Introduction to Statistical Inference
Magdalena Niewiadomska–Bugaj, Robert Bartôszyński · Wiley series in probability and statistics · 2021
This chapter introduces two approaches to statistical inference: frequentism and Bayesianism. While they are different and were considered competing for many years, they are becoming more connected in the contemporary statistical methodology. While graphs are necessary to examine the content and the structure of the data it might also help checking assumptions in statistical models. It is also hard to overestimate importance of the graphs in communicating results of statistical data analysis. The chapter shows some of the possible “traps” one may encounter in implementing statistical methods in practice. Data sets can be of different types, have different characteristics, and different types require different statistical methods. The chapter also introduces some concepts related to the level of measurements. A sampling bias closely related to the bias from importance sampling is connected with the phenomenon, which caused some controversy before it became properly understood. To explain it, the chapter considers an anecdotal example.