Statistics 101

Before do the math, here’s things you should consider to avoid bias:

• The magnitude of axis. $\log$? 2 quantities use separate axises? Dramatic?
• Duration of time. “Every Olympic is the best Olympic in the last 3 years.”
• Demography. Age, wealth, gender, race, etc. Does it cover all the target demography?
• From what media or institution. Different media have different biases themselves. (Although this statement is also a kind of bias.) For a more detailed research, see Machine-Learning media bias.

Here’s some examples of bad (or maybe good in some sense?) axis:

Also, Simpson’s paradox is interesting, see 辛普森悖论.