Marketing Science

Gensch, Dennis H.

doi: 10.1287/mksc.6.3.223pmid: N/A

A two-stage disaggregate attribute choice model is proposed and empirically implemented. The first stage of the model is attribute processing to screen the number of alternatives down to a lesser number. The second stage is brand (alternative) processing which considers the attributes simultaneously while allowing for tradeoffs among the attributes.This two-stage approach is then applied to the same real world data set as two single stage disaggregate models, logit and Maximum-Likelihood-Hierarchical (MLH) which are state of the art models representing the alternative and attribute processing approaches, respectively. The predictive accuracy of the two-stage approach compares favorably to the single stage models. In addition, it seems to offer diagnostic information that can provide managerial insights not found in the output of the single stage model.

Tellis, Gerard J.; Wernerfelt, Birger

doi: 10.1287/mksc.6.3.240pmid: N/A

We present an analysis of equilibrium in markets with asymmetrically informed consumers. Some consumers know both price and quality of all sellers, whereas others know neither but may search among sellers. The equilibrium correlation between price and quality generally increases with the level of information in the market and can be negative when this level is sufficiently small. A meta-analysis of the available empirical studies strongly supports the model's predictions.

Vanhonacker, Wilfried R.; Day, Diana

doi: 10.1287/mksc.6.3.254pmid: N/A

Empirical results in marketing research are often derived from linear additive models estimated on cross-sectional data. An underlying assumption of these model specifications is that each exogenous variable contributes an additive effect to the endogenous variable. Measuring these additive effects is problematic if some components of the endogenous variable are not observable to the researcher. Reliance on observable correlates to capture those aspects of the endogenous variable might require reverse regression to ensure unbiasedness of the estimated effects. In contrast to direct regression used exclusively in marketing research, reverse regression first estimates the effect of the endogenous variable on those correlates and then derives unbiased estimates of the additive effect of other exogenous variables specified in the basic model. This paper discusses the reverse regression approach relying on recently published analytic results. A statistically powerful and analytically simple test is developed which allows the marketing researcher to assess a priori whether direct or reverse regression will provide unbiased parameter estimates. An empirical illustration focuses on the measurement of potential market share rewards for pioneering businesses contained in the PIMS data base. Other areas in marketing where the issue of direct versus reverse regression is pertinent are mentioned.

Norton, Seth W.

doi: 10.1287/mksc.6.3.268pmid: N/A

Analytic and institutional approaches to channel coordination have developed independent of each other. This paper attempts to integrate both approaches by using the Coase Theorem, which has had wide application in the economics of natural resources and the law. Suboptimization means that total channel profits are lower than maximum attainable channel profits. Applying the Coase Theorem suggests that total channel profits can be maximized if property rights in the channel are well defined. Channel coordination is viewed in terms of defining rights, and several hypotheses regarding channel behavior are developed.

Buchanan, Bruce; Morrison, Donald G.

doi: 10.1287/mksc.6.3.286pmid: N/A

Frequency questions (How many times did you do X in the last month?) are often used to measure respondents' behavior rates. Under suitable assumptions, frequency data can be analyzed in the context of the Negative Binomial Distribution model. Recency questions (How long ago did you last do X?) can also be used to assess usage rates, but have not been studied as much. In this study we compare the reliability and statistical efficiency of frequency and recency questions. We show that the relative performance of these two formats can be summarized succinctly. Validity issues are also considered. Our findings suggest that recency questions tend to perform better for low usage, low salience events.