Variance of a Normal distribution and the F-test

Variance of a Normal distribution and the F-test

Spec 6.1–2 · Edexcel A-Level Further Maths: Revision notes

Studying this course? Keep your progress in one place.

Key facts

  • For a Normal sample, (n−1)S2σ2∼χn−12\frac{(n-1)S^2}{\sigma^2}\sim\chi^2_{n-1}: the degrees of freedom are n−1n-1.
  • To test H0:σ2=σ02H_0:\sigma^2=\sigma_0^2, compare (n−1)s2σ02\frac{(n-1)s^2}{\sigma_0^2} with χn−12\chi^2_{n-1}.
  • A confidence interval for σ2\sigma^2 is (n−1)s2χupper2<σ2<(n−1)s2χlower2\frac{(n-1)s^2}{\chi^2_{\text{upper}}}<\sigma^2<\frac{(n-1)s^2}{\chi^2_{\text{lower}}}.
  • The F-test compares two variances: F=s12s22∼Fn1−1, n2−1F=\frac{s_1^2}{s_2^2}\sim F_{n_1-1,\,n_2-1}.
  • Assume independent random samples from Normal populations.

The chi-squared distribution

For a Normal population, (n−1)S2σ2\frac{(n-1)S^2}{\sigma^2} follows χn−12\chi^2_{n-1}.

If X1,…,XnX_1,\ldots,X_n is a random sample from N(μ,σ2)\mathrm{N}(\mu,\sigma^2), then (n−1)S2σ2∼χn−12\frac{(n-1)S^2}{\sigma^2}\sim\chi^2_{n-1}. This needs a Normal population.

The chi-squared distribution χν2\chi^2_\nu is positively skewed, takes only positive values, and depends on its degrees of freedom ν=n−1\nu=n-1. Critical values come from tables: the upper 5%5\% point of χ152\chi^2_{15} is 24.99624.996.

24681012140.020.040.060.080.10.120.140.160.180.2xyupper 5% pointχ²₄
The χ²₄ curve: upper 5% tail shaded from 9.488.

A sample of size 16 is taken from a Normal population. Which distribution does 15S2σ2\frac{15S^2}{\sigma^2} follow?

Testing a variance

Calculate (n−1)s2σ02\frac{(n-1)s^2}{\sigma_0^2} and compare it with the upper or lower point of χn−12\chi^2_{n-1}.

To test H0:σ2=σ02H_0:\sigma^2=\sigma_0^2, calculate (n−1)s2σ02\frac{(n-1)s^2}{\sigma_0^2} and compare it with χn−12\chi^2_{n-1}. The alternative decides which tail to use. Sketch the skewed curve and shade the tail or tails you need.

σ2>σ02\sigma^2>\sigma_0^2

  • Reject if the statistic exceeds the upper critical value

σ2<σ02\sigma^2<\sigma_0^2

  • Reject if it is below the lower critical value

σ2≠σ02\sigma^2\ne\sigma_0^2

  • Split the significance level between both tails

Worked example

n=20n=20, s2=6.8s^2=6.8, σ02=4\sigma_0^2=4. Test H0:σ2=4H_0:\sigma^2=4 against H1:σ2>4H_1:\sigma^2>4 at the 5%5\% level.

n=12n=12, s2=2.5s^2=2.5, σ02=4\sigma_0^2=4. What is the test statistic?

Interval for a variance

Divide (n−1)s2(n-1)s^2 by the upper and lower χ2\chi^2 points; the upper point gives the lower limit.

A (100−α)%(100-\alpha)\% confidence interval for σ2\sigma^2 uses the lower and upper α2%\frac{\alpha}{2}\% points of χn−12\chi^2_{n-1}. Because χ2\chi^2 is in the denominator, the upper point gives the lower limit.

The interval is not symmetrical about s2s^2. It matches a two-tailed test at the α%\alpha\% level.

  • Lower limit(n−1)s2χupper2\dfrac{(n-1)s^2}{\chi^2_{\text{upper}}}
  • Upper limit(n−1)s2χlower2\dfrac{(n-1)s^2}{\chi^2_{\text{lower}}}

Worked example

n=10n=10, s2=8.4s^2=8.4. Find a 90%90\% confidence interval for σ2\sigma^2.

Which χ2\chi^2 point gives the lower limit of the interval for σ2\sigma^2?

The F-test

Divide the two sample variances and compare with the F-distribution.

For independent samples from two Normal populations with equal variances, S12S22∼Fn1−1, n2−1\frac{S_1^2}{S_2^2}\sim F_{n_1-1,\,n_2-1}, the F-distribution with numerator and denominator degrees of freedom.

To test H0:σ12=σ22H_0:\sigma_1^2=\sigma_2^2, calculate F=s12s22F=\frac{s_1^2}{s_2^2}. For H1:σ12>σ22H_1:\sigma_1^2>\sigma_2^2 use the upper α%\alpha\% point. For a two-tailed test put the larger variance on top and use the upper α2%\frac{\alpha}{2}\% point, with the degrees of freedom in matching order. State the assumptions: independent random samples from Normal populations.

  • F-statisticF=s12s22∼Fn1−1, n2−1F=\dfrac{s_1^2}{s_2^2}\sim F_{n_1-1,\ n_2-1}

Worked example

s12=14.4s_1^2=14.4 (n1=12n_1=12), s22=5.6s_2^2=5.6 (n2=9n_2=9). Test H0:σ12=σ22H_0:\sigma_1^2=\sigma_2^2 against H1:σ12≠σ22H_1:\sigma_1^2\ne\sigma_2^2 at the 10%10\% level.

For a two-tailed test, s12=3s_1^2=3 and s22=6s_2^2=6. What is FF?

Try an exam question

The masses of bags of sugar are Normally distributed. The supplier claims the variance is 44 g2^2. A random sample of 1111 bags has sample variance 7.27.2 g2^2. Test, at the 5%5\% level, whether the variance is greater than 44 g2^2. The upper 5%5\% point of χ102\chi^2_{10} is 18.30718.307.

[4 marks]

That's the notes covered.

Carry on to the next subtopic.

Try a question

Under H0:σ2=1.6H_0:\sigma^2=1.6, which distribution does (n−1)S2σ2\frac{(n-1)S^2}{\sigma^2} follow?

1 mark, marked against the mark scheme

Revision notes