Sampling methods: random, systematic, stratified and pragmatic

Overview

Population and the sampling frame in an investigation, random sampling and systematic sampling.

What you need to know

Answer each one before you reveal it.

Questions from this subtopic

Each one with the model answer underneath.

1. 1

   Why do geographers sample rather than measure everything?

   Model answer

   Because measuring the whole population is almost always impossible within the time, money, access and manpower available - you cannot interview every resident of a town or measure every pebble on a beach. A sample is a manageable subset chosen so that it represents the whole. If it is chosen well, conclusions drawn from the sample can be applied to the population with a known degree of confidence; if it is chosen badly, the whole investigation is unsound however carefully the data were measured.

2. 2

   What is meant by the population and the sampling frame in an investigation?

   Model answer

   The population is the complete set of things you are interested in - every household in a village, every pebble on a spit, every field in a parish. The sampling frame is the practical list or grid from which the sample is actually drawn: an electoral register, a numbered list of streets, a grid laid over a map of the beach. A sample can only be as representative as its frame, so a frame that omits part of the population, such as a list that misses new housing, builds bias in before any data are collected.

3. 3

   What is random sampling, and how is it actually carried out in the field?

   Model answer

   Random sampling gives every member of the population an exactly equal chance of being chosen, with the choice made by chance alone rather than by the investigator. In practice you number the members of the sampling frame - the houses in a street, the squares of a grid laid over a study area - and use random number tables or a random number generator to decide which numbers to sample. For an area survey you generate pairs of random numbers and use them as coordinates on the grid to fix each sampling point.

4. 4

   What is the great strength of random sampling?

   Model answer

   It removes investigator bias completely: because chance and not the surveyor chooses the sites or the respondents, there is no possibility of unconsciously picking the easy, the convenient or the attractive. It is also the assumption on which most statistical tests are built, so a genuinely random sample allows the results to be tested for significance and generalised to the population with confidence.

5. 5

   What is the weakness of random sampling?

   Model answer

   Chance can produce an uneven sample. Random points may cluster in one corner of the study area and leave another part unsampled, so that a real spatial pattern is missed, and a small but important subgroup - one land use, one age group, one rock type - may not be picked up at all. Random points can also fall in places that are inaccessible, dangerous or on private land, and generating and then locating each point in the field is time-consuming.

6. 6

   What is systematic sampling, and how is it carried out?

   Model answer

   Systematic sampling takes samples at regular, predetermined intervals rather than by chance: a reading every ten metres along a transect, every fifth house along a street, a quadrat at every grid intersection, a measurement every hundred metres downstream. You fix the interval in advance, choose a starting point, and repeat the interval to the end of the study area.

Other ways to work this

The same content, four other ways.

What is inside Levo

Six ways to work the same material.