卷烟感官评价中小组得分的两阶段统计法  

Two-stage Statistics Method for Calculating Panel Scores in Sensory Evaluation of Cigarettes

作　　者：王明锋[1] 赵建华[2] 兰海强[2] 宫玉鹏[1] 朱保昆[1] 

机构地区：[1]红云红河烟草(集团)有限责任公司技术中心,昆明市红锦路367号650231 [2]云南财经大学统计与数学学院,昆明市龙泉路237号650221

出　　处：《烟草科技》2014年第5期13-16,共4页Tobacco Science & Technology

基　　金：国家烟草专卖局资助项目"天然植物多孔材料在‘云烟’品牌减害增香中的应用研究"(110201001006);云南中烟公司资助项目"天然香原料在云烟低焦产品开发中的应用研究"(2013FL03);国家自然科学基金"双线性因子分析及其拓展研究"(11361071)

摘　　要：为准确计算卷烟感官评价小组的计分,分别采用简单平均法和两阶段统计法对两次不同时间的感官评吸结果进行了对比分析。结果表明:①受异常评价的影响,简单平均法得到两次不同时间的评价结果存在显著差异的错误结论;②采用两阶段统计法能够消除异常评价带来的影响,得到较为准确的小组评价结果,进而得到两次不同时间的评价结果差异不明显的合理结论。To accurately calculate the panel scores in sensory test inclusive of abnormal evaluation values, the results of two panel sensory tests carried out at different time were compared on the basis of simple average method and two-stage statistics method separately. The findings showed that： 1） Due to the presence of abnormal values, the two results of simple average method differed significantly. 2） There was no significant difference between the two results of two-stage statistics method, which indicated that two-stage statistics method could eliminate the effects of abnormal values, and obtain more accurate evaluation results.

关 键 词：中式卷烟 简单平均法 感官评价 方差分析 多重比较 

分 类 号：TS452[农业科学—烟草工业]