"Fixed mindsets" might be why we don't understand statistics



  • Article from Ars Technica on why people fail to understand statistics.

    Reading this, it feels to me that this really applies to IT. The article says that 96% of the population struggles to understand statistical math that we need for normal tasks. In IT, the amount that we need statistics is way, way higher than in an average career path. We need stats for failure rates, reliability, performance, cost, risk, you name it. To do IT well, we have to practically live statistics.

    But I see IT practitioners struggling with these things all of the time. Example, if you tell someone that something is 99% going to fail, and they produce an anecdote of it not failing, they actually think that that somehow disproves the point. Misunderstanding statistics can be so extreme that people think that 99% is the same as 100%, and it makes you wonder what "five nines" must mean to them.



  • One of the examples in the article is that in a jury case, someone needed to explain 10% as meaning "one in ten". I think people who really get stats do this for everything in their heads all of the time - take anything and put it into the most meaningful terms. And I think those same people would be surprised to find out that other people do not do this.