JC2 · H2 Mathematics · Notes
Carry out and interpret one- and two-sided mean tests, estimate variance, find critical regions and solve reverse-test inequalities.
Sampling and Hypothesis Testing summary and formula sheet: every method on 4 pages
Mr Ian Ang’s summary of the whole sampling and hypothesis testing chapter on four pages.
Page 1 of 4 Notation, unbiased estimates and the CLT Part 1: sampling
H2 Math summary notes
Sampling and Hypothesis Testing formulas and methods, section by section
Page 1 of 4Notation, unbiased estimates and the CLTPart 1: sampling
Notation and unbiased estimates: the mean, s squared, and the 1-Var Stats on the GC
- The population has mean and variance . The unbiased estimates are of and of , where , and (sample variance).
- With , and : and .
- With , sample mean and sample standard deviation : .
- Warning: write the whole phrase "Unbiased estimate of population variance, ".
- GC, data in a table (L1 , L2 frequency): STAT, CALC, 1: 1-Var Stats. Read and ; is .
- For hypothesis testing, use , not .
Combining two samples: add the sums and the sizes first
- Sample 1: , , . Sample 2: , , . Combine them.
- New mean .
- New .
- Warning: add the sums and the sizes first; never average the two .
The CLT: sum and mean, with sigma squared given or estimated by s squared
- is not normal, with mean and variance , and is large (). Then and approximately. If is given, use it; if not given, use .
- Mean mass g, standard deviation g, sample of . Find and . As is large, and .
- and .
- With , find the least so that . Here . Enter and use the GC table: at and at .
- So the least .
CLT with a binomial variable: one box is binomial, the mean of 60 boxes is approximately normal
- is the number of faulty items in a box. For 60 boxes, find .
- and .
- is large, so and .
- Warning: the CLT is for the mean of 60 boxes; one box is binomial, not normal.
Page 2 of 4Setting up and carrying out a testPart 2: hypothesis tests
Setting up the hypotheses: H0 is always an equality, and the claim decides H1
| the claim | ||
|---|---|---|
| at least 40 (overstated) | ||
| at most 40 (understated) | ||
| is 40 (or has changed) |
- is always an equality: (the population mean and the claimed value).
- "At least" and "at most" keep the equality in .
- or gives a one tail test; gives a two tail test.
- "Increased" and "faster" mean you want .
The five steps and the GC Z-Test, with a one tail test at 5%
- The mean time is claimed to be at most minutes. , and . Test at whether it is more. Let be the time (min) and the population mean.
- Unbiased estimate of population mean: . Unbiased estimate of population variance: .
- Step 1: and . Step 2: at level of significance. Step 3: is large, so by the CLT, under , and .
- Step 4: using GC, -value . Step 5: -value , so we reject and conclude there is sufficient evidence, at significance level, that the mean time exceeds minutes. If the -value : do not reject ; insufficient evidence (same sentence).
- Z-Test on the GC: STAT, TESTS, 1: Z-Test. Inpt: Stats, : , : , : , : , : , Calculate. For , key the root of the variance only, with no .
A two tail test, sigma known: H1 is not equal, and the GC Z-Test gives the two tail p-value
- The mean is claimed to be , , and . Is the mean different? Test at .
- , , level. Under , .
- Using GC, -value . Since , do not reject .
- Insufficient evidence, at significance level, that the mean is not .
- The GC Z-Test with gives the two tail -value.
One tail and two tail: the p-value is doubled or halved, and the total area stays 5%
- With the same , , and : one tail to two tail, the -value is doubled; two tail to one tail, the -value is halved. For the significance level, the total area stays .
- , , , . Test , then , at . For : -value , reject .
- For : -value , so do not reject .
- A test of is significant at . Is it significant at ? Is significant at ? At : necessarily true, since -value .
- : not necessarily, because doubles to , but doubles to .
Page 3 of 4Rejection regions, unknowns and the words
Rejecting H0: p-value and z-value, for a left tail, right tail and two tail test
- Reject if -value significance level. The -value and the significance level are both areas.
