Sampling Methods

When choosing where to sample, it should be done by generating two random numbers to use as co-ordinates. If you choose the random areas, it could introduce some bias to the results. In a natural environment, you could measure distances from a set starting point in the corner of the sample area or, when working in an urban environment, the random number pairs could be OS grid references. You can use the co-ordinates in different ways - point sampling, line sampling or area sampling. For point sampling, you would make your observations at the location described by the co-ordinates. For line sampling, you would generate a second set of co-ordinates and sample along a line joining the two points. For area sampling, you would use the co-ordinates to position your quadrat.

Where there is a difference across a sample area (termed an environmental gradient), systematic sampling is the method used. An example of this would be a school hockey pitch. Usually, the area around the goalmouth is trampled and mainly bare of plants. The most common methods are to use a line transect or a belt transect and sample either continuously or at regular intervals (interrupted transect). In the case of the hockey pitch, using continuous or interrupted belt transects that include the goalmouth areas or stratified systematic sampling (see next paragraph) would be appropriate. Other forms of systematic sampling would be every 10^th house or every 5^th person in the street. It could also be time based e.g. every 15 minutes during the day

Where there are significantly different parts to the study area (known as subsets), it is important to make sure that the samples you take are representative of each part. This is known as stratified sampling and can be either systematic or random. Examples of subsets in population studies would be age group or social group. An example of a subset in an environmental study would be a clearing in a woodland. The number of samples taken from each subset needs to be proportional to the whole survey.