Population and Sample: How to Define Them with Thesis Examples (2026)

The population and sample subsection is where a committee decides whether your findings can mean anything beyond the people or records you happened to reach. Students usually know the sampling techniques by name; what they get wrong is the chain of definitions that comes before the technique: who the findings are supposed to apply to, who you could actually reach, what list you drew from, and who ended up in the study. This guide explains that chain, sets out the five elements the subsection needs, and gives eight complete worked paragraphs across disciplines.

Quick answer

The target population is the whole group your research question is about. The accessible population is the part of it you can realistically reach. The sampling frame is the actual list or source from which units were selected. The sample is the set of units you studied. A complete subsection names all four, states the inclusion and exclusion criteria, names the sampling technique with a reason, and gives the sample size with a justification. Creswell and Creswell’s Research Design (6th ed., SAGE, 2023) lists these as required elements of a quantitative methods section.

The Four-Link Chain: Target, Accessible, Frame, Sample

Link	Definition	Nursing example
Target population	Everyone or everything the research question is about	Adult patients boarded in US community-hospital emergency departments
Accessible population	The part of the target population you can reach with your access and approvals	Boarded adult patients at one community hospital during the study period
Sampling frame	The list, register or data source units were drawn from	The emergency-department admission log for the twelve study weeks
Sample	The units actually studied	Every boarding episode meeting the criteria, expected to be roughly 300

The gap between each link is a threat to generalisation, and naming the links is how you show the committee you have seen the threats. A sample drawn from an incomplete frame under-covers the accessible population; an accessible population that differs from the target (one hospital, one region) limits how far findings travel. You do not have to eliminate the gaps in a master’s thesis; you have to state them. The techniques for drawing the sample are a separate decision, covered in our guide to sampling methods in research; this article is about defining what you draw from.

The Five Elements the Subsection Must Contain

Target population, defined by the characteristics in your research question, with a size only if a citable source gives one.
Accessible population and sampling frame, naming the sites and the list or record system, with the period covered.
Inclusion and exclusion criteria, each with a one-clause reason.
Sampling technique, named precisely (stratified random, purposive maximum-variation, consecutive, total population) with the reason it fits the design.
Sample size and its justification, by power analysis for tests of hypotheses, by saturation logic for qualitative designs, or by the size of the accessible population for a census.

The criteria in element 3 must match the boundaries you set in Chapter 1. If the scope and delimitation section excludes intensive-care units, the exclusion criteria here must say so in the same words.

Eight Population and Sample Examples by Discipline

Each paragraph contains all five elements in order and is written as it would appear in the methodology chapter. Sizes and sites are illustrative; power calculations are described, not reported, because the numbers depend on your effect size.

1. Nursing (MSN, United States; chart audit)

The target population is adult patients admitted through the emergency department and boarded there before transfer to a medical-surgical unit in US community hospitals. The accessible population is boarded adult patients at one 220-bed community hospital during a twelve-week period in autumn 2026, identified from the emergency-department admission log, which records every admission decision and transfer time. Episodes are included when the patient is 18 or older, has a documented pressure-injury risk score at admission, and boards for at least two hours, the threshold at which the first bundle element becomes due; episodes are excluded when the patient is transferred to intensive care, because a different risk tool applies. Because the accessible population is small enough to audit in full, a total population sample of all eligible episodes is used rather than a subsample. Based on the previous year’s admission log, approximately 300 eligible episodes are expected, which exceeds the sample required to detect a moderate correlation between boarding hours and adherence at conventional power.

2. Education (MEd, Canada; qualitative)

The target population is Grade 4 teachers in Ontario public school boards who teach mathematics in English-language classrooms. The accessible population is the Grade 4 teachers of one board, reached through a board-approved invitation sent by the curriculum department to all schools; the board’s staffing list serves as the sampling frame. Teachers are included if they have taught Grade 4 mathematics for at least two years and have attended the board’s Indigenous-perspectives professional development, so that all participants share a baseline exposure; teachers in French-immersion classrooms are excluded because their instructional materials differ. Purposive maximum-variation sampling is used to select six teachers who differ in school size, urban or rural setting and years of experience, because the study aims to describe the range of instructional decisions rather than the typical one. Six participants with two observed lessons and one interview each is consistent with case-study designs in this field, and recruitment would be extended if new decision types were still emerging after the sixth case.

