|Home > News & Events > MITRE Publications > The MITRE Digest >|
Affective Cognition: Making Rational Decisions About Emotions
Editor's Note: This article and others on emerging technology research at MITRE and around the world can be found in the latest issue of our publication, Envision.
Bias in Decisions
Following the 9/11 attacks, Americans found themselves reevaluating their feelings about travel. Was air travel really as safe as they thought? Was it worth risking exposure to a terrorist act if it meant shorter travel times? Or did it make more sense to travel by car? The result of this outwardly rational analysis revealed a surprisingly large influence of emotion on decision making. Americans began flying less and driving more, although statistics show that flying would still be 10 times safer than driving even if terrorists were hijacking and crashing one passenger jet per week.
In fact, a study showed that an additional 1,595 Americans died in car crashes as a direct result of the switch from flying to driving. This is six times higher than the total number of people on board the doomed flights of 9/11.
While we like to think that we conduct the bulk of our decision making through rational analysis, research has shown that when we're making decisions, our emotions often overrule our logic, particularly when evaluating and responding to perceived risk. This means that when the goal is to influence behavior—as is the case with many governmental communication efforts—the best strategy is to engage both the emotional and rational parts of the decision-making process. Understanding the interaction between our emotional and analytical processes is the goal of the emerging science of affective cognition.
The Decisive Brain
Research suggests that there may be physiological underpinnings for this interaction between emotion- and logic-driven decision making. When confronted with a need to make a decision, two systems are engaged. The first, referred to by psychologists as "S1," rapidly evaluates a large amount of sensory data to produce an intuition or "gut feeling" about the situation. The second, called "S2," carries out a slower, evaluative analysis over a subset of the data.
These systems are supported by different parts of the brain. S1 is associated with activity in parts of the brain linked with emotional processing, like the insula (which seems to be involved in monitoring overall feelings and emotions) and the anterior cingulate cortex (which is specifically sensitive to social emotions). S2 is associated with activity in frontoparietal regions of the brain that deal with reasoning processes and with processes that require the action of working memory—the short-term memory system that allows you to briefly remember and compare important concepts.
Although there is physiological evidence that emotions strongly influence decision making and psychological evidence—like the behavior of travelers after 9/11—that rational analysis is not the sole driver for decisions, many communication tools still emphasize an appeal to the rational part of the decision-making process. In order to improve these tools, it is important to understand the nature of the emotional contribution to decisions.
The Trajectory of Emotion
Researchers frame emotional experience in three dimensions: salience (surprise), valence (value), and dominance (control). Preliminary studies suggest that emotional experiences are not simply states but rather trajectories through the space defined by these dimensions.
For example, a surprising event is arousing (salience) and at first feels negative (valence). This is because you failed to predict the event, and failure to predict is generally bad for survival. This emotional response by S1 will brighten if further cognitive appraisal by S2 concludes the surprise is a welcome one. But if analysis reveals an unwelcome surprise, a negative emotion will result: sadness if the surprise is small, fear if the surprise is large. Cognitive appraisal then continues as you assess your ability to deal with the situation that produced the negative surprise.
Based on this framework of salience, valence, and dominance, MITRE researcher Kevin Burns is developing formal models of affective cognition. These models would be useful in designing tools or techniques that present information for human consumption. For instance, accurate models of the relationship between emotions and deliberative analyses might allow MITRE sponsors to craft messages that appeal to both decision processes.
As an example, engineers are developing "sentiment analysis" systems. These systems can compute statistics on word occurrence frequencies in surveys, letters, emails, and other documents, in order to classify their valence. But such systems will be of limited use unless they also address salience and dominance; a negative sentiment could imply fear or anger, which, as we've seen, can produce opposite behaviors depending on the appraisal of dominance.
Further research with formal models is needed so that future systems can usefully detect sentiment and accurately predict behavior. Advanced systems might then adapt the substance of messages, like medical test results and government policies, to the styles of consumers—achieving desired effects by addressing affective desires.
—by Julia High
Articles and News
Technical Papers and Presentations
Page last updated: March 10, 2011 | Top of page
Solutions That Make a Difference.®