The Law of Small Numbers
Hi,
This week’s email is about a common bias that causes us to overestimate how representative a few limited examples are of the bigger picture.
The information here is from my article on the topic.
Here are the key practical points you should know:
The law of small numbers (LOSN) is the incorrect belief that small samples are likely to be highly representative of the populations they come from (similarly to large samples).
For example, the LOSN could cause someone to assume that the way one person behaves necessarily represents the way everyone from that person's country behaves.
This bias revolves around people's expectation that the characteristics of a parent population will be represented locally in all of its parts.
This bias can influence both people’s perception of how representative current samples are and their prediction of what samples will look like.
To reduce this bias in yourself and others, you should identify situations where it might occur, question related reasoning (e.g., "is this sample big enough to be representative?"), illustrate the issues with it using examples, and use general debiasing techniques (like slowing down the reasoning process).
I also published a short paper about this if you’re interested.
As always, I'm happy to hear your thoughts.
Have a great week,
Itamar