3. Business and Management (MBA, Australia; survey)

The target population is employees below partner level in mid-sized Australian accounting firms with hybrid working policies. The accessible population is the staff of three Sydney firms that agreed to participate, with each firm’s staff directory as the sampling frame. Employees are included if they have been with the firm for at least six months and hold a role for which hybrid working is permitted; partners and contractors are excluded because they set their own arrangements. A census of the accessible population is attempted, with the survey distributed by each firm’s people-and-culture team to all eligible staff, because the firms are small enough that subsampling would reduce power without saving cost. With approximately 420 eligible employees and a response rate consistent with recent internal surveys at these firms, the expected sample exceeds the number required to detect a small-to-moderate interaction effect in a moderated regression with three controls.

4. Psychology (MSc, United Kingdom; trial)

The target population is final-year undergraduates completing a dissertation at UK universities. The accessible population is final-year students enrolled on dissertation modules at one university in the 2025 to 2026 academic year, invited through the module virtual learning environments, which list all enrolled students and serve as the sampling frame. Students are included if they are 18 or older and not currently receiving psychological treatment, on the advice of the ethics committee; students who have previously completed a structured self-compassion programme are excluded to avoid ceiling effects. Volunteers are allocated to intervention or waitlist by simple randomisation using a computer-generated sequence, because random allocation is what allows the group difference to be attributed to the intervention. An a priori power analysis for a between-groups comparison of post-test scores, using the effect size reported in prior brief-intervention trials, gives the target sample, inflated by an allowance for attrition based on those trials.

5. Computer Science (MS, United States; units are repositories)

The population in this study is not people but software projects. The target population is open-source Python repositories with developer-written test suites. The accessible population is the repositories included in a public benchmark suite that packages projects with their test suites and dependency environments, which serves as the sampling frame. Repositories are included if they have at least 80 percent developer-written line coverage and build reproducibly in the benchmark environment; repositories whose tests require network access are excluded because mutation testing must run offline. Stratified random sampling by repository size (small, medium, large by lines of code) is used to select twenty repositories, because fault-detection results are known to vary with size and the comparison should not be dominated by one stratum. Twenty repositories is the number the mutation-testing budget allows within the project timeline, and each repository contributes many mutants, so the analysis unit is the mutant nested within the repository.

6. Public Health (MPH, United Kingdom; records)

The target population is adults eligible for the health-check programme in England. The accessible population is eligible residents of one local authority who booked a health check in the financial years 2024 to 2025 and 2025 to 2026, with the authority’s booking system export as the sampling frame. Records are included if the resident’s postcode maps to the most or least deprived quintile of the authority’s index and the booking has a recorded outcome; records from general practices that joined the programme mid-way through the first year are excluded so that every practice is observed for both years. All records meeting the criteria are analysed, since the study uses routinely collected data and no subsampling is needed. The expected number of records, based on the authority’s published annual programme activity, is large enough that the comparison of attendance proportions between quintiles is not limited by power, and the analysis instead attends to practice-level clustering.

7. Engineering (MASc, Canada; instrumented houses)

The target population is detached houses in prairie Canadian cities retrofitted with cold-climate air-source heat pumps under provincial rebate programmes. The accessible population is rebate recipients in Winnipeg whose installations were completed in the two years before the study and who responded to an invitation sent by the utility on the researcher’s behalf; the utility’s rebate register is the sampling frame, although the researcher receives only the responses. Houses are included if the heat pump is the primary heating system and the owner consents to a season of monitoring; houses with wood or gas supplementary heating are excluded so that measured performance is attributable to the heat pump. Twelve houses are selected purposively from the respondents to span the range of construction decades represented in the register, because envelope condition is the moderator of interest. Twelve is the number of monitoring kits available; the study is descriptive and comparative rather than inferential, and the sample size is stated as a limitation.

