By Phil La Duke
I really need to stop reading LinkedIn discussions. Recently Alan Quilley (a smart guy and truly thought provoker) posted a link to his article Risk Analysis and Management in a discussion thread. True to form the mouth breathers and water heads attacked the article like howling rabid jackals. I don’t, no let’s make that I WON’T, rehash the argument, but sufficed to say, there was a lot of trash slung by smug academics who have never once set foot in anything approaching a medium-to high-risk workplace. Their claim was that the equation Quilley proffered, Risk = Probability x Severity x Exposure, was too simplistic and that Quilley didn’t understand the math.
There was some validity to the argument, which Quilley freely admitted. Both sides agreed that the average safety person doesn’t really understand the nuances of risk, and for that matter probability. I may not have accurately captured Alan’s arguments, and if you’re reading this Alan, my sincerest apologies, but my intent as I have stated is not to rehash the argument, but it did get me thinking about how irresponsible some safety practitioners apply concepts they only sort of understand; strike that; that they don’t understand at all.
Let’s take probability; pretty easy right? If you’ve ever flipped a coin or shot craps you understand probability, right? Wrong. There is a lot more to understanding probability than calculating the odds. There is sample size, and margin of error, and so on. But let’s deal with the simplest definition of probability I could find (I won’t site a source because it appears in about nine different places and none of them site a source so I won’t give credit to someone who is plagiarizing, and before anyone accuses me of doing the same, I freely and wholeheartedly admit that I am not the author of this definition): “Probability is the chance that a given event will occur divided by the number of possible outcomes.” (Feel free to argue amongst yourself). Probability isn’t subjective; it’s absolute. The probability of flipping a normally weighted coin heads side up is 1 in 2 or 50:50, but even in this simple example there is a third, extremely remote possibility that the coin will not flip on either side but land on its side; (let’s just chock that up to “margin of error”) even so, there is a remote possibility that the coin will land on its edge. So while it is generally accepted that the chance of a flipped coin is 50:50 it really isn’t. If we further consider that all coins are not the exact weight and shape (whether because of minor deviations in the minting process, wear from the time and condition of the circulation of the coin, or some other reason for which I can’t imagine) then there is even less certainty of the 50:50 probability. The point is, if we can’t even count on the purity of the odds of a coin flip how can we expect to calculate the odds of an injury.
We tend to think of the probability of injuries as fairly binary—there are two possible outcomes: injury or no injury. This thinking sounds reasonable but it is deeply flawed. Take a look at a person completely a task as part of his or her job. There are more than two outcomes, clearly there is the chance that the worker will be uninjured, we cannot treat the employee being injured as a single outcome because there are multiple causes for a worker injury. Not only that, there are several other outcomes we may forget to consider. For example, a worker could be killed, or at the other extreme the worker could suffer a near miss.
Just as the weight and shape of an imperfect coin can artificially impact probability so can things like the worker’s capability (training, natural aptitude, risk taking, behavioral drift, performance inhibitors, etc.) the process capability (process tolerance, reliability, etc.) in fact there are so many variables at play in worker injuries it’s a wonder we try to calculate probability at all.
To be fair to Quilley, his formula was never meant to be a scientific predictor of a given outcome, rather it is a workable formula for prioritizing injuries, and yes to be fair to the academics Quilley has over simplified probability. So what are we to do with all this? I for my part agree with Alan. The safety practitioners and frontline supervisors shouldn’t have to work differential equations to calculate risk. We need a practical, usable, and simple way to determine whether or not a given task is too risky to perform, which risks on which to concentrate, and which risks are more likely to cause the most severe injuries. A safety practitioner should not have to be Euclid to calculate probability, but then again one should also know that his or her calculation of probability is little more than a guess. A good guess to be sure, but a guess nonetheless. It’s an educated guess based on years of experience
To some extent it comes down to just plain sense. We know drunk driving is dangerous because we have seen too many tragic accidents caused because a driver was drunk, and you don’t have to be Pythagoras to foresee that a teen (or worse an elderly) driver texting is a high risk behavior.
What I am saying, in my round about way, is that arguments over whether or not a safety professional can accurate calculate probability of injuries is of far less important than whether or not we can prioritize the correction of hazards. Someone once said, “if you can’t say something nice about someone, then don’t say anything at all”. To the academics who went to such pains to argue against Quilley’s points I say, “shut up”. Not that I want to stifle freedom of speech, but yammering on and on about how wrong someone is without offering some useful counter suggestion is tantamount to bullying, and as much as I enjoy bullying, I say to the academics, take your theories back to the ivy towers where you can poison the minds of tomorrow’s leaders; they’re not welcome here and we’ve got work to do.