The accuracy of an analysis is a measure of how closely the estimated concentration (or content) is to the real concentration. In practice accuracy is determined by comparing estimated values with certified or agreed values.

An important parameter in determining accuracy in analysis is the confidence
limit (*CL*):

where *t* is the Student t term, normally estimated for 95% confidence
(i.e. *a* = 0.95), *s _{e}* is the standard error of estimate,

*n*is the number of samples in the calibration,

*m*is the number of replicates (measurements) of the analyte sample,

*C*is the analyte content,

_{j}*C*is the mean content of the calibration, and

^{-}_{i}*C*is the known content of each calibration sample.

_{i}If we use a large number of calibration samples (say more then 15) and we analyse near the
mean of the calibration curve, then *CL* simplifies to:

Thus if we make 4 measurements on the analyte sample, the 95% confidence
limit is comparable to the standard error of estimate of the calibration curve.

For more details see: R Payling, in R Payling, D G Jones and A Bengtson
(Eds), *Glow Discharge Optical Emission
Spectrometry*, John Wiley & Sons, Chichester (1997),
pp 428-439.

First published on the web: 15 December 1999.

**Author**: Richard Payling