In our data matrix we have one variable of particular interest and we wish to see how it is affected by movements in a single explanatory variable. For example, the variable we might be interested in is Sales in £mn and we may wish to know how it is affected by advertising expenditure in £000's. The rows of our data matrix may consist of the last seven years of data for these two variables.
Discriminant Analysis Models are very similar to regression models, but differ in one important respect. In Discriminant Analysis the key variable of interest Y is now categorised (eg, Buyer = 1, Non-buyer = 0) rather than being continuous like sales.
Using behavioral data, two alternative segmentation methodologies, regression and cluster analysis, were compared on the basis of rigor, reliability, parsimony, interpretability and strategic implications for retailers. Regression analysis is the preferred methodology on all comparative performance measures. In addition, both psychographic and demographic descriptor variables make substantial contributions to the analysis of consumer shopping activities across retailers. Finally, the analysis reported here demonstrates high levels of cross-shopping, both within and across competitive retail sectors. The overall results lead to important strategic and tactical implications for apparel retailers.
This paper describes the results of a study designed to measure the effects of different promotional activities in stores. Regression analysis, using dummy variables, was used to analyse data obtained from an experimental design incorporating twelve stores over an eight week period. The experimental design consisted of embedding a "cross-over" design, with various combinations of promotional activities being randomly assigned to the remaining cells.
The purpose of this paper is twofold: First, to report the results of a research study which in turn may be compared to researchers' experience elsewhere; second, and perhaps more importantly, to present a thorough discussion of an application of multiple regression techniques which could find profitable use in the tourism and travel industry when traditional marketing research questions concerning demand and market share need to be answered.
In 1970, we, at IFOP/ETMAR, worked on perfecting the computation of indexes from opinion surveys. By experience, we knew how important psychological phenomena reveal to explain the evolution of economic realities such as the successful launching of a new product or the general economic situation. The indexes are built up with a method using factor and multiple regress analyses, which made it possible to select a limited number of questions as the synthetical index computation basis. The fluctuations of these indexes as time passes are a valuable source of information to better understand present time as well as the future. They make it possible to anticipate on the actual facts and take appropriate action at the right time. In the following paper, IFOP/ETMAR describes its main indexes and explains their purpose.
In 1970, we, at IFOP/ETMAR, worked on perfecting the computation of indexes from opinion surveys. By experience, we knew how important psychological phenomena reveal to explain the evolution of economic realities such as the successful launching of a new product or the general economic situation. The indexes are built up with a method using factor and multiple regress analyses, which made it possible to select a limited number of questions as the synthetical index computation basis. The fluctuations of these indexes as time passes are a valuable source of information to better understand present time as well as the future. They make it possible to anticipate on the actual facts and take appropriate action at the right time. In the following paper, IFOP/ETMAR describes its main indexes and explains their purpose.
Covariance analysis is a combination of regression analysis and the analysis of variance. From the latter's standpoint it can be used with any of the above experimental designs to increase the precision of the experiment by removing one or more uncontrolled variables from the error term. From a regression standpoint it can be used to compare several regression lines to test whether they differ significantly from each other and to identify the sources of variance. This paper will illustrate the use of covariance analysis, primarily from a regression standpoint using an example taken from industrial marketing. At a later stage it will be suggested that covariance analysis can be used in the analysis of variance context to examine the competitiveness of price-setting policies by companies,
The purpose of this article is to present a comparison of three alternative methods for sales forecasting - trend analysis, time series analysis and regression analysis. Rather than looking at the theoretical differences, the focus is on practical differences that can be easily understood by the marketing executive. Thus this comparison includes a discussion of the assumptions about the sales forecasting situation inherent in each method, a look at the use of each technique in a number of real situations, and perhaps most importantly, the reliability and accuracy of the forecasts produced using each technique.
Our fundamental aim was to develop a new and more refined technique of image research which would be so simple to apply that the test persons to be questioned would not have to undergo too much strain, yet the results obtained from the interviews should lead directly to new avenues of approach and hence towards more effective marketing and product policies. Our test series finally and after long and thorough experimenting produced a new brand image test method, the so-called MRK Test. This new test method relies on the mathematical concept of multiple regression and correlation. The abbreviation name of the test "MRK" consists of the first letters of "multiple regressions and correlations". What the multiple regression approach is basically trying to achieve is this: we are seeking to find out to which degree we are able to predict the actual buying behaviour from variables that are within our reach, and then seek to determine which weight these variables have in the ultimate buying decision of consumers.