Source

Adapted from my own handwritten and hand-typed notes from Year 1 to Year 3 of my EEE degree

To test a claim about the population variance (or standard deviation), we use the chi-square statistic built from the sample variance. The population must be approximately normal.

Hypotheses and statistic

against one of , or .

Under this is chi-square with d.f. (Table A2). A claim about is the same as a claim about .

Critical region

Remember has area to its right:

Reject if
or

The two tail cut-offs are different numbers because chi-square is not symmetric.

Method

Steps:

  1. Assume is normal and state ,
  2. Compute
  3. Read the critical value(s) for d.f.
  4. Reject if is in the critical region, then conclude in words

Worked example (two tail)

Test against at , using a sample of 15 with .

Solution:

  • , with 14 degrees of freedom
  • and
  • , so the statistic is not in the critical region
  • no significant evidence that differs from 1.25

The interval version of this idea is the Confidence Interval for the Variance.