Teaching Behavior: Can Technology Succeed Where Humans Have Failed?

From the days of Math Blaster all the way through the current iteration of Khan Academy, the nexus of learning content and technology has generally been full of math, science, reading, and other subjects on every fourth grader's report card. Naturally, these also happen to be areas where teachers are most successful in using their words to directly build student knowledge. When a teacher explains the rules for multiplying fractions, their authority on the subject matter is generally unquestioned.

There are, however, plenty of educational domains where students are less likely to heed what their teachers say. Specifically, when it comes to teaching positive behaviors in areas like health, finance, and social interactions, students often have personal experiences give them an inkling that they know better. For example, advice from teachers like "Improving your social status is not a good reason to do drugs," is unlikely to alter a student's decision-making. The result is that the bar set by conventional teacher-student dialogue falls low enough that technology has the potential to make a major impact.

It's not just that humans are less effective when it comes to teaching behavior. There is also reason to believe that technological tools may be good at it. Though being emotionless and anonymous is generally a bad idea while teaching, these characteristics could be advantageous during behavior lessons that normally induce knee-jerk skepticism of authority figures.

Many students don't want to be lectured by elders about life choices, but they may be more accepting if those lessons come from someone or something with which they have no prior relationship. Similarly, teenagers often fear being publicly judged for their feelings or decisions. In such circumstances a machine can make a welcoming partner for thinking about emotions and social pressure.

Teaching positive behaviors may also be a good fit for technology because such lessons tend to focus on evaluating potential outcomes. Decisions about health, finance, and social behavior are all based on understanding the consequences of actions. Compared to class discussions or paper-and-pencil activities, computers are uniquely capable of generating an unlimited number of detailed simulations. Seeing the consequences of your actions in a computer simulation can not only be more appealing, it can also be more engaging and informative.

Research is beginning to demonstrate the promise technology holds for improving behavior. One innovation worth keeping an eye on is computerized facial simulation. For example, a 2011 study (pdf) led by NYU's Hal Hershfield found that allowing people to interact with aged renderings of themselves led them to allocate more money for the future. By establishing a stronger connection to a person’s future self, the computer rendering made it easier for people to see the value in saving money. Given the weak performance of financial literacy programs, these simulations could prove to be a valuable tool for enhancing lessons on financial decision-making.

In addition, a new study forthcoming in Computers in Human Behavior suggests that facial simulations can have a beneficial impact on attitudes about smoking. The study was conducted using a video game that allowed participants to learn facts about smoking by controlling an avatar based on a photo of themselves or another person. To illustrate the effects of smoking, some players who reached the game’s second level had their avatars visibly aged so that the avatars appeared more decrepit. The researchers found that players whose avatars aged reported attitudes about social smoking that were significantly more negative than players with avatars that didn't age. In addition, players whose aged avatars were of their own face perceived themselves to be more susceptible to the consequences of smoking than players whose avatars featured a random face.

Even when technology provides an inferior learning experience, the ability to provide that experience to large numbers of people can still make it a useful tool for teaching behavior. At the moment the biggest roadblock to learning better behavior is access, not quality. Many children simply never get any kind of formal instruction on important life lessons. A lot of attention was recently given to new research showing strong positive effects of cognitive behavioral therapy (CBT) on violence and achievement among Chicago school children. If a computer tool can provide CBT that's 25% as effective to a thousand times as many students, isn't that something worth getting excited about?

Thus far, research indicates that technological tools can be effective at ameliorating social and emotional problems. For example, a new study from a group of Austrian researchers found that online trainings using elements of cognitive behavioral therapy (CBT) can enhance social skills and reduce social phobias. Similarly, a 2010 study led Penn State’s Michelle Newman examined over 100 studies on technology-based treatments for anxiety or depression. Newman and her team concluded that these computerized interventions were a cost-effective way of delivering treatment. Were these programs as good as a human therapist? No. But their value is in the fact that they can reach people who will never have a human therapist.

Entrepreneurs and researchers are constantly developing new tools for teaching kids math and reading, but let's not forget that educational technologies may ultimately have their largest impact in areas where many students either don't listen to their teachers, or don't even have teachers to begin with. We’re still at the stage where words like "preliminary" and "pilot" seem to preface every finding, but someday educational technologies could play a key role in significantly improving human behavior and decision-making.