Not sure which lesson to do first? Find your level and get a plan →

Math · Statistics & Probability · Grade 6-8

Sampling Methods

Sampling methods are the systematic ways to select a subset of a population: simple random (SRS), stratified, cluster, and systematic are probability methods; convenience sampling is a biased non-probability method.


Quick Answer

Sampling methods are the systematic ways to select a subset of a population: simple random (SRS), stratified, cluster, and systematic are probability methods; convenience sampling is a biased non-probability method. Use them whenever you can't ask everyone and must decide whom to ask: students for a school survey, light bulbs to test, voters for a poll. The method decides whether the sample looks like the population. Random methods give every member a known chance of being picked; grabbing whoever is handy does not, and the result can be biased no matter how many people you ask.

Why This Matters

A survey of 50 students can describe a school of 2,000 well, but only if those 50 were chosen fairly. Pick them from the gym after a basketball game and you learn about basketball fans. The right method also saves effort: a stratified sample guarantees every grade is included, and a cluster sample lets you survey whole classrooms instead of tracking down scattered students.

Intuitive Explanation

A school has 1,000 students and wants to know how long they spend on homework, using a sample of 50. Asking the first 50 students through the door at 7:30 a.m. is easy, but early arrivals may be more organized, or may be there for morning sports practice. That is a convenience sample, and its answer can be off however carefully the questions are asked.

Random methods fix this in different ways. A simple random sample (SRS) numbers all 1,000 students and lets a random number generator pick 50, so every group of 50 is equally likely. A stratified sample first splits the school into grades, the strata, and picks randomly within each: if sixth graders are 30% of the school, 15 of the 50 come from sixth grade. A cluster sample randomly picks a few whole homerooms and surveys everyone in them. A systematic sample takes a random starting point on the class list and then every 20th name.

Stratified and cluster sampling sound alike but make opposite moves. Stratified takes some students from every group, which guarantees each grade appears. Cluster takes every student from a few groups, which is cheaper but works well only if each homeroom is a mix of the whole school. Neither is about treatments: deciding by coin flip which sampled students try a new study app is experimental design, not sampling.

Think of it like this. Sampling is tasting soup. Stir the pot and one spoonful tells you about the whole pot, as a random sample does. Skim only the top and you taste the oil that floated up, like a convenience sample. A spoonful from each layer is a stratified sample. Whether the pot was stirred matters far more than the size of the spoon.

Core idea. Sampling methods are the systematic ways to choose a subset of a population — SRS, stratified, cluster, systematic — versus biased convenience sampling.


When to Use

Use sampling methods when a question asks how to choose participants, whether a sample is representative, or why survey results might be biased. Pick SRS when you have a full list and no key subgroups, stratified when subgroups such as grades or regions must all appear, cluster when the population comes in natural groups that each look like the whole, and systematic when working down a list. If the question is about assigning treatments to people already in a study, that is experimental design.

Tip. Ask: Is the focus on the RULE for choosing who enters the sample (rather than how to assign treatments or compute a statistic)?

How to Recognize It

Before you name or judge a sampling method, look at how the population is arranged and how each member got its chance to be picked.

  1. Did every member of the population have a chance to be picked, decided by chance rather than by the person collecting data?

    If yes, it is a probability method: simple random, stratified, cluster or systematic. If people volunteered or were simply easy to reach, it is a convenience or voluntary sample and likely biased, no matter how many responded.

  2. Is the population split into groups, such as grades, with the sample taking some members from every group?

    If yes, it is stratified. If instead a few whole groups, such as three homerooms, are chosen and everyone in them is surveyed, it is cluster sampling; the two moves are opposites.

  3. Is the sample made by taking every kth name or item from a list after a random start?

    If yes, it is systematic. Check the list for a repeating order, such as every 10th seat on a bus being a window seat, because then the sample can miss part of the population.

  4. Could some members of the population be missing from the list the sample is drawn from, such as households without a phone?

    If yes, even a perfectly random draw from that list leaves them out; this is called undercoverage, a form of sampling bias. Fix the list first, or the sample will not represent the people it misses.

Sampling Methods compared with Experimental design, Sampling bias, and Stratified vs cluster

These rows all touch how the people in a study are chosen or used, which is why they blur together. Ask what is being decided: who gets in (a sampling method), what each person receives (experimental design), or what went wrong when the choosing was unfair (sampling bias).

