Non-random Sampling
Not every sample needs a sampling frame or a random number generator. Non-random methods trade the guarantee of a known selection chance for speed and low cost — a trade that's sometimes worth making, and sometimes not.
Why "non-random" doesn't mean "unplanned"
A sampling method is non-random if at least one unit in the population has no fixed, statable chance of being selected — often because selection depends on who happens to be available, or on a researcher's own judgement, rather than on chance alone. Both methods on this course are still deliberate, structured processes — they simply can't guarantee every unit an equal, known opportunity.
Quota sampling
The population is first split into groups based on a characteristic (age, gender, occupation, and so on), and a quota — a target number to interview — is set for each group, usually in proportion to that group's size in the population. An interviewer then approaches people, assesses which group each person belongs to, and interviews them if their group's quota isn't yet full. If someone refuses, or their group is already full, the interviewer simply moves on to the next person. The process continues until every quota is filled.
Both methods split the population into groups first — the difference is entirely in the final selection step. Stratified sampling selects randomly within each group; quota sampling lets the interviewer select non-randomly (typically by whoever is available) until the quota is full. This single difference is what makes stratified sampling random and quota sampling not.
Opportunity sampling
Also called convenience sampling, this method simply takes the sample from whoever is available and willing at the time and place the study happens to be carried out — for example, the first 25 people to leave a train station on a given morning. No groups, quotas, or sampling frame are involved at all; it's the simplest sampling method to carry out, and also the most exposed to bias.
Ask: "could I state, before sampling begins, the exact probability that a specific named individual ends up in the sample?" If the answer depends on chance and can be calculated in advance, the method is random. If it depends on who happens to turn up, or on a judgement call, it isn't — regardless of how organised or "fair-sounding" the method appears.
Worked example 1
The calculation ($48$, $42$, $30$) alone isn't the whole method — a full answer must also describe how the interviewer fills each quota, since that non-random final step is what defines quota sampling.
Worked example 2
"It's not random" is true but too vague for full marks. A strong disadvantage names who is systematically excluded by the time and place of sampling, and links that to why it could bias the specific quantity being measured.
Worked example 3
Neither non-random method can promise the freedom from bias that a truly random method can — but between the two, quota sampling at least controls for known, chosen characteristics, while opportunity sampling controls for nothing at all.
Worked example 4
No sampling method is "best" in the abstract — the right choice always balances statistical rigour against time, cost, and how the results will actually be used. A non-random method used knowingly, for an appropriately modest purpose, is a legitimate choice rather than a shortcut.
Don't describe quota sampling as "random within each group" — the whole point is that the final selection is not random; it's whoever the interviewer approaches until the quota is filled. If the final step were genuinely random, the method would be stratified sampling instead.
Check yourself
A radio station wants 60 listeners' opinions and sets quotas of 24 under-25s, 21 aged 25–49, and 15 aged 50+, based on their known audience breakdown. State the name of this sampling method, and explain why it is not a random method.
(Hint: think about how listeners within each age group are actually chosen once the quotas are set.)