Instructions: Click on the sections below to see the explanations.
Once you know what research hypothesis your experiment will be testing, and how, your next step is to determine who your participants will be.
Picture 1: Human Population, Target Population, Sampling Frame and Sample
Sampling Technique (Click)
You already know what target population your experiment is about. However, your actual participants will only be a sample, i.e., a limited number of people drawn from this target population.
To draw participants from your target population, you will need two things:
- A sampling frame
- A sampling technique
First, while a target population is a rather abstract class of people, a sampling frame is the mechanism used by researchers to get concrete access to members of the target population. For instance, even though your target population might be the entire world population, you will likely only be able to access members of this target population by finding participants in your school. Thus, your sampling frame would be: students enrolled at your school.
NB. Your participants do not have to be (limited to) students from your school, however. For instance, you could use services such as Amazon’s Mechanical Turk to recruit participants. Likewise, you could test adults from your school, or relatives. You could also use social media to find volunteers from different circles.
Second, while the sampling frame defines the population from which the participants in a study can actually be selected, this sample is drawn using a further sampling technique. For instance, if your sampling frame is “students enrolled at your school”, you will not be testing all of them, but only some.
Although there are many sampling techniques available, only a few can realistically be used for your IA. Of all the different techniques available, random sampling is considered the gold standard, because it has the best chance of leading to unbiased, representative samples that reflect the composition of the sampling frame, thus increasing external validity. However, random sampling is often technically difficult to use. “Randomly” asking people around you if they would accept to take part in your experiment is not random sampling (it is convenience sampling.) Random sampling implies using an exhaustive list and a blind technique (such as an online random selector) to select participants in such a way that everybody in the sampling frame has the same statistical chance of being selected. Whenever obtaining such a list or actually gathering randomly selected participants is not technically feasible (which is usually the case), other sampling techniques must be considered.
Most IAs use convenience sampling, which means that students draw their sample with the technique that is the most readily available and easily implemented. It could be asking people assembled at a certain time and place (e.g., at lunch in the cafeteria), or using pre-existing groups, such as classes (which is sometimes called “opportunity” sampling). However, other techniques exist that might be better suited to your experiment. When choosing your sampling technique, consider both advantages and disadvantages, including feasibility, and this with regard to your particular experiment.
NB. Neither your sampling technique nor your participants have to be similar to the original experiment you are replicating.
Table 13: Advantages and Disadvantages of Different Sampling Techniques
(1) A sampling bias is an individual characteristic that:
- Is relevant to the research hypothesis
- Is shared by all participants in the sample because of the sampling technique used
- Is not shared by all members of the target population
(2) External validity is the extent to which the findings of a study can be generalized from the sample to the target population. This depends on how representative the sample is.
(3) Purposive sampling might be necessary if your research question only applies to people with a certain profile (e.g., if your target population is “student athletes”). However, make sure that this profile is not the IV, as your study would then be a quasi-experiment. All of your participants (in both conditions) should have this same profile.
Participant Characteristics (Click)
You will not be able to describe participant characteristics until you actually conduct your experiment. However, when you do, make sure to collect all relevant information. Examples include the following, but “relevant” information will depend on your particular research question:
- Sample size
- Age range
- Gender breakdown
- Cultural background
- Socio-economic status (e.g., students)
One thing you should plan ahead of time, however, are exclusion criteria. As we mentioned earlier, you might need participants who do not know certain psychological theories, or who are not English as a second language learners, etc. To be ethical, you must think of ways you will be able to respectfully avoid or exclude such participants. For instance, you could collect their results, but not include them in the analysis. In this case, data collection cannot be anonymous.
QUESTIONS: Click on the question below to see the worked example.
Answer the question below on your individual Experiment Creation Form.