Hypothesis testing practice reps
Which box does each situation belong in?
Goal
This is a rapid-fire practice for getting comfortable with the language of hypothesis testing, including Type I and Type II errors.
The null hypothesis
H₀ is usually the "no effect" claim (for example, no effect, no difference)
We assume H₀ until the evidence is strong enough to reject it. It works like the courtroom example you just went through: a criminal trial starts from the assumption that the defendant is innocent.
Now click the box where this situation belongs.
| Underlying reality | |||
|---|---|---|---|
| H₀ is false | H₀ is true | ||
| Conclusion from test | Reject H₀ | Correct decision | Type I errorfalse positive |
| Fail to reject H₀ | Type II errorfalse negative | Correct decision | |
All 10 rounds done
The same four boxes show up in every hypothesis test in this course. One last step: the school attendance study.
In these reps, each story told you the underlying reality. In a real study, we usually don't know it.
The school attendance study
A randomized trial tests whether text-message reminders to parents raise school attendance. (The data are hypothetical.)
H₀: text reminders make no difference to attendance.
Estimate: 3 percentage points · p-value ≈ 0.02 · α = 0.05
| Underlying reality (we don't know this) | |||
|---|---|---|---|
| H₀ is false | H₀ is true | ||
| Conclusion from test | Reject H₀ | Correct decision | Type I errorfalse positive |
| Fail to reject H₀ | Type II errorfalse negative | Correct decision | |