103i - Data Collection Techniques

Lesson 103 — Explicit Instruction: Census, Sampling, Experiment and Observation

Strand: Statistics | Descriptor: AC9M8ST01 | Duration: 45 minutes

Learning Intentions

• To distinguish census, sampling, experiment and observation as techniques for collecting data.
• To identify real Australian examples of each technique.

Success Criteria

I can:

1. Define census, sample, experiment and observation.
2. Classify a described data collection scenario by technique, with a justification.
3. Give an Australian real-world example of each technique.
4. Explain the difference between observing and experimenting.

Warmup

(6 minutes — sort and justify, pairs, mini whiteboards)

Sort these into two groups using a rule of your own choosing, then state your rule:

1. Asking every student in the school their favourite subject.
2. Asking randomly chosen students their favourite subject.
3. Counting cars that pass the school gate between and am.
4. Giving one group of plants fertiliser and another group none, then comparing growth.

Expected sorts: many students separate “everyone” (1) from “some people” (2); others separate “just watching” (3) from “changing something on purpose” (4). Both rules are valid and both matter today — they define two different distinctions.

The two questions data collection always answers:

• How much of the population is involved? — all of it (census) or part of it (sample).
• How is the data produced? — by watching what already happens (observation) or by deliberately changing something and measuring the effect (experiment).

Activities

Activity 1 — Explicit Instruction: the Four Techniques (14 min)

I do — definitions with a real Australian example for each:

Technique	Definition	Australian example
Census	Data collected from every member of the population	The Australian Census of Population and Housing — every person in Australia, conducted by the ABS every five years (most recently, tonight’s date puts us in the middle of the 2026 Census)
Sample	Data collected from part of the population, used to represent the whole	A Newspoll or Roy Morgan opinion poll of around – voters, used to estimate the views of over million enrolled voters
Experiment	The researcher deliberately changes something (a variable) and measures the effect	An agricultural trial: CSIRO scientists give one wheat crop a new fertiliser and a matching crop none, then compare yields
Observation	Data recorded by watching what naturally happens, without intervening	A wildlife count: rangers record koala sightings along a set track once a month, changing nothing about the koalas’ environment

The key test to say aloud: “Did the researcher deliberately change something to see what would happen, or did they just watch and record?” — this is the observation/experiment line. “Did they measure everyone, or only some?” — this is the census/sample line.

We do — classify together, justifying with the test above:

1. The ABS counts every household in Australia on Census night.
2. A council installs a speed camera and simply logs the speed of every car that passes for a week.
3. A school nurse weighs a random of students to estimate the average weight of the whole school.
4. A pharmaceutical trial gives half of a group of volunteers a new medicine and the other half a placebo, then compares recovery times.

(Answers: 1. Census — the whole population, no intervention, so it is also an observation-style census; 2. Observation of the whole population passing — a census-style observation, since it records everyone who passes, not a subset; 3. Sample — only part of the population is measured; 4. Experiment — the researcher deliberately intervenes by assigning the medicine.)

Teaching point from Q1 and Q2: the two distinctions are independent. A data collection method can be, for example, a census that is also an observation (Q1, Q2) or a sample that is also an experiment (a drug trial run on a sample of volunteers, Q4). Naming both aspects gives the fullest description.

Activity 2 — Classify and Justify (14 min)

Pairs. For each scenario, name the technique (or combination) and justify using the test questions.

1. Bureau of Meteorology automatic weather stations record temperature every hour at every station across Australia.
2. A market research company phones randomly selected households to ask about breakfast cereal preferences.
3. A Year 8 class tests whether background music affects test scores by having half the class do a quiz in silence and half with music playing.
4. A birdwatcher records every bird species seen in their backyard each morning for a month.
5. The ABS collects income and employment data from a sample of households each month (the Labour Force Survey) rather than everyone.
6. A council counts every vehicle crossing a new bridge on its opening day using an automatic sensor.
7. A scientist exposes one set of seedlings to extra carbon dioxide and a matching set to normal air, then compares growth after four weeks.

(Answers: 1. Census + observation — every station, no intervention; 2. Sample — a subset of households; 3. Experiment, on a sample (the class) — music is deliberately manipulated; 4. Observation, likely a sample of the birds present (not necessarily every bird visits); 5. Sample — deliberately smaller than the full population, monthly; 6. Census + observation — every vehicle on that day; 7. Experiment — CO₂ is deliberately manipulated.)

Activity 3 — Inquiry: Choosing a Technique (7 min)

Pairs, then share.

A local council wants to know: “Would a new pedestrian crossing near the school reduce near-miss incidents with cars?”

1. Suggest one technique (or combination) the council could use, and describe exactly what they would do.
2. Is your suggestion a census or a sample? An observation or an experiment (or does it not fit neatly)?
3. What is one practical limitation of your suggestion?

Discussion targets: most workable answers are observational (recording near-misses before and after installation) rather than a true experiment, since researchers cannot ethically or practically “cause” near-misses on purpose. This previews Lesson 104’s focus: the practicalities of a technique often rule out the “ideal” method.

