Random sampling means every part of the study area has an equal chance of being sampled. It is used to avoid sampling bias, for example choosing quadrat positions that look interesting or are easy to reach. It is most suitable when the area looks fairly uniform, or when comparing two different areas, such as a grazed and an ungrazed field.Method
1. Lay out two long tape measures at right angles along two edges of the study area, to form the axes of a grid.
2. Use a random number generator to produce a pair of numbers.
3. Use these as coordinates, for example (7, 12) means 7 m along one tape and 12 m along the other, and place the quadrat there.
4. Count or estimate the species in the quadrat, recording density, frequency or percentage cover.
5. Repeat for a large number of quadrats and calculate the mean.
Using a large number of quadrats makes the sample more representative of the whole area and reduces the effect of any unusual results. One way to decide how many to use is to plot a running mean against the number of quadrats: once the mean levels off, adding more quadrats changes little.Worked example
A random number generator gives the numbers 3 and 15. On a grid with 1 m intervals, the quadrat is placed 3 m along the first tape and 15 m along the second.Random sampling is not suitable when conditions change gradually across an area, because random quadrats may miss the change.
1. Lay out two long tape measures at right angles along two edges of the study area, to form the axes of a grid.
2. Use a random number generator to produce a pair of numbers.
3. Use these as coordinates, for example (7, 12) means 7 m along one tape and 12 m along the other, and place the quadrat there.
4. Count or estimate the species in the quadrat, recording density, frequency or percentage cover.
5. Repeat for a large number of quadrats and calculate the mean.
Using a large number of quadrats makes the sample more representative of the whole area and reduces the effect of any unusual results. One way to decide how many to use is to plot a running mean against the number of quadrats: once the mean levels off, adding more quadrats changes little.Worked exampleA random number generator gives the numbers 3 and 15. On a grid with 1 m intervals, the quadrat is placed 3 m along the first tape and 15 m along the second.Random sampling is not suitable when conditions change gradually across an area, because random quadrats may miss the change.