NCJ Number
215434
Journal
Canadian Society of Forensic Science Journal Volume: 39 Issue: 2 Dated: June 2006 Pages: 55-64
Date Published
June 2006
Length
10 pages
Annotation
This study compared the precision between "truncated" (rounding unit values of one through nine down to zero) and untruncated (using an estimation for the third digit to the right of the decimal point) in duplicate breath tests for the presence of alcohol conducted in the field by qualified breath technicians of the Toronto Police Services (Canada) over a 4-year period, using the Intoxilyzer 5000C.
Abstract
The findings show that qualified breath technicians operated the Intoxilyzer 5000C with a high degree of precision, resulting in 97 percent of all subject tests being within 20 mg/dL with the untruncated approach to calculations. Following truncation, only 2 percent of breath tests with absolute differences of 21 to 29 mg/dL were within 20 mg/dL or "good agreement." No instances of breath tests 29 mg/dL apart were determined to be "in good agreement" over the 4 years (1995-1998) of the study. The study concludes that truncation in calculating Intoxilyzer 5000C results in no way invalidates the scientific acceptability of these results nor does it unfairly prejudice drinking drivers. A total of 8,585 breath tests were analyzed. The Intoxilyzer 5000 series has sufficient memory to hold the results of 50 paired breath tests. The data was analyzed to determine the number of tests that were within 20 mg/dL of each other truncated compared with untruncated. The truncated values with an absolute difference of 21-29 mg/dL were then analyzed for an effect of blood alcohol concentration and whether there was any trend for the first or second test to be the lowest. 1 table, 6 figures, and 11 references