Here’s part 3 of Mary’s series on research basics for non-researchers. You can read parts 1 & 2 here. Stay tuned — a new post comes out every fortnight!
There are a number of things that you have to consider beyond simply your research design and types of research.
Ethics approval is needed for all research involving human participants (if you want to publish it, or if you are asking participants to perform some activity – generally it’s best to check if you need ethics regardless!). In order to obtain ethics approval, an application must be lodged with the appropriate panel. For this information, visit your institutions website for more information. However, there are a few exceptions. Ethics approval is not required for research on existing data or for market research. It is important to consider, that although ethics approval is not sought in these cases, guidelines must still be followed and/or evidence of your research must be lodged. For more information, have a read of the Australian Market and Social Research Society (AMSRS) Code of Professional Behaviour.
How big should your sample be? Sample size depends on a number of factors including, the size of your population, resources, time, and availability of participants. There is no rule for sample size; rather it is dependent on the study.
From where do you draw your sample? There are four types of sampling methods that (I can think of immediately): convenience, purposive, snowballing, and quota sampling. Much like your choice of research design, your sample can greatly affect the accuracy of your results. Remember, the aim of research is to use a sample to accurately infer the population (you would use the entire population as if you could). Of course, there are also different ways to randomly sample the population as well — but that’s a post all on its own.
Convenience sampling is exactly what it sounds like. It is subject to the availability and proximity of subjects. Although one of the easiest sampling methods, there is a high risk of obtaining a biased sample that does not and cannot accurately depict the population.
Purposive sampling is where you have a list of criteria/characteristics that participants must adhere to in order to be a valid candidate for your research. For example, gender, age, media consumption, etc…
Snowball sampling is very similar to the marketing term snowballing (editor: and something commonly discussed as one of the benefits of social media!). This sampling style involves a chain effect of referrals. That is, you locate individuals with a wide network and they pass on the survey/research design to other individuals, and so your sample grows.
Quota sampling can involve any number of the other sampling types. However, once a certain number of participants from a category, say female, have taken part in the research no more females are eligible.