Stand up!

Heels together. Toes out. Hands at your sides. Raise the leg of your choice right in front of you, six inches off the ground, leg straight, toe pointed. Keep your eyes on your raised toe and begin counting aloud from 1,001 until I say stop.

Do you understand? Begin.

One thousand one. One thousand two . . .

Keep going.

Some dark night on the side of the road, police lights flashing in your peripheral vision, your freedom may depend on how well you do this.

Did you sway? Raise your arms for balance? How about hop? Or put your foot down? If you did any two, a police officer will conclude with 65 percent accuracy, as stipulated in the prevailing science of inebriation diagnostics, that you may be too drunk to drive.

And if you bent your leg, stared straight ahead instead of at your foot or began before I said so, you may be in trouble. Police officers are taught that people under the influence of alcohol don't follow directions well.

If you made it through 30 seconds ramrod straight, congratulations! You may not be drunk.

This is the one-leg stand -- OLS in cop-speak. It is one of the three scientifically researched standardized field sobriety tests, blessed by the National Highway Traffic Safety Administration, that officers call "the holy grail" and give on the side of the road to help them decide whether to make a drunk driving arrest.

NHTSA says police officers who use scores from all three tests combined will be 91 percent accurate in making an arrest. However, in Washington last May, Officer Dennis Fair arrested Debra Bolton, an Alexandria attorney, after determining that she'd failed all three. Bolton, who'd had a glass of red wine with dinner, hotly disagreed. She said she distinctly remembered mentally thanking her yoga instructor for her steady balance during the one-leg stand, and challenged the charges until they were dropped in late August.

The most accurate of the three, according to NHTSA, is the horizontal gaze nystagmus test -- or the "jerking eyeball" test. A police officer will hold a penlight or small flashlight in front of you and ask you to track it visually from side to side. If you've had too much to drink, your eyeballs will begin shaking at about 45 degrees from center.

The two other tests, the one-leg stand and the walk and turn (nine steps forward and back on a straight line) are "divided attention" tests that require both mental concentration and physical coordination. The one-leg stand has its skeptics and its court challenges, and plenty of them, but, according to NHTSA, the test is "easily performed by most unimpaired people."

Oh, really?

On a recent sunny afternoon in Dupont Circle, Franklin Urena, 32, a waiter at Chevy's, couldn't do it. "I have flat feet," he explained as he started hopping at the count of 1,021.

His friend, Henry Van Dyke, 50, didn't make it past the count of 1,003. "Maybe if I had a glass of wine I could do better because I wouldn't be so self-conscious," he said. "I have no coordination."

Christine Ju, Elisa Catalano and Justin Sullivan, all in their early thirties, balanced well but didn't follow directions. "What an absurd test," sniffed Catalano, a yoga teacher.

Forget Alex Filatov, 46, a visitor from Russia, who admitted he was "drunk to [expletive]" and reluctantly put down his brown paper bag with two bottles of zinfandel-Shiraz to take the test. He failed. "But I'm not driving," he said with a giggle. "I take taxi to the airport."

In a completely unscientific test of 14 random people, five passed, seven failed and two were judgment calls -- one because he counted quickly in Italian and the other because it was unclear whether her wobbles would count as a sway.

Antonio Thomas, 28, a bellhop in Georgetown, performed by far the steadiest one-leg stand, even though the former high school football player has had surgery on both knees. He said the test wasn't difficult at all. "And I just came from a bar. I just had a beer," he said sheepishly. "Miller Genuine Draft."

Marcelline Burns, 77, is a research psychologist who calls herself the "grandma guru" of the standardized field sobriety tests. She has made a nice living in a long career developing them, testifying in courts across the country as an expert witness and training police officers how to use them correctly. She scoffed at our unscientific experiment.

"Well pardon me, but I don't think your use of the test is good evidence," she said on the phone from Southern California. "And I do think the fact that someone is confronting a uniformed officer may have an effect on how hard they try."

The average person, she insists, should be able to balance on one leg for 30 seconds. She can. She practices the one-leg stand every few weeks. And she boasts another advantage: "I'm very fit."

