## Want to keep learning?

This content is taken from the University of Birmingham & Chartered College of Teaching's online course, Education Research that Matters: Doing Research in Your Learning Community. Join the course to learn more.
2.6

## University of Birmingham

Selecting an appropriate sample is important for ensuring that your research question can be validly addressed.

The ‘sample’ refers to the members of the population that is selected for the research study. There are different ways to approach sampling depending on the purpose of your study, and there have been important debates about sampling in educational research, particularly the subset of educational research that seeks to understand the effects of educational practices or interventions.

Random sampling is a form of probability sampling seen by some as required for the production of highly generalisable claims about the effectiveness of educational practices or interventions. Others argue that it is not really possible to make such generalisable claims about educational interventions, as the complex contexts in which interventions take place have important impacts on the data. For those arguing the latter view, non-probability forms of sampling emphasise that evaluative research has a better knowledge of the context in which an intervention took place, and so can achieve a better understanding of in which contexts it might be effective or how it might be made effective in different contexts. While non-probability forms of sampling might be better suited to the claims made and value arising from small-scale teacher research projects, it’s worthwhile looking at both probability and non-probability sampling.

## Probability sampling

Probability sampling means that you’ll be seeking a sample of pupils or staff that are representative of the whole. A suitable sample can be achieved in any of the following ways:

### Simple random sampling

Have a sample size in mind and select the individuals that will make up the sample using a random method: a random name generator, pulling names from a hat etc. until the desired size is reached.

### Systemic sampling

This is still a form of random sampling but the list is first randomised before every 3rd/5th/9th (whatever number you choose based on the sample size you want) individual in the list is selected.

### Random stratified sampling

This can be used where you want to ensure your sample comes from across different groups (ability groups, year groups, gender groups…). Split the full list into the groups you’d like to be represented in your sample, and then use the systematic sampling method on each of these separate lists.

## Nonprobability sampling

Nonprobability sampling is any form of sampling where your research will be focused only on a specific group. You may have several criteria for inclusion or exclusion from the sample. A suitable sample can be achieved in either of the following ways:

### Convenience sampling

This is as it sounds. This sample are those that have volunteered or those who are available at the convenient time, location and so on.

### Purposive sampling

Selecting a sample based solely on the characteristics you’re interested in for your research, for instance EAL speakers, boys, early career teachers, and so on.

### Snowball sampling

A snowball sampling approach starts with an initial individual or group of individuals, and then expands its sample based on people proximate to the members of this original group (the name referring to the image of a rolling snowball gathering size). This form of sampling may combine elements of purposive sampling, combining proximity with other inclusion or exclusion criteria.

Take some time to reflect on each of these approaches. Which of these sampling methods might be most suitable for your practice-focused research project?