Data Collection Techniques — AC9M8ST01

• Distinguish population, census and sample.
• Compare random and non-random sampling.
• Distinguish experiment from observation.
• Evaluate bias, precision, error, ethics and practical constraints.

Recall population, sample, variable, categorical/numerical data, basic graphs and the idea that data are collected to answer a question.

A population is the whole group of interest. A census attempts to collect data from every member; a sample studies a subset. A census can still suffer coverage, non-response, wording or measurement bias.

Random sampling uses chance to select members. Non-random methods such as convenience sampling can be practical but may systematically over-represent easy-to-reach groups. Sample size does not repair systematic selection bias.

An experiment deliberately imposes a condition or treatment; an observation records what already happens. Observational association does not by itself establish causation.

Digital devices and simulations can improve consistency, but precision is not the same as accuracy. Calibration, resolution, environmental conditions and rounding affect error.

Ethical collection considers consent, privacy, inclusion and respectful use. Sampling decisions also matter in artificial intelligence: biased training data can produce biased models.

• A census is automatically unbiased. Coverage, non-response and measurement bias can remain.
• A huge convenience sample becomes representative. More observations do not remove systematic selection bias.
• Observation proves cause. Uncontrolled variables may explain an association.
• More decimal places guarantee accuracy. Precision and accuracy are different.
• Ethics is separate from statistics. Consent, privacy and inclusion affect responsible, trustworthy data.

1. Classify four scenarios as census, sample, experiment or observation.
2. Identify one likely bias in a shopping-centre survey.
3. Explain one benefit and one limitation of random sampling.
4. Compare precision and accuracy for two measuring devices.

1. Design a random sample of 80 students from a school roll.
2. Explain why a voluntary online poll can be biased.
3. Give a context where a census is preferable to a sample.
4. Give a context where an experiment would be unethical or impractical.
5. Explain why environmental conditions can bias repeated measurements.
6. Describe how biased training data can affect an AI classifier.
7. Rewrite a leading survey question to reduce response bias.

A company claims that a voluntary survey of 20,000 website visitors proves 82% of all Australians prefer its product. Evaluate the claim by separating sample size, sampling method, population coverage, response bias and the strongest defensible conclusion.

1. What is the difference between a census and a sample? A census targets the whole population; a sample studies a subset.
2. Why can random sampling help? It reduces systematic selection preferences by using a chance mechanism.
3. What can a well-designed experiment support? Stronger evidence about causal effects than observation alone.
4. Why can precise data still be wrong? A precise instrument can be miscalibrated or used under biased conditions.
5. How can sampling affect AI? Unrepresentative training samples can reproduce bias in model behaviour.

1. A council surveys only weekday library visitors. Identify the sampling method, likely bias and a better design.
   Review hint: Separate convenience from random sampling and identify who is under-covered.
2. Design a safe investigation of whether a revision routine changes quiz performance.
   Review hint: Specify treatment/comparison, fair assignment, measurement and consent.
3. A sensor reports 23.417°C each minute. Explain what is needed before trusting the measurements.
   Review hint: Consider calibration, resolution, placement and consistent conditions.

• I identify population and sample.
• I select census/sample/experiment/observation deliberately.
• I recognise selection and measurement bias.
• I distinguish precision from accuracy.

Exit ticket: Why can a very large sample still be poor evidence about a population?

Australian Curriculum v9.0 — AC9M8ST01: investigate techniques for data collection including census, sampling, experiment and observation, and explain practicalities and implications.

Victorian Curriculum F–10 Version 2.0 — Level 8, VC2M8ST01: Exact.

NSW Mathematics K–10 Syllabus (2022) — Stage 4, MA4-DAT-C-01; MA4-DAT-C-02; MAO-WM-01: Partial. NSW Stage 4 classifies/displays and analyses datasets, but does not reproduce the full national collection-method and sampling-design descriptor as one outcome.

• AC9M8ST02 — sampling and distributions
• AC9M8ST04 — statistical investigations

• Australian Curriculum v9.0
• Victorian Curriculum F–10 Version 2.0 Mathematics
• NSW Mathematics K–10 Syllabus (2022)

Curriculum equivalents for Investigate techniques for data collection including census, sampling, experiment and...

Mapped skill: investigate techniques for data collection including census, sampling, experiment and observation, and explain the practicalities and implications of obtaining data through these techniques

These references identify matching or closely related learning. Curriculum sequence, terminology and depth vary, so teachers should use the mapped skill and lesson difficulty to confirm suitability.

Australian Curriculum v9.0 is the canonical source for this SkillrHub lesson. Victoria and NSW entries name the closest published state codes or outcomes; international entries are planning references rather than claims of identical curricula.