“If I go long enough without getting a compliment, I compliment myself, and that's just as good because at least then I know it's sincere.” – Ruth Gordon

I've been writing quite a bit recently about cultivating good habits. (Previously I wrote about balancing your experiencing and remembering selves and crafting personal policies.) In this post I'll talk a little bit about the benefits and dangers of rewards and self-reinforcement.

Ice Cream Sandwich by ralphandjenny on flickr

Rewards are powerful motivators and we should use every tool available to us in the quest to become better versions of ourselves. So how about rewarding yourself for going to the gym? After a good workout it's tempting to indulge yourself with a rich meal or an ice cream sandwich. Won't the positive reinforcement make you want to go to the gym more often?

Setting aside the obvious caloric difficulties, this kind of policy can lead to problems. You might end up going to the gym more often, at least in the short term, but only at the expense of your experiencing self. The pure joy of exercise will fade and you’ll find you’re only doing it for the eventual reward. Besides, if you're only going to the gym to justify eating ice cream, intuitively it feels like you're doing it for the wrong reasons.

Indeed, ongoing research since the 1970's has shown that rewards can undermine existing motivations, even after the reward stimulus is removed.

Intrinsic Motivation and the Overjustification Effect

Alfie Kohn argues in Punished by Rewards: The Trouble with Gold Stars, Incentive Plans, A’s, Praise, and Other Bribes that external incentives have the power to sap intrinsic motivation for an activity. Take a child who loves playing piano and give her a reward for practicing, and you may end up with a child whose enthusiasm for music is permanently dampened.

This is known as the overjustification effect and, though counterintuitive, explains why the things we expect to be motivating, like salaries and bonuses, are often ineffective or even harmful to work performance.

Experimental psychologists Mark Lepper and Richard Nisbett demonstrated this in the early 1970's with the following experiment: they divided a population of young children into three groups. They told Group A that they would get a ribbon as a reward for drawing. They rewarded Group B in the same way, but didn't tell them about the ribbons ahead of time. Group C was simply allowed to draw for the same period of time and was not rewarded. Later on, during free play time, children from Group A spent significantly less time drawing than children from the other groups.

For the Group A kids, drawing changed from an enjoyable activity into an opportunity to receive praise. They interpreted (or justified) their own drawing behavior as happening because of the reward.

This effect has been replicated across a wide range of age groups and activities.

Notably, overjustification happens most strongly with larger rewards, and almost not at all (or even in reverse) with smaller ones: the former provides a plausible story for justification, while the latter leaves experimental subjects reinterpreting even boring activities as fun or interesting.

Effective Self-Reinforcement

Despite the dangers, I do recommend experimenting with self-reinforcement. A great reference for DIY reinforcement training is Karen Pryor's Don’t Shoot the Dog!: The New Art of Teaching and Training. (Even if you don't have time to read the whole book, take a few minutes to look through this excellent set of notes.) Pryor suggests that most of us are severely under-reinforced for the habits we'd like to have:

“[Self-reinforcement] is something we often neglect to do, partly because it doesn't occur to us, and partly because we tend to demand a lot more of ourselves than we would of others.… As a result we often go for days at a time without letup, going from task to task to task unnoticed and unthanked even by ourselves. Quite aside from reinforcing oneself for some habit change or new skill, a certain amount of reinforcement is desirable just for surviving daily life.”

Here are some guidelines for effective reinforcement gleaned from Don't Shoot the Dog! and my own experimentation:

  • DO use small rewards (eat an M&M, smile to yourself, do a power pose, give yourself a compliment, or step out into the sunshine).
  • DO use reinforcement for tasks that aren't fun in themselves.
  • Once you establish a consistent routine, try flipping a coin and only reward on heads (this is a very common technique in game design and might be familiar if you've ever been to a casino).
  • Keep records and make graphs so that improvement can be seen at a glance and lapses are put into context.
  • Reinforcements are most effective when they immediately follow the behavior you are trying to shape.
  • The entire context is reinforced, not just the habit you are shaping: rewards can change the way you feel about the people you spend time with and the places you frequent, so be careful.
  • DON'T punish yourself. Not only is it unpleasant, it's surprisingly ineffective.

Before you set up an incentive scheme to “get yourself” to do something, think about the behavior you’re really reinforcing and think about whether you are willing to sacrifice the joy of the experience itself.

Also, be aware of the inherent limitations of self-reinforcement. The most effective reinforcements are those that convey information, and self-administered rewards don't carry any new information about your performance.

donut_8101 by mamojo on flickr

Why all this talk about rewards and habits on a game blog? A big part of my motivation for producing Double Dynamo in the first place was curiosity: what makes people tick? how can I work most effectively? It's certainly been rewarding working on my own habits as the project has gone from conception to implementation, and very soon, to launch.


Photo credits:


Curious about Double Dynamo? Watch the trailer.


comments powered by