If one in 1, 000 people have a disease and the false-positive rate for a test is 5%, then the probability that a person testing positive has the disease is at most 2%.

ECONOMIST: Letters | The

The probability of a false positive in a chemical assay for a drug or poison (something appearing to show up when it isn't really there) is reasonably well known.

ECONOMIST: Whose body of evidence? | The

That probability could be seen in the action of gold on Friday after release of the very positive employment report.

FORBES: Even After Jobs Report, Fed Cannot Stop Runaway QE2

If the true-positive rate is significantly less than 100%, as it would be for cheaper tests, then the posterior probability could be only 1% or even less.

ECONOMIST: Free-trader Barshefsky