Checks for Understanding

(5 minutes — exit ticket, collected)

1. Define census and sample in your own words.
2. Define experiment and observation in your own words.
3. Classify: “A researcher records the shopping habits of every customer entering a store for one day, without asking any questions.” Name both the population-coverage technique and the intervention technique.
4. Classify: “A teacher tries two different revision methods with two different classes and compares test results.”
5. Reasoning. Explain why a technique can be both a “sample” and an “experiment” at the same time, using an example.

Answers: 1. A census collects data from every member of a population; a sample collects data from only part of it; 2. An experiment deliberately changes a variable to measure its effect; an observation records what happens naturally without intervening; 3. Census (every customer that day) + observation (no intervention); 4. Experiment (the revision method is deliberately manipulated) on a sample (two classes, not every student in the world); 5. The two distinctions are independent — a drug trial run on volunteers (a sample of all possible patients) where the researcher deliberately gives the medicine (an experiment) is both at once.

Common Misconceptions

Misconception	How to pre-empt it
Treating “sample” and “survey” as the same word.	A survey is a tool (a set of questions); a sample is the group it is given to. A census can also use a survey (the ABS Census is a survey given to everyone).
Believing observation and experiment are opposites of census and sample.	Explicit statement: the two distinctions are independent — Activity 1’s Q1/Q2 and Q4.
Assuming a census is always better because it covers everyone.	Flagged for Lesson 104: censuses are accurate but often slow, expensive or impractical.
Thinking any data collection without asking questions is “observation”.	An automatic sensor counting cars is still observation — the key test is whether anything was deliberately changed, not whether people were asked anything.
Calling any comparison of two groups an “experiment”.	Only counts as an experiment if the researcher deliberately creates the difference (e.g. assigning fertiliser); comparing two naturally occurring groups is still observation.
Believing observation cannot involve counting or measuring, only watching.	The BOM weather stations and traffic sensors are observation and precise measurement.

Enrichment — Competition-Style Problems

E1 (Kangaroo style). A school records the exact number of students absent every day for a term, without changing anything about attendance policy. Classify this by both distinctions.

Answer

Census (every student, every day) and observation (nothing is deliberately changed).

E2 (AMC Junior style). A city has residents. A council mails a form to all of them but only respond. Is the intended method a census or a sample? What has actually been achieved?

Answer

Intended: a census (every resident was targeted). Achieved: effectively a sample, and worse, a self-selected one — only people motivated enough to respond are included, which can bias the results (this previews Lesson 106’s non-random sampling ideas).

E3 (Challenge). A researcher wants to know if a new road sign reduces speeding. Explain why “measuring speed at the same location before and after installing the sign” is not a perfectly clean experiment, even though something was deliberately changed.

Answer

Other things could also have changed between the two time periods (weather, time of year, traffic volume, driver awareness campaigns), so it is hard to be sure the sign alone caused any difference. A cleaner experiment would compare two similar locations at the same time, one with the sign and one without.

E4 (Challenge). Explain why the Australian Census of Population and Housing is run only once every five years rather than every year.

Answer

A census of every person in the country is extremely expensive and labour-intensive to run (practicalities explored fully in Lesson 104); five years balances the cost against the need for reasonably current data on population, housing and services.

Homework

1. Define, in your own words, all four techniques: census, sample, experiment, observation.
2. Classify each scenario by both distinctions (population coverage and intervention): (a) The ABS counts every dwelling in a suburb for the Census. (b) A phone company samples customers about network satisfaction. (c) A gardener compares plant growth using two different watering schedules. (d) A council installs sensors that record noise levels at all times on a busy road.
3. Give one Australian example each of a census, a sample, an experiment and an observation, different from those used in class.
4. Explain in one sentence why an opinion poll of voters is a sample and not a census, even though it might feel “big”.
5. A supermarket wants to know whether moving snacks near the checkout increases sales. Describe a data collection approach and classify it using both distinctions.
6. Reasoning. Explain why “asking everyone in your class” is a census of your class, but not a census of “all Year 8 students in Australia”.
7. Reasoning. A weather station observes rainfall every day for a year without changing anything. Explain why this still counts as data collection, even though nothing was “tested”.
8. Challenge. A school wants to know if a new canteen menu improves how much fruit students eat. Design one approach that is a genuine experiment (something is deliberately changed) rather than pure observation, and explain what makes it an experiment.

Answers: Q2 — (a) census + observation; (b) sample; (c) experiment (likely on a small sample of plants); (d) census (all traffic/noise on that road) + observation. Q4 — is a small part of the roughly million enrolled voters, so it does not cover the whole population. Q5 — e.g. record sales for two weeks with snacks in their usual spot, then two weeks with them near the checkout, and compare — this is an experiment (the layout is deliberately changed) run as effectively a census of all sales in that period. Q6 — a census is defined relative to the population you are trying to describe; your class is the whole population when the question is about your class, but only a tiny sample when the question is about all Australian Year 8 students. Q8 — e.g. deliberately introduce a new menu for two weeks while recording fruit taken, compared with the old menu for two weeks; it is an experiment because the menu (the variable) is changed on purpose to measure its effect, rather than simply watching an already-existing situation.