Advanced Methods of Multivariate Analysis: Profiling Female Political Candidates in Greece

Τhe proposed paper presents the use of advanced of multivariate methodology combining Hierarchical Cluster Analysis (HCA) and Multiple Correspondence Analysis (MCA) in two steps (Chadjipadelis, 2015). The methodology is used to analyze the political profile of women who stand for election in Greece and important characteristics who play significant role in this. In the first step, HCA is applied to assign subjects into distinct groups according to the response patterns of the sample. HCA provides us with a group or cluster membership variable, reflecting the partitioning of the subjects into groups. For each group, the contribution of each question (variable) to the group formation is also investigated, in order to reveal a typology of behavioral patterns. Τhe number of clusters is determined upon the empirical criterion of the change in the ratio of between-cluster inertia to total inertia, when moving from a partition with r clusters to a partition with r-1 clusters. In the second step, the group membership variable from the first step, is jointly analyzed with the existing variables via Multiple Correspondence Analysis on a symmetric, generalized contingency table, which cross-tabulates all variables against each other, the Burt table (Greenacre, 2007). The MCA output is a set of orthogonal axes or dimensions, that summarize the associations between variable categories into a space of lower dimensionality, with the least possible loss of the original information of the Burt table. HCA is then applied again, this time on the coordinates of variable categories on the factorial axes, clustering at this point the variables, instead of the subjects. The groups of variable categories can reveal abstract discourses. The composition of the two prior analyses produces and reveals behavioral patterns and abstract discourses which construct a joint semantic map for the variables and the subjects. In this context, the proposed methodology offers an in-depth analysis and comprehensive understanding on women’s political behavior and the important factors which determine it. Geometric data analysis allows us to construct this semantic map of political behavior, in this case study examining female candidates political profile, revealing all the latent relationships patterns with various factors such as social, economic and demographic characteristics. REFERENCES Chadjipadelis, T. (2015). Parties, Candidates, Issues: The Effect of Crisis. Correspondence Analysis and Related Methods. CARME 2015: Napoli, Italy Benzécri, J. P. et al (1980). Introduction la classication automatique d’ après un exemple de donees medicales. Les cahiers de l’ analyse des données. 5 (3):311-340. Benzécri, J. P. et al (1973/1976). L’ analyse des données. Tome 1: La taxinomie. Tome 2 : Analyse des Correspondances,. End ed. Dunod: Paris. Greenacre, M. (2007). Correspondence Analysis in Practice, Chapman and Hall/CRC Press, Boca Raton. Karapistolis, D. (2010). Software Method of Data Analysis MAD.