Error and Treatment of Data

A chemical analysis is only as good as the numbers that go into calculating and expressing the result. Therefore, data analysis is a crucial aspect of chemical analysis. All analytical measurements are subject to greater or lesser degrees of error. So every reasonable effort is made to reduce the amount of error in an analytical measurement. Since some error is inevitable, it is important to know the degree of error and express it correctly in the final result. 

However, just because the results of a set of analyses are in close agreement does not necessarily mean that the values are correct. This is because of the possibility of determinate errors. Such errors have a definite cause (although it may be unknown to the analyst), and each type of determinate error is always in the same direction. For example, if an analytical chanist is using a pipette rated to deliver 25.00 mL of solution that through a manufacturing mistake actually delivered 25.35 mL, a determinate error would be introduced into the analysis in this case, it could readily be detected by calibrating the pipette. 

FIGURE 18.3 Illustration of precision and accuracy in chemical analysis. 

In doing analytical calculations and expressing analytical results, it is important to know and correctly handle and express the uncertainties of the numbers used. For example, a skilled analyst can read the volume delivered by a laboratory burette to the nearest 0.01 mL. Therefore, a volume expressed as 36.27 mL implies that the volume is within 0.01 mL of 36.27 mL and has an uncertainty of 0.01 mL. It would be incorrect to express the volume as 36.270 mL because it is not known to the nearest 0.001 mL as that number would imply. It would also be incorrect to express the volume as 36.3 mL because the value is known more accurately than 0.1 mL. 

---