I was listening to a podcast recently where the hosts were discussing the idea of little experiments one conducts in their daily life.
For instance, say that I want to lose weight. I might spend a month doing the keto diet, and then another month eating vegan and another month just staying under a calorie limit, all the while tracking my progress through measurements like waist circumference and body weight. The idea is I could compare the different months to see what was most effective.
If you’ve ever done these kinds of experiments, then you know it is very easy to trick yourself into thinking that the insights you gleaned from your experiments, while really fascinating, only apply to you. In statistics, the idea of n is that it represents the number of samples/subjects being analyzed in a study. They claim, that in order to ensure the validity of a study, that n should be 30 (though some textbooks say 20 or 50). That if you aren’t looking at least 30 or so samples, you may not have a good representation of the whole population, and therefore your results might be off.
But if you’re experimenting on yourself, that’s a n of 1. And the beauty of that is that while you can’t claim truth for everyone in the world (think about fad dieters), you can at least gain insight into what works for you.
When you’ve got n=1, correlation is all that matters. n=1 is 100% successful at uncovering information about how your brain works. And really, what else do you need?