You are viewing a free preview of this lesson.
Subscribe to unlock all 10 lessons in this course and every other course on LearningBro.
Sampling is the process of selecting a subset of individuals or cases from a larger population for study. Because it is almost never possible to study every member of a population — there are millions of secondary-school pupils, prisoners or churchgoers in Britain — sociologists must select a smaller group that, ideally, stands in for the wider population from which it is drawn. The way a sample is selected has profound implications for the representativeness, and therefore the generalisability, of research findings: a study based on a biased sample can mislead however carefully it is conducted afterwards. Closely bound up with sampling is the question of access — how the researcher physically and socially reaches the group or setting they wish to study. The two issues are inseparable, because the practical difficulty of accessing certain groups frequently forces a researcher to abandon the ideal probability sample and fall back on a non-probability technique. This lesson sets out the main sampling methods and access problems, and — crucially — shows that sampling decisions are never purely technical: they flow directly from the positivism/interpretivism debate about what good sociological knowledge is.
Sampling and access sit within Research Methods in the Theory and Methods component of the AQA A-Level Sociology specification (7192). They are examinable on Paper 1 (7192/1) — Education with Theory and Methods and on Paper 3 (7192/3) — Crime and Deviance with Theory and Methods, where they can feature in a 10-mark "Outline and explain two…" question (for example, two problems of using snowball sampling) and as developed evaluation within the 30-mark essay on Paper 3. Sampling and access are especially load-bearing for the Methods in Context question on Paper 1 (20 marks), because the characteristics of the educational group under study — pupils who are minors, dispersed across classes, gathered behind a school's gatekeeping procedures — directly determine which sampling techniques are feasible and how representative the resulting sample can be. Whenever you are asked to evaluate a method for a specific group, the question of who you can actually reach, and whether they are typical is part of the answer.
Sampling and access are not abstract; the choices they involve have shaped the substantive studies you meet across the course.
Key Definition: Representativeness — the extent to which a sample accurately reflects the characteristics of the wider population, so that findings from the sample can be generalised to that population with confidence.
Before examining specific techniques, several terms must be precise, because examiners penalise loose use of "sample", "population" and "sampling frame".
| Term | Definition |
|---|---|
| Population | The entire group of people or cases the researcher is interested in (e.g. 'all secondary-school students in England') |
| Sample | A subset of the population, selected for study |
| Sampling frame | A list of all members of the population from which the sample is drawn (e.g. the electoral register, a school roll, a GP patient list) |
| Sample size | The number of individuals or cases included in the sample |
| Representativeness | The extent to which the sample accurately reflects the characteristics of the wider population |
| Generalisability | The extent to which findings from the sample can be applied to the wider population |
| Bias | A systematic (not random) error in sampling that results in certain groups being over- or under-represented |
Key Definition: Sampling frame — a complete list of all members of the target population from which a sample can be drawn. The quality and completeness of the sampling frame directly affects the representativeness of the sample; where no adequate frame exists, probability sampling becomes impossible.
The logic of sampling is best understood as a sequence: the researcher defines a population, locates or constructs a sampling frame that lists it, chooses a technique to draw a sample from that frame, and then hopes the sample is representative enough to generalise back to the population. The diagram below makes this flow explicit; note that a flaw at any stage — an incomplete frame, a biased technique, a low response rate — breaks the chain of generalisation.
flowchart TD
A[Define target population] --> B[Obtain or construct sampling frame]
B --> C[Choose sampling technique]
C --> D[Select the sample]
D --> E[Collect data from sample]
E --> F[Generalise findings to population]
B -. no adequate frame exists .-> G[Non-probability sampling]
G --> D
F -. weakened by bias / non-response .-> A
Probability sampling methods give every member of the population a known — and usually equal — chance of being selected. They are the positivist ideal because, by removing the researcher's discretion from selection, they minimise bias and maximise the prospect of a representative sample from which findings can be generalised. They all, however, depend on a good sampling frame.
Every member of the population has an equal chance of selection, typically by numbering everyone on the sampling frame and using a random-number generator. If a population of 1,000 yields a sample of 100, each person has a selection probability of 1000100=0.1.
| Advantage | Disadvantage |
|---|---|
| Every member has an equal chance of selection, minimising researcher bias | Requires a complete and accurate sampling frame, which may not exist |
| Straightforward to understand and implement | May produce an unrepresentative sample by chance (e.g. 100 drawn from a school might happen to contain 80 boys and 20 girls) |
| Results can in principle be generalised to the population | Time-consuming and costly with very large populations |
The researcher selects every nth person from the frame (e.g. every 10th name), with a randomly chosen starting point.
| Advantage | Disadvantage |
|---|---|
| Quick and easy to implement | Still requires a sampling frame |
| Produces an even spread across the frame | If the list has a hidden periodic pattern (e.g. alternating male/female names), the method can produce a systematically biased sample |
The population is divided into strata (sub-groups) on a key characteristic (gender, age, ethnicity, class), and a random sample is drawn from each stratum in proportion to its share of the population. This is generally the most representative technique available.
| Advantage | Disadvantage |
|---|---|
| Ensures key sub-groups appear in correct proportions, improving representativeness | Requires detailed prior information about the population to define strata |
| Allows reliable comparisons between sub-groups | More complex and time-consuming than simple random sampling |
| Reduces the risk of a freak unrepresentative sample | Requires an adequate sampling frame for each stratum |
Example: if a school is 60% female and 40% male, a stratified sample of 100 takes 60 girls and 40 boys, each drawn randomly from their stratum, guaranteeing the gender balance the population actually has rather than leaving it to chance.
