Breathalyzers are wrong more often than most people assume. The margin of error on breath testing devices can range from 5% to as high as 50% depending on the device, the person being tested, and whether proper procedures are followed. Even under ideal conditions, every breathalyzer relies on a biological assumption that varies significantly from person to person, which means the number it displays is always an estimate rather than a precise measurement.
The Built-In Assumption That Creates Error
Every breathalyzer converts the amount of alcohol in your breath into an estimated blood alcohol concentration (BAC). To do this, the device assumes a fixed ratio between alcohol in your blood and alcohol in your breath: 2,100 to 1. That means it treats every person as though 1 milliliter of their blood contains exactly 2,100 times more alcohol than 1 milliliter of their breath.
The problem is that this ratio varies widely between individuals. A study of 100 subjects published in Forensic Sciences Research found that actual blood-to-breath ratios in the post-absorptive phase (after drinking has stopped and alcohol is being eliminated) ranged from 2,125:1 to 2,765:1, with a mean of 2,382:1. No single subject in the study had a ratio as low as the 2,100:1 that breathalyzers assume. This means that for most people, a breathalyzer will slightly overestimate their true BAC because the assumed ratio is lower than their actual ratio. Different countries have acknowledged this uncertainty by adopting different conversion factors. The UK uses 2,300:1, while the US, Canada, and Australia use 2,100:1. There is no international consensus on the correct number.
Fuel Cell vs. Infrared Devices
Not all breathalyzers use the same technology, and the type of sensor matters for accuracy. The two main categories are fuel cell sensors and infrared spectrometry.
Fuel cell breathalyzers are the handheld units officers use during roadside stops. They work by oxidizing alcohol on a platinum electrode and measuring the electrical current produced. These devices are portable, inexpensive, and fast. Their accuracy at a BAC of 0.100 typically falls within plus or minus 0.005 to 0.01, depending on the model. They perform best in the range around legal limits (0.050 to 0.100 BAC) because that is the range their manufacturing standards are designed around. At higher BAC levels, error increases.
Infrared breathalyzers, like the Intoxilyzer 8000 used in police stations, identify alcohol by measuring how it absorbs infrared light at specific wavelengths. These machines are larger, more expensive, and require 5 to 10 minutes to recalibrate between readings. That longer cycle time actually helps accuracy: the wait gives the person being tested time to return to a normal breathing pattern, which reduces sampling errors. Infrared devices are also less susceptible to contamination from alcohol lingering in the mouth.
Despite their differences, a psychometric evaluation published in PeerJ found “little difference in BAC readings between fuel-cell and infrared spectroscopy based” instruments when both were properly used. The key distinction is practical: fuel cell units are more vulnerable to environmental conditions and mouth alcohol, while infrared units are more controlled but require trained operators and a station setting.
What Causes False High Readings
Several common situations can inflate a breathalyzer result beyond your actual BAC.
Mouth alcohol is the most frequent culprit. If you burp, belch, or vomit shortly before a test, alcohol from your stomach rises into your mouth and gets measured alongside the alcohol in your deep lung air. Acid reflux (GERD) can do the same thing silently, pushing stomach contents upward without you even noticing. Fuel cell sensors cannot distinguish between alcohol coming from your lungs and alcohol sitting in your mouth, so the reading spikes.
Mouthwash can produce dramatic false positives. A study measuring breath alcohol after rinsing with Listerine (26.9% alcohol) found readings averaging 0.240 BAC two minutes after use, three times the legal limit, in people who had consumed no alcohol at all. Scope (18.9% alcohol) produced average readings of 0.170. These readings decay rapidly, falling well below 0.080 within 10 minutes for all brands tested. Using mouthwash immediately before a test, perhaps to mask the smell of a drink, can significantly inflate results.
Ketosis from low-carb diets, fasting, or diabetes creates a subtler problem. When the body burns fat for fuel, it produces acetone. Acetone itself doesn’t trigger fuel cell sensors, but under certain conditions the liver converts acetone into isopropanol, a different type of alcohol. Fuel cell breathalyzers respond to isopropanol the same way they respond to ethanol, producing a false positive reading in someone who may not have had a drink at all.
Why the 15-Minute Wait Matters
To guard against mouth alcohol contamination, officers are required to observe a suspect for a minimum of 15 minutes before administering a breath test. During this period, they watch for drinking, burping, belching, or vomiting. If any of these occur, the clock resets. The scientific rationale is straightforward: mouth alcohol dissipates within roughly 15 minutes, so waiting ensures the device measures only deep lung air.
In practice, this observation period is one of the most commonly challenged aspects of breathalyzer evidence. Officers transporting a suspect in a patrol car cannot effectively monitor the person’s behavior during the drive. Courts have suppressed breath test results in cases where the observation was conducted during transit rather than in a controlled setting. If the observation period is skipped or interrupted, the results may be legally invalid regardless of what the device displayed.
Calibration and Maintenance Failures
Breathalyzers drift out of accuracy over time and require regular calibration to stay reliable. Under U.S. Department of Transportation rules, every evidential breath testing device must have a manufacturer-submitted quality assurance plan that specifies calibration check methods, acceptable tolerances, and the intervals at which checks must be performed. Users of the device, whether law enforcement agencies or employers, are required to follow these manufacturer schedules.
When calibration lapses, the device may read consistently high or low without any outward sign of malfunction. Defense attorneys routinely request calibration logs, and missing or overdue records can be grounds to challenge a breath test result. The machine itself won’t tell you it’s out of calibration. It will simply return a number with false confidence.
Breathing Patterns and Lung Capacity
How you breathe into the device also affects the result. Breathing patterns are a recognized source of biological sampling error. A person who hyperventilates before blowing can produce a lower reading, while someone who holds their breath or blows hard at the end of a long exhalation can produce a higher one, because the last air out of the lungs has a higher concentration of alcohol.
Lung capacity matters too. Smokers, people with asthma or COPD, and those with smaller physical lung volume may not produce a breath sample with adequate volume for the machine to analyze properly. If the sample is too small, the device may either reject it or attempt a reading from insufficient data, both of which introduce error.
How Often This Affects Real Cases
Putting exact numbers on “how often” breathalyzers are wrong depends on what counts as wrong. If you define error as any deviation from true BAC, the answer is essentially always: the partition ratio assumption alone guarantees some degree of imprecision for every person tested. If you define it as a result that could change whether someone is over or under the legal limit, the number is smaller but still significant. Someone blowing a 0.08 or 0.09 is within the device’s margin of error of the legal threshold, meaning the reading alone cannot reliably distinguish between legal and illegal.
The American Motorists Association has cited study data suggesting breath testing devices carry a margin of error as high as 50%, though this figure represents worst-case scenarios combining multiple sources of error rather than typical performance. Under controlled conditions with proper calibration, a well-maintained device tested on a cooperative subject in the post-absorptive phase will typically land within 0.005 to 0.01 of the true BAC. The gap between that lab-condition accuracy and real-world performance, with all its variables of reflux, breathing, mouth alcohol, and calibration drift, is where most errors live.