Null-Hypothesis Testing with the T-Table

Abstract

In the previous chapter we discussed that the patterns of Gaussian curve from biological data have a constant frequency distribution and that this phenomenon is used for making predictions from your data to future data. However, this is only entirely true with large samples like samples >100. In practice many studies involve rather small samples and in order for your data from small samples to adequately fit a theoretical frequency distribution we have to replace the Gaussian normal distribution with multiple Gaussian t-distributions which are a little bit wider.

In the twenties the USA government employed wives of workless citizens in the Work Project Administration US. With the help of Monte Carlo Models, and a pocket calculator not yet available at that time, the numerical characteristics of the best fit Gaussian curves for any sample size was calculated. These characteristics were summarized in the famous t-table, which is still the basis of any statistical software program. This chapter is to assess how the t-table can be used for null-hypothesis testing.