How to quit coffee?

A marathon-runner named Matt has the (ambitious?) goal of “quitting coffee.” And, he doesn’t just want to stop drinking coffee – he wants change his belief that “coffee makes [him] happier and more creative.”

How? Well, it’s pretty simple, and arranged like an experiment. Each day, he’ll drink a cup of coffee – but in each cup, there will be some regular, and some decaffeinated. Furthermore, his wife will be randomizing the proportions. And each day, he’ll record his performance, and how he feels – with the hope of proving to himself that on the days he does have caffeine, he feels fatigued or less effective.

But there’s a problem.

Before that, an interlude:

Psychologists in the middle of the last century spent a lot of time studying “learning.” Learning is divided into two types; there’s classical conditioning, which is Pavlov’s dog salivating in response to a bell, and operant conditioning, which is an individual (or rat) choosing to do something. A classic example of operant conditioning is pulling a lever to get some food; once the creature learns they pull the lever to get food (as opposed to whining piteously).

In operant conditioning, one thing is really important: the reinforcement schedule. A continuous reinforcement schedule – where you “reward” the individual every time they complete the action – is standard, and the fastest way to get something to learn something. The alternative is partial reinforcement schedules – where the individual is rewarded only some of the time. Partial reinforcement doesn’t cause people to learn as fast as continuous; but it makes the “response more resistant to extinction.”

That is, under partial reinforcement, people won’t eventually stop. If a rat in a cage presses a button, and food comes out sometimes – randomly – then the rat is going to push the button more than under continuous reinforcement, and will also keep pushing the button long after a rat under continuous reinforcement has given up on it. Humans are worse, if anything – they try to create a pattern to predict when the reward will come, even if it’s completely random.

So why could Matt’s plan backfire?

Matt’s basically putting himself on a partial reinforcement schedule. He’ll drink a cup of coffee in the morning, and sometimes he’ll receive caffeine, and sometimes he won’t; and the amount of caffeine he receives will vary.

This plan could make him drink more coffee, in the end.

The only way to oppose that – which Matt is also doing, to some extent – is to have a negative reaction to drinking coffee. He has a “generally negative” view already (which is why he’s embarked on this plan), and is buying a book to convince himself of the evils of coffee.

Still, I don’t think the negative steps Matt is taking are enough.

There are also, additionally, other problems – e.g. his wife is randomizing the proportions. The way she randomizes the proportions will be very important, and could have a significant effect on the results (she should use something like Excel to generate really random proportions, and not pseudo-random).

Overall, he’d probably be better switching to decaffeinated outright, and after a long time (6 months? However long to kill, or at least diminish, the learned behavior that drinking coffee makes you feel good) stopping entirely. Even better would be having his wife give him just decaf for this entire experiment, and not tell him.