What Sampling Strategy to Use?

Now the focus of our geography lesson today was mainly about Environmental Quality Assessments, which I’ll talk more about at the end of the week. What I did realise was that people seem to find it difficult to chose which sampling strategy to use and what you do for each of them. So I thought it would be pretty beneficial to go through them and outline the advantages and disadvantages. 

Random Sampling

Exactly what it says, this strategy is completely random. To complete this strategy the area being surveyed needs to be divided into a grid formation. Then squares should be picked at random until you have the number of samples needed, the easiest way of picking at random is to number each grid square and then use a random number generator (easily found on the internet or on a calculator) to chose numbers which then relate to a grid square. The numbers picked are then your site locations. 

Positives:

• It isn’t biased

Negatives:

• Can form clusters of samples
• Can result in a poor representation of a area/population
• Inaccessible areas may be selected by the sampling 

Systematic Sampling

This strategy is relatively straight forward to understand and implement. Systematic sampling is choosing an area/street every x metres/roads. If needed an area can be split into grids like for random sampling and a sample taken every x number of grid squares. 

Positives: 

• Prevents clustering 
• Reduces bias associated with pragmatic sampling 

Negatives:

• Can result in a poor representation of an area/population 
• Doesn’t guarantee a wide range of samples e.g. all the samples may be extremely urbanised areas with little green space
• Pattern used for sampling may coincide with a pattern in population distribution which could cause skewed results.

Stratified Sampling

This is possibly the hardest one to understand and one most people find hard to understand. Stratified sampling provides an equal representation of an area by providing a sample size in proportion to the overall population. For example imagine I want to to survey 5 streets in an area which has 10 streets in it. 3 of these streets are Protestant, 7 are Catholic. Ideally I want my sample size to be in proportion of the population so to conduct a stratified sample I need to figure out how many Protestant streets I need to survey and how many Catholic streets I need to survey. To do this some basic maths is required (total number of protestant (or catholic) streets/total number of streets x sample size). So to figure out the sample sizes for the area you would do these equations 

P = 3/10 x 5 = 1.5 streets need to be sampled 

C = 7/10 x 5 = 3.5 streets need to be sampled

In reality it would be difficult to sample half a street but it’s just an example. The streets can then be chosen as needed. 

Positives:

• Provides an equal representation of the population/area
• A wide range of samples can be gained 
• Eliminates the chance of ‘freak’ sample sites

Negatives: 

• May be bias whilst choosing streets

Pragmatic Sampling 

The final strategy you need to know (thank god for that, this post has taken me hours to write and figure out…). Now it would probably help to know the definition of pragmatic; “dealing with things sensibly and realistically in a way that is based on practical rather than theoretical considerations” well that’s according to my mac’s dictionary anyway. Basically this method is where you pick the locations you want based on where you think the results will be for your survey (EQA). 

Positives: 

• You only (theoretically) get the data you need from the area(s) you need
• It saves a lot of time instead of trawling through useless data
• You can ensure you get an equal representation of the population/area and a wide range of samples can be taken
• In theory you can mix both stratified and pragmatic sampling creating something I’d like to name Pragmatic Stratified sampling but don’t quote me on that

Negatives: 

• Hugely biased, you’re unlikely to pick an area which there’s a high chance you’ll get stabbed 

If there are any questions you have through then feel free to ask and one of us will get back to you as soon as we can.