Sampling

Where the relevant population is too large for a cost-effective census to be conducted, a sample of that population must be selected. Advantages of sampling are:

• Costs are reduced
• The collection time is reduced
• Overall accuracy is improved

For several types of population, sampling is the only method of data collection; for example, infinite populations, or testing procedures that entail the destruction of the item being tested, like tests determining the life of a lightbulb.

Sampling is only justified if reasonable results can be obtained, that is if valid inferences about the population can be drawn from the sample. Therefore, sampling must be based on two laws:

• The law of statistical regularity holds that a reasonably large number of items selected at random from a large group of items will, on average, have characteristics representative of the population. It is important that the selection of the sample is random so that every item in the population has an equal chance of selection. The size of the sample should be large enough to minimise the influence of abnormal items on the average.
• The law of large numbers holds that large groups of data show more stability than small ones.

Sampling procedures can be divided into two broad categories: those where elements are selected by random method and those where elements are non-randomly selected.

Non-probability samples: Any sampling technique in which the selection of the sample items is not determined by chance, but rather by personal convenience, expert judgement, or any type of conscious researcher selection.

Probability/random samples: A probability sample is one on which the items are chosen are based on known probabilities. The choice of items is left to some form of chance or random procedure.

Simple random sampling is where every item in the population has an equal chance of selection at each successive stage of the selection process. Two major techniques are:

• ‘Goldfish bowl’ technique which is similar to drawing names from a hat, and
• 'Random number tables', which consist of rows and columns, and a computer is programmed to scramble numbers. Any series of numbers read across or down the table is considered random.

Microsoft Excel can be used to select random items on an excel spreadsheet. It is done by using the term RAND = followed by your random type selection in the dialogue box.

The Random Number Generation analysis tool fills a range with independent random numbers that are drawn from one of several distributions. You can characterize the subjects in a population with a probability distribution. For example, you can use a normal distribution to characterize the population of individuals' heights, or you can use a Bernoulli distribution of two possible outcomes to characterize the population of coin-flip results.