Application of a model to normalize human olfactory response to odours

Eight environmental and two n-butanol samples were delivered to four olfactometry laboratories across the North America and measured at approximately the same time. The n-butanol samples were from certified commercial gas and were used as reference. By applying the normalization model developed in previous research, the standard error of the measurements among the four laboratories is decreased by 36%, the variation decrease by 38%, and the extremely value of the Max/Min ratio decreased by 55%. Three-year's data of two olfactometry laboratories were reviewed. The data set included measurements of 1512 environmental samples and 222 n-butanol references within 119 panel sessions. After normalization, the measurement values of environmental samples approach their true values. The log standard error of environmental samples is decreased by 62.8%. This result indicates that the consistency of members' measurements within a panel on the same sample is significantly increased.