Sampling Methods

Meaning
Use this when you are deciding how to select individuals from a population so the sample can fairly represent it.
Key test
Is the focus on the RULE for choosing who enters the sample (rather than how to assign treatments or compute a statistic)?
Example
A school of 10,000 students wants the average GPA from a sample of 50. The school wants each grade (9-12) fairly represented. Which method?

Experimental design

Meaning
Concerns assigning treatments to subjects already in a study, not selecting who enters it.
Key test
Use when imposing treatments to test cause and effect.
Example
Randomly giving plants fertilizer or water

Sampling bias

Meaning
The distortion that results from a bad method; the methods aim to prevent it.
Key test
Use when describing the error a poor method causes.
Example
Online poll over-representing internet users

Stratified vs cluster

Meaning
Stratified samples a few from EVERY group; cluster samples ALL from a few groups — opposite moves.
Key test
Use stratified to guarantee subgroup coverage, cluster for logistical ease.
Example
Stratified: 10 per grade; cluster: all of 5 random classrooms

Formula & Notation

Worked Examples

Example 1 — Sampling student GPA

Easy

A school of 10,000 students wants the average GPA from a sample of 50. The school wants each grade (9-12) fairly represented. Which method?
Show the steps and the answer
  1. Notice the requirement that each grade be fairly represented.

    An SRS of 50 could, by chance, include very few 12th graders; stratifying rules that out.

  2. Split the school into four strata: grades 9, 10, 11 and 12.

    GPA may differ by grade, so each grade should appear in its proper share.

  3. Allocate the 50 by grade size. With grades of 3,000, 2,600, 2,400 and 2,000 students, take 15, 13, 12 and 10.

    Each grade's share of the sample matches its share of the school; for 9th grade, 300010000×50=15.

  4. Select randomly within each grade.

    Random choice inside each stratum keeps favoritism out of the sample within every grade.

Answer: Stratified random sample

Takeaway: When named subgroups must all appear, stratify first, then sample randomly inside each one.

Example 2 — Surveying the cafeteria line

Standard

Instead, the school surveys the first 50 students in the lunch line. Is this a fair sample of all 10,000?
Show the steps and the answer
  1. Ask whether every student had a chance to be picked.

    Students with a later lunch period, students who bring lunch elsewhere and students who skip lunch had no chance at all.

  2. Ask whether the first 50 in line differ from the rest: they probably share one lunch period and perhaps one grade.

    If the earliest lunch period is mostly 9th graders, the sample over-represents them and their GPAs.

  3. Name the method: a convenience sample, chosen because it was easy.

    Ease of access, not chance, decided who was included, so the bias can't even be measured.

Answer: No — it's a biased convenience sample

Takeaway: If some members of the population had no chance of being chosen, the sample can't be trusted to represent them.

Example 3 — Find the mistake: big but biased

Application

A town of 40,000 adults wants to know how many support a new skate park. A website poll gets 4,000 responses, 70% in favor. A phone survey of 400 randomly chosen adults finds 45% in favor. Lucas writes: "The website poll is ten times bigger, so 70% is the better estimate." Find the mistake and correct it.
Show the steps and the answer
  1. Ask how each sample was chosen: the website poll took anyone who chose to answer; the phone survey picked adults at random.

    A self-selected sample attracts the people who care most, and skaters and their families care a lot.

  2. Compare the sizes: 4,000 is 10 times 400.

    Size reduces random error but cannot remove bias; 4,000 eager supporters are still eager supporters.

  3. Check how far a random sample of 400 usually misses: by about 1400=0.05, 5 percentage points, or less.

    So true support is likely between 40% and 50%, and the website's 70% is far outside that range.

  4. Choose the estimate from the random sample: 45%.

    A smaller random sample beats a larger biased one.

Answer: The phone survey's 45% is the better estimate; the website's 70% comes from a self-selected sample, and its size doesn't fix that bias.

Takeaway: A big sample chosen badly stays biased; how you choose matters more than how many you choose.