8. Social Work (MSW, United States; interviews)

The target population is unlicensed kinship caregivers in the US child-welfare system. The accessible population is unlicensed caregivers known to the child-welfare agency of one Midwestern county, reached through agency caseworkers who share the invitation; the agency’s kinship caseload list is the sampling frame, held by the agency and never seen by the researcher. Caregivers are included if they have cared for a related child under an agency-involved placement for at least six months and are 18 or older; caregivers who are currently licensed or in the licensing process are excluded because their access to training and stipends is the comparison point. Purposive sampling is used to recruit caregivers who vary in relationship to the child and in placement duration, because the study seeks the range of support experiences. Recruitment continues to fifteen interviews or until two consecutive interviews add no new type of support or strain, whichever is later, and the point reached is reported.

Draft Your Methods Chapter With the Chain Already in Place

Tesify structures the population and sample subsection around the four links and five elements, and keeps the criteria aligned with your Chapter 1 scope. More than 15,000 chapters have been written with it, every word by the student.

Start your thesis free

Fill-In Template

The target population is [group defined by the characteristics in the research question]. The accessible population is [that group at these sites during this period], identified from [list, register or data source], which serves as the sampling frame.

[Units] are included if [criterion 1, with reason] and [criterion 2, with reason]; [units] are excluded if [criterion, with reason].

[Sampling technique, named precisely] is used because [reason tied to the design].

The target sample is [number], based on [power analysis with named inputs / saturation logic / the size of the accessible population], with [allowance for attrition or non-response, with basis].

Justifying the Sample Size in One Paragraph

Committees accept three kinds of justification and reject one. They accept an a priori power analysis with its inputs stated (test, effect size and its source, alpha, power), a saturation argument with a stopping rule declared in advance, and a census of a small accessible population. They reject “the sample size was determined by the time available” as the only reason, although time may be stated alongside a real justification, as in examples 5 and 7. For the power route, our sample size and power analysis guide walks through the calculation; the hypothesis form you chose changes the inputs, as our research hypothesis examples explain.

Five Mistakes Committees Flag

Naming the target population as the sample. “The population of this study is nurses in the United States” followed by a sample from one ward. Write all four links so the distance is visible.
Criteria without reasons. Every inclusion and exclusion criterion needs a clause explaining it; unexplained criteria look arbitrary or, worse, convenient.
A “random” sample with no frame. If there was no list to draw from, the sample was not random. Say what it was: convenience, consecutive, volunteer.
Population figures without a source. A number for the target population must come from a citable statistic. If you do not have one, describe the population qualitatively.
Criteria that drift from Chapter 1. The delimitations, the population in the hypotheses and the criteria here must use the same words; the variable roles you defined depend on it.

Frequently Asked Questions

What is the difference between a population and a sample?

The population is the whole group your research question is about; the sample is the subset you actually studied. In a thesis the population is further split into the target population (the whole group) and the accessible population (the part you could reach), with the sampling frame as the list between them and the sample.

Does a qualitative study need a population section?

Yes, though it is often headed “Participants.” It still names the group the study speaks about, the setting, the criteria, the purposive strategy and the stopping rule, as examples 2 and 8 show.

Can the population be documents, records or organisations rather than people?

Yes. The unit of analysis can be a repository, a record, a house or a firm, as in examples 5, 6 and 7. The same four links apply: the whole set the question is about, the set you could access, the list you drew from, and the units studied.

How do I report the sample size I actually achieved?

State the target, the number invited, the number who responded or were eligible, and the final analysed sample, with reasons for each loss. A flow diagram is expected in health disciplines and welcome elsewhere.

Where does the population and sample section go?

In the methodology chapter, after the research design and before the instruments. Chapter 1 carries a one-sentence version in the scope, and the discussion chapter returns to it when discussing generalisability.

What if I could not get a sampling frame?

Say so, name the recruitment route you used instead (social media, snowball, gatekeeper referral) and treat the absence of a frame as a limitation on representativeness. Committees penalise concealment, not constraint.

Next Step

Write the four links as four sentences, add the criteria with their reasons, name the technique and justify the size. Then read it against your Chapter 1 scope for consistency. To draft the methods chapter with the chain structured for you, start your thesis in Tesify.