How Were Participants Actually Selected for the Study?

How did these particular people end up in the study?

A paper tells you that 600 university students completed a survey, 240 patients entered a clinical study, or 3,000 adults were included in a cohort. The next question is easy to overlook: how did those particular people become the participants?

Participants do not simply materialize in a dataset. Researchers define a population, establish eligibility criteria, create or use a sampling frame, identify potential participants, invite or recruit them, and obtain participation. Sometimes probability mechanisms are involved. Sometimes participants volunteer, are recruited consecutively from a clinic, respond to advertisements, or enter through an existing database. Understanding that pathway helps you determine what population was actually studied and what selection processes may have shaped the resulting sample.

Trace the pathway from potential participants to the final sample

Reconstruct participant selection as a process

Begin with the source from which participants could be selected

Before asking how participants were sampled, determine who could realistically have entered the study. A national research question does not imply a national recruitment pool.

Potential participants might come from patient lists at participating hospitals, students enrolled at selected universities, households in sampled geographic areas, members of an online panel, employees of cooperating organizations, respondents to public advertisements, or records contained in an administrative database.

This starting pool matters because no sampling technique can select people who were never represented in the recruitment source or sampling frame.

Separate sampling from recruitment

The terms are sometimes used loosely, but they describe conceptually different parts of the process.

A study could randomly sample 2,000 people from a population register and still obtain responses from only 700. Random selection at the sampling stage does not make participation random. Conversely, researchers may recruit every eligible patient presenting consecutively at selected clinics without drawing a random sample.

Identify whether selection was probability-based or non-probability-based

In probability sampling, units are selected through a probability mechanism from a defined sampling frame. Designs can include simple random, systematic, stratified, cluster, or multistage sampling. Their details matter because selection probabilities and clustering can affect both estimation and analysis.

Non-probability approaches do not give every eligible unit a known selection probability. Convenience samples, volunteer samples, purposive samples, respondent-driven approaches, and various forms of quota recruitment fall into this broader category, although their purposes and methodological implications differ considerably.

The distinction should not be turned into a crude good-versus-bad judgment. Probability sampling can be valuable when estimating characteristics of a defined population, but feasibility, research purpose, design, nonresponse, and sampling-frame quality also matter. Qualitative research, for example, may deliberately select information-rich participants rather than attempt statistical representation of a population.

Do not accept the sampling label without checking the procedure

If authors say they used “random sampling,” ask what was randomized. Was there a complete list of eligible individuals? How was the random selection performed? Were institutions selected first and individuals selected later? Were participants merely approached in an arbitrary order?

Similarly, “convenience sample” tells you relatively little by itself. Were participants students in one researcher's classes, people responding to social-media advertisements, patients available during specified clinic hours, or members of an online research panel? These mechanisms can produce different selection patterns.

The procedure is more informative than the label.

Eligibility criteria create another selection step

After identifying the recruitment source, determine who could qualify. Age restrictions, diagnoses, language requirements, treatment histories, geographic residence, institutional membership, availability during the study period, and numerous other criteria can narrow the eligible population.

STROBE reporting guidance asks authors of observational studies to describe eligibility criteria and the sources and methods of participant selection. CONSORT guidance similarly requires trial reports to describe participant eligibility criteria as well as study settings and locations. These details help readers reconstruct how the observed sample emerged from a broader population.

Participation itself can be selective

Being invited is not the same as participating. People may refuse, ignore invitations, fail to attend appointments, abandon online questionnaires, or be unreachable.

If participation is associated with characteristics relevant to the study, respondents may differ systematically from nonparticipants. For example, a voluntary survey about workplace stress could disproportionately attract employees with particularly strong experiences, although the direction and magnitude of such differences cannot simply be assumed.

Whenever possible, look for numbers at each stage and information about nonparticipants. STROBE recommends reporting numbers potentially eligible, examined for eligibility, confirmed eligible, included, completing follow-up, and analyzed, with reasons for nonparticipation at relevant stages.

Selection bias is not simply another name for non-random sampling

Selection bias has a more specific methodological meaning than “the sample was not random.” Broadly, it concerns systematic processes of selection into a study or analysis that distort the relationship being estimated.

A convenience sample may limit population inference without necessarily producing a biased estimate of every association within that sample. Conversely, selection bias can arise even in a study that began with probability sampling if participation, attrition, or inclusion in the analysis depends on variables related to the exposure and outcome.

Selection into the study is different from allocation within the study

This distinction is especially important in randomized trials. Random allocation determines which intervention enrolled participants receive. It does not ordinarily mean that participants themselves were randomly sampled from the wider population.

A trial may recruit volunteers from a small number of specialist clinics and then randomly allocate those volunteers to treatment groups. Randomization can support the internal comparison between intervention groups, but it does not automatically make the trial participants representative of all people with the condition.

Selection may continue after recruitment

The selection pathway does not necessarily end at enrollment. Participants can withdraw, become lost to follow-up, lack required measurements, or be excluded from particular analyses.

Suppose 1,000 people enter a cohort but only 720 have complete exposure, outcome, and covariate data for a regression model. The estimate from that model is based on those 720 observations, not simply on the original 1,000 participants.

That is why participant selection eventually connects to the question of whether all recruited participants were included in the analysis.

Following the selection pathway from universities to survey respondents

Common mistakes when evaluating participant selection

Draw the selection pathway before judging the sample

A practical way to appraise participant selection is to reconstruct a small flow from the population that could have been reached to the participants who generated the result. At each transition, ask what mechanism allowed some people to continue while others did not.

This approach is more informative than attaching a single sampling label to the paper.

The central question is not merely “What sampling method did they use?” It is “What process produced these observations from the people who might otherwise have been studied?” That formulation usually reveals much more about the evidence.

Before accepting the sample at face value, reconstruct how it was formed

Questions about how research participants are selected

The final sample is the product of a selection pathway

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.