One-sided versus Two-sided tests
It is vital to understand the differences between one-sided and two-sided statistical hypothesis testing. In general, statistical users apply two-sided tests and will be rare that one will actually have a good basis to apply a one-sided test.So, as a rule: apply a two-tailed test.
We that work in statistics, we often refer to statistical testing as an “intimate stranger”. I will take the opportunity in this lecture to also revise some of the concepts of underlying statistical hypothesis testing under the lenses of understanding the differences between one- and two-sided statistical tests.
This will be a small lecture but critical to understand further the concepts underlying statistical inference via statistical testing.
A little review on statistical testing and revisiting previous tests under an one-sided test. Although we revisit this cases in theory so that you understand the concepts and mechanics, they should not be applied as one-sided tests.
A fictional case that makes clear when to use an one-tailed test.