26.37 p-value, power and sample size

• There is a relationship among the type I error rate, type II error rate (or power), and sample size

• Ideally, the type I error rate is very low and the power very high

• One must trade the type I error rate and power for a smaller sample size.

  • Increasing the sample size costs time and money and increases the study complexity

    • It also may entail patient morbidity

  • Diminishing returns (rate of improvement of power) with each additional sample