Wayne S. DeSarbo

GENFOLD2. A Set of Models and Algorithms for the GENeral UnFOLDing Analysis of Preference/Dominance Data

A general set of multidimensional unfolding models and algorithms is presented to analyze preference or dominance data. This class of models termed GENFOLD2 (GENeral UnFOLDing Analysis-Version 2) allows one to perform internal or external analysis, constrained or unconstrained analysis, conditional or unconditional analysis, metric or nonmetric analysis, while providing the flexibility of specifying and/or testing a variety of different types of unfolding-type preference models mentioned in the literature including Carolls (1972, 1980) simple, weighted, and general unfolding analysis. An alternating weighted least-squares algorithm is utilized and discussed in terms of preventing degenerate solutions in the estimation of the specified parameters. Finally, two applications of this new method are discussed concerning preference data for ten brands of pain relievers and twelve models of residential communication devices.

Three-way metric unfolding via alternating weighted least squares

Three-way unfolding was developed by DeSarbo (1978) and reported in DeSarbo and Carroll (1980, 1981) as a new model to accommodate the analysis of two-mode three-way data (e.g., nonsymmetric proximities for stimulus objects collected over time) and three-mode, three-way data (e.g., subjects rendering preference judgments for various stimuli in different usage occasions or situations). This paper presents a revised objective function and new algorithm which attempt to prevent the common type of degenerate solutions encountered in typical unfolding analysis. We begin with an introduction of the problem and a review of three-way unfolding. The three-way unfolding model, weighted objective function, and new algorithm are presented. Monte Carlo work via a fractional factorial experimental design is described investigating the effect of several data and model factors on overall algorithm performance. Finally, three applications of the methodology are reported illustrating the flexibility and robustness of the procedure.

Optimal variable weighting for hierarchical clustering: An alternating least-squares algorithm

This paper presents the development of a new methodology which simultaneously estimates in a least-squares fashion both an ultrametric tree and respective variable weightings for profile data that have been converted into (weighted) Euclidean distances. We first review the relevant classification literature on this topic. The new methodology is presented including the alternating least-squares algorithm used to estimate the parameters. The method is applied to a synthetic data set with known structure as a test of its operation. An application of this new methodology to ethnic group rating data is also discussed. Finally, extensions of the procedure to model additive, multiple, and three-way trees are mentioned.

Criteria for Corporate Venturing: Importance Assigned by Managers

Little is known about what really makes ventures succeed or fail and therefore, what one should consider when deciding whether or not to back a corporate venture. What is required are many more systematic studies of the venturing process. In this study we look at one small part of the process; the way in which managers go about evaluating a venture and what importance they attach to the various criteria they use to assess corporate ventures as they decide whether or not to support them.The other issue of interest is whether, with venture decision-making experience, there is a shift in the importance of criteria. Do managers inexperienced in venturing, start off with one set of weightings on criteria and then learn by experience to weigh criteria differently? Nearly every venture decision will be reviewed or have to pass some form of limited approval by, or get logistical support from, at least some managers who may be inexperienced. The degree of their support and approval will depend on the inexperienced managers model of what constitutes an appropriate venture. Therefore, it is important to study both the novice and the experienced venture manager.This study used conjoint measurement procedures to quantify the importance of several factors to managers making go/no-go decisions as to whether they would support a series of hypothetical corporate ventures.The results indicate that there is a very high correlation between the judgements of inexperienced managers and those that have had some involvement in venture decision making. In virtually every case the direction of the preference for levels is identical. The effect of experience is not to change the model that the manager uses, but rather to crystallize the preferences and tradeoffs involved.The most interesting result was the overriding importance attached to corporate fit. The message is clear: do not expect support from any managers, inexperienced or otherwise, if there is no perceived fit between the firm and the venture.There were also relatively high levels of importance attached to seven other variables: size of investment, presence of an experienced venture champion, corporate experience with product, low threat of competition, utilization of proprietary technology, rate of return, and gross margin.For the sample in this study an optimal venture proposal can be described via the following criteria:-High corporate fit-Low initial investment-Experienced venture champion-Experience with product/service-Low competitive threat-Proprietory technology-High gross margin-High rate of return

A New Stochastic Multidimensional Unfolding Model for the Investigation of Paired Comparison Consumer Preference/Choice Data

This paper presents the development of a new stochastic multidimensional (scaling) unfolding (Coombs 1964) methodology which operates on paired comparison consumer preference or choice data and renders a spatial representation of both consumers and the products or brands they choose. Consumers are represented as ideal points and products as points in a T-dimensional space, where the Euclidean distance between the product points and the consumer ideal points provide information as to the utility of such products to these consumers. The econometric and psychometric literature concerning related models which also operate on such paired comparisons data is reviewed, and a technical description of the new methodology is provided. To illustrate the versatility of the model, a small application measuring consumer preference for several actual brands of over-the-counter analgesics, utilizing one of the optional reparametrized models, is described. Finally, future areas of further research are identified.

Least squares algorithms for constructing constrained ultrametric and additive tree representations of symmetric proximity data

A mathematical programming algorithm is developed for fitting ultrametric or additive trees to proximity data where external constraints are imposed on the topology of the tree. The two procedures minimize a least squares loss function. The method is illustrated on both synthetic and real data. A constrained ultrametric tree analysis was performed on similarities between 32 subjects based on preferences for ten odors, while a constrained additive tree analysis was carried out on some proximity data between kinship terms. Finally, some extensions of the methodology to other tree fitting procedures are mentioned.

A Customized Market Response Model: Development, Estimation, and Empirical Testing

A customized, stepwise, log-linear, distributed lag, restricted market response model is proposed to estimate the effects of various elements of promotion expenditures on sales in the presence of potentially significant effects due to trend and/or seasonality when using time-series data. As distinct from standardized software packages, the customization offers management several benefits: (a) an (optional) imposition of prior restrictions on the directions of the coefficient variables; (b) an empirical determination of the lag structure for selected variables; (c) the detrending of the data to allow for the assessment of incremental marketing mix effects above trend; and (d) a simplified sensitivity analysis. The model is empirically tested and validated using sales data for a brand where the impact of several marketing mix variables is estimated and investigated via policy simulations. A comparison of these results with those obtained from a corresponding unrestricted model illustrates the advantages of this approach. Finally, the limitations of this procedure and directions for future research are discussed.