Sampling

When selecting or sampling your research participants, you need to be aware of the similarities and differences between the population (Flick, 2009; Robinson, 2014) in terms of: demography, geographical location, physical attributes, psychological attributes or life experiences.These considerations determine who or what can be included. However, the purpose of sampling is closely related to the research question(s) of your study. Consider the following case:

Imagine if you want to study the views of students and staff in EdUHK towards the newly established cafeteria, you need to consider:

  • the sample size (how many people do you want to or can ask?);
  • the selection criteria (how do you select the sample – the people to be interviewed?); and
  • the feasibility (can you finish collecting the data within a particular period?).

If you have problems in making these decisions, you may want to re-consider carefully about what and who to be included.

Sampling methods

We only highlight several commonly used methods here, including: snowball sampling, critical case sampling, and purposive sampling.

Snowball sampling

  • It is a type of non-probability sampling, as researchers ask participants/informants to introduce potential contacts.
  • It is very useful when your research groups are difficult to reach, e.g. refugees, drug addicts, sex workers and professionals etc.
  • Nevertheless, researchers may fail to identify the right person to be studied. By making use of his/her connections, the researcher may only recruit those participants who are well connected.

 

 WARNING

You want to use snowball sampling not because you want to save time or to interview your friends. You need good justifications: why those interviewees matter; how those interviewees contribute to your research findings!

Purposive sampling

  • It is a sampling technique which relies on your own judgment (and criteria) in selecting members to participate in your research (think about your research question and topic carefully!)
  • It is effective when primary data sources are limited (e.g. a few insiders) due to the nature of research design and aims and objectives.
  • Example: When it comes to studying teachers’ opinion of whether Liberal Studies should remain as a core subject in high school education, you may target Liberal Studies teachers as they have the knowledge of teaching the subject and interacting with students. 

Critical Case sampling

  • It is a type of purposive sampling
  • For pragmatic concerns, you can consider critical case sampling especially when time and funding is limited.
  • When selecting cases, researchers should always bear in the mind of these: Why is this case worthy of study? Can the case yield the most information possible? (Patton, 1990)
  • Even though a few cases cannot provide generalized findings, researchers could still deduce observations based on in-depth study.
  • Usually, around ten participants will be appropriate for a small scale research.
  • Example: Umbrella Movement can be used as a case to study youth’s participation of social movements in Hong Kong. There are different social movements in the history of Hong Kong in which youth participated, but Umbrella Movement could be a good case in terms of its scale, duration and influences to youth’s political participation.

How much is enough 

The typical principle to determine a sample size is data saturation: continue the sampling process until there is redundant (saturated) information. This indicates that no new information and themes can be collected.

 

TIPS

Be practical! You need to consider your time, finances and academic level. Always discuss with your supervisor and estimate the amount of time needed to collect sufficient information from interviewees.