Alternative hypothesis
Alternative hypothesis is the claim a researcher seeks to establish in a hypothesis test. It is accepted only when the test statistic falls beyond the critical threshold, providing sufficient evidence to reject the null.
Also known asH1 · Ha
FrameworkHypothesis testing
See it move
A two-column comparison sets the two-tailed alternative hypothesis (H₁: μ ≠ 200) against the one-tailed form (H₁: μ > 200). The two-tailed version splits rejection across both tails at α/2 each side, giving a critical value of ±1.96 at α = 0.05; the one-tailed version concentrates the full significance level in one tail and uses a critical value of 1.645. Both forms require the choice to be made before data are seen — switching direction afterwards is p-hacking.
Check yourself
A quality manager wants to determine whether a new machine reduces the mean defect rate below the current level of 5%. Which of the following correctly formulates her alternative hypothesis?