Hypothesis tests for a Normal mean

Hypothesis tests for a Normal mean

Spec 5.3 · Edexcel A-Level Maths: Revision notes

Studying this course? Keep your progress in one place.

Key facts

  • If X∼N(μ,σ2)X\sim N(\mu,\sigma^2) then the sample mean Xˉ∼N(μ,σ2n)\bar X\sim N\left(\mu,\frac{\sigma^2}{n}\right).
  • The standard error σn\frac{\sigma}{\sqrt n} shrinks as nn increases.
  • Hypotheses use the population mean μ\mu, never xˉ\bar x: H0:μ=μ0H_0:\mu=\mu_0.
  • Standardise: z=xˉ−μ0σ/nz=\frac{\bar x-\mu_0}{\sigma/\sqrt n}, then use a p-value or critical value.
  • Always conclude in context, and never claim proof.

Distribution of the sample mean

The sample mean of nn Normal values is Normal with the same mean and variance σ2n\frac{\sigma^2}{n}.

If X∼N(μ,σ2)X\sim N(\mu,\sigma^2) and a random sample of size nn is taken, the sample mean has Xˉ∼N(μ,σ2n)\bar X\sim N\left(\mu,\frac{\sigma^2}{n}\right).

Its standard deviation σn\frac{\sigma}{\sqrt n} is the standard error. It is smaller than σ\sigma and decreases as nn increases.

Example: σ=4\sigma=4, n=25n=25 gives variance 0.640.64 and standard error 0.80.8.

4904924944964985005025045065085100.10.20.30.40.5xysample mean, n = 25single value
The sample mean (n = 25) is narrower than a single value (dashed)

X∼N(μ,36)X\sim N(\mu,36) and a sample of 9 is taken. What is the standard error?

Setting up the test

The test is about the population mean μ\mu, with a known variance and a random sample.

The test is about the population mean μ\mu with claimed value μ0\mu_0: H0:μ=μ0H_0:\mu=\mu_0 against H1:μ>μ0H_1:\mu>\mu_0, μ<μ0\mu<\mu_0 or μ≠μ0\mu\ne\mu_0.

Assume a Normal population with known variance σ2\sigma^2 and a random sample. Under H0H_0, Xˉ∼N(μ0,σ2n)\bar X\sim N\left(\mu_0,\frac{\sigma^2}{n}\right).

1351401451501551601650.020.040.060.08xyμ₀ = 150distribution of X̄
Under H₀, X̄ ~ N(μ₀, σ²/n). Illustrative: μ₀ = 150, σ = 20, n = 16, so the standard deviation of X̄ is 5.
  • H0H_0μ=μ0\mu=\mu_0
  • Under H0H_0Xˉ∼N(μ0,σ2n)\bar X\sim N\left(\mu_0,\dfrac{\sigma^2}{n}\right)

A manufacturer thinks the mean mass of a product has changed from 250 g. What is H1H_1?

Test statistic and p-value

Standardise the sample mean, then find the tail probability in the direction of H1H_1.

The test statistic is z=xˉ−μ0σ/n∼N(0,12)z=\frac{\bar x-\mu_0}{\sigma/\sqrt n}\sim N(0,1^2).

Then either find the p-value and compare with the level (halved if two-tailed), or compare zz with the critical value: ±1.645\pm1.645 at 5% one-tailed, ±1.96\pm1.96 at 5% two-tailed, ±2.326\pm2.326 at 1% one-tailed.

−4−3−2−112340.10.20.30.4xyz = -2.5N(0, 1)
Standard Normal: the shaded tail beyond z = -2.5 has area 0.0062

Worked example

σ=4\sigma=4, n=25n=25, H0:μ=500H_0:\mu=500, H1:μ<500H_1:\mu<500, xˉ=498\bar x=498. Test at 1%.

When standardising xˉ\bar x, what do you divide by?

Critical region for the mean

The critical values are μ0±zσn\mu_0\pm z\frac{\sigma}{\sqrt n} for the relevant zz.

You may find the critical region for Xˉ\bar X directly. Under H0H_0 the critical value is μ0±zσn\mu_0\pm z\frac{\sigma}{\sqrt n}.

A larger sample gives a smaller standard error, so the critical value moves closer to μ0\mu_0.

For a lower one-tailed 1% test with σ=12\sigma=12, n=40n=40, μ0=150\mu_0=150: Lˉ<150−2.326×1240=145.6\bar L<150-2.326\times\frac{12}{\sqrt{40}}=145.6.

1111.211.411.611.81212.212.412.612.81311.511212.49

Worked example

σ=1.5\sigma=1.5, n=36n=36, H0:μ=12H_0:\mu=12, H1:μ≠12H_1:\mu\ne12, 5% level. Find the critical region for Tˉ\bar T.

In the example, tˉ=12.3\bar t=12.3. What is the conclusion?

Concluding in context

State what happens to H0H_0, then what that means for the mean in context.

  • Reject H0H_0: 'there is evidence at the 5% level that the mean lifetime of the batteries is less than 150 hours.'
  • Do not reject H0H_0: 'there is insufficient evidence at the 5% level that the mean has changed.'

Do not say the claim is proved. The test assumes a Normal population with the stated variance and a random sample.

1351401451501551601650.020.040.060.08xycritical value 141.8distribution under H₀
One-tailed test at 5% (same illustrative distribution): if the sample mean is in the shaded region, below 141.8, reject H₀.

Reject H0H_0

  • Evidence at the stated level
  • State the direction, in context

Do not reject H0H_0

  • Insufficient evidence
  • Not proof that H0H_0 is true

Which is a complete conclusion?

Try an exam question

The mass of a loaf is N(μ,16)N(\mu,16) g. A baker claims μ=800\mu=800. A random sample of 16 loaves has mean 797. Test at the 5% level whether the mean mass is less than 800 g, given the critical value is −1.645-1.645.

[5 marks]

That's the notes covered.

Carry on to the next subtopic.

Try a question

If the null hypothesis μ=500\mu=500 is true, what is the distribution of Xˉ\bar X?

1 mark, marked against the mark scheme

Revision notes