In replicating Sir Ronald Fisher's famous experiment with tea tasting, it is essential for the subjects in the experiment to be blindfolded or, alternatively, presented with the cups of tea after they have been poured (Binomial coefficients..., 2006). However, given that when cold milk is poured into hot tea, the milk does not curdle (and when hot tea is poured into milk, the reverse occurs), the subjects would need to be blindfolded so that they could not "see" the curds and therefore guess correctly the order in which the tea and milk were combined.
             In Fisher's experiment, the tea-tasting lady was given 8 cups of tea with milk, four of each version and was told that she would be tasting four of each variety (Preece, 1990). The experiment could be varied with respect to the number of each version of the beverage presented to subjects and with respect to random presentation of the samples; ideally, however, each version should be presented as frequently (if randomly) as the other. In any approach, it is likely that the subjects will identify the method used to combine the liquids correctly at least 50 percent of the time (Binomial coefficients..., 2006). The null hypothesis of the experiment reflects Fisher's original hypothesis: The lady cannot tell the difference between milk poured into tea and tea poured into milk." The problem with the experiment is that there can be more than one mull hypothesis and, as significantly, the need to prevent subjects from viewing the cup with the beverage (Preece, 1990).
             Small samples that are statistically representative of a larger population are viable in experiments because one can extrapolate or project based on such studies to the larger population (Preece, 1990). When testing two populations with small sample sizes, the samples should be as nearly identical as possible if one is concerned with validity and reliability matters and if the lar
             ...