Most of the time in my Stats class I'm given a sample and the sample usually gives a proportion greater than or less than the null hypothesis proportion. So I assume we use a one-sided sig test. What's an example of using a two-sided significance test?
For example, in my class we had to do a hypothesis test on whether TVs were cheaper the day before or on Black Friday. We got sample data for Black Friday TV prices and the average was lower than the average for the day before.
The above is an obvious example of using a one-sided sig test. What's an example problem that would require a two-sided?
Suppose we have a data set of the wind velocities of the past 30 years.
We know that the wind velocity was $x$ 60 years ago.
Has the average wind velocity increased or decreased over the past 30 years?
Null hypothesis: $\mu = x$
Alternative hypothesis: $\mu\neq x$
We need two sided testing.
I'm sure you can think of other examples.