In a number of previous posts I’ve discussed the dubious efficacy, statistics and constitutionality of DUI roadblocks, aka "sobriety checkpoints". However, I’ve only briefly mentioned the potential for abusing the right to set up these roadblocks — the potential for government to use them as a pretext to violate citizen’s rights. Consider the following story from yesterday’s Washington Post:

Safety Stops Draw Doubts

D.C. Police Gather Nonviolators’ Data

Lisa Davis had done nothing wrong. She was wearing a seat belt, was obeying the speed limit and produced a valid driver’s license when D.C. police pulled her over one recent night at a traffic safety checkpoint in a crime-plagued neighborhood.

Even so, an officer jotted down some basic information before letting her go, including her name, address and the time and location of the stop for a police database used for crime solving.

"I’ve got some serious constitutional issues with that," Davis said as she sat in her idling Acura at the checkpoint at Kansas Avenue and Shepherd Street NW in the Petworth neighborhood. "I feel like it’s a violation of my rights. It’s a slippery slope to Big Brother."

The details about Davis and the stop will be fed into the database, which is linked to a computer that includes arrest records and mug shots of criminals….

Civil liberties advocates aren’t the only ones questioning the practice. The policy is sparking concern among some officers who conduct the checkpoint stops, most of which are made in areas where crime, not traffic safety, is the primary concern.

"That’s an invasion of privacy, demanding information from a citizen and putting that in a database," said Officer Gregory I. Green of D.C. police, who is assigned to represent the police union….

The officers jot down motorists’ information on a file-card-size form called a PD-76, which is recorded into the database. The forms also are used for routine traffic stops, and the information will also be used for a racial profiling study, (D.C. Police Chief Charles) Ramsey said.

D.C. homicide Detective Paul Regan said the collection of such data has been "a great intelligence tool".

In the 6-3 United States Supreme Court case holding that DUI checkpoints were permissible, Justice Rehnquist admitted that the stops constituted "seizures" within the meaning of the Fourth Amendment. However, he said, this was justified because "No one can seriously dispute the magnitude of the drunken driving problem or the States’ interest in eradicating it."

Would the Court would have ruled as it did had the justices known what these supposed "sobriety checkpoints" would eventually become?

(Thanks to Steve Oberman of Knoxville, TN.)

One of the dirty little secrets in DUI law enforcement is that breathalyzers are susceptible to error caused by radio frequency interference (RFI), sometimes called electromagnetic interference (EMI). Put simply, any electronic device in the vicinity of the breathalyzer can emit electrical energy which can interfere with the circuitry of the machine, causing false test results. (A common example of the problem can be found in restaurants, where signs saying "Warning: Microwave in Use" alert customers to the danger of radio frequency interference with heart pacemakers.)

The police station where the tests are usually given is, of course, a veritable jungle of devices emitting electromagnetic energy — computers, cell phones, fax machines, police dispatch transmitters, teletypes, AM-FM radios, copy machines, hand-held "walkie-talkies", radar units, security cameras, microwaves, electronic locks, transmitters in police cars in the parking lot, fluorescent lighting, and so on….And in the middle of all of this sits the breathalyzer.

The problem is not a new one. In 1983, the National Bureau of Standards quietly prepared a preliminary report on tests performed on the various breath testing devices used by police agencies nationwide (Effects for the Electromagnetic Fields on Evidential Breath Testers). Each of the 16 models tested were subjected to four different frequencies typically present in the standard police environment. Of the 16 units tested, 6 showed minimal interference; 10 of the 16 showed substantial susceptibility on at least one frequency. The report characterized the potential effect of RFI on the testing of alcohol as "severe". Those conducting the study noted that the local Washington D.C. Metropolitan Police Department was complaining that breathalyzers were giving erroneous breath alcohol readings in the presence of radio transmissions.

In a field demonstration of the RFI problem for representatives of NBS and the National Highway Traffic Safety Administration, D.C. officers using a breathalyzer in a mobile van showed how handheld radios radically affected the analysis of breath samples. To avoid a loss of public confidence in breathalyzers, the report was kept confidential — until attorney Don Nichols of Minneapolis successfully filed a legal action under the Freedom of Information Act.

