All others bring data.”
Wouldn’t it be nice to take a break from all the bad news making the rounds? The last thing you need is one more writer reminding you of what you already know about what’s now being called “the new normal.”
You know what I mean. Thirty million or so of our friends and neighbors thrown out of work. If you’re lucky enough not to have been one of them, at work you’re now facing a long list of new challenges brought on by social distancing. From the Department of the Obvious, managing from the kitchen table – or by zooming in – does nothing to make managing safety performance easier.
How about the collateral damage to your personal life. School’s closed; so is day care. Kids at home and underfoot. If you’re working from the house, think that’s a distraction? When they’re bouncing off the walls, what are you gonna do? Vacation plans? Scrapped indefinitely. Dining out? That hasn’t been an option.
Throw in what’s happened to your retirement savings, the real worry about your health and that of your family members – that is the point of everything – you’d have to be crazy not to be feeling stressed out.
So you are. Even if you don’t let on that you are. Think those good followers of yours aren’t feeling what you are?
Of course they are. Stress levels in organizations like yours are at levels as high as I’ve ever witnessed. I’m not making this up, just adding it up.
Logic would suggest the impact of all that stress should be showing up at the bottom line of safety: who goes home safe and who does not.
Is that what’s happening?
Assumptions, Hypotheses, and Facts
If you’re a regular reader of the News, I’m betting you’ve already connected up the dots and jumped to your conclusion. You’re thinking, “Paul probably asked for the data, and has the answer.” If you know me well, you probably would add, “That contrarian is sitting on data to prove that conventional wisdom wrong.”
We’ll see about that. But first things first: let’s make sure you understand what you already know on the subject.
As a leader, you routinely operate – by “operate” I mean act, make decisions and view situations – on the basis of certain beliefs and assumptions. A lot more beliefs and assumptions than you recognize and appreciate. You do this all the time, usually without even giving the subject a thought. You just assume these to be true. Everybody does.
Want proof? You should demand proof. Show me the data!
Let’s start with Cognitive Biases. The hundred or so different mental shortcuts we humans take, for the reason everyone takes shortcuts: it’s easier. Why think hard when you can think easy?
For example, instead of looking for data that disagrees with what we think, we look for data that confirms what we already think. That’s called Confirmation Bias. We stick with what we already decided – decide based on what just happened. Anchoring Bias and Recency Bias. Didn’t have a contingency plan for a pandemic? Nobody did, because that’s something that hasn’t happened in your lifetime. Normalcy Bias.
Think you’ll have a COVID contingency plan for next year’s flu season? Of course you will. Circle back to Recency Bias. If you’re tempted not to, you’ll look around and see that’s what everyone else is doing. Uh oh. So you’ll “Get on the bandwagon” and “Run with the herd.”
Two more Cognitive Biases. We are social animals. Who’d want to be a contrarian?
Then there’s the Conventional Wisdom. These are the things considered to be true, but without being put to the test of real data. It’s like doubling down on Cognitive Biases. My personal favorite: “Familiarity breeds contempt.”
That’s why the manager of the baseball team doesn’t hang out with the players after the game. Unless the manager happens to believe, “To know someone is to love them.”
Which wisdom do you accept as true?
Thank you, Professor Frank Miller, for posing that question. In March 1971, to a classroom filled with mostly dozing college undergraduates. Mostly does not mean entirely. That’s the error known by the name of the Some All Fallacy. The next time you use the words “everyone” or “no one” you’re probably stepping into that.
Want some examples specific to managing safety performance? Here are a few I get a lot. When correcting behavior, you have to give someone a compliment as part of the discussion to influence their behavior. When there’s a big change in a policy or procedure, making the decision is the hard part, not to mention the most important part. Rolling out the change is the easy part, and somebody else’s job. As you move up the chain of command towards the top, leaders have more influence, not less.
If you are of the opinion that is true, the best advice I can give you is to stop reading, go back and re-read that last paragraph, identify each Conventional Wisdom, and put them to the test. I promise you will talk yourself into the right answer.
If you’re too busy to do that, send me an email and I’ll do the talking…er…writing for you. Managing safety performance is too important to rely on any assumption that is flat wrong.
Testing The Hypothesis
If nothing else, the process of management would be better served were leaders to simply acknowledge when they’re making an assumption or acting on an untested belief. At least that way, every leader would know if they’re standing on terra firma – or not.
An even better approach would be to do what a good research scientist does: form a hypothesis and test it. Find data that confirms the hypothesis. Seek out data that inconveniently does not fit with the hypothesis. Our natural inclination not to take that second step was discovered (or confirmed) in a famous experiment that led to the naming of Confirmation Bias.
So, to the research question du jour: the correlation between organization stress levels and safety performance.
A few weeks ago, we hosted a webinar on the challenge of managing safety performance in the time of COVID 19. Called it Managing From The Kitchen Table because that’s exactly what a fair number of your peers have been forced to do. Not anybody’s idea of a good idea, but we don’t always have a choice in these matters.
If you missed the call, it’s been saved. In your spare time, feel free to give it a listen. It’s free and there are no catches.
Knowing that a significant number of your peers would be on the call, it was the perfect opportunity to test that hypothesis. I will freely admit when I worked on your side of this deal – as a manager in operations in the chemical industry – I subscribed to that belief.
We ran a simple poll question on the call to test that hypothesis: What’s happening to your safety performance during this crisis?
The choices were:
- It’s gotten worse
- It’s stayed the same
- It’s gotten better
- Honestly, I don’t know
Note the research methodology: the poll was anonymous; no leader was asked to produce any data to support their answer. There were fifty-nine responses.
If you know anything about conducting surveys and sampling data, you know the potential problems with a drawing a conclusion from a small sample and from responders who may not be a representative sample of the general population. This was both.
But with those disclaimers, I have hard data, and it is more than what is known in the research business as “anecdotal data”, a fancy way of saying a few stories.
Bring On The Data
One more thing. Consider what happens by slowing down the thinking causes, reducing an assumption or belief to writing, and figuring out how to put a belief to the test: that process reveals!
It sure as heck revealed a lot to me.
Following that process caused me to think about the situations where I assumed that high stress and disruption causes injury rates to rise. I asked myself, “On what basis did I reach that conclusion?” I tried to match up safety performance data with times of stress and disruption.
The more I thought about the hypothesis, the less I liked it.
See what I mean about the process being revealing? In fact, I actually developed an alternative hypothesis: in times of stress and disruption, safety performance might actually get better, not worse.
If you think I’m making this up, ex post facto, you need to listen to the call. I actually climbed out on a limb and made a prediction as to what the data would prove – after it was collected – but before it was revealed. This really is how hypothesis testing works
My prediction? On average, safety performance in the sample population would stay the same and may very well get better.
I know: bring me the data.
- It’s gotten better: 42%
- It’s stayed the same: 36%
- It’s gotten worse: 10%
- Honestly, I don’t know: 12%
The Last Word
As to why the data might be what the data is, there are any number of explanations. Organizations perform well in crises because a crisis brings out the best in people. I suppose someone might argue it’s harder to get hurt working from the house, and easier to under-report injuries when the boss is sitting at the kitchen table.
It’s just one small study. Someone who does research for a living ought to do serious research on the matter. That will probably happen. In a few years, you can look up the findings in some academic journal.
Thing is, neither you nor I are in the research business. Your job is to lead and manage safety performance, and my job is to help you do your job well – and better.
Right here and right now, there’s a simple and important lesson to be learned and understood: it’s far better to rely on data than trust your assumptions are right.
Paul Balmert
May 2020