Marketing Research的课件- ANOVA_s.pptVIP

ANOVA Session: February 1-5-8; 2010 1. Objectives/Purpose Test whether treatments (categorical variables) lead to different levels for a (set of) metric outcome variables, e.g. Does merchandise juxtaposition affect clothes’ purchase intentions? Does beer bottle shape influence purchase quantity? How does visit frequency and use of samples affect physicians’ prescriptions? How does promo activity affect store sales and traffic? Overview of approaches Q1: Do Store Sales increase with Coupon Activity(no/yes)? T-test (ind. Samples): Store Sales by Coupon Activity Q2: Do Store Sales increase with Promotion Intensity (low, medium, high number of price cuts)? Example: One-Way ANOVA ANOVA or multiple t-tests? 1 test, ?=.05: Probability of decision=‘effect’ while there is none: .05 Probability of decision=‘no effect’ while there is none: .95 3 tests, ?=.05 in each test: Probability of decision=‘no effect’ in each test, while there is none: (.95)3=.9025 Probability of decision=‘effect’ in at least one of the three tests, while there is none: (1-.9025)=.0975.05!!! Probability of erroneously finding effect increases with number of tests Q3a: Do Store Sales increase with Couponing and Promotional Price Cut Activity??2-way ANOVAQ3b: What if we control for region wealth?? 2-way ANCOVA ANOVA or MANOVA? Multiple Univariate Q? Structured Multivariate Q? Intrinsically Multivariate Q? E.g. Impact on store sales and store traffic? 2. Steps in AN(C)OVA Design Assumptions Model Estimation and Fit Interpretation of Results Validation 2.1. Design Sample Size (20/cel) Factorial Design: Treatments (or Blocking Factors) and Interactions Use of Covariates? (continuous, pre-measure, independent of treatment, limited number) Dependent var(s) Example: Two-Way AN(C)OVA Dependent: Store Sales Treatments: Couponing Activity (2), In-store Promo (3) Sample Size Use Region Wealth as covariate? 2.2 Assumptions Independent Observations ? (Repeated-Measures ANOVA?) Equal