This is when the mean of the data collected is quantitatively close to the true value, otherwise known as the reference. There can be a fair amount of variance in a data set and the instrument could still be called accurate.

There are three components to accuracy: Bias, Linearity, and Stability.