# The Order of Things By:: [[Ross Jackson]] 2022-09-27 Sequence matters. The order in which one learns things can influence one’s thinking. Unpacking the impact of one’s learning sequence can be challenging. Learning statistics is one example of this phenomenon. Frequently when one learns statistics one learns about parametric tests (e.g., _t_-test, z-test, ANOVA) before nonparametric tests, if the nonparametric tests are covered at all.  When learning about the parametric tests one learns that there are characteristics of the data that must be met to use the tests. The data must be normally distributed, the observations are independent, and the variance must be homogeneous. If one is just starting as an analyst, and if one only learned the parametric tests, it is not uncommon for one to simply declare these requirements are assumed. At a level beyond this approach, an analyst might assess the data and see if these requirements are met. If the stars align, one can explain that the data were assessed, they conform to the requirements, and one used the appropriate test. Often data are not so accommodating. At some point, an analyst will confront data that do not conform to those requirements, and they would benefit from the application of nonparametric tests. When an analyst takes the time to learn how to use these tests, one discovers something interesting in the process. The nonparametric tests are more generally useful. In short, nonparametric tests can be used both when parametric tests can and can’t be used. The order of things seems to be reversed here. If nonparametric tests are more generally useful, shouldn’t they be learned first?   There are cultural touchstones associated with learning. In economics, supply and demand are central. In philosophy the notions of Plato’s allegory of the cave and Kant’s categorial imperative have prominence. In statistics, it is the _t_-test and z-test. Every discipline attempts to create its disciples. The order in which things are presented conveys something. Perhaps the order is chronological. Other times the order is sequential as topics build upon and incorporate prior concepts. Sometimes the order is of social importance. In statistics the order is suspect. Starting with nonparametric tests provides a basic and general application of assessment. While uncommon in application, nonparametric tests enable one to progress when parametric tests fail. Given how important this would seem it is shocking how seldom those tests are covered. The order of things precludes it. #### Related Items [[Statistics]] [[Non-Parametric Statistics]] [[Sequence]] [[Learning]] [[Technique]] [[Disciples]] [[Thinking]]