Fellow Toastmasters,

This speech is highly scientific, but was inspired by... my cat1. Did you know that as of 2009, US pet owners have 93.6 million cats?

How many of you have cats?

And how many of you allow them to sleep with you in the bed during the night?

And how many of those of you don't have a significant other?... Never-mind that question!

Anyway, I love cats, but I also love my sleep, and cats tend to be active at the most ungodly hours of the night... and they want you to be active too! The problem is that it only takes one time - one time - of paying attention to your cat at 5am, and she will know to bug you for weeks to come. Let's say that for the following days, you roll over, pretend to be asleep, yell at the cat, or otherwise stubbornly resist... but how long can you resist that cute ball of fur? Eventually, you'll give in: you'll wake up, get out of the bed, and feed the cat.

At that point, it's pretty much game over. You've just rewarded the cat's behavior of waking you up, after not rewarding it for a while. The cat has learned that even if you don't feed her at 5am for a while, you eventually will. In other words, historically, her attempts of waking you up in the wee hours of the morning have eventually worked.

What you did is called "intermittent reinforcement", or "random reinforcement", and psychologist B.F. Skinner famously quipped:

Behavior that is reinforced intermittently is much more difficult to extinguish than behavior that is reinforced continuously.

Why? Because once the continuous reinforcement stops, there's actually a good chance that it won't resume; while intermittent reinforcement has, historically, resumed every time.

This is why many undesirable child and teenage behaviors are so difficult to stop. We might be able to resist a child's nagging most of the time, but if we yield every once in a while, the child will persist with it. Random reinforcement works.

Can you guess a very profitable use of the principle of random reinforcement...? Las Vegas slot machines. [pause]

By law, slot machines in Nevada are required to pay out at least 75% of the money they take in. This means that if you stick around long enough, you'll see $7.5 dollars being paid out for every $10 played (by all players), so you'll eagerly await for the other suckers to run out of money, until your turn at the machine comes (to run out of money too).

An even more fascinating instance of random reinforcement in action is the Stockholm syndrome. This is a psychology term for situations in which hostages develop feelings of appreciation and even admiration for their kidnappers, and often end up defending them in court, in spite of having gone through days of captivity, danger and abuse. This may sound odd, but it's not that rare - the FBI's Hostage Barricade Database System shows that roughly 27% of victims exhibit evidence of Stockholm syndrome.2 Here is the explanation that FBI psychologists offer for this behavior:

Kindness serves as the cornerstone of Stockholm syndrome; [...] If the captor is purely evil and abusive, the hostage will respond with hatred. But, if perpetrators show some kindness, victims will [...] concentrate on the captors' "good side" to protect themselves.3

"if perpetrators show some kindness"... Random reinforcement... works.

Now, hostage situations are relatively rare (how many of you have been taken hostage?), and you may rightfully ask, "How is this speech relevant to me? What is the take-away value?"

I'm going to say that quite a number of us may be unaware victims of random reinforcement, and that we can use random reinforcement instead of continuous reinforcement, for good purpose.

Some of us may be in abusive relationships (of any nature - at work or at home). In quite a few abusive relationships, the abused person does not attempt to fight or escape the abuser, nor do they attempt to seek assistance - because the abuser is sometimes nice to them (or even to someone else - "He/she never shouts at the children").

We may watch whatever is on TV (the average American watches 5.1 hours of TV per day4), or read magazines that are overall useless, because once in a while, we do find something interesting. (Yet we don't read National Enquirer, because that one just gives no reinforcement.)

And so on. Playing the lottery (or paying the lottery). Holding on to a losing stock because at some point it did go up. Et cetera.

Now, how can we use the principle of random reinforcement to our advantage?

Children can be taught that chores and homework need to be done anyway, and they'll get a reward at some point. We can get some quality "disconnected time" by not picking up the phone or responding to IM or e-mail all the time. We'll still get calls and IMs and e-mails because we do eventually respond. (On the other hand, if you want to give a bad date the hint... don't ever pick up). Speaking of dates - and I'm getting more and more controversial with the advice here - random reinforcement explains why some women fall for jerks who generally mistreat them. Jerks can get away with being jerks because once in a while they do something nice.

The same idea of doing something nice, this time for yourself, is often mentioned as a productivity technique: "rewarding yourself every time you get something done". The truth (established by rigorous social science experiments)5 is that rewards improve productivity only for mechanical tasks or for tasks that don't require much cognitive ability. Various authors mention this as one of the most consistently reproduced findings of social sciences, and one of the most ignored. Indeed, businesses offer all sorts of incentives for high performance, but those incentives have been shown to reduce performance. Those of us who are managers may want to use random reinforcement in some fashion, instead of continuous reinforcement.

Finally, awareness of the random reinforcement effect can help us see if we are in relationships that are occasionally gratifying, but overall, unhappy.

And in my case, if I want to have a good night's sleep, I must certainly never, ever, feed the cat at 5am again.


  1. I don't really have a cat. ↩

  2. G. Dwayne Fuselier, "Placing the Stockholm Syndrome in Perspective," FBI Law Enforcement Bulletin, July 1999, 22-25. ↩

  3. "Understanding Stockholm Syndrome". FBI Law Enforcement Bulletin. Department of Justice Federal Bureau of Investigation. July 2007. pp. 10. ↩

  4. Americans Watching More TV Than Ever; Web and Mobile Video Up too - "The recent results of Nielsen's Three Screen Report - a quarterly analysis from Nielsen's Anywhere Anytime Media Measurement initiative (A2/M2) - show that the average American watches approximately 153 hours of TV every month at home, a 1.2% increase from last year." ↩

  5. Challenging Behaviorist Dogma: Myths About Money and Motivation, Kohn, Alfie, 1998. ↩

My tags:
 
Popular tags: