Close this window
Study notes icon

Study Notes: Variability, Accuracy and Precision

If we could measure with perfect accuracy down to the molecular level we would be able to determine, for example, the exact concentration of glucose for the sample in position 3 in the analyser printout below, and this figure might look like this - 29.975754302183654000% w/v.

This perfect value is called the True Value. Needless to say, the true value is not measurable as we are unable to measure down to the molecular level for many reasons.

We strive for measurement of the Accepted Value, which approximates the true value and is accurate enough for our purposes. In this case the machine measured 29.98% w/v, which is more than accurate enough for our purposes, as we are only required to give the answer to one decimal place.

If the Accepted Value is close enough to the True Value for our purposes then we say that the answer is Accurate (just like an archer hitting the bullseye). We check accuracy in a run by using Check Standards that have a verified concentration. We then compare the measurement of the check standard with their verified value (accepted value) and if close enough, the accuracy of the run has been established.

1. 0.01 BLK
2. 2.04 Low
3. 29.98
4. 31.33
5. 30.62
6. 30.54
7. 31.18
8. 19.87 Med
9. 52.15
10. 51.95
11. 50.88
12. 51.07
13. 51.54
14. 50.54 High
15. 45.49 SLPC
16. 45.46 SLPC
17. 45.59 SLPC

But what about precision? Each time a sample is measured there will be subtle variations in the result. For example, look at the results for the Precision Checks in positions 15 to 17.

Notice that although it was the same sample, the same run, the same machine and the same operator, there are small differences in the values obtained each time the check was measured. This range of values reflects the Precision of the run. The larger the range of values the lower the precision and the smaller the range of values the higher the precision.


To recap:

Accuracy reflects how close we have come to measuring the true value. Accuracy might also be thought of as correctness.

Precision shows how close a number of replicate measures are to each other, in other words how the individual values are spread or grouped.

Resources and Training Room  >>  Study Notes  >>  Variability, Accuracy and Precision
Close this window