sampling

• Sampling is used when it is impossible, or simply not necessary to collect large amounts of data. Collecting small amounts of carefully selected data will enable you to obtain a representative view of the feature as a whole.
• You cannot, for example, interview ALL the inhabitants of a city but you CAN interview a fraction of those populations and indicate how the whole is likely to behave. You need to choose the one most appropriate to what you are trying to collect.
• · Random Sampling: One that has no bias and in which anything has an equal chance of being selected. This will usually involve random number tables as it is not possible to do it completely randomly if you select the sites/people. Something will always make you choose one over the other.
• Systematic Sampling: Samples are taken at regular/fixed intervals, for example every 3rd person that walks by, or measuring a river at every 100 metre
• Stratified Sampling: Based on knowing something in advance, either about the population of the city or about the river. i.e. if you know the ages of a village, you must have a representative sample of each age group.
• Bias in Sampling: It is possible, through poor choice of sampling method or insufficient evidence, to achieve a result that is unrepresentative of the population/river. Taking samples on a different day/week on a river will distort the results, whilst sampling population of a city, or studying tourism on the same day/week/time of year, will change the results.
• Sample Size: The size of sample usually depends upon the complexity of the survey being used. If it is a questionnaire to people, then you need quite a large sample, around 50 to 100 to be reliable, although it partially depends on the numbers of people who are in the area being sampled and can therefore be less.