A Bias for Confirming Evidence

One reason people try to restore their situations to match previous or expected situations, rather than to update their understanding of the new and unfolding situation, can be seen in an experiment conducted by Mark Snyder and William Swann. Snyder and Swann told the participants in their experiment that they would be interviewing people to find out if the other person was an introvert or an extrovert. On the surface this seems like a relatively easy task: introverts tend to be more distant and quiet, whereas extroverts tend to be outgoing and talkative. Introverts often get energized by a good book or a conversation with a single friend, whereas extroverts get energized by parties and large social gatherings. Traits such as these—it would seem—should be relatively easy to discern when conversing with another person.

Snyder and Swann found, however, that the ease with which the participants in their experiment discerned a person's introversion or extroversion depended significantly on the instructions they were given. For example, rather than ask participants to find out if a particular person was an introvert or an extrovert, Snyder and Swann told half of their participants that they needed to interview people to find out if they were introverts, and told the other half that they needed to interview people to find out if they were extroverts, but they gave these participants no data to indicate whether the people they were going to interview were introverts or extroverts. They then gave the participants a list of ten questions that largely made sense only to ask of people who are already known to be introverts, eleven questions that largely made sense to ask of people who were already known to be extroverts, and five neutral questions.

Snyder and Swann asked the participants to select twelve questions from the list for their interviews. On average, the people who were told to find out if the person they were interviewing was extroverted selected a high number of questions designed for people that are known to be extroverted and the people who were trying to find out if the person they were interviewing was introverted selected a high number of questions designed for people that are known to be introverted. These differences were significant, suggesting that people tend to look for evidence that confirms what they are looking for rather than look for disconfirming evidence, even when there is no data to suggest that a person should look for confirming evidence.

Snyder and Swann followed this first experiment with others. They had participants actually conduct the interviews rather than just pick the questions. They gave the participants data on how likely it was that the person they interviewed would be an introvert or extrovert so that the participants would know to look for disconfirming evidence. They even offered to pay participants money for making more accurate assessments. Even so, in every case, the simple instruction to find out if a person was an introvert or an extrovert led participants to look for—and sometimes to even create—evidence that confirmed the personality trait that had been mentioned in their instructions.

The tendency for people to look for evidence that confirms their expectations has been found in experiments of all kinds. For example, in one experiment, participants were given an article to read that presented evidence for the deterring effect that capital punishment has on violent crime and an article to read that presented evidence suggesting that capital punishment does not deter violent crime. Although both articles were written with equal rigor, people hailed the article that supported their beliefs as a highly competent piece of work and were hypercritical of the article that refuted their beliefs, looking for minor flaws and enlarging them.

In another experiment, scientists presented people with information about the current Republican and Democratic presidential candidates. As they read the information the scientists monitored responses in the participants' brains using functional magnetic resonance imaging. They found that when participants were presented with information that disconfirmed their expectations, the reasoning areas of the brain were hardly activated at all; in other words, people put no mental effort into understanding any evidence that disconfirmed their beliefs, dismissing it out of hand. In contrast, when people were presented with information that confirmed their expectations, the emotional areas of the brain were activated pleasantly; in other words, their brains rewarded them for finding confirming evidence. The tendency for people to seek confirming evidence crosses all political boundaries and all domains of life.

The human tendency to seek confirming evidence is an efficient thing for our brains to do. If evidence confirms our expectations, we can continue acting the way we have been acting, and we will not have to invest the energy that is necessary for thinking consciously about what we should do differently. Our brains need to manage our actions with as little conscious thought as they can, or they would be overwhelmed with energy demands and information overload. We therefore create simple expectations for our experiences and act as if those expectations are correct, sometimes even when we receive overwhelming evidence that they are not. Our unconscious minds cling tightly to our expectations, searching for and even creating data that confirms those expectations and ignoring or discounting data that does not.

The tendency to seek data that confirms our expectations explains why Mindy did not realize that her anger and her intentions were unproductive. Her son's potential expulsion confronted her with two options: she could fight it, or she could change her expectations. The tendency to seek confirming evidence suggests she was not likely to change her expectations. It is more comfortable to solve problems than it is to admit that we have inappropriate expectations. Because of this, Mindy said the organization's people and processes were unfair; she created arguments to back up her claims, and these arguments made it "obvious" to her that she should fight the leaders of the organization over her son's expulsion.