ON THE MULTIPLE SEGMENTATION PROBLEM: AN EMPIRICAL INVESTIGATION OF THE EFFICACY OF THE MULTIPLE LOGIT APPROACH IN PREDICTING SEGMENT MEMBERSHIP.
Item
-
Title
-
ON THE MULTIPLE SEGMENTATION PROBLEM: AN EMPIRICAL INVESTIGATION OF THE EFFICACY OF THE MULTIPLE LOGIT APPROACH IN PREDICTING SEGMENT MEMBERSHIP.
-
Identifier
-
AAI8014991
-
identifier
-
8014991
-
Creator
-
TEMARES, MELVIN LEWIS.
-
Contributor
-
Matthew Goldstein
-
Date
-
1980
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Business Administration, Marketing
-
Abstract
-
The field of market segmentation as an area for investigation has been under scrutiny for many decades. The relevance of this area was denoted by a special issue of the Journal of Market Research in August of 1978 devoted to ten articles relating to market segmentation. In addition, the use of varied statistical techniques as predictive indicators with regard to buyer behavior has centered on the area of segmenting markets. Professor Paul Green has been the leader in this area and with his able co-authors has introduced the use of the logit loglinear model as a tool for analyzing consumer behavior in the market place and has dealt with the logit approach for a two group problem. Also, when dealing with the problem of market segmentation, only Roger Calantone and Alan Sawyer have dealt with segments remaining constant over a period of time. This dissertation combines both of these areas of interest. Market Segmentation is investigated over two periods of time using panel data and analyzed utilizing the logit approach in order to develop a predictive model based upon demographic characteristics. The data are from two separate years for the same consumer panel. Validation of our approach is accomplished through the usual split-sample analysis. Thus, the two group approach of Professor Green is expanded to the multiple group problem and in combination with the panel data over time represents a situation not represented in the current literature.;Chapter One presents the reason for this study, the description of the data set and some background with regard to the market segmentation problem.;Chapter Two develops the background, rationale and mathematics for the use of the loglinear-logit approach to the market segmentation problem. It is a logical choice because discriminate analysis has become one of the most frequently used techniques for classification of data and the loglinear-logit analysis holds the greatest promise for effectively utilizing model building techniques.;Chapter Three involves the data presentation and computer analysis. The panel data is tested and a model utilizing the appropriate interactions of the variables is constructed.;Chapter Four includes the analysis of the data as presented in chapter three. The loglinear-logit analysis technique is utilized in order to predict the classification of the data through the construction of the appropriate discriminant function. A conclusion is reached on the basis of the data with regard to the efficacy of the multiple logit approach in predicting segment membership over time.;The final chapter, Chapter Five, includes a summary of the findings and presents avenues for possible future research. It is hoped that by using this dissertation as a foundation, further growth in the use of mathematical analysis for marketing decisions will be forthcoming. The sparseness of the data leads to the future possibility of an expansion of the current data base or applying the techniques to a different data base. This paper is not designed to test the data base used but uses this data base only to show how these statistical techniques are applied.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
-
Program
-
Business
