The 2-Minute Thought: A Reminder that a Little Knowledge is a Dangerous Thing

It was Alexander Pope who said, “A little learning is a dangerous thing,” and according to Carmen Sanchez and David Dunning in the Harvard Business Review, it’s true.

It’s not complete beginners that need to worry.   According to Sanchez and Dunning, when people are entirely new to a task, they stay alert and cautious.  The problem is when they gain just a little experience.  That is when overconfidence soars, and “unconscious incompetence is . . . something they grow into.”

Sanchez and Dunning cite research that doctors who are learning to do spinal surgery do not usually make errors until their 15th operation.   Likewise, beginning pilots have relatively few accidents, but then accident rates start rising before peaking at 800 flight hours and subsequently dropping off again.

What happens is that learning a new task requires taking in a lot of information – a hard job even for experts.  Because there is a lot of noise in the information we process, it’s very easy to misread cues.  What relative beginners don’t realize is that it takes a large amount of data to be able to sift noise from signal.  That is what Sanchez and Dunning call the “Beginner’s Bubble.”   

The good news is that after more mistakes are made, more experience is acquired, and more data is processed, people start to understand their errors and self-correct.  But then the problem becomes that confidence starts to rise disproportionately a second time. 

Is it possible we go through never-ending cycles of overconfidence and self-correction?

Sanchez and Dunning say that the second peak in overconfidence appears pretty consistently in their research, but they still have much to learn about how confidence levels change after multiple experiences.  That is, they say, “we cannot be certain what would happen to overconfidence after the 60th trial.”

Still, their work suggests that we should be wary of overconfidence regardless of age or experience level – and whether we’re engaging in medical diagnosis, investment, or any other endeavor. 

One real-life example the authors identify is financial literacy.  It’s long been known that people significantly overestimate their financial knowledge.  It’s also long been a puzzle that teaching financial literacy hasn’t proven to result in better financial decision-making.  Some have questioned teaching financial literacy classes at all. 

And yet real financial literacy does improve slowly and incrementally across age groups -- it seems with experience.  Sanchez and Dunning note that self-confidence rises quickly between late adolescence and young adulthood and then levels off until late adulthood, when it begins to rise again. 

For investors, where calibrating confidence to knowledge is key to success, the takeaway may well be that you can get better with age – but slowly and incrementally.

To get a sense of how good you are at knowing what you don’t know, take investor Michael Mauboussin’s test on confidence calibration here (  The objective isn’t to get as many correct answers as you can.  It is to estimate what you don’t know as well as possible.  On average, people think they can answer 70% of the questions correctly, but in reality get only 60% right.