Keith E. Stanovich, an emeritus professor of applied psychology and human development at the University of Toronto, answers:

Decades of research have shown that humans are so-called cognitive misers. When we approach a problem, our natural default is to tap the least tiring cognitive process. Typically this is what psychologists call type 1 thinking, famously described by Nobel Prize–winning psychologist Daniel Kahneman as automatic, intuitive processes that are not very strenuous.

This is in contrast to type 2 thinking, which is slower and involves processing more cues in the environment. Defaulting to type 1 makes evolutionary sense: if we can solve a problem more simply, we can bank extra mental capacity for completing other tasks. A problem arises, however, when the simple cues available are either insufficient or vastly inferior to the more complex cues at hand.

Exactly this kind of conflict can occur when someone chooses to believe a personal opinion over scientific evidence or statistics. When we evaluate a personal opinion, we automatically engage the evolutionarily old regions of the brain, which encourage social interaction and peer bonding. But understanding scientific evidence, a more recent achievement, involves more complex, logical and difficult type 2 processing.

From this dual-processing perspective, we can see several ways in which personal opinion might trump scientific thinking. First, some people may not have learned the rules of scientific thinking. In such cases, type 1 processing will be their default setting. And even if we can evaluate concrete evidence, our tendency to revert to type 1 processing may still lead us astray, ignoring logical reasoning in the face of an emotionally persuasive personal opinion. In other words, even when scientific thinking is compelling, our propensity to be a cognitive miser and conserve mental energy often prevents us from engaging type 2 processes.

The good news is that it is possible to override our tendency toward type 1 processing. To do so, we must practice scientific and statistical thinking to the point of automaticity, eventually making it our go-to option.