ABSTRACT
The hypothesis is about the attributes of the population, not about the attribute of the sample. Because of random events and the vagaries associated with experimental testing, the sample data will not reveal the true mean or true variance of the population. In the coin flipping illustration, the distribution of successes should be the binomial. Unfortunately, the tradition in statistical hypothesis testing, uses “accept” instead of “not reject”. There are alternate approaches to rejecting the hypothesis, to assessing the degree of violation or improbability of a data outcome if the hypothesis is true. The normal distribution describes most data, and more so the average of data, and is therefore the most widely used model for comparing the mean of continuous distributions. The hypothesis is about some data-based attribute of the population that would be expected to be manifest if the supposition is true.
