Next: Inferences with two samples
Previous: Tests on a population proportion

Rather than pre-specify a value for a, we try to compute the smallest value of a (also known as the significance level) for which the hypothesis H₀ would be rejected and report this value (called the P-value) for the test. Thus

• H₀ would be rejected at level a for any a³P-value
• H₀ would not be rejected at level a for any a<P-value

H_a: Upper Tailed q >q₀:   P = 1-F(z)
H_a: Lower Tailed q <q₀:   P = F(z)
H_a: Two Tailed q¹q₀: P = 2*[1-F(|z|)]

Since exact values are not tabulated for every probability, find two probability values such that Value₁< P-value < Value₂. One can then give lower and upper bounds on the P-value.

Next: Inferences with two samples
Previous: Tests on a population proportion