A medication is claimed to reduce blood pressure by 10 mmHg on average. A clinical trial with n=49 patients shows xˉ=8.2 mmHg reduction, s=7 mmHg. Test H0:μ=10 vs Ha:μ=10 at α=0.05.
Example 3
medium
A factory claims defect rate 5%. From n=400, p^=0.08. Test H0:p=0.05 vs Ha:p>0.05 at α=0.05.
Practice Problems
Try each one, then open the solution to compare your method.
Example 1
easy
State null and alternative hypotheses for each scenario: (a) testing if a coin is fair, (b) testing if a new drug reduces fever faster than the standard drug.
Example 2
hard
A teacher claims students average 80 points. A skeptic samples n=25 students: xˉ=76, s=10. Using z=s/nxˉ−μ0, test H0:μ=80 vs Ha:μ<80 at α=0.01.
Example 3
easy
In a hypothesis test, what is the null hypothesis H0 usually a statement of?
Example 4
easy
A test gives p=0.02 with α=0.05. Do you reject or fail to reject H0?
Example 5
easy
A test gives p=0.08 with α=0.05. State the conclusion.
Example 6
easy
True or false: failing to reject H0 proves H0 is true.
Example 7
easy
When must the significance level α be chosen?
Example 8
easy
A claim is 'the mean equals 100.' Write the null hypothesis symbolically.
Example 9
easy
A researcher wants to show the mean is greater than 50. What is the alternative hypothesis Ha?
Example 10
easy
Does a statistically significant result automatically mean the effect is large and important?
Example 11
medium
A test of H0:μ=100 vs Ha:μ=100 has xˉ=104, σ=10, n=25. Compute the z test statistic.
Example 12
medium
A two-sided test has test statistic z=2. Using P(Z>2)≈0.0228, find the p-value.
Example 13
medium
With test statistic z=2, two-sided p-value 0.0456, and α=0.05, state the conclusion.
Example 14
medium
A one-sided test Ha:μ>50 has xˉ=53, σ=12, n=36. Find z.
Example 15
medium
For a right-tailed test, z=1.5 and P(Z>1.5)≈0.0668. With α=0.05, conclude.
Example 16
medium
A test rejects H0 at α=0.05 but a 95% CI for μ is (48,52) and H0:μ=50. Is this consistent?
Example 17
medium
A test of H0:μ=200 vs Ha:μ=200 has xˉ=195, σ=20, n=16. Find z.
Example 18
challenge
Explain why we say 'fail to reject H0' rather than 'accept H0', using the courtroom analogy.
Example 19
challenge
With n very large, a trivial effect (mean differs from 100 by 0.1) becomes statistically significant. Explain the mechanism and the lesson.
Example 20
challenge
A researcher runs 20 independent tests at α=0.05 on data where every H0 is true. What is the expected number of false rejections, and what is this problem called?
Example 21
medium
A two-sided z-test has z=−1.8. Using P(Z<−1.8)≈0.0359, find the p-value and conclude at α=0.05.
Example 22
medium
A test of H0:μ=25 vs Ha:μ>25 has xˉ=28, σ=6, n=9. Find z.
Example 23
easy
A test has p=0.03 and α=0.05. State the decision and what it means.
Example 24
easy
State the null and alternative hypotheses for testing if a coin's heads probability differs from 0.5.
Example 25
easy
A z-test gives test statistic z=0.5 and P(Z>0.5)≈0.3085. Find the two-sided p-value.
Example 26
easy
For H0:μ=50 vs Ha:μ<50 with z=−2.5 and P(Z<−2.5)≈0.0062, state the conclusion at α=0.01.
Example 27
medium
For H0:μ=100 vs Ha:μ=100 with xˉ=106, σ=10, n=25, compute the z statistic.
Example 28
medium
Continuing: with z=3 in a two-sided test, P(Z>3)≈0.00135. Find the p-value.
Example 29
medium
For H0:p=0.5 vs Ha:p>0.5, p^=0.60, n=100, compute the z statistic.
Example 30
medium
With z=2 in a right-tailed proportion test and P(Z>2)≈0.0228, conclude at α=0.05.
Example 31
medium
Compute the z statistic for H0:μ=70, xˉ=67, σ=8, n=64.
Example 32
medium
With z=−3 in a two-sided test, P(Z<−3)≈0.00135. Find p-value and conclude at α=0.01.
Example 33
medium
State which error type is committed: a court convicts an innocent defendant.
Example 34
medium
State which error type is committed: a court fails to convict a guilty defendant.
Example 35
medium
A two-sided test produces p=0.04. What is the smallest significance level at which we reject H0?
Example 36
medium
Match: α=0.05. The 95% confidence interval for μ is (48,52) and we test H0:μ=50. Decision?
Example 37
hard
A researcher runs 50 independent hypothesis tests at α=0.05 under all true nulls. What is the expected number of false rejections and the probability of at least one?
Example 38
hard
With n=100, xˉ=102, σ=10, test H0:μ=100 vs Ha:μ=100. Compute z and the p-value (use P(Z>2)≈0.0228).
Example 39
hard
Why does increasing n (with fixed σ and effect size) tend to make any nonzero deviation statistically significant?
Example 40
hard
For H0:μ=0 vs Ha:μ>0, σ=2, the rejection region is xˉ>1.0. With true μ=1.5 and n=16, find the power.
Example 41
hard
Explain why p-hacking inflates the Type I error rate.
Example 42
challenge
For a one-sided test Ha:μ>μ0 with σ known, derive the sample size n needed to achieve power 1−β at significance α for a specified effect δ=μ1−μ0.
Example 43
challenge
Using the Bonferroni correction for m independent tests, what individual α should each test use to maintain family-wise error rate ≤0.05 with m=10?
What to focus on
Core idea:Hypothesis testing checks whether sample data is surprising enough to reject a default claim about a population.
Common stuck point:The procedure for hypothesis testing is the easy part; the trap is treating 'fail to reject H0' as 'prove H0 true'. Asking "Am I deciding whether sample data is surprising enough to reject a specific stated claim about a population?" first is what keeps a correct-looking calculation from being attached to the wrong concept.
Ask yourself:Am I deciding whether sample data is surprising enough to reject a specific stated claim about a population?
Concept recap
A systematic method to decide whether sample data provides enough evidence to reject a claim (null hypothesis) about a population parameter.
A hypothesis test works like a smoke detector. It assumes "no fire" until the smoke is too thick to blame on toast. The threshold is fixed before anything burns: a more sensitive setting (larger α) means more false alarms over toast, a less sensitive one means it may sleep through a real fire.