Lesson 115 — Planning a Statistical Investigation

Strand: Statistics | Descriptor: AC9M7ST03 | Duration: 45 minutes

Block note. Lessons 115–120 form one connected investigation project: plan (115), collect (116), analyse (117), report (118), review (119), consolidate (120). Students keep a single project folder throughout.

Learning Intentions

  • To understand the statistical investigation cycle.
  • To write a statistical question that data can actually answer.

Success Criteria

I can:

  1. Name and describe the four stages of a statistical investigation.
  2. Distinguish a statistical question from a non-statistical one.
  3. Define the variable, population and sample for my investigation.
  4. Write a practical data collection plan.

Warmup

(6 minutes — which can data answer? pairs)

Decide whether each question could be answered by collecting data, and if not, why not.

  1. How tall is our tallest student?
  2. Do Year 7 students spend more time on screens than Year 9 students?
  3. Is maths more important than art?
  4. How many hours of sleep does a typical Year 7 student get?
  5. What is the best colour?

Answers: 1. Yes, but trivially — one measurement, no variability; 2. Yes — a genuine statistical question; 3. No — a value judgement, not a measurable quantity; 4. Yes — “typical” implies variability; 5. No — a matter of taste.

The distinction to name: a statistical question anticipates variability in the answers. “How tall is Sam?” has one answer. “How tall are Year 7 students?” has many, and needs data to summarise.

Activities

Activity 1 — Explicit Instruction: the Investigation Cycle (12 min)

The four stages — the statistical counterpart to the modelling cycle from Lesson 74:

StageWhat happensKey question
PoseWrite a statistical question; define variable, population, sampleWhat exactly am I asking?
CollectPlan and gather data accuratelyIs my data trustworthy?
AnalyseDisplay and summarise the distributionWhat does the data show?
InterpretAnswer the question; state limitationsWhat can I honestly conclude?

A good statistical question has four properties:

  1. It anticipates variability — the answers will differ.
  2. It names a measurable variable.
  3. It specifies a population.
  4. It is answerable with data we can actually collect.

I do — improve a weak question, live:

Weak: “Do people like sport?”

  • Variability ✓ but “like” is not measurable, “people” is not a population.

Better: “How many hours per week do Year 7 students at this school spend playing sport?”

  • Variable: hours per week (continuous). Population: Year 7 at this school. Sample: our class. Answerable ✓

We do — improve these together:

  1. “Are boys taller?”
  2. “Is homework too much?”
  3. “What do students eat?”

(Sample improvements: “How does the arm span of Year 7 boys compare with that of Year 7 girls at this school?”; “How many minutes of homework do Year 7 students at this school do on a typical weeknight?”; “How many pieces of fruit do Year 7 students eat in a day?“)

Activity 2 — Pose Your Question (16 min)

Pairs. The project begins here; this plan is carried into Lesson 116.

Your investigation. Choose a question your class can answer with data collected in one lesson.

Suggested areas (or propose your own for approval):

  • Reaction times (ruler-drop test)
  • Arm span, height, or hand span
  • Time spent on homework, sleep, or screens
  • Number of letters in names; number of siblings
  • Estimation accuracy (estimate then measure a length)
  • Memory span (digits recalled)

Write your project plan, covering all six points:

  1. The question — precisely worded, anticipating variability.
  2. The variable — what exactly is measured, and its units.
  3. Discrete or continuous — with justification.
  4. Population and sample — who you want to know about, and who you will measure.
  5. Collection method — the exact procedure, including precision.
  6. Prediction — what you expect to find, and why.

Circulating prompts:

PromptPurpose
Would two people measuring the same student get the same number?Forces a precise procedural definition.
Could you collect this in one lesson?Practicality — ambitious questions fail at the collection stage.
Who is your population, and is your sample really it?Sample-versus-population honesty (Lesson 107).
What will your data look like if your prediction is right?Makes the prediction testable, not decorative.
Is your variable a number?Categorical variables limit the analysis to the mode only.

Two comparison questions are worth encouraging — e.g. “Do students who play a musical instrument have faster reaction times?” — because they lead to back-to-back displays in Lesson 117. But warn: comparison needs two samples, doubling the collection.

Activity 3 — Inquiry: what Could Go Wrong? (11 min)

Pairs swap plans and stress-test them.

Read another pair’s plan and answer:

  1. Could you follow their collection method exactly, without asking questions?
  2. Name one thing that would make two measurers disagree.
  3. What is their population, and does their sample fairly represent it?
  4. Name one source of bias in how they plan to collect.

