Reply

If researchers choose a 1-sided test, they must select which of the 2 options they think is most appropriate based on their expectation of the result. Therefore, this decision must be made before any descriptive exploration of the data. The choice  must rely on an assumption of a certain direction or outcome that is confirmed by robust research before the decision is made.1 Furthermore, 1-sided tests have more statistical power if you have a strong rationale for the directionality of a difference/relationship.

Source: Journal of Cataract and Refractive Surgery - May 28, 2019 Category: Opthalmology Authors: Mathias V. Melega, Monica Alves, Rodrigo Pessoa Cavalcanti Lira, Iuri Cardoso da Silva, Bruna Gil Ferreira, Hermano L.G. Assis Filho, Fernando Rodrigo Pedreira Chaves, Alexandre A.F. Martini, Livia Maria Dias Freire, Roberto dos Reis, Carlos Eduardo Leite Tags: Letters Source Type: research