Estimation of Random Deviations in Analytical Methods using Analysis of Variance

Abstract

Analytical methods can vary from extremely simple one-step procedures to those which include a number of steps, often involving complex sample treatments and manipulations, and it can often prove extremely difficult to isolate the variability associated with each step in the method. Proper experimental design can yield more information than an ad hoc approach, and subsequent statistical analysis can reveal the magnitude of random deviations in the steps of a multi-step method, which are not always obvious during a cursory examination. In this exercise, Analysis of Variance (ANOVA) is used to estimate the variance components in a multi-step analytical procedure, namely a sequential extraction of trace elements from soil, with subsequent analysis by flame atomic absorption spectrometry. The exercise will be used to demonstrate the following:

• the principles of good experimental design

• the importance of replication and randomisation

• the use of F-tests and hypothesis testing

• the estimation of random deviations in an analytical method.

The exercise is aimed at final year undergraduate students, or postgraduate students. The students are required to perform the analytical procedure as part of a laboratory exercise, working in a group and pooling the data at the end of the lab. A subsequent tutorial will then be used to analyse the data and draw conclusions about the analytical method.