Modern society requires us to become statistically literate. The right and proper understanding and interpretation of statistical hypothesis testing is essential to statistical literacy, however it is frequently confused with mathematical proof by contradiction. This study explores how statistical hypothesis testing can be distinguished from mathematical proof by contradiction, that is, what is specifically different between them. To achieve the purpose in this study, their logical structures are compared using the analytical framework: argument. Consequently, four differences are found. In statistical hypothesis testing, unlike mathematical proof by contradiction, (a) its premise is mostly non-mathematical statement and not invariable, (b) contradiction in the strict sense of the word does not arise, (c) its conclusion/claim cannot always be supported and defended by its premise, and (d) defending its conclusion/claim is necessary. The analogical approach with proof by contradiction will work effectively when hypothesis testing is taught and learned, but using that approach alone hypothesis testing has the risk of being assimilated into proof by contradiction. To understand the essence of statistical hypothesis testing properly, it is necessary to compare intentionally hypothesis testing with proof by contradiction and characterize the former as not the same as the latter.
View full abstract