Reviewing Causality to Assist in Promoting Crowd Safety
Whilst the study of crowds and mass gatherings continues to develop, there remain gaps in knowledge and application in relation to crowd safety (Cabinet Office 2009). This notion is supported by the fact that disasters continue to occur and sadly people continue to die in crowd related incidents. As Johnes (2004) points out, ‘disasters have powerful emotional, psychological and social impacts. They bring home the realities of risk in a way that abstract possibilities cannot.’ (p. 147) and it is therefore imperative appropriate research is carried out to reduce their likelihood. It is widely accepted that the majority of people who die, do so through asphyxiation caused by excessive density. However, the reasons for a crowd reaching critical densities would benefit from further research to sufficiently answer the key questions posed by Helbing and Mukerji
How and why do such disasters happen? Are the fatalities caused by relentless behavior of people or a psychological state of panic that makes the crowd ‘go mad’? Or are they a tragic consequence of a breakdown of coordination? (2012, p. 1)
Rather than viewing in isolation, there are clear benefits of categorisation, which then allows for a broad and focussed approach to crowd safety. The distinct fingerprint of a disaster, recorded accurately and appropriately, highlights common themes, repeated mistakes, gaps in knowledge or ineffective management. It must also be accepted that some incidents are unavoidable and cannot be prevented, such as the Ukraine air disaster in 2002, and others are down to human error which again are very difficult to prevent such as the Korean train incident in 2003 (Still 2020). However, the vast majority have preventable causes which should be correctly analysed allowing for classification and subsequently, categorisation. The most difficult step is indeed the first one. Identifying the cause of a crowd disaster can be a complicated and subjective process and therefore requires an informed, non-bias approach. Crowded events can be viewed on a cusp catastrophe model and it is suggested that in a significant number of cases the trigger is in fact an avoidable action. Too often causes are attributed to misbehaviour in the crowd. The failure or reluctance to correctly assess and identify the cause, prevents lessons being learnt and crowd management being modified appropriately.,
The DIM-ICE model devised by Professor Keith Still provides a framework for crowd managers and was born from a detailed categorisation of incidents and disasters. This approach is championed in this article and it is suggested a similar concept can be used to classify incidents caused by a preventable human trigger. Rather than contribute to the development of a meta model as with Still’s categorisation, the suggested method can highlight distal causes and contribute towards better understanding and informed training of those charged with managing, or controlling, crowds.
Analysing the history of crowd related incidents should allow for improvements in crowd safety and ultimately a reduction in disasters. However, such incidents are never a result of one single issue but rather a combination and build up of factors which may not be obvious, even to an experienced crowd manager (Turner 1978, Helbing and Mukerji 2012, Murphy 2020). Elements of a disaster can be divided into proximate and distal causes and having an understanding of this categorisation allows for a more appropriate, holistic view. Proximate cause relates to the immediate agent, the obvious, direct reason for the injury or damage such as a stadium collapse or excessive crowd density. Distal causes are the underlying factors such as stadium design, inappropriate management, lack of training or poor communication. (HSE, 2004). For example, the infamous Hillsborough tragedy of 1988 can be divided into the proximate causes such as the opening of a gate, the lack of direction given to supporters on entry, the terrace fencing, and the distal causes such as poor stadium design, inexperienced match commander, stereotyping of football supporters, ignorance of previous issues. All crowd related disasters will result from a number of proximate and distal causes and should be reviewed as such.
Crowd disasters can be divided and categorised in many different ways; numbers of fatalities, location, time, or event type for example. The key is that the applied categorisation assists someway in the development of crowd safety. Still (2004) devised the DIM-ICE meta-model after review of incidents highlighted ‘Design’, ‘Information’ and ‘Management’ as three key causation categories. The model is now used by crowd managers to focus their attention during the planning stage and ensure each heading is considered across all stages of an event (Breen 2017). It is suggested a further categorisation which could help improve understanding and go someway to address distal, root causes, would be to look at incidents whereby a human trigger from those with some authority or responsibility for the crowd, is a significant factor in the disaster. Whilst such incidents would likely be covered in Still’s model under the ‘Management’ heading, the suggested categorisation sharpens the view on human action, allowing for key links to be made with the subsequent crowd reactions and therefore knowledge, understanding and training to be developed. Figure 2 below details a selection of crowd disasters whereby the author has identified a human trigger as a key factor
The human action detailed in the final column of Figure 2 can be considered a key factor when applying the Cusp Catastrophe Theory to these incidents (see Cusp Catastrophe article on this website). It is therefore argued that improved awareness and understanding could significantly reduce similar disasters in the future and prevent avoidable ‘stampedes’ and ‘mass-panic’. Rather than focus on the subsequent crowd behaviour it is remiss not to consider the steps leading to this stage – what are the true proximate and distal causes immediately before the crowd began to ‘panic because they are dying” (Still 2004). Benedictus’ (2015) damning appraisal of the situation supports the author’s suggestion, “Why do crowd crushes happen? Why do they keep happening? That is easy. These are not natural disasters. They happen because no one stops them”. On reviewing the above categorisation of incidents there are glaring similarities in a number of disasters.
