotechnical systems theory to stakeholders through 3 days of workshops, which was potentially insufficient time given the inherent conflicts between sociotechnical systems theory and existing design philosophies. Our participatory approach involved efforts by system stakeholders to reconcile the two approaches, by combining the protective aspects of the safety management approach while incorporating some of the sociotechnical systems theory values and principles. Given the sensitivity of the stakeholders to the political realities of transportation system design, this appears to be a practical and reasonable approach. However, it limits the promised benefits of sociotechnical systems theory in promoting flexible and adaptable systems.
It is reasonable that genuine paradigm shifts would need to occur over a much longer period of time. The use of the CWA-DT in this context can be viewed as the beginning of an ongoing process to introduce the sociotechnical systems theory approach and integrate it into rail level crossing design and evaluation processes. The workshops provided a means to initiate conversations about systems thinking and sociotechnical systems theory among a diverse group of stakeholders from road and rail domains, including engineers, policy officers, human factors researchers, consultants and safety executives. Potentially, this experience might have positive effects on how the participants approach safety issues in their future work, while also strengthening professional networks across the various stakeholder groups.
The two expert-generated designs - the Community Courtyard and Ecological Interface Design crossings - offered more radical change in the functioning of rail level crossing systems. Both designs involved slowing the train through the crossing and did not feature any standard regulatory signage and warnings such as flashing lights. The Ecological Interface Design crossing concept demonstrated positive features in the evaluation process; however, further research is needed to determine whether all aspects of the design (such as mirrors) would be effective in the real world. The Community Courtyard crossing, however, was received quite poorly in user evaluations, with all road users considering it to be unsafe and motorists ranking it as least efficient. Although non-motorised road users rated it higher for efficiency, its overall acceptance by users remains questionable.
The poor response to the Community Courtyard crossing is again indicative of key tensions between sociotechnical systems theory and the traditional safety management approach usually applied in road design. For example, the sociotechnical systems theory value of humans as assets suggests that humans in the system should be given control over their decisions, while the principle of flexible specification (i.e. that design should only specify that which is necessary) advocates that humans should be supported to exhibit flexibility and adaptability in their behaviour. In contrast, the safety management approach encompasses concepts such as the hierarchy of control, which are focussed on separating humans from hazards. When applied within transport systems, these concepts tend to focus on limiting performance variability rather than supporting adaptive variability. Therefore, designs based on sociotechnical systems theory are likely to violate road users’ expectations about what is acceptable in road design, due to their deviation from established conventions.
Further obstacles to embracing sociotechnical systems theory-based designs include law reform and public opinion. Although many international jurisdictions have adopted a ‘safe system’ approach to road safety (Salmon et al. 2013b), none have yet attempted to deal with the law reform issues that would be required to facilitate implementation of a strategy that genuinely embraces shared responsibility. To truly support sociotechnical systems theory approaches to design in public safety domains, many changes to the ‘protective structure’ (Dekker 2011) would be required. For example, the operation of the legal system and the interventions of regulators need to be considered in relation to what organisational behaviour is rewarded and punished.
Moreover, a general trend in public opinion away from individual responsibility for safety towards government responsibility has been noted (Leveson 2004). This shift is evident in cases of civil proceedings against organisations and governments, where it is contended that governments hold a duty of care towards the public. In the rail level crossing context, it appears anecdotally that if active warnings are in place where a collision occurs and the equipment operated as designed (i.e. provided a warning), the public will attribute the incident to the road user as they have breached the road rules. However, if the equipment failed to warn of a train, the railway company and/ or government are deemed responsible. This situation has stifled the implementation of low-cost innovations at rail level crossings (Road Safety Committee 2008).
Focussing on culpability in accident investigations also limits the data collected and, consequently, our ability to learn from past mistakes.
Although have advocated for the sociotechnical systems theory approach in this research programme, the question remains: can we be assured that designs based on this approach would be more successful in preventing accidents than traditional approaches? To answer this important question further research is required to compare the traditional safety management approach with sociotechnical systems theory, to determine which is most successful in preventing accidents. Potentially, modelling approaches such as agent-based modelling and systems dynamics (see Section 11.5.2) could provide answers.