We’ve all heard about the terrorist with high explosives in his pants who got through airport security despite being on the TSA’s watchlist. If the null hypothesis is that the passenger is not a terrorist (innocent until proven guilty,) then this is a case of failing to reject the null hypothesis when it is actually false, a Type II error. As you might suspect, the other type of error occurs as well, non-terrorists are subjected to additional security. The situation is summarized in the table.
|H0: Passenger is not a Terrorist
HA: Passenger is a Terrorist
|Passenger is Not a Terrorist (H0 correct)||Passenger is a Terrorist (HA correct)|
|Passenger is Passed Through Security||Correct||Type II Error|
|Passenger is Frisked at Security||Type I Error||Correc|
More than one list is involved here. One is the well-known “no fly list,” which includes about 2,500 names, 90 percent of which are not US Citizens. The other is the Transportation Security Administration’s (TSA) “automatic selectee” list — its list of about 13,500 names of people who are not permitted to board an aircraft without being given the once-over by the agency’s machines and uniformed, latex-gloved personnel — is based on people’s names, not on physical factors like age. One such automatic selectee, 8-year old Cub Scout Mickey Hicks, has been on the list, apparently, since birth. For the Hicks family, which travels a lot, this is something of a hassle.
Statistical risk is unavoidable. Whenever decisions are made there is a chance that the wrong decision will be made. The questions involved are extremely difficult, and extremely important. In this case the question is about the tradeoff between liberty and safety. However, according to the NY Times,
For every person on the lists, hundreds of others may get caught up simply because they share the same name; a quick scan through a national phone directory unearthed 1,600 Michael Hickses. Over the past three years, 81,793 frustrated travelers have formally asked that they be struck from the watch list through the Department of Homeland Security; more than 25,000 of their cases are still pending. Others have taken more drastic measures.
It is one thing to err on the side of safety when adding names to a secret list, but there is room for debate about the correct error rate. Hundreds to one seems a bit excessive, to me anyway.
The problem is exacerbated by the extreme difficulties encountered in getting one’s name off of the automatic selectee list. Thousands of people who aren’t terrorists cannot prevent the list from misidentifying them, causing them delays and embarrassment when trying to board commercial aircraft. In a free country it is incumbent on the authorities to make every effort to correct their errors once they have become known.
Another issue is that the TSA’s list is also used for purposes other than airport security. Sen. Frank Lautenberg (D-N.J.), encouraged by New York City mayor Michael Bloomberg, wants to prohibit anyone on the FBI’s terrorist watchlist from possessing a firearm. Yet, the list and its criteria are secret, and Lautenberg’s bill would criminalize the exercise of a constitutionally protected right while denying a person the opportunity to clear himself of accusations in a fair and open hearing before a court of law. Besides that, one can argue that the tolerable error rates for the purpose of airport security would be different.
Also, it makes me very nervous to hear about lists of citizens being kept by governments. These are not people who have been accused of any crime. They are not being sought to be brought in for questioning or for trial. There are no warrants out for these people. History tells us that such lists will inevitably be misused by those in power, often to terrible ends.
The TSA’s process for doing this is obviously broken and in need of quick and drastic improvement. Six Sigma can help the TSA do this. One can only hope that they take advantage of it soon. Our safety and liberty depends on it.