3. Controls, Replication, Sample Size, and Bias

A control group is not a useless group. It is the baseline that shows what happens without the treatment or under normal conditions. Without it, you may see a change but not know whether the treatment caused it.

Good experiments also use replication, meaning the test is repeated. One seedling might grow strangely by chance. Twenty seedlings give a clearer picture. If a result appears again and again, confidence increases.

Sample size matters for the same reason. A study of two people can be distorted by one unusual person. A larger sample is usually more reliable because random differences tend to balance out.

Bias is a systematic unfairness in how data is collected or interpreted. If a survey about sleep habits only asks night-shift workers, it may not represent everyone. If a researcher measures only the healthiest plants and ignores the dead ones, the result is biased.