Dynamic Olfactometry Variability in Determining Odor Dilutions-to-Threshold

When conducting research on odor abatement technologies, a weak link has been the inability to place aprecise quantifying number on odor strength that can be accurately reproduced within the same laboratory and betweenlaboratories. Olfactometry is one of the most accepted means for evaluating odor samples. However, insufficient efforthas been made to determine variation between or among individual panelists. This study was designed to analyze threeodor sample strengths over a three-day period by two odor panels, each composed of eight panelists. Three samplestrengths were presented randomly to each panelist three times within each session. Whole-panel variation ranged from a22% to 50% difference in reported odor units for the same sample, depending on sample strength. Using two differentairflow rate calibrations resulted in a 9% to 28% difference in odor units for the same sample, depending on samplestrength. Panelist variation ranged from 4.3- to 7.1-fold, depending on sample strength, although panelist standardizationslightly reduced this variation. Sample order had no effect on odor strength determination. A learning curve for individualpanelists appeared to exist, as odor unit evaluation for later observations were 3.8 and 4.1 times greater than similarevaluations for the first samples. Variation between panelists is approximately the same as variation among evaluationsdone by the same individual; efforts should be made to reduce both these variations. To be able to detect differences inodor concentrations between control and treatment samples, large reductions in odor concentrations with several airsamples are needed to accurately detect a significant reduction in odor.