The objective of this study was to compare the number of drivers with drug concentrations above the legal cutoffs for driving under the influence of illicit substances in paired samples of blood and oral fluid. Between January 2008 and September 2009, 2,949 randomly selected drivers participated in a roadside survey. Each was asked to provide blood and oral fluid. Samples were analyzed for 11 illicit substances or metabolites by ultra-performance liquid chromatography–tandem mass spectrometry and gas chromatography–tandem mass spectrometry. Out of the 2,750 drivers who gave both blood and oral fluid, 28 (1.0%) had drug concentrations above the legal cutoff in blood and 71 (2.6%) were above the legal cutoff in oral fluid. Fifteen (7.5%) of the 199 drivers who gave an oral fluid sample but refused to provide blood tested positive, significantly more than drivers who provided both samples. Based on oral fluid analysis, 2.6 times more subjects tested positive for drugs compared to blood analysis. Those that refused to give a blood sample were 3 times more likely to test positive for drugs. Even in a survey that guaranteed total anonymity, people fearing a positive test result might have been more likely to refuse to give a blood sample.