True distribution (revealed at debrief): Red: 40% · Blue: 30% · Yellow: 20% · Green: 10%
| Color | 10-Bead Draw Count | 10-Bead Draw % (÷10 ×100) | 30-Bead Draw Count | 30-Bead Draw % (÷30 ×100) | True % | Closer: 10 or 30? |
|---|---|---|---|---|---|---|
| Red | 40% | |||||
| Blue | 30% | |||||
| Yellow | 20% | |||||
| Green | 10% | |||||
| TOTAL | 10 | 100% | 30 | 100% | 100% |
Bar Graph: Compare your 10-bead and 30-bead results to the true distribution (use two different colors).
Y-axis = Percentage (0–100%). Color key: _____ = 10-bead draw _____ = 30-bead draw _____ = True %
Draw bars for each color using both sample sizes and the true %. Label your bars.
Identify the sampling method used in each scenario. Write: Random / Convenience / Stratified
| Scenario | Method | Why might this cause bias? |
|---|---|---|
| A researcher surveys the first 50 people who walk into a shopping mall. | ||
| Names of all 500 students are put in a hat; 50 names are drawn. | ||
| A school surveys 5 students from each of its 10 grade levels. | ||
| A researcher only asks their friends and family for the survey. |
Which sample size gave you a more accurate picture of the jar — 10 beads or 30? Use your actual data (percentages) to explain your answer.
Predict: What would happen to your results if you drew 50 beads? 90 beads? Explain your reasoning.
Could a 10-bead sample ever give a perfect result? Would that mean 10 beads is always good enough? Explain.
Formula: count ÷ sample size × 100 = %
1. You drew 30 beads: 14 red, 9 blue, 5 yellow, 2 green. Calculate the % for each color:
Red: ____% Blue: ____% Yellow: ____% Green: ____%
2. The true distribution is 40% red. How far off was your red result? _____percentage points
3. If you surveyed 200 students from a school of 2,000, what percentage of the population did you sample? _____%
4. Is 10% a good sample size? What else would you need to know to answer that question?
Find one news article, advertisement, or statistic that uses a survey or study. Analyze the sampling.
My conclusion about whether this sample should be trusted: