Rockfordgreeneinternational's Blog

Much has been made of my criticisms of Behavior Based Safety.  For some, my questioning of BBS is essentially a personal affront.  Others accuse me of offering criticisms without providing viable alternatives. For the record, I don’t feel I have to offer any alternatives—and yes, I am publicly saying that if you knowing sell, promote or advocate a systems (safety or otherwise) that you know doesn’t work or that costs more than it will ever recoup, you are a conman, a thief, and yes a snake oil salesman and you know how you are so spare me your false indignation.

Despite all the hoopla about my criticisms of BBS and my repeated requests to those who would shout me down as charlaton who knows nothing of BBS to define the quintessential elements of a BBS program no one has responded.  There seems to be little consensus on the definition of BBS. So either people are selling something that they can’t define, or they are too fearful that another pretentious gas bags will decry them as heretics.

So deafening was the silence that I was forced to turn to the most dubious of sources: Wikipedia. It’s easy enough to dismiss Wikipedia as a useless source, but given the article is so blindly pro-BBS and virtually anyone can change the entry, I have to assume that the advocates would have turn into me with the same petty snappishness with which they have lit into me. According to Wikipedia a good BBS process contains four elements:

• Observation at site.  This practice can be boiled down to openly and honestly watching the worker and providing feedback on the safety with which the worker has performed the required tasks.  There are some obvious flaws associated with Behavior Observations. First, the success of the process assumes that the work performed during the observation is generally representative of the work as it is typically performed.  In other words, it assumes that the worker will perform the task while he or she is being observed in much the same way he or she would when unobserved.  This assumption is made despite reams of research that shows that any observation will change the behavior of that which is being observed.  It also ignores the Hawthorne Effect, which is the tendency for things to improve simply because there is an intervention of any stripe simply because the subjects know they are being watched.  The improvement is not permanent, however, so little or no benefits are derived from this short-term improvement.
  Secondly, even if the observer were to observe an unadulterated, and fairly representative performance of a task, it is fraught with problems.  Process variability and variability in human behavior means that for the observer to glean statistically valid information about the job the observer would have to devote considerable time to the observation and watch tasks being performed over and over again. This commitment just isn’t practical.  If we assume that a typical supervisor has 10 direct reports and does one observation a week, and said observation lasts 10 minutes we have a very small sample with which to work, a sample so small that any inference would be invalid—even if the process is in control. If the process is out of control and is not reliable than any observations are isolated events from which no meaningful inference can be drawn.
  Many organizations employing a BBS system expect a supervisor to do more than 10% observation rate and that in itself because problematic. Supervisors are increasingly scarce and are expected to do more and more with less resources. The pressure to conduct observations that consume resources while producing little in tangible results when supervisors are evaluated and compensated according to their productivity is far more likely to provide and incentive to complete the observation quickly instead of completing fewer, high-quality observations.
  Workers who are being observed understandably feel judged by supervisors and so BBS encourages supervisors to begin the feedback by focusing on the positive behaviors before turning to the areas requiring improvement.  But what BBS advocates tend to ignore is that supervisors will require substantial interpersonal skills to provide accurate and timely feedback.  BBS assumes a good faith effort by over taxed supervisors, and it further assumes a fairly mature training and performance appraisal system that provides and reinforces feedback skills. Certainly, these skills are important in the success of any supervisor but in my experience organizations where first line supervisors universally have mastered and routinely apply these skills are exceedingly rare.  This a foundation that is critical to the success of BBS and one seldom identified before the sale.
  Instead of requiring supervisors to observe an individual in isolation, supervisors should be encouraged to conduct layered process audits where the principle objective is to assess the overall effectiveness of  the department and to identify, contain, and correct the hazards that cause injuries.  It should not matter if the hazard is behavioral, procedural, or systemic. In fact, the walk thru need not be limited to safety—hazards often place more than workers in jeopardy.  A single hazard could threaten safety, quality, delivery, morale, profitability, and the environment.  Why burden supervisors with the requirement to laboriously observe workers one at a time when a more holistic approach would take far less time and yield much more substantial results.  When I say there is a better way, it is to this to which I am referring.
• Data gathering and preliminary reports. This element of BBS really is no different than any other approach.  Except that BBS relies on the information gathered in the observations.  While there is nothing explicit in BBS that would forbid including Total Productive Maintenance (TPM), 5S Tours, Process Layered Audits, Job Safety Analysis, or ergonomic data in this activity, I would submit that doing so muddies the water to such an extent those doing so are no longer following a strict BBS discipline.
• Report analysis and recommendation. Again, there is very little difference in this element of BBS than an element of any responsible approach to worker safety.  Unfortunately there is the GIGO phenomena which pertains to everything from accounting to databases.  GIGO stands for Garbage In Garbage Out and in the case of BBS GIGO means that if the information is flawed in what, where, when, and how it is collected than it is not likely to result in useful results.
  Let’s assume, though, that behavior observations produce perfect snapshots of workers doing performing there jobs.  Unless a process is under control, and by this I mean in the statistical sense, any inferences made based on this data is specious.
  But beyond the efficacy of the information, there is even more problems.  One principle problem is the whole concept of “recommendation”  BBS is a safety-professionalcentric process.  Safety professionals craft recommendations instead of a committee lead and staffed by Operations professionals who can make decisions and effect real changes in the workplace.

The Wikipedia article fails to identify several philosophic and procedural commonalities of many BBS initiatives that I believe should be addressed:

• Recognition and Rewards for a Lack of Injuries.  Most BBS programs have at least some program that recognizes a good safety record and rewards workers for not getting injured.  This element of BBS is dangerous and destructive.  Far from achieving its stated goal of rewarding safety, it instead rewards people for not reporting injuries and near misses.  Instead of having Safety BINGOs and bonus for days worked without injury employers should reward participation in making the workplace safer.  Most companies have some sort of suggestion program and should divert the resources currently expended on safety incentives to incentives for making suggestions that make the workplace safer. Workers who are injured often become pariahs and outcasts because they deprive the other workers of the reward.
• A Belief That Most Injuries are Caused by Intentional Unsafe Behaviors and At Risk Activities.  Nobody wants to get hurt and our business processes are not intended to hurt people so no amount of behavior modification will change these things.  Further, there are many types of hazards that may cause injuries that have little or no cognitive behavioral element—poorly maintained equipment, low process reliability, ergonomically unsound practices, inappropriate workspace layout, ineffective training, human error and a host of other failure modes can cause injuries.  The belief that any of these causes, up to and including deliberate risk taking, can be universally said to be the single greatest threat to worker safety is either the height of arrogance or the height of stupidity.  Instead of building a process based on questionable 80-year old research organizations need to do systematic research on the causes of injuries in their work environments, pareto the causes, and bring their limited resources to bear on the causes most likely to pose the greatest threat.  Organizations need to understand where they are at risk and focus their attention on the things that will help them reduce their risks.

There are elements of BBS that can be very effective, but when we go into a problem assuming we already know the majority of the answers (in this case we assume that most injuries are caused by unsafe behaviors in most work environments) we are highly likely to inappropriately attack the problem with prefabricated solutions.  People who make money on a given methodology will always find ways to manipulate research to support an off the shelf solution when a custom-solution (perhaps a far cheaper and simpler one) will suffice.  We can’t continue to spend what we can’t defend. Does BBS get results? sure.  Are BBS results real? Sometimes.  Are they sustainable? Maybe. Even a broken clock is right twice a day.  But selling a solution without completely knowing my situation just because some insurance inspector in 1931 asked around the shop and the supervisors told him that most workers got hurt because they needed to be more careful is, again I say, snakeoil.