09 March 2009
Brian Toft Tim Marsh

Imagine you're going on holiday with children somewhere hot and exotic, where the kids will spend half their time in the sea. Apart from drowning, would you be most concerned about sharks or coconuts? Unless you've heard this before, you'll probably say the sharks. But in truth, deaths because of falling coconuts are about 10 times as common as deaths from shark attacks - around 200 and 20 fatalities worldwide each year respectively.

This erroneous risk perception is probably driven in part by a primeval fear of being eaten alive by another animal. On the other hand, it's also a good example of the "availability heuristic" (a heuristic is an intuitive judgement). Understanding this, and other common errors in risk perception, can help us understand why we make so many mistakes in safety management decision-making, and how tools as simple as benchmarking can help employers mitigate the effects of flawed reasoning and a lack of objectivity.

Invincibility cloak

The availability heuristic says that we'll give more weight to evidence that is most readily available to us to recall. For example, we'll be very aware of the risk of cancer if a loved one has recently been a victim and it's at the forefront of our minds. Similarly, we'll be very aware of a risk factor that's been plastered all over the media; SARS and bird flu are good examples, as are shark attacks, which will make the headlines when coconuts will not. The experts say that risk perception is somewhere between extremely subjective and entirely subjective.

Another major problem can be what might be termed the "invincibility cloak". Whether it's Freudian denial, optimistic overconfidence or just plain foolishness, some people and organisations seem genuinely to believe that nothing bad will ever happen to them. This is because they have a deep-rooted confidence that "it'll be OK".

Technically this is known as "emotional reasoning"; a belief that what you feel must be true. This is the flip side of psychologists' work persuading people that they can safely travel by plane because they hardly ever crash and it's actually much safer than driving.

Another systemic problem we can face is that we much prefer to learn from the positive than from the negative, even though we all know the truth of the maxim that ends "and the wisest one learns from the mistakes of others".
So we find in practice that, in an attempt to filter out information that doesn't match our pre-existing mindset, we can actually walk past several negative learning opportunities until we find just one positive one that we declare is "most like us". Here this combines with another classic mistake of reasoning - over-generalisation, which is to come to a general conclusion on limited data.

Perhaps the most problematic mindset we see is the one that has us believe it's always the person not following the rules that's causing all the problems, and not perhaps the rules themselves - the "fundamental attribution bias". Studies show that while we all have our own particular biases, we all consistently overestimate the importance of the person and underestimate the importance of the environment when assessing a situation. And the worse the outcome, the greater the bias. So when things go wrong, we put too much blame on the person and can give too little emphasis to the time pressures, or the (lack of) tools and equipment, or the (absence of) training, or the way supervisors routinely praise efficient but slightly risky work. This can turn into a rather long list.

An interesting and important variation on this mindset is hindsight bias. We all know how easy it is to be wise after the event and many people will find themselves insisting they knew things were about to go badly wrong but no one listened to their warnings. Or perhaps they didn't give any warnings because they felt they would just be ignored.

A problem here is that this hindsight bias becomes reality in people's heads and we can find ourselves judging the individuals involved too harshly. Naturally, where we all agree that it was X's fault and that, frankly, we wouldn't have done that in their shoes, it can really hinder learning if the truth is that actually we would have.

"Cognitive dissonance" is basically the formal term for where our stated attitudes and our actual behaviours diverge. The problem is that we can find it is far easier to fudge than to deal with this dissonance objectively.

This can be problematic at work at an individual level where a new worker might say: "This behaviour doesn't seem safe to me - it makes me uncomfortable. I think of my family first and don't put myself at risk." They could then go the long way around every day and ignore the smirks and teasing of their colleagues. But it's much easier for the individual to get back into equilibrium by "deciding" instead that: "This lot have been around since year dot and they seem OK ... So I guess it's not that unsafe after all." Of course, it still is. This can apply at an organisational or industry-wide level too.

Blind agreement

Another phenomenon is "groupthink", where a group of like-minded individuals get together and pretty much blindly agree on some matter, with little regard to alternative views or hard facts. Every reader will have a hatful of examples where this has proved costly in some way. The failed Bay of Pigs invasion of Cuba in 1961 when the CIA persuaded the Kennedy administration to let it send in a doomed cadre of Cuban exiles to try to overthrow Fidel Castro, was the first example officially labelled a groupthink problem.

Studies suggest that the more cohesive the group, the more likely groupthink is and the more pressure any dissenting voice will come under. One of the most influential papers in the world of social sciences was published by Ben Schneider, who argued we need to manage individual diversity actively, as organisations will naturally attract a certain type of person then select those who best match the company's profile. Those who fit in least well are most likely to leave. Schneider called this ASA theory (attraction, selection, attrition), and it automatically makes organisations natural environments for groupthink.

It's been suggested that groupthink played a significant part in the US Challenger space shuttle disaster. Despite the engineers at booster rocket makers Morton Thiokol insisting that it was too cold to launch the shuttle and (initially) having the support of their management team, the decision was changed when NASA leant on them to toe the party line. The key moment was when the most senior person in the debate commented, "I won't overrule the contractors' decision - but frankly I'm appalled at their no-fly position."
After a furious debate, it was agreed that maybe they were being overly cautious after all. Dissenting voices were now ignored as the others closed ranks.

Open your eyes

As well as managing diversity and bringing in fresh eyes, organisations need to get into the habit of benchmarking with similar organisations. That said, much can also be learned from organisations and industries that are not so similar.

Recently I was at an industry association event where a perceived weakness that all delegates agreed on was the difficulty they had learning from each other - but they listened intently to a speaker from a different industry. Perhaps "not invented here" syndrome applies less strongly where it's "not invented anywhere near here".

It's useful to carry out "what if...?" scenarios in training, where delegates deliberately debate the very worst that might happen. Checking assumptions with both in-depth quantitative and qualitative assessments is also essential, so the likes of safety culture benchmarking are key here, as are lead measures and good old-fashioned problem-solving-in-teams training (by which we mean systematic coverage of the above, as well as - or even instead of - "Let's have a great laugh building a raft out of twigs and drink cans then get drunk back at the hotel" training).

A simple brainstorming exercise involves writing down the pros and cons of your point of view and then the pros and cons of that of an opponent. You'll find that the two boxes that fill up quickly are your pros and their cons. Learning to fill all boxes equally is a really useful technique, if only in preparation for a negotiation. Another interesting exercise is to ask one member of a team of five or so to do nothing but strategic overview and monitor evaluation work in a problem-solving exercise. Nine times out of 10 you'll find they outperform a basic team of five.

It is a scientific truth that, outside a lucky fluke, the limits of effective safety management are set by the limits of objectivity - because how else can you most effectively and accurately identify risks and target resources? Maximising objectivity is far harder than might be apparent at first glance. As ever, the way forward is some well targeted training for key individuals, then systematically applying the principles learned.

More information

• ASM Handbook, T Marsh (IIRSM, 2008).
• "The people make the place" in Personnel Psychology, 40, pp437-453, B Schneider (1987).
• Learning from Disaster: A Management Approach, B Toft and S Reynolds (Palgrave, 2005).

Dr Tim Marsh is a consultant and trainer specialising in behavioural safety, www.rydermarsh.co.uk, Brian Toft is professor of patient safety, Department of Health and Life Sciences, Coventry University

Categories:
Risk assessment, Risk assessment, Article