Test Patterns

Before Marcelline Burns and the one-leg stand, police officers were on their own.

Some threw coins on the ground and ordered that only nickels or quarters be picked up as a way of figuring out a driving-while-intoxicated arrest. They would have a driver lean back and touch one finger to his nose. Recite the alphabet without singing. Count backward from 100 by threes. Trace a paper maze. Rapidly tap a telegraph key. Some gave tongue twisters such as "Methodist, Episcopal, sophisticated statistics." Texas Rangers just chatted for a bit before making a judgment call.

Chuck Hayes, a state trooper for 30 years in Oregon, remembers fellow officers putting their flashlights down on the ground and telling drivers to run around them five times. "I mean, how are you going to tell if a person is impaired that way?" he said. "There was just some weird, weird stuff."

So in 1975, NHTSA put out a request for proposals to develop a valid standardized test battery easy to use on the side of the road. Burns and her colleagues at the nonprofit Southern California Research Institute in Los Angeles got the job.

They recruited 238 subjects from the local unemployment office -- anyone over 21 with a driver's license and who admitted to imbibing a few -- and paid them $3 a day to be part of the experiment. Subjects came to the lab, starting at 8 a.m., and were "dosed" with either a placebo of orange juice or a screwdriver. ("I did this research for 30 years and we always used screwdrivers," Burns said. " . . . Most everyone will drink a screwdriver. ")

The subjects, in varying states of sloshedness, were then led to small rooms where 10 California police officers awaited them. The officers administered six different sobriety tests, then made a determination on the subject's blood alcohol content and whether they would make an arrest.

Burns's final report, "Psychophysical Tests for DWI Arrest," was published in 1977. She wrote that while all six tests were sensitive to alcohol -- meaning that drunk subjects tended to perform worse than sober ones -- the "best" tests were the three that are now in use today.

Burns did not test the drunken subjects when sober to see how well they could balance on one leg naturally.

"The evidence that it's an easy task comes from the placebo people," she said. "They could do it fine."

Let's look and see just who those placebo people were. Here's a bar chart from her study: Hmm. Most of them were men. Young men between the ages of 22 and 29.

One could argue the placebo people didn't look much like America.

So hundreds of thousands of drivers have been arrested -- no doubt many deservedly so -- on the basis of a 30-year-old study that, critics argue, has never been published in a peer-reviewed, scientific journal, never tested on a large scale with a control group and, perhaps more astonishing, has nothing to do with actual impairment from alcohol. Burns is quite upfront in admitting the tests are designed only to gauge blood alcohol content, not whether you're a menace on the road.

Steady as They Go

Consider how imbalanced we are. Forty percent of Americans will at some point in their lifetime experience a balance disorder, and dizziness and vertigo are the third leading cause of doctor visits, behind lower back pain and headaches, according to the National Institutes of Health.

The research on how humans stand on their own two legs indicates that the older you get, starting at about age 50, the less likely you will be able to stand on one leg for very long, and the more likely you will be to fall. As for what's the "normal" ability to balance, that's anybody's guess.

Even Burns writes that people 50 pounds overweight, over age 65 or who have had injuries may not be able to balance well.

"The human equilibrium is a very complex system," said Richard E. Gans, founder and director of the American Institute of Balance in Florida. "Some people refer to it as the sixth sense, and that doesn't mean 'I see dead people.' "

The body must use three different systems in order to balance -- the vestibular system that works like "finely tuned gyroscopes" in each inner ear, the visual and the skeletal systems, particularly the muscles, joints and tendons in the lower half of the body.

Ever had encephalitis, meningitis, shingles, chickenpox, ear infections, cardiovascular problems, numbness or tingling in the extremities or migraines? You may not be able to balance. Diabetes? You may not be able to feel your feet well enough to balance.

John Schumacher, founder and director of Unity Woods, the largest yoga center in the Washington area, said he's seen beginning balancers all over the map. Some get better with practice and some don't. The ones who totter have weak muscles, bad ankles, fallen arches or are having a bad day. "Sometimes people are distracted or upset emotionally -- that makes balancing harder," he said. "Just the stress of being observed by a police officer in a stressful situation like that could throw anybody's balance off."

