What should be done if not all animals can be measured?

• A. Measure only the biggest animals.
• B. Random sampling.
• C. Ignore unmeasured animals.
• D. Replace with simulated data.

Answer: (B)

When you can’t measure every animal, the goal is to get a subset that represents the whole population without bias. Random sampling does exactly that: every animal has an equal chance of being chosen, so the measured group mirrors the larger group in its average, variation, and genetic makeup. This lets you reliably estimate population means, variances, and genetic parameters from the data you collect, while avoiding systematic favoritism toward certain animals.

Choosing only the largest animals would skew results and overstate what the population as a whole is like. Ignoring unmeasured animals wastes valuable information and introduces bias, since you lose the ability to generalize beyond the measured subset. Replacing real measurements with simulated data can be useful for method development or exploring scenarios, but it can’t stand in for actual data when you’re estimating true population characteristics.

So, random sampling provides the best balance of feasibility and statistical validity, giving you representative data to base conclusions on.