Manufacturers of the various breath testing machines, which had long claimed RFI was simply the invention of defense lawyers, suddenly started offering "RFI detectors" as an option on their products. Predictably, these "detectors" have proven relatively ineffective. First, as repeated tests have demonstrated, there are segments of the frequency band to which the detectors are blind. Second, the detectors are rarely calibrated correctly, if at all. This type of calibration must be done at the factory, but most law enforcement agencies are unwilling to take their machines out of service. Instead, the detector is "calibrated" by a police officer simply holding a hand-held radio next to the machine; if the detector is activated, it is considered "calibrated". Of course, this only indicates that the detector worked one time at the one frequency.

Further, the "calibration" is rarely done during an actual capture and analysis — that is, during actual operating conditions — and so the all circuits are not tested during all phases of the operation. The only real guard against false blood alcohol readings due to RFI is to require duplicate analysis — that is, running two separate tests. This does not eliminate RFI, as a constant source of electromagnetic energy can cause duplicate false results, but it reduces the likelihood. Many states now require duplicate breath tests; many others, however, still do not.

Many law enforcement agencies use videotapes to record a DUI suspect’s driving, appearance, demeanor, slurred speech and/or performance on the field sobriety tests. This taping may be done with a camera mounted in the police car, or with one at the police station (if there are no dedicated recorders at the station, experienced defense attorneys often try to obtain tape from the security cameras).

It is common to encounter situations where a suspect was videotaped but the tape was later erased (the legal term for this erasure is spoliation). This is sometimes done accidentally. Unfortunately, it is often done for a more insidious reason: the tape shows that the arrested person may not have been under the influence — his driving was not erratic, his speech not slurred, his balance and coordination on the field sobriety tests not impaired.

There is a string of United States Supreme Court decisions which deals with the consequences of lost and destroyed evidence generally (Brady-Agurs-Trombetta-Youngblood). Roughly, these decisions require any material evidence to be turned over to the defense if it is requested by the defense. Even without a request (the defense may not know about it), evidence must be turned over if it is exculpatory — that is, if it could have played a material defense role. The loss or destruction of exculpatory evidence constitutes a denial of due process. If the evidence was not clearly exculpatory but was still “potentially useful”, it is a denial of due process only if the loss or destruction was done in “bad faith” — that is, intentionally or for the purpose of denying the defendant access to it.

The burden of proof is on the defendant. The Catch-22, of course, is: How do you prove the erased tape was exculpatory if it has been erased? Or that the erasure was “potentially useful”? Or that it was erased in “bad faith”? Because of these usually insurmountable hurdles, some police officers continue to erase videotapes when their content contradict the damning descriptions in their arrest reports.

For many years now, MADD has focused much of its considerable manpower (over 600 chapters), resources (revenues of $48 million a year) and political influence on the proliferation of DUI roadblocks (or, to use the politically correct phrase, "sobriety checkpoints"). To justify this invasion of our privacy, we have been repeatedly assured that "checkpoints" are extremely effective in reducing alcohol-related traffic fatalities — and these assurances have been accompanied by "statistics". Let's take a closer look at the statistics….

According to MADD's own website, 40 states have checkpoints and 10 do not. Well, it would be interesting to compare the states with the highest percentage of alcohol-related fatalities with the list of states not using checkpoints: If MADD is correct, the states with the highest fatality rates will be the no-roadblock states. Fortunately, another section of MADD's website provides such statistics for each of the states. The 5 states with the highest alcohol-related fatality rates:

Hawaii
Nevada
North Dakota
Rhode Island
South Carolina

According to MADD, all 5 states should be non-checkpoint states. In fact, however, 4 of these states use checkpoints; only Rhode Island does not. Well, what about the 5 states with the lowest fatality percentages? They are:

Georgia
Kentucky
Indiana
Iowa
New York

If MADD is correct about the effectiveness of checkpoints, these should all be checkpoint states. But as with the previous list, only 4 of the states permit the use of sobriety checkpoints; Iowa does not. As with the previous list, the percentage is what one would expect from pure random incidence: 20% of the states (10 of 50) do not have checkpoints — and 20% of the states on each list (1 of 5) do not use checkpoints. There appears to be no correlation between fatality rates and the use of checkpoints.

