Polygraph failure rates vary widely depending on the agency and polygraph program involved. For example, the failure rate for the Department of Defense's counterintelligence-scope screening program is very low: it seems the only persons who "fail" are those who make substantive admissions. By contrast, the failure rate for the FBI's pre-employment screening program is about 50%. But the failure rate for the FBI counterintelligence-scope screening program (of current employees) is reportedly close to 1%.

Abe,

I don't know what the false positive rates of any particular polygraph screening programs are. Determining a false positive rate would be difficult because those who "fail" typically cannot prove they didn't lie (one cannot prove a negative). In addition, agencies that rely on polygraph screening don't publish estimated false positive rates.

However, one can reasonably infer that an agency's false positive rate will be directly proportional to the overall failure rate and inversely proportional to the base rate of guilt among the population being screened for a particular behavior.