Third Helix is now Kickbomb Entertainment! Please visit kickbomb.us for all future updates.


13 Mar 2010

In my last session Thursday, Chris Hecker energetically explored the current state of knowledge about the psychological effects of achievement-type reward structures. He opened the talk with a plea to the academic community to do more work in this field, because the research is spotty at best and there's a severe lack of consensus. Hecker kept coming back to that lack of consenus throughout the talk, insisting to the audience that despite his personal opinions on the subject, the research doesn't currently lead to a specific conclusion.

He cited works by B.F. Skinner and Alfie Kohn, both of whom have written about the adverse effects of particular reward structures on motivation. Kohn in particular has railed against "pop behaviorism", which put simply is the idea of "do this, and you'll get that".

Hecker drew a distinction between intrinsic and extrinsic motivation. Intrinsic motivation comes from within: a desire to do a thing for the sake of the thing itself (Csikszentmihalyi terms this an "autotelic" task). Extrinsic motivation comes from without: a desire to do a thing for the sake of an expected, associated reward -- like an achievement!

Hecker's thesis was that "the research shows that in most cases, extrinsic motivators are ineffective and actually decrease intrinsic motivation on interesting tasks." He noted that intrinsic motivation appears to be superior to extrinsic motivation with respect to creativity, problem solving, quality and speed of work, and overall happiness.

He referenced a study where two groups of kindergarteners were told to draw a picture. One group was told they'd be rewarded with candy (and indeed, this occurred), while the other group was promised (and received) no reward. The next day the experiment was repeated. The group that had been rewarded the day before didn't draw anything the second day, but the other group did.

He cited another study in which Pizza Hut gave school children a free pizza for each book they read. When the pizzas stopped coming, the kids stopped reading. Hecker argued that what should have happened is the kids should've been given a book for each pizza they ate. ;)

He then spent some time discussing various classifications of rewards: tangible vs. verbal, expected vs. unexpected. He discussed "contingent" rewards: those that require engagement in or completion of a task. The vast majority of rewards in games are contingent on something.

He reminded the audience that the few bits of research in this field are still incomplete, and still in conflict. But there are two points, he says, that psychologists appear to generally agree on. First, that "for interesting tasks, tangible, expected, contingent rewards reduce free-choice intrinsic motivation". Second, that "for interesting tasks, verbal, unexpected, informational feedback increases free-choice and self-reported intrinsic motivation". Even just these two results, he says, are "huge".

It's worth distilling that a bit. Basically, the research appears to show that if you hand out achievements (an example of "tangible, expected, contingent rewards") for in-game behaviors, players will soon stop engaging in those behaviors unless you continue handing out achievements. Conversely, if you simply *acknowledge *player behaviors in a non-contingent way ("verbal, unexpected, informational feedback") players may actually become *more interested *in playing the game for it's own sake, achievements be damned.

Hecker invoked Jesse Schell's by-now-known-to-everybody DICE 2010 talk, and highlighted a response to that talk by Raph Koster: "We don't have any issues with rewards that are even more tangible, like giving your kid money for a straight-A report card... as long as we are comfortable with who is setting up the reward and what criteria." Hecker argued that the research suggests giving your kid money for a good report card -- a tangible, expected, contingent reward -- may actually make future report cards *worse *if you stop paying.

He highlighted another comment, this one by Mike Jackson of CVG UK: "It would be brilliant if you got, say, money credited to your PSN account for finishing games. You know, a real-life reward. That'd get you playing." Hecker responded fiercely: "Why are we making games if we're having to pay people to play them?!" The audience responded with uproarious applause.

Hecker insists, again, that while the current research *appears to suggest *that the comments he just highlighted are wrong, the evidence is not yet conclusive. He argues that we *need *to study this in more detail, to keep our medium healthy.

He concluded with two major points. First, that "the industry needs to start studying the long term impact of achievements on players." He noted that Microsoft's usability research group is in a fantastic position to do this, and laments that said research is, apparently, not being done. Second, that "developers need to be better versed in the literature and more thoughtful about the consequences of extrinsic motivators."

While the talk was brilliant, energized, and provocative from start to finish, I think Hecker actually delivered his best line during the Q&A. Asked whether we should employ a reverse-psychology approach of awarding achievements for dull tasks that we don't want players to perform, Hecker noted that this approach would likely still erode intrinsic motivation by definition and is thus undesirable. He then posed a question of his own: "Are we disincenting people from playing games just for the beauty of play?"

Thought-provoking material indeed.

Posted In:

game-design gdc video-games