Two random samples from the same population can contain different members, so their proportions, means, medians and ranges can differ. This is sampling variation, not automatically an error.
Repeated samples show the pattern of possible results. A single sample estimate should therefore be reported as evidence about the population, not as the exact population value.
When selection remains random and fair, larger samples usually fluctuate less from sample to sample because each estimate uses information from more population members. This does not guarantee that every larger sample is closer to the population value, and it does not fix selection bias.
Digital simulation is useful because many repeated samples can be generated quickly. At Year 8, the goal is to compare variation visually and numerically, not to introduce formal confidence intervals or hypothesis tests.
Sampling ideas have real uses: school-uniform opinion polling, heights and arm spans, weather forecasting, visitor proportions, biodiversity monitoring by First Nations Ranger Groups and other groups, and data-driven/AI decision-making.