Firstly, by providing a significant reward for being at the front of a crowd, organisers are encouraging the crowd to move toward the cusp of catastrophe. The sad irony of mass fatalities during charitable handouts of clothes, food or other essential supplies must be viewed in this light rather than simply blaming an out of control mob. Such an event requires careful consideration and planning to prevent crowd density and movement combining with disastrous consequences. In Morocco in 2017 a larger than expected crowd gathered before one such event, at this point, with high density and low movement (BBC 2017). With appropriate understanding it would have been clear that beginning to hand out food could trigger movement in the dense gathering and therefore cause catastrophe. Other events with perceived significant reward, often referred to as ‘craze’ events (Fruin 1981, Ranieri 2005, Seabrook 2011), such as store openings (Ikea Saudi Arabia 2004), significant discounts (Wal Mart 2008, Cape Town 2010) or free gifts (Aracaju 2001, Manilla 2006) must also be reviewed without simply blaming the unruly mob when their actions are evidently foreseeable..
Secondly, the categorisation in also highlights how those with authority over the crowd must understand crowd behaviour. Inappropriate actions can clearly have catastrophic effects costing many lives and it is suggested these are borne out of lack of knowledge as well as confirmation bias. Whether it is the direct action of firing tear gas into a crowd (e.g. Lima 1964, Harare 2000, Oromo 2016) inappropriately influencing crowd flow (e.g. Moscow 1982, Hillsborough 1989, Duisburg 2010) or locking / failing to open gates (e.g. Buenos Aires 1968, Kathmandu 1988, Buenos Aires 2004, Luanda 2018) actions by police, security or crowd managers must be reviewed appropriately to prevent repeated failings. The apparent cycle of (a) expecting a crowd to behave as an irrational mob, (b) inappropriate crowd management (potentially moving the crowd towards catastrophe), (c) the crowd reacting (potentially becoming disordered) and (d) the incident then being viewed as a crowd panic or stampede must be broken to prevent the continual strengthening of the conditional bias. Le Bon’s theory can become a self-fulfilling prophecy if the crowd themselves continue to be seen as the reason for disaster. Those with authority and responsibility in crowded environments must have the knowledge and awareness to provide the highest standard of safety possible, let alone cause significant disasters by inappropriate, avoidable actions. As Berlonghi 1995 points out ‘every profession from dentists to architects, has a world of distinctions that they must learn and then apply and constantly practice’ (p. 240). Indeed, incompetence should not be masked by outdated theory (such as Le Bon and McDougall), a biased view of crowds or sensational headlines (such as The Sun 1989 cited BBC 2016). Effective, regular review and subsequent categorisation will assist in improving crowd safety
Contribution to Crowd Safety
Despite the reoccurrence of disasters throughout the past century, event management and crowd science are relatively new academic disciplines (Murphy 2020). Efforts must continue to be made to develop awareness in this field and the author’s categorisation could influence training to allow lessons to be learnt and applied. Still’s DIM-ICE meta model (2004) is a welcome addition to crowd safety planning and is an example of how appropriate categorisation can highlight consistent themes. This model is used by event professionals during the planning phase and it is suggested the further categorisation discussed should be used to focus attention on the direct impact of inappropriate, incompetent crowd management. Berlonghi (1995) argued some time ago ‘those involved in crowd management…cannot be excused from the significant responsibility of providing the public with the highest standard of safety and security that is both possible and feasible’ (p. 239). Yet figure 2 provides a powerful illustration of how, too often, this standard is not reached. An accurate and appropriate categorisation, avoiding the all too common confirmation bias, should be encouraged to allow any crowd related professionals to access and review such information. Being presented with a comprehensive group of major disasters, largely attributable to incompetence, should be an impactive training tool for crowd managers, police officers and security professionals. A concise reference tool, such as shown in figure 2, should feature as part of a crowd professional’s knowledge base and the fundamental factors likely to push a crowd towards the cusp of catastrophe should become widely known and accepted.
Conclusion
Interestingly, the Latin word ‘turba’ can be translated into English as ‘uproar’, ‘disturbance’ or ‘crowd’ and classical crowd theory follows this ancient line of thought that the three words are interchangeable. However, whilst there is indeed a ‘systemic instability’ (Helbing and Mukerji 2012, p 19) in crowded events, many occur frequently around the world without issue (Fruin 1993). It is agreed widely by academics that the repeated reporting of disasters as the fault of the Latin-esque , classical theory, crowd, is both unhelpful and more importantly most often inaccurate (Sime 1995, Cocking et al 2009, Helbing and Mukerji 2012, Seabrook 2011, Cabinet Office 2009). Instead crowd disasters should be accurately and appropriately reviewed and categorised to develop understanding and highlight reoccurring themes. The potentially catastrophic relationship between crowd density and movement should then become clearer as well as the significant impact crowd professionals can have through inappropriate behaviour. Supplementary to Still’s (2004) DIM-ICE model, it is suggested the deeper categorisation focussing on the avoidable human trigger disasters should be undertaken and developed in the training of crowd safety professional. The Cabinet Office (2007) claim crowd management is ‘a safety management challenge that must be addressed holistically, systematically and continuously.’ (p. 14). Learning from past disasters must play a key role in this challenge and misconceptions about them must not continue to hamstring improvement. (Helbing and Mukerji, 2012).
There are occasions whereby disasters cannot be avoided by the professionals charged with managing them, however those which are avoidable, must indeed be avoided.
References
BBC (2016). Sun’s Hillsborough front page ‘symbol of lies and cover-up’. [online]
[accessed 8th December 2020] ww.bbc.co.uk/news/uk-36149489
BBC (2017). Morocco food stampede kills 15. [online] [accessed 28th November 2020]
www.bbc.co.uk/news/world-africa-42044887
Benedictus, L. (2015). Hajj Crush. How crowd disasters happen and how they can be
avoided. The Guardian. Sat 3 Oct 2015. [Online] [Accessed on 20 November 2020].
www.theguardian.com/world/2015/oct/03/hajj-crush-how-crowd-disasters-happen-and-howthey-can-be-avoided
Berlonghi, A. E. (1995). Understanding and planning for different spectator crowds. Safety
Science, 18, pp. 239-247.
Cabinet Office (2007) Public Safety in Complex and Built Environments
Cabinet Office (2009) Understanding Crowd Behaviours: Supporting evidence
Cocking, C., Drury, J. and Reicher, S. (2009). The psychology of crowd behaviour in
emergency evacuations: Results from two interview studies and implications for the Fire &
Rescue Services. Irish Journal of Psychology.
Elliott, D., Smith, D. (2006). Football stadia disasters in the United Kingdom: Learning from
tragedy? In D. Smith and D. Elliott (Eds.), Key Readings in Crisis Management: Systems and
Structures
Fruin, J. J. (1993). The causes and prevention of crowd disaster. In R. A. Smith and J. F.
Dickie (Eds.), Engineering for Crowd Safety (pp.99-108). Amsterdam: Elsevier.
Helbing, D., Mukerji, P. (2012). Crowd Disasters as systematic failures EPJ Data Science
2012, 1:7 http://www.epjdatascience.com/content/1/1/7
Johnes, M. (2004). ‘Heads in the Sand’: Football, Politics and Crowd Disasters in Twentieth‐
Century Britain, Soccer & Society, 5:2, 134-151,
Murphy, I. (2020). To prevent disaster, now is the time to study crowd safety. [online]
[accessed on 2nd December 2020]. http://isemurphy.com/to-prevent-disaster-now-is-the-timeto-study-crowd-safety/
Pearl, T. H. (2015). Crowd Crush: How the Law Leaves American Crowds Unprotected
(February 23, 2015). 104 Kentucky L.J. 1 (2016), [online] [Accessed on 2nd December 2020]
https://ssrn.com/abstract=2569373
Ranieri, A. (2005) The causes and prevention of serious crowd injuries and fatalities at
outdoor music festivals
Santayana, G. (1905) Reason in Common Sense, p. 284, volume 1 of The Life of Reason
Seabrook, J (2011). Crush Point. When large crowds assemble, is there a way to keep them
safe? [online] [accessed on 2nd December 2020]
http://www.newyorker.com/magazine/2011/02/07/crush-point
Still, G. K. (2000). Crowd Dynamics. PhD Thesis, University of Warwick, UK.
Still, G. K. (2014). Introduction to Crowd Science. CRC Press.
Still. G. K. (2020). Crowd Safety and Risk Analysis. [Online] [Accessed on 20 December
2019] http://www.gkstill.com/ ExpertWitness/CrowdDisasters.html
The Guardian (2008). Walmart Worker Killed in Black Friday Stampede. 29th November
2008. [online] [accessed on 28 November 2020]
http://www.theguardian.com/world/2008/nov/29/wal-mart-blackfriday-shoppersstampede
Zheng, X., Sun, J., Zhong, T. (2010). Study on mechanics of crowd jam based on the cusp catastrophe model, Safety Science, Volume 48, Issue 10, Pages 1236-1241,
Zeeman, E. C. (19676). Catastrophe Theory. Scientific American, Vol. 234, No. 4, pp. 65-
83. [online] [accessed on 27th November 2020].