📋 Teacher Cheat Sheet — Session 3: Designing Better Surveys

Data Science for Young Minds · Grade 4 · Ages 9–10

~60 min Ages 9–10 Session 3 of 8 ND-Friendly

⏱ Session Agenda

Time	Block	What's Happening
0–5	🎯 Hook	Show a badly worded survey question. Students try to answer it — and get confused. Discuss why.
5–20	📖 Lesson	Likert scales · Multiple-choice vs. open-ended · Recognizing leading and loaded questions · Neutral wording
20–42	🎮 Activity	Redesign 3 poorly written surveys; then pilot revised questions with a partner
42–52	💬 Debrief	Share redesigns — class votes on which version is clearest; discuss piloting experience
52–58	✍️ Write	Students draft 2 original survey questions for their capstone topic
58–60	👋 Close	"What's one thing a good survey question never does?" Exit ticket.

Pacing note: The piloting step (students asking their partner the revised questions) is where the real learning happens. Give it at least 8 minutes — hearing confusing answers live is more powerful than any lecture.

📦 Materials Needed

Prepare before class:

Printed worksheets (one per student) Pencils 3 "bad survey" cards (projected or printed) Sticky notes for class redesign gallery (optional)

💡 Frame revision as professional: "Real researchers always test their questions before launching a survey. Today you are real researchers."

📚 Key Vocabulary

Likert scale — a rating scale (e.g., 1–5) measuring agreement or frequency

Pilot — to test a survey on a small group before full launch

Revision — improving a question based on what you learned from testing

Neutral — wording that doesn't push the respondent toward any answer

Leading question — a question that hints at or pushes toward a particular answer

💬 Discussion Questions + Teacher Notes

"Don't you think homework is too long?" — Is this a good survey question? Why not?"
→ It's leading — it tells respondents how to feel before they answer. Neutral version: "How do you feel about the amount of homework you receive each night?" This introduces "leading question" concretely before defining it.
"What's the difference between a Likert scale and a yes/no question?"
→ Likert gives a range of opinion (strongly disagree → strongly agree). Yes/no forces a binary. Likert captures nuance — useful when feelings exist on a spectrum. Ask: "What might you miss if you only asked yes/no?"
"Why do researchers pilot their survey before sending it to everyone?"
→ You discover confusing words, ambiguous phrases, questions people interpret differently than you intended. Finding out early = fixing it before bad data ruins your study. Framing: "Piloting is not a sign the survey was bad — it's a sign you're being careful."
"What's wrong with an open-ended question on a survey of 200 people?"
→ Hard to analyze — every answer is different, you can't count or graph them easily. Open-ended = rich but hard to summarize. Multiple-choice = limited but easy to organize. Both have trade-offs.
"Who did you think of when you wrote your question? Who might answer differently than you expect?"
→ This builds perspective-taking. A question clear to the author may confuse a reader. Piloting forces you to consider the respondent's point of view — the same skill needed in interpretation.

🎮 Survey Redesign Activity — The 3 Bad Surveys

Students work in pairs to redesign each bad question. Then pilot their revision by actually asking their partner and recording the answer.

❌ Bad Survey 1:
"Don't you think recess should be longer?"
Problem: leading question

✅ Better version:
"How long do you think recess should be? (circle: 15 / 20 / 30 / 45 / 60 minutes)"
Neutral, gives options

❌ Bad Survey 2:
"What do you think about food?"
Problem: too vague, open-ended without purpose

✅ Better version:
"On a scale of 1–5, how satisfied are you with the school lunch menu? (1=very unhappy, 5=very happy)"
Specific, uses Likert scale

❌ Bad Survey 3:
"Do you sometimes maybe occasionally read books at home?"
Problem: double-barreled / confusing

✅ Better version:
"How many books did you read at home last month? (circle: 0 / 1 / 2 / 3 / 4+)"
One clear question, specific answer choices

🎯 Opening Hook

Project or read aloud: "Do you agree that our terrible, boring school lunches should be completely changed immediately?"

Ask students to try to answer it. Watch the confusion. Then ask: "What's wrong with this question?"

→ Students will identify the loaded language, the assumption ("terrible"), and the forced implication. This gives you all the vocabulary hooks you need.

📏 Likert Scale Reference

5-point agreement scale:
1 = Strongly disagree
2 = Disagree
3 = Neutral / not sure
4 = Agree
5 = Strongly agree

5-point frequency scale:
1=Never · 2=Rarely · 3=Sometimes · 4=Often · 5=Always

Post this on board. Students reference it during their redesigns.

🧠 ND-Friendly Tips

Compare old vs. new side by side — always show the bad version and the improved version together. The contrast is the lesson.
Frame revision as professional — "Real researchers always test their questions. You are doing exactly what scientists do."
Structured pilot script — Give students these words: "I'm going to read you a survey question. Please tell me your answer and whether the question was clear." Reduces social anxiety.
Allow written pilot responses — If verbal piloting is stressful, the partner can write their answer and circle "Clear" or "Confusing."
One redesign at a time — Do Bad Survey 1 together as a class before releasing students to redesign 2 and 3 independently.