return of the free breathalyser

We report the results of the second generation BAC test. Alcohol consumption was standardised between two subjects and BAC measured at regular intervals. We found the BAC varied widely between the two subjects, but when scaled to take into account the masses of the subjects, good agreement in BAC was found, leading us to the conclusion that BAC is determined, to first order, by body mass, rather than any metabolic differences.

Brought to you by Bob & Alice1

After the immense success of the first experiment (and successive failure of any commercially available consumer breathalysers), it was decided that the only way to conduct a follow-up experiment was to return to the scene of the crime: Parkes.

This experiment aimed to explore a different aspect of alcohol absorption; namely, the effect of alcohol on different people. Our two subjects, Alice and Bob, boldly agreed to drink the same amount of alcohol at set intervals until the conclusion of the experiment. This experiment would then probe how the BAC of Alice and Bob, who were different genders and masses, would evolve as a result. Due to Alice's lack of appreciation for the fine taste of beer, Bob made the noble sacrifice of agreeing to drink some form of rosé. And so the experiment began.

You're probably wondering about the specific details of the experiment. For example, did they eat the same food during the time interval over which the experiment took place? Almost - Alice isn't a fan of vegetables and Bob wasn't a fan of seafood. So there were marginal differences in their food consumption. What did they drink? What percentage alcohol? In all there were four bottles of wine between the two subjects, two of 6.5% rosé and two of 9% rosé. A glass was consumed every 20 minutes for 3 hours, when Alice caved2. Another question you may or may not be wondering is whether they both went to the bathroom equally. The answer is no - Alice's bladder was clearly inferior and untrained compared to Bob's.

Here are the plots, a) data only (with cubic interpolation in colour), b) the linear fits through the data points, and c) an alternative visualisation of the data, including data taken after the 3-hour cutoff:

A note on the plots: the data are coloured as Alice (green), Bob (blue) and scaled-Bob (red). An improvement in this experiment compared to last time was the use of averaging on the data point to minimise the effect of machine error. Each data point represents the sum of three separate measurements of the BAC. The error bars on the plots represent the extent of this error. In general we find no difference between the interpolated cubic data and the raw data, evident in the linear fits as well. We were able to probe the effect of mass on BAC, noting that Bob's weight is 1.7931 times Alice's weight. The additional line, labelled on the plots, represents this scaled data which we refer to as scaled-Bob. We find that the mass-scaling agrees quite well with the data, although it deviates from Alice's BAC at later times. This may suggest that Alice absorbs alcohol at a slightly fast rate than scaled-Bob, or that mass is not the only factor to be taken into account when determining BAC absorption rate.

We note that at low BACs, the spread in measured data points is larger than at higher BACs. The data only becomes very stable at points corresponding to a BAC of about 0.01. Since there was never a BAC measured lower than 0.08 on the Coachman breathalyser, we suggest that any measurements of BAC below 0.015 should be regarded with skepticism, as the noise/error is likely to be around 0.005 or higher. We have no explanation for the sudden large error in Alice's BAC measurements at 1.5 hours.

Overall, we found that the scaling of BAC is in general agreement with a mass proportionality, although the slope of the scaled-Bob data is still distinct from the slope of the Alice data. We note that there was a slight time delay between Bob's measurements and Alice's (a consistent 5 minutes), but that shifting by this time offset does not make a difference to the gradient nor the agreement in data, suggesting that the scaling is genuinely different. As a minor sidenote, there was some evidence separate to this experiment that indicated the effectiveness of hyperventilation in temporarily reducing BAC. Other areas of interest for the future involve probing the high-time-resolution BAC domain and monitoring the decay of BAC for different test subjects.

The results of this experiment have been very useful in testing alcohol uptake during a standardised drinking phase. However, the results cannot be used to determine how the alcohol was processed in the bodies of Alice and Bob once consumption had ceased. We propose a further experiment where a high concentration of alcohol is ingested over a short amount of time (eg. a middy of vodka), followed by regular monitoring of subjects' BAC, as the alcohol is processed3.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License