Mathematics › 🔄 Grade 7 › Sampling: How a Small Group Can Reveal a Big Population

Sampling: How a Small Group Can Reveal a Big Population

Draw random samples from a hidden population and watch the estimate settle down.

Grade 7Middle School

Sampling: How a Small Group Can Reveal a Big Population — illustration

💡

The big idea: A well-chosen random sample lets you learn about an entire population without checking every member. The larger and more random the sample, the more reliable — and less variable — your estimate becomes.

🎯 By the end, you'll be able to

Distinguish a sample from the population it's drawn from
Explain why random sampling tends to produce a representative sample, while convenience sampling can be biased
Use a sample statistic, like the sample mean, to make an inference about an unknown population characteristic
Recognize that estimates from different random samples vary, and that increasing sample size reduces that variation
Judge whether a described sampling method is likely to produce a representative sample

📎 You should already know

Mean and measures of center (Grade 6)
Range, IQR, and MAD (Grade 6 Data Spread)

Why sample at all?

Suppose an online test-prep course wants to know the average score of everyone who took its practice numerical reasoning test last month — tens of thousands of people. Grading and averaging every single attempt is possible but slow and expensive.

Instead, statisticians study a smaller group — a sample — and use it to make an informed guess, or inference, about the entire population. This works surprisingly well, as long as the sample is chosen carefully.

Population vs. sample

The population is the entire group you actually want to know something about — every test-taker, every 7th grader in a district, every lightbulb a factory makes today. The sample is the smaller subset you actually measure.

You use the sample to calculate a statistic (like a sample mean), and then you use that statistic to estimate the matching value for the whole population.

🔑 Random samples tend to represent the population

A random sample gives every member of the population an equal, fair chance of being chosen — for example, picking 30 student ID numbers with a random number generator. Because no group is systematically favored or excluded, a random sample tends to look like a smaller version of the whole population.

A convenience sample (whoever happens to be nearby) or a voluntary sample (whoever chooses to respond) is usually biased — it systematically over- or under-represents certain groups, no matter how large it gets.

Using a sample to estimate a population

If a random sample of 40 test-takers has a mean score of 71, the best estimate for the entire population's mean score is also around 71 — but it's an estimate, not an exact match. A different random sample of 40 people would likely give a slightly different mean, like 69 or 73.

That sample-to-sample bounce is normal. The question isn't whether it happens — it's how big the bounce tends to be, and how to shrink it.

🎮 Random Sampling — Watch the Estimate Settle LIVE

This starts as a random sample of 5 practice-test scores (out of 100) drawn from a much larger population. Click 'Add random point' repeatedly to grow the sample and watch the mean bounce less and settle down as n increases. Then try manually adding several low (or high) scores with the slider + 'Add point' — see how a biased handful of choices drags the mean away from where the random sample was heading.

✨ Bigger random samples bounce less

As a random sample grows, unusually high and unusually low values tend to balance each other out, so the sample mean lands closer and closer to the true population mean each time you repeat the process. This is why a random sample of 400 voters gives a far more dependable estimate of an election result than a random sample of 10.

Simulating many samples of the same size — and looking at how much their means vary — is exactly how statisticians gauge how much to trust a single sample's estimate (7.SP.2).

⚠️ A bigger biased sample is still biased

Sample size only fixes variability — it does nothing to fix bias. If a pollster only calls landline phone numbers, calling 20,000 landlines instead of 200 still systematically misses everyone who only uses a cellphone. The result is confidently wrong, not just noisy.

Before trusting any sample statistic, ask: was this sample random, or could the method itself have left certain groups out?

📝 Worked example: A student wants to know the average number of hours 7th graders at her school sleep on a school night. She surveys the 12 students on the volleyball team. Is this sampling method likely to give a representative sample? Why or why not?

The population is all 7th graders at the school; the sample is the 12 volleyball players.
Volleyball players share a practice schedule and may have systematically different sleep habits than 7th graders in general.
This is a convenience sample based on team membership, not a random sample of the whole population.

✓ No — it's likely biased. A random sample drawn from the entire 7th grade (for example, using randomly selected student ID numbers) would better represent the population.

📝 Worked example: Four different students each take a random sample of 10 scores from the same set of practice-test results (the population mean is unknown). Their sample means are 71, 74, 69, and 76. What do these results tell you about the population mean?

Each sample mean is only an estimate — no single sample of 10 will exactly match the population mean.
The four estimates cluster between about 69 and 76, so the true population mean is likely somewhere in or near that range.
Averaging the four sample means, (71+74+69+76) ÷ 4 = 290 ÷ 4 = 72.5, gives an even better combined estimate.

✓ The population mean is likely close to 72–73. No single sample of 10 is exact, but combining several random samples narrows the estimate.

📝 Worked example: One class takes a random sample of 5 students' heights to estimate the average height of all 400 students in the school; another class takes a random sample of 100 students. Which estimate is more likely to be close to the true school-wide average, and why?

Both samples are random, so both are unbiased in method.
A sample of 100 captures far more of the natural variation in height across the school than a sample of just 5.
Larger random samples produce sample means that vary less from sample to sample, so they tend to land closer to the true population mean.

✓ The sample of 100 is more likely to be close to the true average — bigger random samples reduce the variability in the estimate.

Check your understanding

1. A city wants to know the average commute time for all 50,000 residents. Inspectors survey 500 randomly chosen residents. What is the population, and what is the sample?

The population is the entire group you want to know about (all 50,000 residents); the sample is the smaller group actually measured (the 500 surveyed).

2. Which sampling method is most likely to produce a representative sample of a school's students?

Randomly selecting from the complete enrollment list gives every student an equal chance of being chosen, avoiding the systematic bias of convenience or self-selected groups.

3. As you increase the size of a random sample (keeping the sampling method random), what typically happens to the variation between different samples' means?

Larger random samples tend to average out unusual individual values, so sample means from larger samples cluster more tightly around the true population mean — the variation decreases (though it doesn't vanish entirely).

4. A pollster wants the average opinion of an entire state but only calls landline phone numbers, missing most young adults who use cellphones only. Increasing this sample from 200 to 20,000 landline calls will...

A biased sampling method stays biased no matter how large it gets — a bigger sample just repeats the same systematic exclusion (here, of cellphone-only young adults) more times.

5. A random sample of 40 lightbulbs from a factory's daily production has a mean lifetime of 812 hours. What is the best conclusion?

A random sample statistic like the mean gives a reasonable estimate — not a guarantee — of the corresponding population value. A different random sample would give a similar, but not identical, estimate.

✅ Key takeaways

Statisticians study a sample to learn about a population that's too large or costly to measure completely.
A random sample gives every member of the population a fair chance of being chosen, which tends to make it representative.
Convenience or voluntary samples are usually biased because they systematically leave out or over-include certain groups.
A sample statistic, like the sample mean, is only an estimate of the population's true value — different random samples give somewhat different estimates.
Larger random samples produce estimates that vary less from sample to sample, making them more trustworthy.