- Critical at : invNorm. Two tail: invNorm.
- (left tail): -value , reject if . (right tail): -value , reject if .
- (two tail): -value , reject if .
- . Reject when is in the critical region.
Critical region of x bar: use invNorm, not z, and answer as a range of x bar
| key in | critical region | |
|---|---|---|
| , LEFT: | ||
| , RIGHT: | ||
| , CENTRE: and | or |
- , , , level. Find the critical region. Under , .
- "Critical region" is a range of : use invNorm, not . Key in invNorm(area, , , Tail).
- For key in , LEFT: , so . For key in , RIGHT: , so .
- For key in , CENTRE: and , so or .
- Reject when is inside the region. Answer as a range of .
Finding an unknown: k, n and the level, from the rejection condition
- , , , , . is rejected at . Find . Reject: , so , giving .
- , , , . is not rejected at . Find the greatest . Not rejected: , so and .
- So and the greatest . is an integer: keep the inequality.
- The smallest level of significance at which is rejected is the -value.
- For the -value in section 6, .
What the words mean: p-value, significance level, unbiased estimate, random sample and assumptions
- -value: the probability that the sample mean time is more than or equal to minutes when the population mean time is minutes; or, the smallest level of significance at which is rejected.
- Significance level : with probability we wrongly conclude (the mean time exceeds ) when in fact is true (the mean is ).
- Unbiased estimate: , the mean of the estimator equals the population parameter.
- Random sample: items selected independently, each with an equal chance of being selected.
- Assumptions: large means none about , and you write "since is large, the sample mean can be approximated by a normal distribution using the CLT". small: assume is normal.
Page 4 of 4Pictures: the normal curves for the tests
The five step test, drawn: the shaded tail is the p-value
- The test has and , with sample mean .
- Under , . The curve is centred at and is marked on the axis to the right.
- The shaded area to the right of is the p-value, .
- : is in the right tail, so reject .
Critical region, one tail: invNorm gives the boundary of the 5% right tail
- The alternative is and , at the level.
- The shaded right tail has area and starts at .
- .
- The critical region is .
Critical region, two tail: the 5% is split 2.5% in each tail
- The alternative is , so the is split each side.
- Each shaded tail has area , with ends at and on either side of .
- gives both ends.
- The critical region is or .
Two tail test, drawn: both tails shaded and the p-value doubled
- The test has and , with .
- The curve is centred at , with on the left and its mirror on the right. Each shaded tail has area .
- Both tails are shaded: the p-value is .
- : do not reject .
One tail p-value doubled: add the mirror tail for a two tail test
- The test has , with , and .
- The curve is centred at . The orange tail below has area and the blue mirror tail above has area .
- One tail: orange, , reject.
- Two tail: add blue, .
Worked examples
Hypothesis testing worked examples
Mr Ian’s worked examples for the chapter, with the full working.
Choose the alternative from the claim
Let denote the population mean of the quantity being investigated, including its units. The question’s claim decides the alternative before the sample result is used. “Decreased” gives , “increased” gives , and “changed” gives . In the usual JC test write the null at the boundary .
Test a directional claim
Has a new method reduced the filling time?
Filling times have population standard deviation 12 s. A random sample of 36 independent fills using a new method has mean 46.4 s. Previously the population mean was 50 s. Test at the 5% level whether the population mean time has decreased.
Let be the population mean filling time with the new method, in seconds.
Because n = 36 is sufficiently large for a CLT approximation in this setting, under the null hypothesis the sample mean is approximately normal:
Since , reject . There is sufficient evidence at the 5% significance level that the population mean filling time has decreased.
For a small sample, the same known-variance z-test would require a normally distributed population. A large sample supports a CLT approximation; it does not establish independence or repair biased selection.
A two-sided test counts equally extreme results in both directions
Has the mean changed?
Use the same sample, but now investigate whether the population mean has changed from 50 s. The null remains ; the alternative is . The observed z is still −1.8.
