Waiting for the Mexican Elevator

Actually, these elevators are in the Empire State Building, not Mexico City, but that’s where the CC-licensed photo was taken. Don’t tell.

As I write this, I am sitting in my hotel in Mexico City, taking a break from teaching clicker training to instructors from all over Mexico and as far as Guatemala and Colombia. It’s been a great time thus far! Saturday Alena and I spoke 8 hours on aggression, and this week we’ve been working on clicker mechanics, foundation work, shaping, etc.

With students of varying levels of clicker experience, we’ve been pulling a lot of everyday examples of operant and classical conditioning at work. There’s a perfect discussion topic waiting in the form of the elevator in our hotel.

Our hotel is perfectly nice (except for an odd war we’re running with the maid service over towels), with a doorman and a convenient taquería and everything. But the lobby elevator call button’s light is out — when you push the button to call the elevator, the light doesn’t come on.

Now, this isn’t terribly obvious; the hotel’s glassy architecture includes avante-garde call buttons and there isn’t a visible dead bulb. But it means there is no feedback upon pressing the button. The elevator has been called, but we can’t see it.

And so, frustrated behaviors begin. Some people hit the button repeatedly; some look around for another button to push. Some press, fidget, press, fidget; some wander around the corner to make sure stairs are available. No one simply hits the button and waits quietly — not even those who know the light is out and the elevator is coming. Why?

Behaviorally speaking, the lack of feedback (reinforcement) is creating uncertainty. Even though more valuable reinforcement (the elevator) may eventually be coming, the person pressing the button isn’t sure of that. Consequently, other behaviors appear — extinction bursts (repeatedly pressing the button, or pressing it harder and harder), seeking (looking for another button), displacement behaviors (fidgeting), or abandoning the behavior (looking for stairs).

This is why we always reinforce good behavior in our dogs, even if only verbally. I don’t want displacement behaviors showing up in my carefully trained behaviors! so even if I’m not going to treat, I still praise or otherwise positively reinforce (primary or conditioned).

That’s great, you say, but what about the people who know the elevator is coming? Surely if they know the light is out, they can wait quietly? Why do they still fidget or try alternate behaviors?

That’s superstition at work! After the uncertainty of the initial button push created alternate behavior, the elevator eventually arrived, of course. This reinforced not only the original button pushing, but all behavior subsequent, including the unnecessary behavior. So upon returning to the elevator, even though one knows the button will work, one unconsciously includes the fidgeting behavior which occurred before the appearance of the elevator.

How powerful is reinforcement and superstitious behavior? I’ve been in this hotel four days, I know exactly what is going on behaviorally, and I still press the call button more than once. I’m ashamed and fascinated at the same time.

Is lack of prompt reinforcement diluting your trained behaviors with unwanted fidgeting or experimentation? Are superstitious behaviors invading your training? Take a look at your reinforcement timing!

UPDATE, 7-22:

Since writing this, I’ve continued to observe, and I’ve found one more interesting behavior which appears very often — in the quest for feedback, I often see people hit BOTH buttons, even for the direction they don’t want, just so some light would appear! Doesn’t this sound like a dog or child performing some unwanted behavior in order to get attention or feedback of some kind? Hmm….!