The Disturbing World of Fallacious Conclusions and Specious Arguments

Almost a month ago I was engaged in a spirited and contentious debate, again about the supposed merits of Behavior Based Safety.  Once again I was shouted down on line for having the unmitigated audacity to question the long-term impact of Behavior Based Safety.  It started when I made the admittedly blunt assertion that the contention that “people either choose to work safe or unsafe” is an unsupportable position. I thought the statement was clear, concise, and accurate. I certainly wasn’t trying to cultivate controversy.

Almost immediately the forum filled with people who questioned my experience, knowledge of safety and credibility.  After all, who was I to call the emperor naked? Millions have been made on systems that seek to make the workplace safer using basic, Skinner-based behaviorism.  How dare I question all of that?

Behavior-Based Safety proponents point to a study conducted by the National Safety Council that found that 90% of all worker injuries were caused by unsafe behaviors.  Most safety professionals hold this study sacred; it makes sense.  We’ve all seen instances where worker injuries and even tragedies could have been easily prevented had workers just acted with more care, professionalism, or plain common sense.  Even though this study is over 30 years old and to my knowledge never been independently confirmed in the safety community to question it is to commit the worst kind of blasphemy (Some of the more staunch allies will point to Heidrich’s Injury pyramid that also found (over 70 years ago) that a high percentage of injuries are caused by unsafe acts).  But I don’t question these findings.  I believe that 100% of all injuries have some behavioral causation, in so much as if no one is doing something than no one is likely to get injured.

My point is who cares? I don’t want to quibble with statistics, my point is, okay, so now what? BBS providers seem to have had this “aha” moment that holds that since behavior played a role we now have the magic bullet to prevent all injuries because all we need to do is remind people to be more careful or motivate them to be more safe, or use basic behavior modification to “fix” people. Many BBS theorists never asked “why” the people behaved the way they did. And without Root Cause Analysis to understand not only why they behaved the way they did, but also why they believed what they were doing was safe.

The crux of my view of safety is this: nobody wants to get hurt and you system wasn’t supposed to hurt them se we had ought to fix the problem not the blame. Recent, well-documented repeatable studies on how the brain works indicate that human error is an inevitable (and in an evolutionary sense desirable) human characteristic. So we can only train people to work safely to an extent, and we MUST be more proactive, which allows us to apply controls that are higher and more effective on the Hierarchy of Controls.

In most cases, I believe that too much emphasis is placed on individual behaviors and not enough on organizational behaviors, in other words too many people worry about modifying individual behaviors (while ignoring human error research, Maslow’s work, Fredrick Taylor’s work, Edward Deming’s work, the entire fields of organizational psychology, neurology, anthropology, and more) at the expense of modifying systemic issues that cause people to make bad decisions that ultimately get them hurt.

Safety is a qualitative measurement that most companies treat as a quantitative measurement. If company A has less injuries than company B it does not necessarily follow that company A is a safer place to work because as you point out there is an element of luck. Certainly there is a correlation between risk and injuries, but safety is a relative term used to describe risk. And as long as we view safety as a body-count things will continue to erode.

Safety is the probability of a worker doing his or her job without getting hurt. There needs to be a paradigm shift within the safety community.  At the foundation of this paradigm shift is the question of exactly how do we calculate risk. I’ve been working on an answer to that question for sometime and still don’t have a formula I’m happy with, but in broad strokes, I believe that the more hazards you have the more likely you are to injure workers. But there are other factors at play that raise the overall risk of injuries in the workplace.  Factors that directly influence the likelihood that people will make poor choices: ineffective communication practices, weak incident investigation, high amounts of nonstandard work, processes that aren’t in control, production bottlenecks, unstable work levels (layoffs/hiring), etc. The presence of these factors increase risk of injuries and safety can only be increased if we identify and manage these risk factors. Yes, behaviors play a key role in risk and should be managed, but not necessarily through behavior modification.

Millions are spent on Behavior Based Safety where the goal is to reward people for working safely, and yet most decisions made regarding safety are made in a reactive microsecond where there isn’t really time to make a conscious decision as to how the best and safest way to respond. Even in cases of mechanical failure, the true fault can often be traced back to a behavior: poor maintenance, poor inspection process, improper installation, etc.” I agree. But generally speaking, BBS has been bastardized to the point where organizations are only looking at behavior at the production level. They do a shoddy job of root cause analysis and if they do go far enough up the decision tree they far too often cop out without addressing the organizational system flaw that provides the “why” behind the behavior.