Socratic scaffolding — the bias discussion:

PromptPurpose
If you ask only your friends, who is missing?Selection bias — the sample is not representative.
If students report their own screen time, what might happen?Under- or over-reporting — self-report bias.
If you measure reaction times only after lunch, what varies?A confounding factor — time of day.
If the loudest students volunteer first, what happens?Volunteer bias — willing participants may differ systematically.
Can bias always be removed?Rarely. It can be reduced and must be reported — the honesty standard from Lesson 76.

Then: each pair revises their plan in response to one comment received, and notes the revision.

Checks for Understanding

(5 minutes — exit ticket, collected with the plan)

  1. What makes a question statistical rather than not?
  2. Name the four stages of a statistical investigation.
  3. Rewrite as a statistical question: “Do students like reading?”
  4. For your own investigation, state the variable, its units, and whether it is discrete or continuous.
  5. Reasoning. Name one source of bias in your collection plan and how you will reduce it.

Answers: 1. It anticipates variability in the answers and names a measurable variable and a population; 2. Pose, collect, analyse, interpret; 3. E.g. “How many minutes per day do Year 7 students at this school spend reading for pleasure?”; 4–5. Student’s own, marked against the plan.

Common Misconceptions

MisconceptionHow to pre-empt it
A question with one answer is statistical.The warmup’s Q1 — no variability, no investigation.
Vague variables (“happiness”, “how much”).The measurable-variable requirement and the two-measurers test.
Sample and population treated as the same.Named separately in the plan; Lesson 107’s vocabulary.
Believing bias can always be eliminated.It is reduced and reported, not erased.
Planning a question that cannot be collected in the time available.The practicality prompt during circulation.
A prediction with no reasoning behind it.Point 6 requires “and why”.

Enrichment — Competition-Style Problems

E1 (Kangaroo style). Which is a statistical question: “How many students are in Year 7?” or “How many siblings do Year 7 students have?”

Answer

The second — it anticipates variability. The first has a single fixed answer.

E2 (AMC Junior style). A survey asks “Do you agree that our canteen is excellent?” Name the problem and rewrite it neutrally.

Answer

It is a leading question — the wording invites agreement. Neutral version: “How would you rate the canteen on a scale of to ?”

E3 (Challenge). A student surveys the school’s fastest runners about weekly exercise and concludes Year 7 students exercise for hours a week. Identify the flaw and its effect.

Answer

Selection bias — the sample is drawn from an unusually active subgroup, so the estimate is far too high for the whole population. The conclusion is about fast runners, not Year 7 students.

E4 (Challenge). Two students investigate screen time. One asks students to estimate; the other collects phone screen-time reports. Which data is more trustworthy, and what does the other risk?

Answer

The phone reports — they are measured, not recalled. Estimates risk under-reporting (social desirability) and simple memory error. But phone data misses other screens, so neither is complete; the limitation should be stated either way.

Homework

  1. Explain in one sentence what makes a question statistical.
  2. Rewrite each as a good statistical question: (a) “Is our library popular?” (b) “Do people sleep enough?” (c) “Are Year 7s good at estimating?”
  3. For each of your Q2 questions, state the variable, its units, and whether it is discrete or continuous.
  4. For your class investigation, write out the full collection procedure so precisely that a stranger could follow it. Include the precision to be used.
  5. Name the four stages of a statistical investigation and write one sentence on each.
  6. Identify the bias in each: (a) surveying only students in the computer lab about screen time (b) asking “Don’t you agree homework is excessive?” (c) measuring reaction times only in the last five minutes of a lesson.
  7. Write a prediction for your investigation, with a reason.
  8. Reasoning. Explain the difference between a population and a sample, using your own investigation as the example.
  9. Reasoning. Why must a collection procedure state the precision to be used?
  10. Challenge. Design a study to answer: “Do Year 7 students underestimate short lengths?” Specify the question, variable, procedure, and one source of bias with a mitigation.

Answers: Q2 — e.g. (a) “How many times per week do Year 7 students visit the library?” (b) “How many hours of sleep do Year 7 students get on a typical school night?” (c) “How close are Year 7 students’ estimates of a cm length to the true value?” Q6 — (a) selection bias: computer-lab students likely have higher screen time (b) a leading question (c) a confounding factor: fatigue or restlessness at lesson’s end. Q9 — without it, different collectors round differently, so values are not comparable and the data set mixes precisions. Q10 — e.g. show a cm strip, ask each student to write an estimate in centimetres before measuring; variable: estimate minus true length (continuous, cm); bias: students who see others’ answers are influenced — collect estimates privately and simultaneously.