The Segmentation of a Sample Population by Methods "Decision Tree" and "Conjoint Analysis: the Comparison of the Results"

<p>This project examines the main characteristics of such methods of segmentation as &quot;Conjoint Analysis&quot; and &quot;Decision Trees.&quot; Although a wide range of theoretical and empirical works related to certain methods of segmentation can be found, there is a lack of studies focusing on comparison of various methods. Moreover, the methods considered are very popular in the field of marketing, however, they are quite rarely used for segmentation in the field of sociology. In social studies researchers usually apply cluster analysis to divide people into groups, but it has at least one serious disadvantage, which consists in a lack of a unified strategy of clustering. Therefore it is necessary to enrich our knowledge in this area and look for other methods capable of producing more accurate results. &quot;Conjoint Analysis&quot; and &quot;Decision Trees&quot; are widely used for segmentation in marketing, so the research has suggested, they may be useful in sociology. Thereby, two main objectives of this study were to find out whether methods applied can obtain interpretable segments in sociological research and to compare the methods by comparing the results obtained. The theoretical analysis of the literature on the subject has detected that both methods have their advantages and limitations, that&rsquo;s why researchers should choose one of them with regard to the objectives of their studies. Empirical analysis has discovered that both methods can be successfully used for segmentation in sociological studies, because they provide interpretable and understandable results that reflect the real situation. Also it was found that, despite the differences in the procedure of applying methods the final results are quite similar. Thus it can be concluded that this project provides a great impulse for further, more detailed studies of using the methods considered in the field of sociology.</p>