Sponsoring Section/Society: ASA Section for Statistics in Marketing
Session Slot: 2:00- 3:50 Monday
Estimated Audience Size: 45
AudioVisual Request: none
Session Title: New Methods for Marketing Research
Theme Session: No
Applied Session: No
Session Organizer: Elrod, Terry Univ. of Alberta
Address: 3-23 Faculty of Business; Univ. of Alberta; Edmonton, Alberta; Canada T6G 2R6
Phone: 403-492-5884
Fax: 403-492-3325
Email: Terry.Elrod@UAlberta.ca
Session Timing: 110 minutes total (Sorry about format):
Opening Remarks by Chair - 0 minutes First Speaker - 25 minutes Second Speaker - 25 minutes Third Speaker - 25 minutes Fourth Speaker - 25 minutes Floor Discusion - 10 minutes
Session Chair: Elrod, Terry Univ. of Alberta
Address: 3-23 Faculty of Business; Univ. of Alberta; Edmonton, Alberta; Canada T6G 2R6
Phone: 403-492-5884
Fax: 403-492-3325
Email: Terry.Elrod@UAlberta.ca
1. Manifest versus Latent Representations in Multiple Category Choice
Russell, Gary J., Univ. of Iowa
Address: Univ. of Iowa; College of Bus. Admin.; 108 Papajohn Bus. Admin. Bldg.; Iowa City, IA 52242-1000
Phone: 319-335-0993
Fax: 319-335-3690
Email: gary-j-russell@uiowa.edu
Abstract: Multiple category choice is a "pick-any" selection task in which consumers purchase a subset of n alternatives. Correlation among individual choices can be induced using autoregressive models defined with observed indicator variables or with latent utilities. We contrast the two approaches using both statistical and marketing theoretic criteria.
2. Modelling Fuzzy Data in Qualitative Marketing Research
Rust, Roland T., Vanderbilt University
Address: Owen Graduate School of Management; Vanderbilt University; 401 21st Avenue South; Nashville, TN 37203
Phone: 615-343-6732
Fax: 615-343-7177
Email: rustrt@ctrvax.vanderbilt.edu
Varki, Sajeev, University of Auckland
Cooil, Bruce, Vanderbilt Univ.
Abstract: In qualitative marketing research the data items to be classified are often fuzzy (belonging to more than one category), which complicates the judgment task. Latent class models, besides being misspecified in the presence of fuzzy data, often fail to converge. We adapt the grade of membership model of Manton, Woodbury, and Tolley to the context of judgment-based classification. This unconditional model is a generalization of the latent class model. We also propose a conditional model that is asymptotically equivalent to the unconditional model, but is much faster.
3. Weak Models Versus Weak Data: Diagnostic Aids for Ambiguous Results
Stewart, David W., University of Southern California
Address: Department of Marketing; Marshall School of Business; University of Southern California; Los Angeles, CA 90089-1421
Phone: 213-740-5037
Fax: 213-740-7828
Email: dstewart@bus.usc.edu
Abstract: Numerous occasions arise on which the results of a model fitting exercise produce ambiguous results. Such ambiguity may arise either because the data are insufficient to provide a strong test of the model or because the model under test is inherently weak. Although the source of ambiguity has important implications, it is not always obvious whether the problem resides in the data or the model. The present paper provides a discussion of this problem and offers two diagnostic procedures for differentiating cases where data are weak from cases where the model itself is weak.
4. Obtaining Product-Market Maps From Pairwise Brand Preferences
Elrod, Terry, Univ. of Alberta
Address: 3-23 Faculty of Business; Univ. of Alberta; Edmonton, Alberta; Canada T6G 2R6
Phone: 403-492-5884
Fax: 403-492-3325
Email: Terry.Elrod@UAlberta.ca
Abstract: This talk introduces product-market maps and explains the advantages of estimating them from consumer preferences for existing brands. A product-market map is fit to pairwise preferences obtained from users of air traffic management systems. An additional analysis of brand perceptions assists in interpretation and verification of the map.
List of speakers who are nonmembers: None