The sample size should be large enough so that after the sample is split twice there will be a sufficient testing sample size.

• A. Not specified
• B. Not described
• C. True
• D. False

Answer: (C)

When planning sampling for validation, you need enough data in each subset to produce reliable performance estimates. If you split the sample twice, the final testing portion becomes smaller, so starting with a larger overall sample ensures that the testing set after both splits still has enough observations to be meaningful. This is why the statement is true: a sufficiently large initial sample is required to maintain a representative and reliable testing subset after two splits. In practice, you weigh how many data points you need in the final test to achieve your desired precision and then choose an overall sample size that accommodates both splits to meet that requirement.