Aug 12 2008

How Confident Are You?

Published by under Uncategorized
Comments: 14

Confidence in our beliefs is just another neurological function. We all have knowledge and we mentally assign a level of confidence in each piece of knowledge we think we have. Sometimes the assignation of confidence level is conscious – we may be asked how sure we are about something. And sometimes it is not conscious but it affects our behavior.

Often different individuals may even confront each other on their relative levels of confidence, when they each believe contradictory information. This conflict is sometimes encapsulated in the phrase, “You wanna bet?”

For the critical thinker it is important to avoid unjustified confidence, and to understand what factors may lead us to a high degree of confidence when we are in fact wrong. For example, we often assign a high degree of confidence to memories that are vivid and accessible, but the evidence shows that these factors do not predict accuracy. A memory may feel completely real, but be completely wrong.

Skeptics often encounter the “confidence” defense – we are asked to believe anecdotal accounts based upon the reporter’s confidence in their accuracy. I have been told many times by believers that they were absolutely certain about specific facts of a case that turned out to be demonstrably wrong. In these cases it seemed to me that confidence derived from a desire to believe, and therefore have a clean story in support of the belief. But confidence is an effective form of persuasion – we tend to trust other people’s confidence, even when we shouldn’t.

Science has largely quantified confidence – that is, after all, was science is largely about: having a high degree of confidence in our explanations about the natural world.

Neuroscientists are now looking at confidence as a specific neurological function. The work is still preliminary, but appears to be bearing fruit. A new study published in Nature in entitled: “Neural correlates, computation and behavioural impact of decision confidence.” They used a rat model of confidence in which they exposed rats to two different odors and if they chose the stronger of the two they were given a reward. The two odors either varied greatly in strength, in which case the distinction was easy, and it is assumed that the rat’s confidence in their choice was therefore high; or the odors were not very different, making the choice difficult and confidence low.

What they found was that activity in the orbitofrontal cortex correlated with uncertainty. This suggests that this part of the brain is involved in making us feel less confident (or more cautious) about our conclusions. Of course, as the title of the paper admits, this is a correlational study. Causation in neuroscience is often a very complex question, and much more study will need to be done to put this correlation into a proper context.

They also showed that the rats in this study were more likely to abort a trial and start a new one when the odor difference was small, and their confidence uncertain. So their model of confidence did correlate with a measurable behavior as well.

These kinds of studies, along with psychological studies, have the potential to greatly inform our thinking about confidence.  What I am specifically interested in, as a skeptic, is the conscious and subconscious calculations that go into our assessment of confidence in what we think we know. If those calculations are flawed then they can result in unfounded confidence, which is a barrier to doubt, questioning, and investigation – all things critical to skepticism.

What “confidence heuristics” do we rely upon day to day? We already know that we trust our own memories more than we should. Also, we share in other people’s confidence – if they seem confident then we believe them. We also seem to hold a higher confidence in beliefs in which we are emotionally invested.

But I bet there are more subtle heuristics, or rules of thumb, that we evolved to follow in arriving at our confidence levels. This is an area of research with which I am not very familiar, and I am not sure how much there actually is. I plan to explore this further, but if anyone is familiar with such research please provide some links.

Meanwhile, it’s a good idea to at least question our own confidence – to ask why we are so confident in a particular belief, and if that justification is rational.

14 responses so far