| Exploratory Analysis of Marketing Data: Trees vs. Regression (2008) | |||||||||||||
Abstract | |||||||||||||
| This article compares the predictive ability of models developed by two different statistical methods, tree analysis and regression analysis. Each was used in an exploratory study to develop a model to make predictions for a specific marketing situation. The Statistical Methods The regression model is well known and no description is provided here. Tree analysis, however, is less well known. To add to the confusion, it has been labeled in a number o £ ways – e.g., multiple classification, multilevel cross-tabulations, or configurational analysis. Whatever the names, the basic idea is to classify objects in cells so that the objects in the cells are similar to one another yet different from the objects in other cells. Similarity is judged by the score on a given dependent or criterion variable (which differentiates this method from cluster or factor analysis, where the similarity is based only upon scores on a set of descriptive variables). Tree analysis is an extension to n variables of the simple cross-classification approach. Consider the following example: a researcher is studying the factors which determine whether a family owns two or more automobiles. He finds that income may be used to classify respondents. Illustrative results for his sample are provided in Figure 1. He then decides that the number of drivers in the family may also be important for high-income families. | |||||||||||||
Publication details | |||||||||||||
| |||||||||||||