Many years ago, I was flipping through stations on the radio and came across a talk show. I don’t remember the topic of discussion, but something the host said stuck with me.

“Scientists,” the host blustered (and I paraphrase), “are 95-percent confident of these results! Wow!”

The sound you’re now hearing is that of mathematicians tearing out their hair.

Sure, there’s the old saw about lies and statistics. But would it hurt pundits to know a little bit about statistics before they misuse them?

No two real measurements are exactly the same, whether they are from a telephone poll of political opinions or a set of apparent magnitudes of variable stars. This begs the question: How does one tell which measurement is correct?

It’s pretty rare—I won’t say impossible—to know the “correct” answer ahead of time. What scientists and mathematicians do instead is to take many measurements (not just two), find an average result, and estimate the precision of the result.

 …