Success: make an inference limited by how the sample was chosen. (AC9M8ST03, AC9M8ST04)
- Recall · 2 min. Ask why A and B's bicycle proportions differ by one person even though both have 20.
- Model · 4 min. Say: “In these two simulated random samples, bicycle estimates were 15% and 20%; another fair sample could differ too.” Show the optional repeated simulation: across 1000 seeds, the average error was 8.95 percentage points for samples of 10 and 3.92 for samples of 40. Do not claim every larger sample must be closer.
- Guided claim test · 5 min. Sort “sample A had 3 cyclists” as direct model fact; “all students cycle at 15%” as overreach; “larger fair samples often vary less, but do not guarantee exactness” as a qualified pattern.
- Independent brief · 7 min. Learners draft three sentences for a fictional principal: two source figures, one method limitation and one ethical next action. A source label must accompany each number.
- Peer check · 4 min. Partner asks how a simple random sample would be drawn and whether a census is necessary or appropriate.
- Exit · 3 min. “Can two equal-size random samples give different percentages?” Key: yes; this model has 15% and 20% bicycle.