In the field of statistical data analysis in research, there is considerable discussion about the details of statistical methods and statistical modeling applied to the actual data. However, the fact that philosophical and conceptual issues in statistical science lie behind a variety of practical problems is not often discussed in the open. Meanwhile, this paper discusses the relative comparison of competing hypotheses, model selection by AIC (Akaike’s Information Criterion), and the nature of statistics as “abduction” as nondeductive reasoning, which is based on the statistical criterion of “simplicity.”