Since , do not reject . There is insufficient evidence at the 5% level of a change in the population mean. This does not prove the population mean is 50 s.
In original units, the upper equally extreme sample mean is 53.6 s, reflected from 46.4 about 50. The p-value is the probability of a sample mean at most 46.4 or at least 53.6, assuming the population mean really is 50 s.
The same sample, two different pre-specified questions
Changed: two-sided p = 0.0719. At 5%, do not reject H₀.
The tail must follow the question asked before seeing the data.
Explain p-value and significance level in context
For the lower-tail test above, the p-value 0.03593 is the probability of obtaining a sample mean of at most 46.4 s in 36 fills when the population mean is 50 s, under the model assumptions. It is not the probability that the null hypothesis is true.
A 5% significance level means the procedure has a 0.05 probability of concluding that the population mean filling time has decreased when the population mean is actually 50 s. It describes the risk of a false rejection under the null, not the percentage of individual fills below 50 s.
Use the decision rule reject when p ≤ the significance level. The boundary is retained consistently in the reverse inequalities below. Calculate with unrounded values before comparing.
Unknown variance: estimate it before testing
Coded sample summaries
A new independent random sample of 40 filling times gives and . The population variance is unknown. Use a large-sample z-test at 5% to investigate a reduction from 50 s.
Let again be the population mean filling time. Test against . With a sufficiently large sample, the estimated standard error gives the approximate z statistic:
Reject , since the p-value is below 0.05. There is sufficient evidence of a decrease in the population mean at the 5% level.
If a question gives the sample variance using denominator n, multiply it by first. If it explicitly gives an unbiased variance estimate , use that directly. A small unknown-variance sample does not justify this large-sample approximation by itself.
Find a critical region in sample-mean units
Which sample means reject a claimed minimum?
A maker claims mean content is at least 100 g. Contents are normal with known standard deviation 4 g. Ten independent items are sampled. Find the critical region for a 5% test of the claim.
To challenge “at least 100”, use and , where is the population mean content in grams. Under ,
Reject when the lower-tail probability is at most 0.05:
The critical region is g, with the unrounded boundary understood. A claim of “at most 100” would be challenged using an upper-tail alternative instead.
Work backwards from rejection or non-rejection
Find the possible population standard deviations
A random independent sample of 80 visits has mean 16 min. The population standard deviation is . A large-sample test of against rejects at 5%. Find the range of .
Rejection requires z to be at least the upper 5% critical value. Since is positive,
A smaller standard deviation makes the same one-minute difference more significant. The endpoint follows the stated p ≤ 0.05 decision rule.
Find null means that are not rejected
A normal population has known standard deviation 0.6 min. A random independent sample of 25 has mean 5.9 min. Find the values of for which is not rejected against at 1%.
Non-rejection places the statistic strictly between the two critical values:
The bounds are approximate; the exact values use . Keep both sides of the inequality when solving for the unknown null mean.
Variation: least sample size to reject
With known population standard deviation 12 s and observed sample mean 46.4 s, find the least n giving rejection of against at 5%. Assume normal observations so the test is valid for every n.
The least n is 31. At n = 30, z ≈ −1.6432 does not reach the critical value; at n = 31, z ≈ −1.6703 does. This holds the observed mean and known variance fixed; it is not a statement that collecting more data will leave the observed mean unchanged.
Recover the significance level from a rejection rule
A fixed cutoff and a different null mean
A sample mean has distribution under the null. A test rejects when . Find its significance level. What happens to that level if the null mean becomes 51, with the same standard error and cutoff?
The significance level is about 2.28%. With null mean 51,
The cutoff lies farther below the null mean in standard-error units, so the false-rejection probability is smaller. The significance level is a probability under the null distribution; it is not calculated using the observed mean as the distribution’s centre.
Before you finish
Define the population mean, choose the tail from the claim, justify the sampling model, show the standard error and p-value, compare with the stated level, then conclude in context. “Do not reject” means insufficient evidence for the alternative.