Power of a test equals 1−β. If β=0.20, what is the power?
Example 3
hard
A factory's mean weight is supposed to be μ0=100 g. A sample of n=25 has xˉ=98 and s=5. Compute the t-statistic for H0:μ=100.
Example 4
hard
An 80% confidence interval for μ is (48.5,51.5). Does a two-sided test of H0:μ=50 at α=0.20 reject?
Example 5
medium
A coin is flipped 100 times and lands heads 60 times. Test H0:p=0.5 vs. Ha:p=0.5. Compute z.
Example 6
challenge
Two researchers test the same hypothesis at α=0.05. The chance both make a Type I error (when H0 is true and tests are independent) is approximately what?
Example 7
medium
A coin is flipped 100 times and lands heads 60. Use H0:p=0.5 vs Ha:p=0.5 and SE =0.5⋅0.5/100=0.05. Compute z.
Example 8
medium
A sample has xˉ=52, s=8, n=64. Test H0:μ=50 vs Ha:μ=50. Compute t.
Example 9
medium
We test H0:p=0.3 vs Ha:p>0.3 with p^=0.35, n=100, SE under H0=0.3⋅0.7/100≈0.0458. Compute z.
Example 10
hard
Two independent samples have xˉ1=20,s1=4,n1=50 and xˉ2=18,s2=5,n2=50. Test H0:μ1=μ2 vs Ha:μ1=μ2. Compute t using SE=s12/n1+s22/n2.
Example 11
hard
A paired-sample test of n=25 differences has dˉ=3, sd=10. Test H0:μd=0. Compute t.
Example 12
challenge
Test H0:p=0.5 vs Ha:p>0.5. Out of n=400, 220 successes. Compute z using SE under H0.
Example 13
hard
A company claims its light bulbs last an average of 1000 hours. A sample of 50 bulbs has xˉ=985 hours with σ=40. Set up the null and alternative hypotheses for a two-tailed test.
Example 14
hard
Continuing the light bulb example (H0:μ=1000, n=50, xˉ=985, σ=40), calculate the test statistic.
Practice Problems
Try each one, then open the solution to compare your method.
Example 1
easy
In hypothesis testing, what does the null hypothesis H0 typically state?
Example 2
easy
What is the alternative hypothesis Ha?
Example 3
easy
In a hypothesis test, which hypothesis do we assume true while computing evidence?
Example 4
easy
What do we do when the evidence against H0 is strong enough?
Example 5
easy
True or false: failing to reject H0 proves the null is true.
Example 6
easy
A test statistic measures how far the sample result is from H0 in units of what?
Example 7
easy
Fill in: hypothesis testing decides between two competing claims about a population ____.
Example 8
easy
Order the steps: (a) compute test statistic, (b) state hypotheses, (c) decide. What is the correct sequence?
Example 9
medium
A sample mean is 105, the null claims μ0=100, and SE =2.5. Compute the test statistic z.
Example 10
medium
State H0 and Ha for testing whether a coin is biased toward heads (p = probability of heads).
Example 11
medium
A test gives z=2 for a two-sided alternative. Roughly what is the p-value (use that beyond ∣z∣=2 is about 5%)?
Example 12
medium
A two-sided test gives z=2 (p ≈0.05). At α=0.05, what is the decision?
Example 13
medium
A new drug shows a tiny, clinically meaningless improvement, but the test rejects H0 with n=100,000. What does this illustrate?
Example 14
medium
Why must hypotheses be stated before collecting or examining the data?
Example 15
medium
A one-sided test of Ha:μ>50 gives z=1.7. The upper-tail area beyond z=1.7 is about 0.045. At α=0.05, decide.
Example 16
medium
In the courtroom analogy, what plays the role of the null hypothesis?
Example 17
medium
A sample of n=49 gives xˉ=53, the null claims μ0=50, and σ=14. Compute the test statistic z.
Example 18
challenge
A sample of n=64 gives xˉ=52, σ=8, testing H0:μ=50 vs Ha:μ>50. Compute z and decide at α=0.05 (one-sided z∗=1.645).
Example 19
challenge
Explain the difference between a Type I and a Type II error in hypothesis testing.
Example 20
challenge
A study tests 20 independent hypotheses, each at α=0.05, with all nulls actually true. About how many false 'significant' results are expected?
Example 21
easy
A coffee company claims its bags contain 250 g. State H0 for testing whether the true mean weight differs from 250 g.
