Click image above to embiggen.

From https://chrisholland.ch/correlations-causations-and-coronavirus/.

# Regression visualization widgets

These may or may not work on your computer, but they’re very handy for wrapping your head around what we’re doing when we find the regression line/least squares line/line of best fit.

Here’s an applet for seeing what the line of best fit is doing.

And here’s an interactive explanation of least squares.

# Correlations real and spurious

Take a look at this graphic: what your favorite drink says about your politics in one chart. Then take a look at these amusing and plainly spurious correlations.

# Play “Guess the Correlation”!

# P values and more!

Here are some resources relevant to this week’s discussion of chi square and p values. First, this generator shows the shape of Chi Square sampling distributions at different degrees of freedom. Next, a dandy p value calculator for the tests we’ve learned this semester. (Also see this Xtranormal animation about statistical power and p-values. Xtranormal was the coolest thing on the internet for a month back in 2011.)

# Gender and naming colors

# A timely graph

# Fisher, Bigfoot, and the null hypothesis

Here’s a reprint of a short 2014 *Nautilus* article by Carl Zimmer in which he describes how Ronald Fisher and some colleagues came up with the idea of testing null hypotheses in the 1920s. The story begins with pouring milk in tea, and one upshot is that we can’t definitively rule out the existence of Bigfoot.

# Dear Data

While you are sheltering in place, take a look at this charming project, Dear Data, which I love, and which gives you some idea of just how much you can clearly communicate visually if you’re clever about it.

# Type I vs type II errors

Here’s a nice discussion of type I and type II errors in statistics and in the justice system.