4 ms·
I think a "statistics" class could still help a lot of people without ever touching continuous distributions (which as you mention are more naturally built off
by mycologos 3y ago
I think a "statistics" class could still help a lot of people without ever touching continuous distributions (which as you mention are more naturally built off calculus) or multivariate distributions (which as you mention are more naturally built off linear algebra). Simple stuff like basic conditional probability, Bayes' rule, confidence intervals, basic hypothesis testing, and general "statistical thinking" don't really require calculus or linear algebra. Just getting people to a point where they understand some of the really basic tools seems pretty valuable, and pretty far from the current state of things.
- lahvak 3y agoA mandatory nitpick: you cannot do conditional probability without multivariate distributions, but you are right, you do not need linear algebra for a basic introduction. Even continuous distribution can be introduced, you just start with discrete cases, introduce areas in a histogram, and then approximate the histogram by a continuous curve. You can even get to the Riemann integral that way if you want to. And nobody actually calculates normal probabilities by integrating the Gaussian function, it's kind of hard to do.