Quote:
Originally Posted by instarx
You got that backwards. Precision is when your readings are grouped around a small point. Accuracy is when your readings are near the true value.

I'd actually take the correction one step farther; I define it in my Physics class around how we actually use measurements in the real world:
Precision is the exactness of a measurement:
28.50 psi indicates a measurement made to the nearest 1/100th of a psi.
31 psi indicates a measurement to the nearest 1 psi.
The difference between these numbers is the amount of precision implied in the measurement.
Accuracy is the closeness to the actual correct value.
If the actual pressure in a tire is exactly 29.994 psi, the gauge reading 31 was more accurate, although it implied less precision.
The classic example of shots at a target and having grouped results isn't my favorite; I would consider getting the same results from a tool when measuring multiple times "consistent" rather than "precise".
The distinction with precision comes when you combine measurements in multiple calculations. If you know your tire loses 1 psi per month, that's a value precise to the nearest 1 psi. If you measured it as 28.943 psi last month, and you use your 1 psi/month figure, you can't say that your tire is now at 27.943 psi. You say it's at either 28 or 30 psi, depending on how exactly you use significant figures. In this case, I'd say 28 psi.
I recently heard about a traffic accident reconstruction expert from my wife (she was watching the trial) and the "expert" went through all his figures and said he determined with certainty that the vehicle had to be traveling 56.554 mph. I wish I was on the jury or on the opposing side of court; I'd have raised the BS flag on that one.