What Do Coin Flips Have to Do With EHS Risks?

Many EHS professionals still struggle to identify and correct risks in the workplace. In fact, as we’ve recently discussed in this blog, both the number and rate of occupational fatalities have been rising in recent years, which points to problems with risk assessment and control.

One reason for this poor performance may be that risks contain a strong element of statistical probability, and we often don’t think about statistical matters correctly. To help understand some of the common mistakes we make, let’s take a look at a classic real-world statistical event – a coin flip – and then try to see what lessons we can apply to risk analysis and EHS management.

Let’s Flip

Suppose we take a coin and flip it four times in a row, and then repeat, writing down the results each time. Take a look at the coin flip results below.

Series 1

Flip 1 2 3 4
Outcome H H H H

 

Series 2

Flip 1 2 3 4
Outcome H T T H

 

Which of the two series seems more likely to occur to you, assuming we have a fair coin and no tricks?

If you answered that Series 2 seems more likely, because its result has exactly two heads and two tails, you’re not alone. Most people know that the odds of a coin flip coming up either heads or tails is 50/50, so they look at Series 2, see the expected equal number of “heads” and “tails” outcomes, and conclude it must be more probable than Series 1. And they are wrong.

Why? Because the statistical likelihood of a series of outcomes is just the product of the likelihood of all the individual outcomes. In layperson’s terms, we multiply all of the probabilities together. As we’ve already mentioned, the likelihood of a coin flip coming up either heads or tails is 50%, or 0.5. So in Series 1, the probability of the series is (0.5x0.5x0.5x0.5), or 0.0625.

What is the probability of Series 2? It’s also (0.5x0.5x0.5x0.5), or 0.0625, so contrary to what we may have first thought, the probabilities of the two series are exactly equal. This might seem counterintuitive because we know that “heads” and “tails” outcomes are equally likely, so a series that has the same number of each seems “right” to us. But we’re confusing the likelihood of an overall number of heads and tails, in any order, over a very long sequence of flips, with the likelihood of a specific sequence of “heads” and “tails,” especially in a short run of flips.

Over many flips, we’ll see many runs of a few heads or few tails in a row. The overall numbers of each even out to something close to 50/50 over the long run. We shouldn’t be surprised when any shorter sequence of flips doesn’t reflect the overall expected numbers of “heads” and “tails” over a longer time period.

What Does This Have to Do With EHS?

Great question!

To see the relevance, remember that probability is inherent in the concept of risk. Hazards have certain statistical likelihoods to actually result in a safety incident. And just as a short series of coin flips doesn’t necessarily reflect the probability of outcomes over the long run, safety performance over a limited period of time doesn’t necessarily tell you much about actual risk probabilities.

Let’s consider an example. Suppose we have two different facilities, which have essentially identical operations and numbers of employees. During a calendar year, Facility A has 5 recordable injuries, and Facility B has none. Based on this, we’re tempted to conclude that Facility B is safer, meaning, there must be fewer risks there. In fact, many real EHS professionals in a similar situation have made exactly that conclusion.

In a past life in EHS consulting and corporate EHS management , I saw something much like this scenario play out several times. Each time, EHS managers and corporate management saw the location with no recordable injuries as the “good child” – the one with a good handle on safety and few risks. Surprise and disappointment would inevitably follow when several recordable injuries within a short period of time would happen – a statistical phenomenon known by the cruel but accurate term “regression to the mean.”

Management would belatedly recognize that there wasn’t any significant difference in EHS culture or program strength that would’ve accounted for the temporary disparity in safety performance. They’d slowly realize the truth that there were risks at the facility all along, and that the time they’d been monitoring safety performance was simply not long enough to be truly representative of risk levels. Much like a short series of coin flips may not be representative of overall probability, a short duration of time may not show us safety performance representative of overall risks.

Or, as I’m fond of putting it: “Absence of incidents does not imply absence of risks.” Given enough time, safety performance will eventually reflect the real probabilities of risks in the work place.

So What Should I Do?

Another great question, not least because it recognizes that there is something that needs doing. Making assumptions based on short-term safety performance, as we’ve seen, doesn’t always work.

The main thing we should do is always assume that there is room for improvement – that there are risks out there we haven’t identified or fully controlled yet. We may have chemical hazards we haven’t properly evaluated, non-routine tasks we haven’t risk assessed, or temporary workers we haven’t adequately trained. Be proactive about looking for these issues, and encourage employees to participate in your safety culture by bringing hazards and safety observations to your attention. Even better, show them that you’ve listened to their feedback and acted on it, because that will make it more likely that they’ll participate in the future.

Coin flips are subject entirely to chance – make sure your EHS programs aren’t.

Let VelocityEHS Help!

Prevention of accidents starts with awareness of hazards, and that awareness relies on accurate assessment of risks. Our Risk Analysis software simplifies Job Safety Analysis (JSAs) for the job tasks in your workplace, and makes it easy to schedule, assign and track corrective actions to provide your employees with the visibility of risk management activities they need work more safely.

Want to be more proactive about occupational safety? Try our powerful and simple Audit & Inspection solution, which allows you to build and deploy specialized checklists for equipment, storage locations and operations, and to instantly generate and assign actions needed to correct identified issues.

To effectively manage Hazard Communication, you need to know what chemicals are in your inventory and understand the exposure hazards associated with them. Our MSDSonline SDS & Chemical management solutions give you the visibility you need to improve your hazard communication practices and properly assess chemical risks, including established exposure limits, and to help you determine the controls needed to protect your workforce.

Finally, always remember that your employees are your best resource for safety. Tap into the diversity of risk perceptions among your workers by holding regular safety meetings, and use the feedback from those meeting to improve safety for all employees. Our Safety Meetings solution will help you schedule your meetings, manage your agendas and log your minutes so you can make the most out of your safety team’s suggestions, and show your employees that their feedback matters.