One sample non-parametric tests

Abstract

The previous chapters discussed situations where we made an hypothesis about a population parameter. We hypothesize that the population mean or proportion is a specific value and then determine the likelihood of drawing from this population a sample with a different mean or proportion. We do this by calculating a z-score or a t-score and looking up the corresponding probability in the appropriate table. If the difference between the sample statistic and the hypothesized population parameter is large, the corresponding probability that it is drawn from this population will be low. In short, the question boils down to whether an observed difference between a sample statistic and population value is ‘big enough’.