As I have said, all injuries are caused by behaviors if we look hard enough. But there is a huge leap to the conclusion that we can therefore use behavior modification to prevent all injuries. The debate between process solutions and behavior modification is pointless. We need to take a more holistic view of safety and focus both on improving processes, providing good training in how to correctly do a job (I’m talking core skills training not safety training—a welder who understands how to weld is infinitely less likely to injure himself or others than someone who has been given 40 hours of safety training but can’t weld.), reduce nonstandard work, and yes provide feedback on behavior.

Ergonomics, fork-truck accidents and lockout violations cost industry $100s of millions. And yes, there are significant behavioral elements to each of these but telling drivers to be more careful is like telling me to be taller. I’d like to be, but having my supervisor observe me won’t change anything for long. Do you ever speed? If you see a policeman do you slow down? do you resume speeding? This is because enforcement changes the climate of safety but does little to change the culture.

I was recently asked “why don’t we start to dissolve the artificial distinctions between protection and production in favor of effective total performance. (sic)” Why? because there are a LOT of people whose livelihoods depend on one methodology at the exclusion of all others. We can’t put all our proverbial eggs in one basket. We need to balance the need to protect workers against the need to produce at a competitive rate. But the first step in doing so is to recognize the quantitative “safety” does not exist, and that only by viewing safety as a comparative, qualitative measurement can we ever hope to recognize the apparent dichotomy between safe and productive. If we view both these terms as qualitative in nature we can begin to see them as supportive of one another, look at correlations between improved safety and improved productivity, quality, cost, and morale.

Behavior management has a place in any good safety program, but it needs to be counter-balanced by mistake proofing (which really isn’t about not making mistakes but in most cases its about ensuring that mistakes don’t kill anyone.

BBS alone will always fail because it ignores the fact that people don’t WANT to get hurt or always CONSCIOUSLY make bad decisions that lead to injuries. A fair amount of unsafe behavior is not a conscious decision so trying to modify it by rewarding a “good” decision is pointless.  Process based safety alone will always fail because…well you can’t bubble wrap the world.


#attitude, #attitudes-toward-safety, #behavior-based-safety, #behaviour-based-safety, #contingency-planning, #culture-change, #dont-hurt-yourself, #fabricating-and-metalworking-magazine, #increasing-efficiency, #increasing-productivity, #just-cause, #phil-la-duke, #phil-laduke, #philip-la-duke, #philip-laduke, #rockford-greene, #rockford-greene-international, #safety-culture, #safety-incentives, #stop-trying-to-prevent-every-possible-accident, #variability-in-human-behavior

Opening a Dialog About Safety

The Editor of Fabricating and Metalworking suggested that in an effort to ensure that i don’t rehash topics that I’ve already covered in prior installments of The Safe Side my monthly column devoted to worker safety.  This is what I submitted as the introductory article.

From the outside, the world of safety looks pretty simple—either a workplace is safe or it’s not. And if a workplace is unsafe it’s reasonable to expect that people will agree what specifically is making it unsafe.  But for those of us who work within the safety profession nothing could be more complex or hotly contended than what constitutes a safe workplace and what is the biggest reason that workplace is unsafe.  In fact, it’s hard to get safety professionals to agree on the very definition of the word “safe”. So instead of a column this month, I’m beginning a series on safety.  In it I will explore different ideas about safety and what can be done to make the workplace safer.

Measuring Safety

It’s a widely held belief among system thinkers, lean gurus, and Quality Operation System enthusiasts that you get what you measure and if you can’t measure something it may as well not exist. This gets dicey when you try to measure safety, because, like quality, safety isn’t seen as the presence of something rather than the absence of injuries.  And it is near impossible to proclaim the absolute absence of something.  So let’s begin our discussion with some context and some definitions of basic terms.  The absence of injury does not denote the presence of safety; rather “safety” is an expression of probability and a calculation of risk.  Because safety is a probability nothing can ever be pronounced completely “safe” instead, safety is relative.  We can accurately describe a situation or condition as “safer” than another, however.

