If I want to determine if people like tea or coffee more and conduct a simple random survey of about 50 people, I have some confusion about how to phrase the question to the people and how that affects the Null Hypothesis.

If I say, "Do you like tea or coffee more?" then I'm not sure how to state the null. If I say, "Do you like tea or coffee equally well?", then it is clear that the null would be mu = .5 and the alternative would be mu does not equal .5.

How would I set up the null for the first question of "Do you like tea or coffee?

I have read some articles that state journals don't like one sided tests (compared with two sided) unless the question is yes or no. So, I'm confused about how to set the null and alternative for this simple experiment.

If I say, "Do you like tea or coffee more?" then I'm not sure how to state the null. If I say, "Do you like tea or coffee equally well?", then it is clear that the null would be mu = .5 and the alternative would be mu does not equal .5.

How would I set up the null for the first question of "Do you like tea or coffee?

I have read some articles that state journals don't like one sided tests (compared with two sided) unless the question is yes or no. So, I'm confused about how to set the null and alternative for this simple experiment.