· H2 Mathematics — Ten-Year Series, worked ← All 11 questions

2024 A-Level · Paper 2 · Question 8

2024 H2 Maths Paper 2, Question 8

Hypothesis testing 8 marks test of a mean
Download Paper 2 solutions (PDF)

Part (a)

1 mark

Lee should carry out a one-tailed test (lower-tailed).

Explanation: Lee suspects specifically that the mean number of birds has decreased, which is a one-directional claim about μ\mu. A two-tailed test is used only when the claim is that the mean has simply changed, so here a one-tailed test is appropriate.

Part (b)

2 marks

Let μ\mu be the population mean number of birds visiting the new feeder in the 20-minute interval while Lee eats his breakfast.

H0:μ=17.3H_{0}:\mu = 17.3
H1:μ<17.3H_{1}:\mu < 17.3

Part (c)

3 marks

We first find unbiased estimates of the population mean and variance from the sample.

x‾=∑xn\overline{x} = \frac{\sum x}{n}
=51232= \frac{512}{32}
=16= 16
s2=1n−1(∑x2−(∑x)2n)s^{2} = \frac{1}{n - 1}\left( \sum x^{2} - \frac{(\sum x)^{2}}{n} \right)
=131(8702−512232)= \frac{1}{31}\left( 8702 - \frac{512^{2}}{32} \right)
=131(8702−8192)= \frac{1}{31}(8702 - 8192)
=51031= \frac{510}{31}

Since n=32n = 32 is large, we do not need to assume the population is normal: by the Central Limit Theorem, X‾\overline{X} is approximately normal, and we use the test statistic

z=x‾−μ0s2/nz = \frac{\overline{x} - \mu_{0}}{\sqrt{s^{2}/n}}
=16−17.3510/3132= \frac{16 - 17.3}{\sqrt{\frac{510/31}{32}}}
=−1.813 (4 s.f., by GC)= - 1.813\ \left( 4\ \text{s.f., by GC} \right)

For H0H_{0} to be rejected at the α%\alpha\% level (one-tailed, lower), the α%\alpha\% significance level must exceed the pp-value:

p-value=P(Z<−1.813)p\text{-value} = P(Z < - 1.813)
=1−Φ(1.813)= 1 - \Phi(1.813)
=0.0349 (3 s.f., by GC)= 0.0349\ \left( 3\ \text{s.f., by GC} \right)

H0H_{0} is rejected when α100>p-value\frac{\alpha}{100} > p\text{-value}, so

α100>0.03491\frac{\alpha}{100} > 0.03491
⇒α>3.491\Rightarrow \alpha > 3.491

Since α\alpha is a positive integer, the least value satisfying this is α=4\alpha = 4.

To check your answer: at α=4\alpha = 4, p=0.0349<0.04p = 0.0349 < 0.04, so H0H_{0} is rejected ✓; at α=3\alpha = 3, p=0.0349>0.03p = 0.0349 > 0.03, so H0H_{0} would not be rejected — consistent with αmin⁡=4\alpha_{\min} = 4. ✓

∴\therefore x‾=16\overline{x} = 16, s2=51031 (≈16.5)s^{2} = \frac{510}{31}\ ( \approx 16.5), and the least possible value of α\alpha is 44.

Part (d)

1 mark

Since H0H_{0} is rejected, there is sufficient evidence, at the 4%4\% significance level, to conclude that the population mean number of birds visiting the new feeder during the 20-minute interval while Lee eats his breakfast has decreased (is less than 17.317.3).

Part (e)

1 mark

No.

Explanation: the population distribution of the number of birds visiting the feeder is not stated (or known) to be normal, and n=10n = 10 is too small for the Central Limit Theorem to guarantee that X‾\overline{X} is approximately normal. Without either of these, the test in (c) could not be carried out validly on only 10 mornings.

This is how we mark, and how we teach.

Every Ten-Year Series question, worked in full, with the reasoning written out and the traps named.

Find out more