The population is divided into naturally occurring clusters (schools, GP practices, postcode areas); a random sample of clusters is chosen, and then all — or a random sub-sample of — members within them are studied. Large national surveys often combine stages (randomly selecting regions, then schools, then pupils).
| Advantage | Disadvantage |
|---|---|
| Practical and cost-effective for geographically dispersed populations | Less representative than stratified sampling — chosen clusters may be atypical |
| Does not require a single complete sampling frame for the whole population | Reduces the effective sample size, widening the margin of error |
Non-probability sampling methods do not give every member an equal, known chance of selection. They are used when probability sampling is impractical or impossible — typically because no sampling frame exists (you cannot list all drug users) or because the research aim is depth rather than breadth. They are therefore the interpretivist's usual recourse, prioritising access to rich, meaningful data over statistical generalisation.
The researcher sets quotas — targets for the number of participants in each category (so many women aged 18--30, and so on) — and then fills each quota, but selection within a category is not random.
| Advantage | Disadvantage |
|---|---|
| Quick and cheap — no sampling frame needed | Not truly random, so the sample may not be representative |
| Ensures key groups are represented | The researcher's choice within each quota introduces bias (e.g. approaching only approachable-looking people) |
| Widely used in market research and opinion polling | Findings cannot be generalised with the confidence of a probability sample |
The researcher uses their own judgement to select participants particularly relevant to the question — common in qualitative work where specific knowledge or experience matters more than typicality.
| Advantage | Disadvantage |
|---|---|
| Targets individuals with the precise knowledge, experience or characteristics required | Highly subjective — the researcher's judgement decides who is included |
| Useful for studying rare or hard-to-reach populations | Cannot be generalised to the wider population |
| Efficient — no time wasted on irrelevant respondents | Introduces researcher bias |
The researcher finds one or a few initial participants who refer others, who refer others again — the sample grows like a snowball. It is the standard solution where the population is hidden.
| Advantage | Disadvantage |
|---|---|
| Reaches hidden, deviant or hard-to-reach populations (drug users, sex workers, undocumented migrants, professional criminals) | Not representative — people refer others like themselves, so whole sub-groups may be missed |
| Does not require a sampling frame | The researcher has limited control over who ends up in the sample |
| Builds trust — referral by a known person eases cooperation and disclosure | Early contacts have a disproportionate influence on the final sample's composition |
Example: Laurie Taylor used snowball sampling to study professional criminals, beginning with the former armed robber John McVicar, who introduced him to others — a vivid illustration of how access drives method when no frame of "criminals" can exist.
The researcher simply selects whoever is available and willing — the least rigorous technique.
| Advantage | Disadvantage |
|---|---|
| Quick, easy and cheap | Highly unlikely to be representative |
| Useful for exploratory or pilot work | Findings cannot be generalised |
| Strong bias — only those who happen to be in a particular place at a particular time are included |
The researcher advertises and those who come forward form the sample (common in media-recruited and some survey research, and the root of the Hite reports' problems on the questionnaire).
| Advantage | Disadvantage |
|---|---|
| Easy to recruit; participants are cooperative and engaged | Volunteer bias — volunteers differ systematically from non-volunteers (more motivated, more opinionated), so the sample is unrepresentative |
Access is the researcher's ability to reach and study the target group, and it is frequently the single hardest practical obstacle in a project — so much so that it often dictates the sampling method (forcing snowball or opportunity sampling) and therefore caps representativeness from the outset.
A gatekeeper is a person or body with the power to grant or deny access to a group, setting or information.
| Examples of Gatekeepers | Context |
|---|---|
| Head teachers | Access to schools and pupils |
| Prison governors | Access to prisons and prisoners |
| Gang leaders (e.g. JT for Venkatesh) | Access to criminal or deviant groups |
| Managers | Access to workplaces and employees |
| Parents | Access to children (also required for consent) |
| Community / religious leaders | Access to ethnic-minority or faith communities |
| Ethics committees | Formal approval to research human participants |
| Issue | Explanation |
|---|---|
| Selective access | Gatekeepers may open some doors and shut others, biasing the sample (a head may offer only well-behaved classes) |
| Control over the research | They may steer what the researcher sees, whom they meet and what may be asked |
| Conditions and restrictions | They may demand approval of questions or the right to review findings, compromising the study's independence |
| Power dynamics | The researcher depends on the gatekeeper's goodwill, an unequal relationship that can shape conclusions |
| Ethical implications | A gatekeeper consenting on behalf of others (a head teacher for pupils) raises the question of whether participants gave genuine informed consent |
Key Definition: Gatekeeper — a person or body that controls access to a group, institution or setting. Gatekeepers can facilitate or obstruct research, and their influence on what the researcher can see and whom they can sample must be weighed when judging the quality and representativeness of the data.
The AO3 point to carry into essays is that access and representativeness are linked in a chain: where access is hard, the researcher is pushed toward non-probability methods (snowball, opportunity, purposive), which sacrifice representativeness; and where access runs through a gatekeeper, the gatekeeper's selectivity can bias even a nominally careful sample. Recognising this mechanism — rather than merely listing "access is a problem" — is what lifts an evaluation.
Subscribe to continue reading
Get full access to this lesson and all 10 lessons in this course.