The same principles will also apply to other subjects where you conduct primary research on people.
What is sampling?
Sampling means selecting people to take part in your research. There are two key steps:
- First, define the group of people that you want to study. Perhaps you live in London and you want to do a survey on elderly people. This is know as the 'population' for your research. However, you need to define it specifically. Instead of 'elderly people', you could say 'adults aged 70+ living in London'.
- Now you have to pick a smaller group from this population, and conduct the research on them. Unless the population is very small, you can't study all of them! This smaller group is called your 'sample'.
If you have done Geography/Geology at school and collected rock samples, the concept is very similar. Rather than taking the whole rock/mountain, you collected a small piece and studied that.
The problems are also similar. What if the area of the mountain that you collected the rock sample from just happened to have a different type of rock than 99% of the mountain? Perhaps that rock was moved there by a glacier, and originated in a different place. The problem is that the sample is not representative of the whole. The results of the study on that rock sample can't be generalised to the whole mountain!
|The concept of sampling applies to other subjects. Image: luigi alesi|
It is very similar in Psychology. In the above example, if you selected a sample of elderly people by asking your grandmother's friends, they might not be representative of the whole population of people aged 70+ in London. Perhaps they are more affluent, in better health, or better educated than the average? Perhaps they are not as ethnically diverse as the whole population?
If your sample is not representative of the population, then it is difficult to generalise the findings of your research. What you find out might be true of your sample, but it may not be true of the whole population.
To put it another way, your sample is biased!
Types of sampling
The example above - sampling family or friends - is called an 'opportunity' or 'convenience' sample. Other sampling methods can get a much more representative sample, and, therefore, better data! Here are the main types to be aware of:
- Random sampling: this means that everyone in the population studies has an equal chance of being chosen. This is a lot more difficult than it sounds! How do you ensure that everyone has exactly the same chance? Putting all the names in a hat might work, but not if there are 1000s of names! Psychologists typically use a random number table, together with a numbered list of members of the population. But what if the people you randomly select don't want to take part in your study? The problems are obvious!
- Stratified: Here, the researcher makes sure that groups within the population are represented fairly within the sample. For example, having a 50-50 mix of males and females. it may depend on what is important to the research. If religious belief was likely to affect the results, then researchers may try to ensure that their sample had the same proportions of religions as the population as a whole.
- Systematic: A systematic sample involves picking people at fixed intervals from a list of the whole population. So if you have a list of 1000 students in your year and pick every 50th name, that is a systematic sample.
These other types are less representative, but often used:
- Opportunity: as described above, this is where participants are chosen on the basis of convenient availability. This might mean approaching people who walk past and asking them to take part. Like fishing in a lake, you take whatever comes along! This tends to lead to a biased sample, but is often used because it is quick and easy to do.
- Self selecting: Also called 'volunteer sampling', this is when people come forward to take part in your study rather than you finding them. Milgram's (1963) study of obedience used people who responded to an advert, and were therefore a self-selecting sample. They may differ from average members of the population in various ways, for example, by being more interested in helping scientific research, or more in need of the money paid to participants!
- Quota sample: similar to stratified sample, but here you simply ensure that you have some from every category (e.g. every religion) rather than keeping the proportions the same as the population. It may be that some groups have a very tiny population, so you set a minimum 'quota' to ensure that they are not missed out. For example, you might select two people from every religion, regardless of how common those religions are.
For Scottish Higher, you should be aware of all of the above types. For A-Level, all except quota are included. At university level, these will cover you well at least for 1st and 2nd year.
In some situations, an opportunity sample might be adequate. A random sample is ideal, but has many practical problems in obtaining one. A good compromise might be to use a stratified sample.
Many thanks to the University of California, for permission to use their images on this post. Read their article on sampling here.
Any questions, feel free to ask in the comments!