Chuck Hayes, now a field sobriety test trainer for the International Association of Chiefs of Police, said that officers are trained to ask about injuries or medical problems and take them into account when scoring the one-leg stand.

"A lot of the American public just has a tough time with balance, period," he said. "If an officer sees that, and if there weren't any clues on the other two tests, that's a no-brainer: This person is not impaired."

But how many other officers would make the same call?

"I can't say. That's why training is so important." (To make sure they have more than a dry classroom experience, Hayes gives "wet workshops": Subjects drink beer or 80-proof liquor and then perform the tests for officers to score.)

To Hawa Waleed, an Alexandria chiropractor, there is too much human variation to attribute the inability to balance solely to alcohol.

"Maybe the best thing you can do is, before you get your driver's license, have everybody take these tests, so at least you have a base line," he said, only half-joking.

What Next?

Troll the Internet, and you will quickly find disparaging reports with titles like "Field Sobriety Tests: Designed for Failure," "Field Sobriety Tests: The Flimsy Scientific Underpinnings" and "Voodoo Science."

Some forensic psychologists and, understandably, a slew of DWI defense attorneys have been assiduously picking apart Burns's research on the standardized field sobriety tests for years.

But she is unmoved. "We're now 30 years past the development of the test. They're widely used by police officers. Why would they use them if they don't help them make a proper decision?" she said. "These defense attorneys write all this stuff, but never once do they suggest an optional test. What do they want the officer to do? Toss a coin?"

Not at all, says Spurgeon Cole, a forensic scientist and consultant in Georgia who has been her chief nemesis in court and expert witness for the defense for years. But maybe videotapes in patrol cars, he argues, would help remove some of the subjectivity.

"We have no idea how well a sober person can perform on the SFST [field tests]. How does age or gender affect performance? How does fatigue or practice affect performance?" he has written. "Without answers to these basic questions, the SFST remain in the same category as tarot cards."

Cole did his own study, administering the tests to 21 of his students at Clemson University in South Carolina -- none of whom had had a drop of alcohol -- and then showing the videotape of their performance to a group of officers. They officers reported they'd arrest nearly half the students.

"And these people had absolutely zero to drink," Cole said in an interview. "These tests are absolutely worthless."

The battle over the one-leg stand and sobriety testing has raged in the courtroom. In Florida, the courts have decided that prosecutors cannot refer to them as "tests" that one can pass or fail. Rather, they are considered subjective "observations." In Ohio and other states, the courts have held that unless the tests are administered in the standardized way, they are "inherently unreliable and thus inadmissible."

And retired police officers, such as former New Jersey state trooper Gil Snowden, have made lucrative careers as expert defense witnesses, showing how often the police get it wrong. "The tests are reliable if the police do it the right way," he said, "but I see so many errors."

In Fairfax County, well, Judge Ian M. O'Flaherty hasn't even let prosecutors get as far as the field sobriety tests. He's been throwing out drunk-driving cases -- five in one recent week -- that automatically presume one is intoxicated with blood alcohol content over .08 on the grounds that the law itself is unconstitutional.

So what's a police officer to do?

Scientists are developing roadside saliva tests to measure blood alcohol, but that's years away. A Texas A&M researcher has come up with a video game-like machine to test memory and physical dexterity and record mistakes, timing and accuracy. A computer then calculates your intoxication level.

Some researchers like the "alternating hand pat" test, in which the subject hits the back of one hand with the palm of the other and then alternates hands. Others believe some kind of driving simulator would be best. Maybe someone driving around with it in a mobile van. A closed driving course, while impractical, would be best measurement tool, everyone agrees. But previous studies have found that about 20 percent of drivers actually improve with alcohol. "I guess it calms them down," said Cole.

That leaves us right back where we started, with the one-leg stand.

One thousand one. One thousand two. One thousand three . . .

You put your right foot in: A Miami police officer administers a sobriety test. Florida courts have ruled that such measures must be referred to as "observations," not "tests."Marcelline Burns defends the field sobriety tests she developed in 1977.A Minnesota policeman tests a driver. Many people have "a tough time with balance," says one former state trooper.