Can Eye Tracking be Used to Predict Strategic Behavior?

Life is full of choices, all of which have payoffs and pitfalls. When you have a goal and make choices or decisions on how to achieve it, you are developing a strategy. Creating and executing a strategy is a highly cognitive process; your mind is looking at available options and evaluating the pros and cons of each.

The way in which people go about developing strategies is a subject that many researchers have investigated. Game theory takes a look at how people behave during strategic situations in which their success is dependent on the choices of another person.

Within this construct, researchers can observe behaviors and attempt to predict how the players will act as they develop strategies to maximize their own personal payoffs.

A research group from Stanford recently used eye tracking to investigate its effectiveness for studying information acquisition behavior during strategic game scenarios.

In the study, researchers hypothesized that the frequency of fixations on relevant information relates to how that information influences choices.

Subjects played 10 rounds of 4 different “normal form” games in which the choices and payoffs for each player are presented in a matrix and each player has to make their best decision based on their potential payoff and the predicted action of their opponent. These games are called “learning games” as the players adjust their strategy each round based on what they learn about their opponent’s strategy.

Eye tracking data was recorded using head-mounted EyeLink II systems as 12 of the subjects completed the games. The number of fixations on the payoff regions of the matrices was the primary data collected with the eye tracker because it was thought to correlate with the players’ decision-making process. They found that the number of times the subjects looked up, the payoff values decreased as they played because they began to learn and memorize the values for both themselves and their opponent.

While it’s interesting news, it wasn’t what they were necessarily hoping to discover. They wanted to see if collecting eye tracking data was a viable method for developing predictive algorithms of behavior during strategic situations. And as it turns out, it’s not. Fixations on payoffs weren’t predictive of players’ actions; concentrated gaze on a payoff square didn’t necessarily mean it was chosen.

As is common in studies which evaluate eye tracking effectiveness for analyzing or predicting human behavior, it was found that eye tracking data is best used in combination with other techniques like think-out-loud methods, tracking mouse behavior, or simply behavioral observation by video surveillance to get a good approximation of true cognitive processes.

STUDYING LEARNING IN GAMES USING EYE-TRACKING

Related articles:

1. Eye Tracking: Social Behavior and How We Look at Faces
2. The Latest in Eye Tracking Web Usability Research pt2
3. Eye Controlled Video Games? Better Late Than Never
4. Eye Tracking: Investigating What Women Notice in Advertisements
5. Eye Tracking Looks at How We Rely on Google for Answers
6. Eye Tracking: Facebook and LinkedIn Usability
7. Eye Tracking: Giving Buyers a Basis for Comparison
8. Eye-Com Eye Tracking Drive Simulator Studies Driver Inattention
9. Neuromarketing Eye Tracking Helps Campbell’s Soup Get a Makeover
10. The Latest in Eye Tracking Web Usability Research pt1