I took statistics for the first time just a few years ago, and it seemed quite difficult. There weren’t right answers; there were confidence intervals. Equations had to be memorized and I didn’t understand why they worked. And instead of just math, there was lots of information like designing experiments and what sort of data is better than other sorts.

But the class was designed quite well, and I got through it. Then I took econometrics (which is statistics for economics), which turned out to be more of a coding class, since there is no reason to do statistics by hand if you can have a computer program figure it all out for you.

Statistics is the backbone of science, really. (And whether or not that is a good thing is debatable.) It’s not entirely useful for everyday living, except to understand what scientists are doing. It is necessary to learn and understand if a person wants to complete graduate-level projects.

I think one of the better ways of learning statistics is to learn about basic concepts and how they are useful (confidence intervals and P values and statistical significance). Then, you learn how to code (using R or similar), in order to avoid doing the math.

But then the only way to really learn statistics is to do statistics: to have a data set and to want conclusions from that data set, and then to figure out how to wrangle those numbers.

It’s a tiny bit like you can’t really learn how to parent until you are parenting. And you can’t really learn how to do statistics until you are doing statistics.