Why the way people are selected matters
A sample is the set of people, objects or cases a study examines. Researchers often use one because studying every member of a larger group would be impractical. But a sample does not automatically stand in for that group. How its members were selected affects both the questions the study can answer and the confidence readers should place in broader conclusions.
Before assessing a sampling method, identify the population of interest: who or what is the study meant to describe? A survey about all customers has a different target from one about customers who recently contacted support. Then ask who had a chance to be included. If some members of the target population could not be reached, the findings may miss experiences that matter, regardless of how many responses the study collected.
Random selection: a route to broader estimates
In a simple random sample, every member of a defined population has a known and equal chance of selection. Researchers might, for example, assign a number to every person on a complete list and use a random process to choose whom to invite. Other probability methods give people different but known chances of selection. The shared principle is that selection follows a specified chance-based process rather than a researcher's preference or a participant's availability.
Random selection can help researchers estimate characteristics of the population and express uncertainty due to sampling. It does not guarantee a perfectly balanced sample. Chance can produce uneven groups, especially in smaller samples. Nor does it solve every problem: an incomplete list can exclude people, and selected people who do not respond may differ from those who do. When reading a study, look beyond the word “random” to find out what list was used, who was selected and who actually took part.
Convenience selection: quick access, narrower claims
A convenience sample includes people or cases that are easy to reach. A feedback form offered to website visitors or interviews conducted with people at a nearby event might use this approach. It can be useful when a team needs to explore a question, test a survey's wording or gather early reactions with limited time and resources.
The limitation is that accessibility can be related to the thing being studied. People who choose to fill in a feedback form, for instance, may feel more strongly than those who ignore it. Because members of the wider population do not have a known chance of inclusion, a large convenience sample is not necessarily representative. Its findings can describe the respondents and suggest questions worth investigating, but they should not be presented as a precise estimate for everyone.
Purposive selection: choosing cases for a reason
With purposive sampling, researchers deliberately select participants or cases that can illuminate a particular question. To understand how people use a complicated service, they might seek first-time users, experienced users and people who stopped using it. The goal is not to imitate the population's proportions. It is to include perspectives or experiences that are especially relevant to the inquiry.
This method is common in work that aims to understand processes, meanings or differences between cases. Its value depends on the selection criteria being clear and appropriate. If a study seeks a range of experiences but recruits only enthusiastic users, important difficulties may remain hidden. Purposive findings can offer useful explanations and ideas, but the number of people expressing a view should not be treated as evidence of how common that view is in the wider population.
Match the method to the claim
No single sampling method is best for every question. If the aim is to estimate how common an opinion is in a defined population, a well-designed probability sample is generally more suitable than a convenience or purposive sample. If the aim is to learn how a particular experience unfolds or to identify issues for further study, carefully chosen cases may be more informative than a small random group.
A useful check is to compare the study's conclusion with its selection process. Ask who the researchers wanted to understand, who could be selected, who participated and why those participants were chosen. Also ask whether missing voices might change the interpretation. Clear reporting of these details does not remove every limitation, but it makes the findings easier to use responsibly: as population estimates where the design supports them, or as insights into selected experiences where it does not.
