Hypothesis question

[Rachyxxx]

Simple question: can you have a one-tailed and a two-tailed hypothesis in the same study. I am running 2 2way ANOVAs and one of them (extant literature is mixed) is just exploring whether there any difference either way (two tailed) and the other I am certain will be different in a certain direction (one-tailed). Is this okay to do? Does that makes sense or am I talking rubbish?
Thankyouu

 

[noetsi]

You can have different hypothesis some of which are one and some two tailed. One hypothesis can not be both, it would not make any substantive sense.

 

[hlsmith]

That would be acceptable, given you have not seen/analyzed your data yet. It is good to document this ahead of time to offset HARKing. I will just point out that claiming a one-sided hypothesis is a very strong assumption.

Another approach, perhaps when you become seasoned, would be to use Bayesian analyses that weights results by your prior knowledge and not conduct hypothesis testing at all, but report the results and let the field interpret them as they wish.

 

[noetsi]

In about 20 years I might be seasoned enough to uses Bayesian methods