World class safety won't come from "accidental success"
May 24, 2011
In 1898, E. L. Thorndike published a seminal paper in which he introduced the “Law of Effect.” In it he demonstrated that behavior resulting in a “satisfying state of affairs” will be repeated in the future in the same or similar circumstances. Moreover, he established that behavior resulting in an “annoying state of affairs” would not be repeated and would eventually go away. These findings were the foundation of modern behaviorism and later theories of reinforcement.
The work initially conducted by Thorndike involved putting cats into “puzzle boxes” requiring that they push a lever in order to escape the box and obtain food. The satisfying states of affairs were obviously escape from the box and the food reward. Without any knowledge of the box cats would first explore randomly and later try various responses that would not work to escape. They would eventually happen upon the correct response--pushing the lever. Thorndike measured the time it took the cats to perform the correct response each of several times they were put into the same puzzle box. What he found was that during each subsequent trial less and less time was required for the cat to push the lever. Ultimately they would perform the correct response right away without any incorrect responses intervening.
Popular psychology and education have called this trial and error learning. However, Thorndike originally described it as “trial and accidental success” learning. This gradual learning process produces what today is known as the standard learning curve. This learning curve is very predictable in human behavior.
So, what can safety professionals learn from Thorndike’s work performed more than a century ago? Put simply, we cannot rely on “accidental success” to keep people from getting hurt. And yet every safety program that focuses on recordable injuries as the primary indicator of safety success is doing precisely this. Let me explain.
Injuries are the byproducts or outcomes of behavior. To the organization, the absence of injuries is equivalent to Thorndike’s satisfying state of affairs. Having an injury would therefore be the opposite. And, although like Thorndike’s cats we should learn from things that do not work (i.e., those things that led to an injury), if this is all we’re doing, we’re going to hurt a lot of people in order to obtain a vision of zero injuries (the equivalent of the cats pushing the lever to escape the puzzle box without first doing other incorrect responses). That is, the gradual learning curve comes at a great cost.
Fortunately, in safety we don’t have to rely on accidental success. In fact, we can choose to take advantage of what over 100 years of behavioral science research has taught us since Thorndike’s law of effect was published. Moreover, in the right hands, we can put the applications of this knowledge to immediate use in our organizations. A well designed and implemented Behavioral Safety Process is intended to put you into control by teaching you the things you need to know from 120 years of rigorous behavioral science taught in a manner that anyone can understand. Below is a quick example of what you will learn to apply.
In Thorndike’s study, the cats had an opportunity to learn only through direct experience. Of course, we must learn from experience, but direct experience with at-risk situations will get people hurt. Thus, we must do more. And the behavioral sciences have a lot to say about what can be done.
Specifically, to speed up the learning process, we must understand and apply a 3-term model of behavior called ABC. That is, in order to change the gradual learning curve associated with trial and accidental success we must first Activate the behavior with specific direction so that the performer knows what to do. Next, we must observe the Behavior occur in its natural setting and evaluate it relative to the expectation we established during the Activation phase. Third, we must provide appropriate Consequences for the behavior observed. Specifically, we must reinforce, with positive feedback, behavior consistent with the expectation (a satisfying state of affairs to the performer) and we must stop behavior not meeting expectations and re-Activate the expected behavior with corrective feedback (an annoying state of affairs to the performer).
If we regularly apply in our organizations what we have learned from 12O years of behavioral science research, we will embrace a journey of continuous performance improvement such that we can drive the numbers for which we’re held accountable rather than relying on chance or accidental success.