ABSTRACT
In clinical research, two-sample hypothesis testing is more frequently used relative to one-sample hypothesis testing. In this statistical inference, the underlying parameters of two different populations are compared but the parameters are not assumed to be known, as observed in one-sample t test for example. We discussed in previous chapters hypothesis testing when a single or one sample is involved. Specifically, we compared the parameters of the population from which the sample was drawn with that of the larger population, whose parameters were assumed to be known. To recap, single-group hypothesis testing provides an estimation of the proportion (binary or categorical measures), mean (continuous scale variable), or a comparison of the observed continuous or ordered/ranked values to a norm or standard. For example, a one-sample hypothesis testing is appropriate in comparing the mean weight of children (0 to 19 years old) with leukemia with that of same age children in the United States (values of US children 0 to 19 years assumed to be known). Also, when a single group is measured twice (paired sample) or more (repeated measures), which allows us to estimate how much the proportion or mean in the single group changes between measurements, a single-sample hypothesis testing technique is used.
