《市场研究：产品口味测试.ppt》由会员分享，可在线阅读，更多相关《市场研究：产品口味测试.ppt（24页珍藏版）》请在三一办公上搜索。
1、,Types of Preference test,偏好A偏好B偏好C,40%30%25%,?,p1-p2Z=,Sqrt p1(1-p1)+p2(1-p2)+2p1*p2/N,What about N=400?,单样本百分数检验,太甜比较甜正好不太甜一点也不甜,.5.4.3.2.1,?,x Z=,What about N=400?,U,单样本平均值检验,/sqrt N,非常喜欢比较喜欢一般不太喜欢一点也不喜欢,.,Z=,What about N1=300；N2=200?,P1-p2,独立样本百分数检验,sqrt p1(1-p1)/N1+p2(1-p2)/N2,男（N1）,女（N2）,25%,1
2、0%,35%,30%,7%,3%,40%,30%,15%,5%,T2B?,Types of Product Test Discrimination Test(判别测试)Duo-trio test(成对三元测试)Triangle difference test(三元差别测试)Preference Test(偏好测试)Monadic test(单一产品测试)Sequential monadic test(3 products)(顺序单产品测试)Proto-monadic test(第一单产品测试）Paired comparison test(配对比较测试)Round Robin Test(循环测试
3、),Round Robin Test,Why Round Robin Test?Time saving Facilitates more discrimination More easy to operate More accurate response,for based on more clear judgement rather than on vague recall or feeling More easy to differentiate longitudinal data,ApplicationAn extension of paired testPaired compariso
4、n test where 3 or more products involvedFocus on sensory attributes,CharacteristicsTest at least 3 products or stimuli in one single projectFieldwork is conducted just like that in a paired test comparison test or proto-monadic testGenerate large sample size through aggregating a number of small pai
5、red samplesCost effective compared with simple paired comparison test,Fieldwork Operation Each respondent tries 2 products and makes comparison after trialTo counteract sequence-related bias,trial order in rotation among respondentsBlind testProminent code must be marked for identity,Sample size det
6、ermination Test Leg（测试组数）:if there is k test products,there should be L=k(k-1)/2 Test LegsNumber of tests(每个产品在所有组数中出现的次数）:defined as(k-1),meaning each product will be involved in(k-1)Test Legs,Calculate sample size needed Determine the no.of respondents for each test product(denoted as C,where empi
7、rically a minimum of 30 and 100+is normal to facilitate most statistical analysis（确定每个产品被多少人测试）Compute the number of respondents in each Test Leg by the formula n=C/(k-1)（每组的样本数）Calculate the sample size by the formula（确定总样本数）N=Lx n=k(k-1)/2xn,An example 1 Assume there are 6 products,and the expecte
8、d sample size for each product is 100,then No.of test=6-1=5n=100/5=20L=15N=20 x15=300,An example 2 Assume there are 5 products,and the expected sample size for each product is 100,then No.of test=5-1=4n=100/4=25L=10N=25 x 10=250,Compared with 3-paired test A little more samples and cost,E.G.assume 1
9、 product is tested by 100 respondents,then N=2-paired 3-paired 5 products250180 6 products 300200,How to facilitate discrimination Statistical toolAction standard focus,or the combination,Tools for Discrimination At Different CL:CL=67%at a High Risk in a ProjectCL=80%at a Medium Risk in a ProjectCL=
10、90%at a Low Risk in a Project 1-tailed vs.2-tailed 1-tailed facilitates significant difference,facilitates significant difference,Tools for Discrimination Overall liking Rating on Critical attribute,say,strength of flavor Purchase intention arousal Preference(either overall or on specific attribute
11、basis,which will be taken as an integral part of the questionnaire design),Tools for Discrimination Overall Critical attribute Purchase intent Preference,67%CL80/85%CL1-tailed90%CL2-tailed,More intention to launch new product,More intention to reduce cannibalization,Depend on objectives!,0,2-tailed,0,1-tailed,谢谢聆听！,
