Number of panellists or number of replicates - how to receive most information for a GC-O-MS study.

In a detection frequency analysis, a high number of panellists is necessary. As welltrained panellists can be difficult to recruit it can be of interest to use the same\npanellists in replicates of the same samples. To gain more information about the samples the variation is required to be between samples and not between individual\npanellists. This study shows that for samples with a low content of odour-active compounds, the variation is often between samples and not between panellists. In\nthis case, when using the same panellist several times, the amount of information about the sample still increases. In samples with a higher content of odour active\ncompounds, the variation is more often between panellists. In this case, using the same panellist in replicate does not increase the amount of information about the\nsample.