Identifying the Determinants of Advertising Intensity in Iran's Industries Using the Multiple Discriminant Analysis Technique

This research is aimed at identifying the main variables in an industry based on which the advertising intensity in that industry can be estimated. Using prior knowledge, six variables, including investment, statism ratio, concentration, sales, added value and direct export were utilized by Multiple Discriminant Analysis, using the stepwise method. It turned out that among these variables, sales, direct export and concentration are variables which can better enhance the discrimination power of the model. Based on the means and standardized canonical coefficients of these three variables in different groups and also the group centroids in each function, we concluded that the first function is better at discriminating the first group from the other two and the second function is better at discriminating the third group from the other two. The correct classification percentage of 82% indicates the highly satisfactory performance of the model in classifying the data.