TABLE OF CONTENTS Anyone taking samples for analysis needs to realise the importance of stating adequately the problem they want to solve or identifying the basic questions that they want the analysis to address. Then they need to assess whether their plan is capable of giving results that will provide the answers. We should ask ourselves:
What should we sample and what do we need to measure?
How many samples would there be for chemical analysis?
Can we ensure that results obtained can be analysed statistically?
What are the limitations imposed by considerations of cost, convenience, accessibility and equipment?
We cannot analyse all the material we are studying. As we have discussed in Chapter 1, we usually want to get a sample that is representative of a bulk material or a fluid or a tissue being studied. The practical solution is to take a number of samples throughout the material. For sampling to be truly random, any portion must have the same chance of being included. We need to avoid introducing bias; if we are taking samples from animals, we should not pick only the ones that don't run away; if we are sampling a forest, we need to avoid concentrating on areas near the roads. Random sampling allows us to use some statistical techniques on our results and to estimate the range about our sample mean within which the true mean probably lies.
You may be taking samples which you will pass on to an analytical laboratory for...
