An application of probability theory to a group of breath-alcohol and blood-alcohol data.
 Many jurisdictions have "per se" driving-while-intoxicated (DWI) status expressed in terms of a blood-alcohol concentration (BAC) standard (in grams per 100 mL or the equivalent).
 Since breath-alcohol (BrAC) analysis is typically employed to determine BAC, there is often challenge to the use of an assumed 2100:1 conversion ratio.
 This concern may be relevant in light of considerable data that show a low percentage of cases in which BrAC greater than BAC, and this concern increases when the BrAC is used to predict BAC in the context of "per se" legislation.
 Probability theory provides a basis for estimating the likelihood of an individual having a BrAC greater than or equal to g/210 L with a corresponding BAC less than 0.10 g/100 mL.
 Actual field data from the state of Wisconsin (n = 404) were evaluated to determine the probability of this occurrence.
 The probability for this occurrence involves the multiplication law for independent events.
 The computed probability from the data was 0.018.
 The actual number of occurrences where BrAC greater than or equal to 0.10 g/210 L and BAC less than 0.10 g/100 mL was 5, resulting in a probability of 0.012.
 The concern of having BrAC greater than BAC at the critical "per se" level has a very low probability of occurrence, which thus supports the reasonableness of "per se" DWI legislation based upon a blood-alcohol standard determined by breath-alcohol analysis.
