Precision and Accuracy

[dyspareunia]

---

Members don't see this ad.

Kaplan says that precision is a necessary, but not sufficient condition for accuracy. This makes sense intuitively, but it is contradictory to what is in FA (i.e. accuracy without precision).

Precision = all bullets hit the target in roughly the same area, but not in the bullseye.
Accuracy = all bullets are around the edges of the bullseye, but not in a tight cluster and not directly in the bullseye.
Precision plus Accuracy = all bullets in the bullseye.

Any input?

 

[SBR249]

I think First Aid has it right.

You do not need precision to be accurate, your measurements are accurate as long as the average of the sample measurements are close to the true value. However If the standard deviation or variance of those measurements are large, then your measurements are imprecise.

This graphic from wikipedia is a very good illustration

Source: http://upload.wikimedia.org/wikipedia/commons/3/38/Accuracy_and_precision.svg

 

[AlejMZ]

Well from my perspective (BS degree in Industrial Engineering and current MS2), statistically/clinically relevant-wise just think of accuracy as a device giving us the ¨correct measurements ¨, where as the precision of a device gives us the ¨ same constant measurements ¨ (regardless if they are correct or not, meaning you can have precision without accuracy), I have been doing quite well in this topic just using that simple concept and not try to get overwhelm with complex explanations... Hope it helps