Statistical hypothesis

-value hypothesis test

The -value test involves finding the probability, under our assumed () model, of getting a test statistic at least as extreme as observed. If it is less than a designated level of significance ¹, the result is said to be statistically significant and is rejected in favour of .

Steps

1. Identify the appropriate distribution of the underlying random variable
2. Formulate a null hypothesis of our assumed parameterisation of the distribution; and an alternate hypothesis of its negation, which may be either one-sided or two-sided.
3. Construct the distribution of the test statistic being used under the assumptions of .
4. Find the probability of a test statistic at least as extreme as observed under .
5. State the conclusion: if the -value is less than the selected level of significance , it is highly unlikely is true and we therefore reject it in favour of ; otherwise we maintain .
6. Interpret these results in the context of the problem.

#state/tidy | #SemBr