Hypothesis testing without the jargon

Hypothesis testing is a fixed recipe wrapped in intimidating language. Once the five steps are automatic, the exam questions become mechanical.

What this covers

  • Null and alternative hypotheses
  • Test statistics: z, t, chi-square
  • p-values and significance levels
  • Type I and Type II errors
  • Confidence intervals

Worked example

Problem

A machine should fill 500 ml. A sample of 36 bottles averages 495 ml, Οƒ = 12. Test at Ξ± = 0.05.

  1. 1
    State hypotheses
    Hβ‚€: ΞΌ = 500, H₁: ΞΌ β‰  500 (two-tailed).
  2. 2
    Compute the test statistic
    z = (495 βˆ’ 500) / (12/√36) = βˆ’5/2 = βˆ’2.5.
  3. 3
    Compare
    Critical values at Ξ± = 0.05 two-tailed are Β±1.96; p β‰ˆ 0.0124 < 0.05.
  4. 4
    Conclude in context
    Reject Hβ‚€ β€” evidence the machine is underfilling.
Answer: Reject Hβ‚€ at the 5% level (z = βˆ’2.5, p β‰ˆ 0.012)

Mistakes that cost marks

  • Saying you 'accept' the null hypothesis instead of failing to reject it.
  • Halving or doubling the p-value for the wrong tail.
  • Concluding statistically without stating the context.

Questions students ask

What does a p-value actually mean?

The probability of data at least this extreme if the null hypothesis were true β€” not the probability the null is true.

Practise statistics until it sticks

Noeta remembers every step you missed and brings it back right before you forget it β€” 10 free questions a day, no card required.

Start free

Related topics