Extremely Small Sample Size in Some Toxicity Studies: An Example from the Rabbit Eye Irritation Test

The conventional sample-size equations based on either the precision of estimation or the power of testing a hypothesis may not be appropriate to determine sample size for a “diagnostic” testing problem, such as the eye irritant Draize test. When the animalsʹ responses to chemical compounds are relatively uniform and extreme and the objective is to classify a compound as either irritant or nonirritant, the test using just two or three animals may be adequate.