School Survey Report and Carnival Game Fairness Check
How can a small sample tell us about a whole school, and is that game really fair?
The mystery: A coin landed heads 7 times in a row. Is the next flip more likely to be tails?
Questions we will figure out together in this unit:
write survey questions that collect useful data, and explain why a sample chosen at random can represent many people.
Phase 1: Spark
Do Now. The student council wants to know what students think about the spring carnival. Which question would give more useful data: “Do you like fun?” or “Which day would you come to a carnival: Friday evening, Saturday afternoon or neither?” Tell a partner why.
The Census Bureau sends its big yearly survey to about 3.5 million addresses. How do you think it chooses them?
The reading opens over the lesson; close it to come back to exactly where you were.
Who our report is for. Our School Survey Report is for the student council, teachers and families. They want facts they can act on: which day to hold the carnival, which games to run, and how students spend their time. A report only helps if the questions are clear and the sample is fair.
Good survey questions. A useful question is clear, asks about one thing, and has answers you can count or measure. “How many minutes did you spend on homework last night?” gives a number. “Don’t you agree homework is too long?” pushes people toward one answer; that is a leading question, and it makes the data less trustworthy.
Two kinds of answers. Some questions give numbers, like minutes or ages; we can find a mean, median and spread for them. Others give categories, like a favorite game; we can count them and find percents.
Worked example 1: fixing a question. Omar wrote, “Isn’t the ring toss the best game?” Step 1: it leads people toward one answer. Step 2: it only allows yes or no. Step 3: rewrite: “Which game would you most like to play: ring toss, bean bag throw, or spin the wheel?” Step 4: now every answer is a category we can count fairly.
Worked example 2: one person stands for many. The Census Bureau chooses about 3.5 million addresses at random. Sample data: if a town has 24,000 households and the survey picks 240 of them at random, Ana finds what each one stands for. Step 1: 24,000 ÷ 240 = 100. Step 2: each chosen home stands for about 100 homes. Step 3: if 60 of the 240 say they rent, 60/240 = 1/4 = 25%. Step 4: about 25% of 24,000, or 6,000 households, may rent.
Why random? When chance does the choosing, every member of the population has the same chance to be picked. Nobody can stack the sample with their friends or with people who agree with them. That is why the Census Bureau does not let people remove themselves once chosen.
Our plan. Over the next lessons we will test sampling methods, draw random samples ourselves and use them to estimate facts about the whole school.
Use the reading and the worked examples. Explain your choice each time.
Draft questions for our School Survey Report.
Write three survey questions the student council could use: one that gives numbers, one that gives categories, and one about the carnival. Label each.
Sample data: 30 of 120 randomly chosen students say they walk to school. What percent is that?
Write two new Wonder Board questions that a survey of our school could answer.
We know our audience and what makes a good question. Next we investigate how the way we choose people can make a survey fair or unfair.
Exit ticket: why does it matter that the Census Bureau chooses addresses at random instead of asking volunteers?
Pacing: Do Now 5 min, reading 8, mini lesson 12, guided practice 10, question drafting and exit ticket 5.
Accuracy: over 3.5 million households contacted, about 3.5 million addresses randomly selected, every state plus DC and Puerto Rico, addresses cannot be removed, a small sample to limit cost, and the 5-year limit are from the Census Bureau page opened 2026-11-20. The 24,000-household town is sample data. Checked: 24,000 ÷ 240 = 100; 60/240 = 25%; 25% of 24,000 = 6,000; 30/120 = 25%.
Likely misconceptions: “more volunteers is always better”; treating a yes/no opinion question as numerical data.
Supports: a question checklist (clear, one idea, not leading, countable answers). Stretch: students write a question with a numeric answer that would be hard to measure honestly and explain why.
Answer guide: a fair rewrite of the line question: “How many minutes would you be willing to wait in line for a carnival game? 0-5, 6-10, 11-15, more than 15.” The exit ticket should say random choice gives everyone the same chance and keeps the sample from being stacked.
U.S. Census Bureau, American Community Survey: Top Questions About the Survey · U.S. government work, public domain
The American Community Survey is a survey run by the U.S. Census Bureau all year, every year. The full census happens only once every 10 years, so this survey keeps information up to date for leaders across the country.
Every year the Bureau contacts over 3.5 million households. It randomly selects about 3.5 million addresses each year, in every state, the District of Columbia and Puerto Rico. Random means chosen by chance, so no person, not even the person who lives there, decides who gets picked.
The Bureau says each chosen address stands for thousands of other households like it. That is why an address cannot be removed from the sample once it is chosen. If chosen homes could drop out, the sample would quickly stop being a fair picture of the country.
The Bureau also tries to keep the sample as small as it can, to save money and to bother fewer people. It works to keep a home from getting the survey more than once in 5 years. Leaders use the results to plan things like schools, hospitals, school lunch programs and bridges.
Source: U.S. Census Bureau, American Community Survey: Top Questions About the Survey (census.gov)