Imagine you are in charge of investigating the health status of a group of animals. Among them, some animals are truly healthy (non-infected, disease -) and other are truly sick (infected, disease +). To estimate how many animals are infected, you apply a diagnostic test to all animals of this population: test results are negative (test -) or positive (test +).

Image source

If the assay you use is never wrong, then it is a perfect test (also called "gold standard" or "perfect reference test"). In that case

• It always gives true positive results, i.e. positive results for all sick (infected) animals
• It always gives true negative results, i.e. negative results for all healthy (non-infected) animals

However, most of the time, the test you use isn’t a gold standard. In consequence

• It sometimes gives false positive results, i.e. positive test results for some healthy animals
• It sometimes gives false negative results, i.e. negative test results for some sick animals