Let's take a look at another set of statistics: the effect of the proliferation of checkpoints on the national rate of alcohol-related fatalities. If checkpoints are effective, we would expect to find that alcohol-related fatalities will have declined since their widespread acceptance in recent years .

Again, the statistics do not support this. To use MADD's own numbers: Since 1982, the number of fatalities nationwide from alcohol-related crashes has declined every year — until about 1993, when it dropped to 17,908. Perhaps coincidentally, this was the year after the United States Supreme Court ruled that sobriety checkpoints were not unconstitutional. In the 10 years since then, sobriety checkpoints have gained widespead acceptance — but the number of fatalities have levelled off, vacilating between 17,908 and 17,013. Far from supporting MADD's position, one could even argue that this proves sobriety checkpoints have actually halted the steady decline in alcohol-related deaths. This would probably be incorrect — but indicative of how statistics can be used to serve a desired objective.

Incidentally, my favorite example of distorting statistics for self-serving purposes is MADD's own oft-repeated claim:

Since MADD's founding in 1980, alcohol-related fatalities have decreased 44 percent (from 30,429 to 17,013) and MADD has helped save almost 300,000 lives.

300,000? Do the math….

In my previous post, I discussed the problem of mouth alcohol – that is, falsely high breathalyzer readings caused by alcohol samples coming from the mouth rather than from the lungs.  And the response from some of our more sophisticated readers (law enforcement and/or lawyers?) has been predictable:  What about mouth alcohol detectors? Some breath machines have what is called a slope detector, commonly referred to as a “mouth alcohol detector”.  This is an electronic circuit designed to detect the presence of mouth alcohol as the breath is being captured by the machine.  It does this by detecting any pronounced negative slope in the alcohol intake curve, since alcohol content from the mouth or throat will decline more rapidly than alcohol from the lungs.  In theory, the presence of mouth alcohol will cause the test to abort. Unfortunately, these “detectors” are simply unreliable, due primarily to a design flaw.  Rather than try to explain the technical defects, I will let Dr. Michael Hlastala, Professor of Physiology, Biophysics and Medicine at the University of Washington School of Medicine, explain:

The slope detector is problematic for all breath instruments and has been misrepresented by the manufacturers. When a subject with alcohol in the blood, with no extra alcohol in the breath, exhales, the breath alcohol continues to increase during exhalation.  It does not reach a “plateau” until the end of airflow.  It continues to rise, giving a positive slope.  If you swish a little alcohol in the mouth (and have no alcohol in the blood), wait awhile and exhale, the breath alcohol will rise until a peak is reached about 1/3 of the way into the exhalation, and then decline gradually.  It is the declining breath alcohol (negative slope) that triggers the slope detector to register the breath as having mouth alcohol. If the subject has alcohol in the blood as well as the mouth, then the normal rising breath alcohol curve will add to the declining mouth alcohol curve to produce what is often a level curve.  Thus, the slope detector is unable to detect the presence of mouth alcohol when some is present in the mouth, yet breath alcohol concentration will be higher than it should be.  The slope detector cannot detect false mouth alcohol under this circumstance.

To make matters worse, the slope/mouth alcohol detectors are never calibrated by the police, as this has to be done at the factory.  The only thing police technicians do is simply rinse their mouth with alcohol and then breath into the machine: if the detector is triggered, it is assumed to be working.  But as Professor Hlastala has observed:

Whenever the slope detector is checked, it is done with alcohol in the mouth, but not in the blood.  Therefore, the slope detector serves no purpose and mouth alcohol frequently affects the breath alcohol reading.

Bottom line:  Despite the claims of manufacturers anxious to sell their machines, these detectors are unreliable and mouth alcohol remains a serious problem in breath alcohol analysis.