3.4
Accuracy
Accuracy is a property of a single measurement result and is influenced by both random and systematic
errors, it is therefore not included as such in the method validation itself. The accuracy of a measurement
result describes how close the result is to its true value and therefore includes the effect of both precision
and trueness (expressed in the form of bias).
Precision relates to the repeatability and / or reproducibility condition of the measurement “getting the same
measurement each time”. In other words the closeness of results of multiple analyses to each other.
Precision is therefore a measure of the spread of repeated measurement results and depends only on the
distribution of random errors – it gives no indication of how close those results are to the true value.
A measurement system can be accurate but not precise, precise but not accurate, neither, or both. For
example, if an experiment contains a systematic error, then increasing the sample size generally increases
precision but does not improve accuracy. Eliminating the systematic error improves accuracy but does not
change precision. However, the validity of a measurement system is dependent on ‘fitness for purpose’ and
hence the accuracy and precision need to be assessed against the validation criteria. There is no absolute
scale.
We can never make a perfect measurement. The best we can do is to come as close as possible within the
limitations of the measuring instruments.
The illustration below is used in an attempt to demonstrate the difference between the two terms, accuracy
and precision.
Suppose you are aiming at a target, trying to hit the bull's eye (the center of the target) with darts. Here are
some representative patterns of darts in the target:
A
B
C
A
Accurate and precise. The darts are tightly clustered and their average position is the center of the
bull's eye.
B
Precise, but not accurate. The darts are clustered together but did not hit the intended mark.
C
Accuracy and precision are poor. The darts are not clustered together and are not near the bull's
eye.
Accuracy is also used as a statistical measure of how well a binary classification test correctly identifies or
excludes a condition. That is, the accuracy is the proportion of true results (both true positives and true
negatives) in the population.
Condition
True
False
Positive
True positive
False positive
→
Positive
predictive value
Test
outcome
Negative
False negative
True negative
→
Negative
predictive value
↓
Do'stlaringiz bilan baham: |