Sampling

Why do we sample?
If the total number of items under consideration (known as the population) is small, say up to 100, then it would be best to measure them all. But often sampling is desirable for several reasons:

(i) It is quicker than measuring every item
(ii) It is therefore much cheaper
(iii) Often it is impossible to measure everything because the population size is so great.
(iv) It may be unnecessary to measure the whole population – a sample might give very similar results
(v) We may not know the population size or precise location.
(vi) We may wish to take a snapshot of the population at one moment in time (e.g. the Met office).

How can our results become biased?
* The data from which the sample is being taken may themselves be biased – e.g. people from a telephone directory (Why?)
* Choice of sample may be unfair – is the number of old/young & men/women etc. representative?
* The time the sample was taken may produce bias – e.g. week day/ weekend/ school holidays/ Sundays etc.How big should my sample size be?
* The size of the sample will usually be directed by the time and resources available. It should be as large as possible but should also be balanced with quality (detail).
* At least 30 samples should be considered in order to be able to draw valid conclusions.
* A sample which is too large could give too many results to handle.

There are three basic sampling strategies:
1. Point sampling
2. Line sampling (transect)
3. Area sampling

Having adopted one (or more) of these strategies, there are 3 main Sampling Methods:
1. Random - use random number tables to select sample site. Each point selected has the same chance of selection as all other sites.
2. Stratified - the most accurate method involves an initial analysis of the population being studied and its division into categories. The sample is allocated to the size of these categories, and specific sites are then selected by random techniques. This method ensures that the results are a true representation of the whole population.
3. Systematic - uses a rule or a set procedure to determine the sample site or person, e.g. to ask every tenth person, or sample every 100 metres.

---