Do the Thing You Want to Do 2nd, Not 1st.

As we’ve discussed previously, the ABC framework states that behavior (B) is caused by its antecedents (A) and its consequences (C). Of these two, consequences typically have the biggest impact on behavior. We tend to repeat behaviors that produce reinforcers and stop behaving in ways that produce punishers.

We have also discussed how using reinforcement tends to be a better approach for changing behavior than using punishment. Although punishment technically “works”—by definition, punishers decrease the behavior that produced them—it also comes with a number of unwanted side effects: aggression toward and avoidance of the person who might be delivering the punishment, for example.

So, for that reason (and others), behavior analysts tend to prefer interventions that use positive reinforcement to change behavior.

We at Gale believe that your behavior-change program should center around positive reinforcement as well.

When you engage in a desired behavior, positively reinforce it as soon as possible!

But one potential problem with using positive reinforcement (or punishment) is identifying what will be reinforcing ahead of time. Remember, we determine whether something is reinforcement not by what we think it should be, but rather by how it actually affects behavior: If it doesn’t maintain or increase behavior, then it’s not reinforcement. Simple as that.

So how can you determine ahead of time what will be reinforcing to you? Will a positive comment do the trick? What about a new pair of shoes? A nice meal? A warm hug from a loved one?

Over the years, behavior analysts have taken steps to determine what will be reinforcing before they begin an intervention.

One approach has been the “preference task” in which people are given an array of options and asked which option they would prefer. The assumption is that whichever options are chosen will likely be reinforcing for the person who made the choice.

Another approach to solving the “what-is-reinforcing” conundrum comes from a psychologist named David Premack, who attempted to determine exactly what makes something reinforcing.

After conducting a number of experiments in the 1950s, Premack concluded that reinforcement occurs when a higher-probability behavior—or, loosely speaking, a “more desirable” behavior—follows a lower-probability, or less desirable, behavior.

So what does this mean?

Imagine you have two behaviors that you can engage in. You can either eat your vegetables or eat a tasty piece of cake. You don’t really enjoy eating vegetables, but you love eating cake, and if given the opportunity to freely pick one or the other, you’ll almost always choose to eat cake.

So, in this instance, eating vegetables is the lower-probability behavior, and eating cake is the higher-probability behavior.

Knowing this, we can now set up a “reinforcement contingency” (a predictable relation between a behavior and a consequence), where we can use cake-eating to reinforce vegetable-eating. In other words, you must eat your vegetables before you eat your cake (hence the reason this idea is sometimes called “Grandma’s Rule”).

If we always introduce this reinforcement contingency, we should find that the lower-probability behavior of eating vegetables increases because the higher-probability behavior of eating cake follows and reinforces it.

Let’s look at one more example. Imagine you have the choice between watching TV, which you do a lot, and going for a 30-minute walk, which you don’t do as much as you probably should. In this case, walking is the lower-probability behavior and watching TV the higher-probability behavior.

We could therefore set up a contingency where you must walk for 30 minutes before you watch TV. If you continue to do this, we should find that walking increases, because watching TV follows and reinforces it.

So, according to Premack, why does this increase in the lower-probability behavior occur? 

It’s because engaging in the lower-probability behavior produces access to the higher-probability behavior. As long as watching TV is higher probability and becomes available only after you walk, you’ll increase your walking to whatever amount is required to watch TV.

How, then, can you capitalize on the so-called “Premack principle” when you’re trying to change your behavior?

Simple.

Think of a behavior you want to increase. This is the lower-probability (or less desirable) behavior. Next, think of a behavior that is higher probability (or more desirable) than the first behavior.

Now set up a contingency whereby you must engage in the lower-probability behavior before you can engage in the higher-probability behavior.

Eat your vegetables before you eat your cake. Go for a walk before you watch TV. Write 200 words before you take a shower. Drink 2 cups of water before you have your morning coffee. Finish that expense report before attending social hour with friends.

If you want to apply the Premack principle even more broadly, start each day by taking the tasks you need to accomplish and, then, as much as possible, order your day from lower-probability tasks to higher-probability tasks.

What you’ll find is that each subsequent task is more desirable than the last and that you’ll end your day doing things that are highly probable and most likely very enjoyable.

Finally, a word of caution. In his research, Premack found that the probability of tasks could change. In other words, what was initially reinforcing may not be after some time. For example, if you’ve been using TV watching to reinforce exercising, you may eventually find that exercising becomes more probable (and possibly more enjoyable) than watching TV. In this case, if you want to further reinforce exercising, you may need to identify a different higher-probability behavior that you can use as a reward.

Ultimately, the Premack principle is a useful way to think about what might be reinforcing ahead of time. And its consistent application provides a relatively easy way to apply positive reinforcement in a behavior-change program.