Example 22
easy
A teacher believes a new method increases mean test scores above the historical mean of 72. State Ha.
Example 23
easy
At significance level α=0.05, a p-value of 0.03 leads to what decision about H0?
Example 24
easy
At α=0.01, a p-value of 0.04 leads to what decision about H0?
Example 25
medium
For Ha:μ>50 with z=2, find the approximate p-value (use P(Z>2)≈0.0228).
Example 26
medium
For Ha:μ=50 with z=2, find the approximate two-sided p-value (use P(Z>2)≈0.0228).
Example 27
medium
A drug company tests whether its pill lowers blood pressure (Ha:μ<130). State the direction of the rejection region for a z-test.
Example 28
medium
A two-proportion z-test compares p^1=0.40 and p^2=0.50 with SE of the difference =0.05. Compute the z-statistic.
Example 29
medium
A claim says 'at least 60% approve.' A pollster wants to test against this claim. State Ha if she suspects approval is lower.
Example 30
hard
For a t-test with n=25, how many degrees of freedom does the t-distribution have?
Example 31
hard
A test has α=0.05 and observed p-value =0.051. State and justify the decision.
Example 32
hard
Doubling the sample size (other things equal) changes the standard error by what factor?
Example 33
hard
A chi-square goodness-of-fit test has 4 categories with no estimated parameters. How many degrees of freedom?
Example 34
challenge
If 20 independent tests are run at α=0.05 when all nulls are true, what is the expected number of false rejections?
Example 35
easy
True or false: if p>α, we reject H0.
Example 36
easy
Which hypothesis carries the burden of proof?
Example 37
easy
If a two-sided test gives z=−2.5, is the result statistically significant at α=0.05?
Example 38
medium
A drug trial wants to detect whether the drug raises the cure rate above 50%. Write H0 and Ha.
Example 39
medium
A test gives z=1.5 (two-sided). The two-tailed p-value is roughly 0.134. At α=0.05, what is the decision?
Example 40
medium
Which is the better description of α: (A) the probability H0 is true; (B) the probability we wrongly reject a true H0?
Example 41
medium
Which decision is consistent with a 95% CI for μ−μ0 that contains 0?
Example 42
medium
If we shrink α from 0.05 to 0.01, what happens to (i) Type I error rate, (ii) Type II error rate?
Example 43
hard
An advertiser tests 20 independent claims at α=0.05 each. If every H0 is true, about how many will be 'significant' by chance?
Example 44
hard
Why is 'fail to reject H0' weaker than 'accept H0'?
Example 45
hard
Why might a researcher use a one-sided test rather than two-sided?
Example 46
medium
True or false: a statistically significant result with p=0.04 always means the practical effect is meaningful.
Example 47
medium
A class average is claimed to be 70. We test H0:μ=70 vs Ha:μ<70. A sample gives xˉ=69 with z=−0.8. Are we 'getting evidence the average is lower'?
Example 48
hard
An experimenter runs the test, gets p=0.07, then collects more data and stops when p first drops below 0.05. Why is this wrong?
Example 49
hard
A school claims students sleep an average of 8 hours. A sample of 36 students has xˉ=7.5 with σ=1.2. State the hypotheses for a one-tailed test (testing if students sleep less) and compute the z-statistic.
Example 50
hard
In a two-tailed z-test at α=0.05, the test statistic is z=2.10. Using critical values ±1.96, should H0 be rejected?
What to focus on
Core idea:Assume nothing is going on, work out how the statistic would vary by chance alone, and see whether the observed result would be rare. A rare result counts as evidence against the null; an ordinary one leaves the null standing but unproven.
Common stuck point:Students often know a procedure related to hypothesis testing but skip the recognition step: Is there a specific claim about a population, such as no effect or a stated value, that the sample data might contradict? That leads to a calculation or graph that looks reasonable but answers a different question.
Ask yourself:Is there a specific claim about a population, such as no effect or a stated value, that the sample data might contradict?
Concept recap
Hypothesis testing is a formal statistical procedure for using sample data to decide between two competing claims about a population parameter. You state a null hypothesis (no effect) and an alternative hypothesis, collect data, compute a test statistic, and determine whether the evidence is strong enough to reject the null.
A criminal trial runs on the same logic. The defendant (the null) is presumed innocent, the prosecution's evidence (the data) must show guilt beyond reasonable doubt (a small p-value), and the verdict is "guilty" or "not guilty", never "innocent". One difference: a test fixes in advance how often it will convict an innocent null, the rate α.