CHAPTER 18What the Experts SayLearning ObjectivesGet This! Fore! Golfers Benefit from Conjoint AnalysisGet This! Fore! Golfers Benefit from Conjoint Analysis – cont’dNow Ask YourselfMultivariate Statistical AnalysisMultiple Regression AnalysisSlide 9Slide 10Multiple Discriminant Analysis (MDA)Multiple Discriminant Analysis (MDA) – cont’dMultiple Discriminant Analysis (MDA)  cont’dFactor AnalysisFactor Analysis – cont’dCluster AnalysisCluster Analysis – cont’dMultidimensional ScalingMultidimensional Scaling – cont’dConjoint AnalysisConjoint Analysis – cont’dChoosing the Appropriate TestDecision Time!Net ImpactChapter 18 End of PresentationMarketing Research, 2nd EditionAlan T. ShaoCopyright © 2002 by South-WesternPPT- 1CHAPTER 18MULTIVARIATE ANALYSISMarketing Research, 2nd EditionAlan T. ShaoCopyright © 2002 by South-WesternPPT- 2What the Experts Say…I’ve never considered myself a ‘quant jock,’ for reasons probably due to genetic ancestry. I find it difficult to get excited about the inner workings of optimization algorithms or exploring the sensitivity of [multivariate] ANOVA to violations in the independence assumption, for example. Rather, my interest in the various multivariate tools arises from their usefulness as a means for examining phenomena that do interest me.--Rob Kleine, A Seminar in Multivariate Statistics, http://www.gentleye.com/research/ multivar/index.html, September 23, 2000.Marketing Research, 2nd EditionAlan T. ShaoCopyright © 2002 by South-WesternPPT- 3Learning ObjectivesDiscuss the basics of multivariate statistical analysesExplain which technique is appropriate given the type of variables involvedDescribe the usefulness of multivariate statisticsMarketing Research, 2nd EditionAlan T. ShaoCopyright © 2002 by South-WesternPPT- 4Get This!Fore! Golfers Benefit from Conjoint AnalysisEvery golfer has two things in common. They’re all looking to drive the ball farther and to hit it with more accuracy.Sawtooth Technologies, a well-known company providing software for research data collection and analysis, uses conjoint analysis to examine the extent to which average driving distance, average ball life, and price are concerns of golfers when selecting their golf balls.Marketing Research, 2nd EditionAlan T. ShaoCopyright © 2002 by South-WesternPPT- 5The trade-offs might deal with paying a little extra for a ball that travels farther. A golfer might value a long drive more than a highly durable ball. And a choice prediction might be that a golfer prefers the long-life ball over the distance ball since it has the larger total value. All of these findings would be based on computations from conjoint analysis.Get This!Fore! Golfers Benefit from Conjoint Analysis – cont’dConjoint analysis encompasses three critical steps:–Collecting trade-offs–Estimating buyer value systems–Making choice predictionsMarketing Research, 2nd EditionAlan T. ShaoCopyright © 2002 by South-WesternPPT- 6Now Ask YourselfDoes conjoint analysis make intuitive sense to you? If so, why is it needed?What other multivariate techniques are available to researchers?Do I need to be a statistical expert to understand multivariate statistical analysis?Marketing Research, 2nd EditionAlan T. ShaoCopyright © 2002 by South-WesternPPT- 7Multivariate Statistical AnalysisMultivariate Statistics: Investigates more than two variables at a time.Many times, multivariate techniques are a means of performing in one analysis what used to take multiple analyses using univariate techniques.The techniques can be used to summarize data and reduce the number of variables necessary to describe the data.Several of the more common multivariate techniques: –Multiple Regression Analysis–Multiple Discriminate Analysis–Factor Analysis–Cluster Analysis–Multidimensional Scaling–Conjoint AnalysisMarketing Research, 2nd EditionAlan T. ShaoCopyright © 2002 by South-WesternPPT- 8Multiple Regression AnalysisThe premise behind multiple regression analysis is consistent with that of simple regression analysis: to determine the association or relationship between dependent and independent variables. In multiple regression analysis, more than two variables are included in examinations. The dependent and independent variables must be interval-scaled to use this technique. The general form of the multiple regression model is as follows:nnXXXY ...22110where = Y intercept of the regression model = slope of the regression model0n...,21Marketing Research, 2nd EditionAlan T. ShaoCopyright © 2002 by South-WesternPPT- 9or the computed multiple regression model isnncXbXbXbaY  ...2211 where computed value of the dependent variablecY a = y intercept when x equals zeroand partial regression coefficients1bnbb ...2nXXX ,...,,21independent variablesPartial Regression Coefficient: Denotes the change in the computed value, , per one unit change in when all other independent variables are held constant. cY1XMultiple Regression Analysis – cont’dMarketing Research, 2nd EditionAlan T. ShaoCopyright © 2002 by South-WesternPPT- 10The association between the dependent and independent variables is referred to as the coefficient of multiple determination, denoted by . It is interpreted in a similar manner as we did when we referred to bivariate data. The coefficient of multiple determination is computed as follows:2R2Rwhere TSS = total sum of squares = RSS = regression sum of squares = ESS = error sum of squares = 2)( YY2)( YYc2)(cYYTSSESSTSSRSSR  12Multiple Regression Analysis – cont’dMarketing Research, 2nd EditionAlan T. ShaoCopyright © 2002 by South-WesternPPT- 11Multiple Discriminant Analysis (MDA)Multiple Discriminant Analysis (MDA): The appropriate tool for predicting the membership of observations in two or more groups. Similar to multiple regression analysis except different types of variables are involved. Appropriate if the dependent variable is nominal, categorical, or multichotomous and the independent variables are interval data. When two classifications are being examined, it is referred to as a two-group discriminant analysis. When three or more classifications are identified, then multiple discriminant analysis is used.Marketing Research,