Journal of Marketing Research

Information for evaluating the statistical significance of stepwise regression models developed with a forward selection procedure is presented. Cumulative distributions of the adjusted coefficient of determination ( R ̄ 2 ) under the null hypothesis of no relationship between the dependent variable and m potential independent variables are derived from a Monté Carlo simulation study. The study design included sample sizes of 25, 50, and 100, available independent variables of 10, 20, and 40, and three criteria for including variables in the regression model. The results reveal that the biases involved in testing statistical significance by two well-known rules are very large, thus demonstrating the desirability of using the Monté Carlo cumulative R ̄ 2 distributions developed by the authors. Although the results were derived under the assumption of uncorrelated predictors, the authors show that the results continue to be useful for the correlated predictor case.

Cohort analysis is used to analyze the effects of aging and cohort succession on consumption of a product class. Cohort analysis is explained and a consumption study is reported. Strong cohort differences are found and age diffences are opposite in direction from those found by traditional methods of analysis.

Data obtained from 375 members of a consumer panel in a two-phase study of consumer experiences with automobile repairs and services were used to examine the antecedents and consequences of consumer satisfaction. The results support previous findings that expectations and disconfirmation are plausible determinants of satisfaction, and suggest that complaint activity may be included in satisfaction/dissatisfaction research as suggested by earlier descriptions of consumer complaining behavior.

The authors propose a simple Bayesian approach which combines self-explicated data with conjoint data for estimating individual-level conjoint models. Analytical results show that, with typical conjoint data, improvement may be expected over the estimation and prediction results obtained with ordinary least squares (OLS). The expected improvement in prediction is confirmed by pilot empirical results.

A comprehensive literature review of techniques used to increase response rates to questionnaires was conducted. Conclusions were based on arithmetic combination of 497 response rates found in 93 journal articles. Response rates were found to be increased by personal and telephone (versus mail) surveys, either prepaid or promised incentives, nonmonetary premiums and rewards, and increasing amounts of monetary reward. Other facilitators that increased responding were preliminary notification, foot-in-the-door techniques, personalization, and followup letters. Several facilitators were found to be ineffective. An estimated magnitude of effect was calculated for each design variation.

The use of repeated measures or within-subjects designs has become common in marketing research. A review of articles employing such designs suggests that researchers may not fully appreciate the nature and assumptions of repeated measures analyses.

The authors discuss a special case of multiple time series models in which the autoregressive and moving average parameter matrices are diagonal. These models, called MTS-D, are appropriate if the random shocks which drive the series are only contemporaneously correlated. The flexibility and usefulness of this approach are illustrated in the context of a competitive market situation with five sales series, where individual components follow different univariate time series models. Theoretical as well as empirical justifications for expecting more precise parameter estimates are given. Forecasts from the model compare favorably with those from the univariate models.

In four sessions over a two-week time period, 32 women and 32 men viewed 128 randomly selected prime-time and daytime television commercials from early 1980. In an adaptation of Schneider's person perception approach, the viewers rated the models on a set of 13 social psychological attributes which were subsequently factor analyzed into three factors: responsible-mature, social image-status, and independence. Perceptions of male and female models differed as a function of time of airing and product, but not as a function of the viewer's sex. Overall, women were not portrayed in a more negative way than men are portrayed in today's (1980) television commercials.

The author extends the research on consumer prepurchase information seeking to a product which is essentially socially symbolic rather than utilitarian. A survey of the purchasers of a fashion item was undertaken, and the data subjected to a novel cluster analytic technique. Several distinct patterns of overt search were derived and related to socioeconomic segments of individuals. The author concludes that to explain consumer information seeking it is necessary to gain a better understanding of both the motives underlying search and the risks inherent in any purchase.

Structural equation analysis was used to replicate and extend previous salesforce role stress research by testing hypotheses about the relationships among sales supervisory behavior, salesforce role stress, and salesforce job satisfaction. On the basis of previous research, four supervisory behavior variables were hypothesized to be related to salesforce role stress—leader consideration, initiation of structure, participation, and feedback. The results indicate sales supervisory behavior is related to salesforce perceptions of role stress. In addition, the estimates indicate role conflict, organizational communication, leader consideration, and leader initiation of structure are significant predictors of salesforce job satisfaction.

The problem of statistically assessing cluster analysis results is discussed; several measures based on both within-group homogeneity and between-group heterogeneity are defined. Permutation distributions for these measures are defined and expressions are derived for easily calculating the parameters of these distributions. Using these measures and their respective distributions, one can test whether given clusters significantly differ from clusters which are randomly determined.

Cluster analysis is widely used in marketing research, especially in market segmentation studies. Though it is well known that cluster analysis approaches can be unreliable, the author's research suggests that unreliability may be a greater problem than has previously been suggested, especially with approaches involving large numbers of decisions about cluster membership.