But even so, the extremely high variation in working conditions and the even higher variation in human behavior make it tough to get an accurate read on the relative risk endemic to a given activity.  In short, for the most part all we can do is guess at the probability that a given activity will result in an injury, and even were we able to predict with statistical certainty that an injury will occur there is little we can do to predict the severity of the ensuing injury.  The difference between a fatality and a near miss (a “close call” where a worker could have been seriously injured but was spared) is little more than luck.

The Probability Gambit

One area of safety where there seems to be little disagreement is in the belief that the greater the number of hazards with which worker interact, and the greater the frequency of that interaction, the higher the probability of injury.  If a worker continues to behave unsafely or perform tasks that have been poorly designed from an ergonomics perspective eventually someone (not necessarily the worker him/herself) will get seriously injured or killed.  But because it’s impossible to say with certainty how much time or how the worker will be injured it is often dismissed as safety bugaboo.  As I used to say (when people would be overly concerned with remote possibilities) “maybe the moon will fall out of the sky.”  Warning of increased danger without being able to quantify probability is useless information—unless your sole intent is to say “I told you so”.

Man Versus Machine

Now that we have a common understanding of the definition of the term “safety” we can now explore the various, hotly contested theories of the best way to improve the safety of the workplace.  The first argument you will likely encounter is the philosophic question “what causes injuries?” One school of thought holds that because most injuries (studies have suggested that as many as 90%) have a behavioral element it follows that injuries are caused by unsafe behaviors and if unsafe behaviors are the most likely cause of injuries the most reasonable way to reduce injuries is to reduce unsafe behaviors.

Many safety methodologies focus on basic behavior modification techniques—carrots and sticks (that is, rewarding desired behaviors and punishing undesirable behaviors)—to increase safe behaviors while decreasing dangerous activities. Advocates hold that time tested research and hundreds of organizations support their techniques, many of which suggest observing workers while they work and offering feedback on the safety of their activities.  These systems encourage employers to offer financial or other incentives for low injury rates and improvements in key safety measurements.


No Life Was Ever Prolonged By Reminding A Person Not To Die

Critics of behavior-based safety systems contend that because nobody wants to get hurt and processes aren’t designed to injure workers the resources expended to improve worker safety would be better brought to bear against system problems. They deride safety observations as expensive and deeply flawed babysitting and incentive programs for encouraging under-reporting of injuries.  These people would have us believe that the road to a safer workplace is through mistake proofing our processes and that if people aren’t intentionally getting hurt no amount of behavior modification will change things.

Proponents of process-based safety believe that efforts to improve worker safety must be at the top of the Hierarchy of Controls (an engineering tool for determining the most effective way to eliminate the risk of process failures) Critics of process based safety counter that we can’t bubble-wrap the world and once a process is in place it is often too costly and impractical to idiot-proof the world.

To a layman the argument over behavior versus process seems pretty basic, maybe even pointless but these philosophies have staunch supporters and bitter critics and the actions you take to make the workplace safer are intrinsically linked to where you stand on this issue.

Individual Responsibility Versus System Responsibility

Maybe you’d prefer to argue the merits of holding the individual responsible over the system, or vice versa.  Some believe that strict policies and dire consequences for incompliance are the way to a safer workplace, while others argue that only an enterprise-wide solution can reduce injuries.  Both sides have their points; after all don’t we need to hold someone accountable for gross negligence and dereliction of duty?  On the other hand, volumes of research prove that populations act very differently from individuals and many experts believe that human behavior is strongly influenced by the systems in which people interact and that the pressure to conform to societal norms manifest in how people behave.

Safety Schools of Thought

Each of the activities in these quadrants represent some activity linked to worker safety in some way.  Some take a more holistic view of worker safety than others and some treat safety as a process outgrowth.

#attitudes-toward-safety, #behavior-based-safety, #behaviour-based-safety, #culture-change, #driver-safety, #fabricating-and-metalworking-magazine, #increasing-efficiency, #injury-prevention, #just-cause, #phil-laduke, #philip-la-duke, #philip-laduke, #process-safety, #rockford-greene, #safety-culture, #safety-incentives, #safety-recognition-programs, #variability-in-human-behavior, #worker-safety