I’m tired of phrases like, “this is not normal” and, “let’s get back to normal.” Were things ever normal? For different people and different cultural memories times were better and worse than today, but the idea of a static, comfortable normality seems to me very much imaginary.
In statistics, the “normal distribution” is one where the median and the mean averages are identical, while the probability of a value falling above or below average descending the further it is from the average. It’s a common pattern in nature and society, but it’s by no means the only pattern. Income distributions in the United States are famously skewed, with most people earning less than the mean; daily vehicle traffic and birdsong follow the day-night cycle rather than peaking during the day. Using the normal distribution to estimate effects and test for differences can be useful in data analysis, but only if the distribution fits the data.
Something that statisticians and data scientists try to do is to move beyond describing observed data to making predictions about what is expected, based on what is observed. When done poorly (for instance, when normality is assumed to be standard for all situations), it can lead to bad predictions and hard-to-trust analysis. When done well, it can lead to better strategies and deeper understanding of what to expect. I wonder if, when talking outside the world of data, it could help to speak of what is expected and what is hoped for, rather than what is “normal.”
Emordnilap 