Where You'll See It

  • The 1936 Literary Digest poll

    The magazine mailed about 10 million ballots, using lists such as car registrations and telephone books, and got about 2.4 million back predicting that Alf Landon would beat Franklin Roosevelt. Roosevelt won in a landslide. George Gallup, polling only about 50,000 people chosen more carefully, called the winner.

  • TV ratings from a panel of homes

    Ratings companies don't track every television. They recruit a panel of homes chosen to match the country by region, age, income and household size, a form of stratified sampling, then estimate the whole country's viewing from those homes.

  • Quality checks on a bottling line

    A factory filling 12,000 bottles a day tests every 200th bottle, a systematic sample of 60. That works unless a fault repeats on the same cycle, for example a filler nozzle that fails every 200th bottle, which the sample would hit every time or never.

Common Mistakes

  1. The slip: Equating a large sample with a representative one

    Instead: a biased method (convenience) stays biased no matter how big it gets.

  2. The slip: Mixing up stratified and cluster sampling

    Instead: stratified takes some from every subgroup; cluster takes everyone from a few chosen subgroups.

  3. The slip: Calling systematic sampling random

    Instead: picking every 10th person is only unbiased if the list has no hidden periodic pattern.


Mini Practice

Try these on your own. Tap Reveal when you want to check.

  1. A school has a numbered list of 600 students and wants a systematic sample of 30. How far apart are the picks? If you start at student 7, what are the first three picks?

    Hint: Divide the list size by the sample size to get the gap.

  2. A school has 300 sixth graders, 200 seventh graders and 100 eighth graders. It wants a stratified sample of 60, in proportion to grade size. How many from each grade?

    Hint: What fraction of the whole school is the sample?

  3. To learn how often adults in a town exercise, a reporter surveys the first 100 people leaving a gym. Name the method and say whether it is biased.

    Hint: Who is easy to reach here, and are they typical?

  4. A school with 20 homerooms wants opinions. Plan A: randomly pick 4 homerooms and survey every student in them. Plan B: randomly pick 3 students from every homeroom. Which plan is stratified and which is cluster?

    Hint: Does the plan take some people from every group, or everyone from a few groups?

  5. A class list alternates boy, girl, boy, girl, starting with a boy. You pick every 10th name, starting with name 1. What goes wrong?

    Hint: Write out the first few positions you would pick. Are they odd or even?

  6. An online poll about a new law gets 50,000 votes. Kai says: "With 50,000 votes, the poll must represent the whole country." Correct Kai.

    Hint: Who chose to take part in this poll?

Want the full set?

50 practice questions for this concept — free to try, every one with a complete worked solution showing the why, not just the answer.

Frequently Asked Questions

How can 1,000 people tell us what millions of voters think?

What matters is how the sample is chosen, not what fraction of the population it covers. A random sample of 1,000 is accurate to within about 3 percentage points whether the population is 50,000 or 300 million people. Think of tasting soup: one spoonful works for a small pot or a huge one, as long as the pot is stirred. A biased sample of a million people can still be badly wrong.

How do you actually pick a simple random sample?

Number every member of the population, then let chance choose the numbers. You could put all the names in a hat and mix them well, use a random number generator, or use a table of random digits. Skip any number that repeats or is out of range. The researcher never chooses by hand, because people tend to pick friendly faces, nearby seats or names they like without noticing.

Why is a voluntary online poll a bad sample?

Only people who care strongly choose to answer, so their opinions are not typical. If a website asks whether homework should be banned, students who hate homework rush to vote while the rest ignore it. The poll might show 85% in favour even if most students do not mind homework. Because people pick themselves, this voluntary response bias does not shrink as more votes come in.

When would you use cluster sampling instead of a simple random sample?

When the population is spread out and reaching individuals is expensive. To survey students in a city with 40 schools, you could randomly choose 4 schools and survey every student in them, instead of travelling to all 40 for a few students each. It works well if each school is a lot like the whole city. If schools differ a lot, for example between rich and poor neighbourhoods, a stratified sample is safer.

Learning Path

Sampling Methods

You are here

You need Sampling Bias, Representativeness and Randomness first: why a skewed selection misleads, what it means for a sample to look like its population, and how chance removes favoritism. This concept adds the four probability methods and when each fits. Next, Experimental Design uses randomness again, this time to assign treatments to the subjects you have sampled.

See Also