Dawes P J D, Dawes M T, Williams S M
Department of Otorhinolaryngology, Head and Neck Surgery, University of Otago, Dunedin, New Zealand.
Clin Otolaryngol Allied Sci. 2004 Dec;29(6):648-54. doi: 10.1111/j.1365-2273.2004.00873.x.
We have examined the repeatability of a perceptual space, or smell map, initially developed in a sample of 20 people. A total of 110 normosmic people performed similarity ratings for pairs of the 11 commonly encountered odours used in the previous study. Multidimensional scaling was again used to analyse the relationships between odours. Some individual plots and an archetype, generated using the data from all participants, produced results similar to the previous study, confirming the ability to obtain reproducible results using a different tester and different participants. Exploration of the source space, that is, the spatial distribution of the individual plots relative to each other, suggests that the current model explains about 50% of the variation between subjects for the two-dimensional solution and 57% for the three-dimensional solution.
我们检验了最初在20人样本中开发的感知空间或气味图谱的可重复性。共有110名嗅觉正常的人对先前研究中使用的11种常见气味的两两组合进行了相似性评级。再次使用多维标度法来分析气味之间的关系。使用所有参与者的数据生成的一些个体图和一个原型产生了与先前研究相似的结果，证实了使用不同的测试者和不同的参与者能够获得可重复的结果。对源空间的探索，即个体图相对于彼此的空间分布，表明当前模型对于二维解解释了约50%的个体间变异，对于三维解解释了57%的个体间变异。
