Web Critique

Correspondence analysis is a statistical technique that was designed as an exploratory technique to analyse two-way and multi-way tables that have a certain measure of correspondence between columns and rows. Correspondence analysis is often compared to factor analysis. Its primary purpose as contrasted to traditional hypothesis testing that is used to verify an a-priori relation between variables is to determine relations within data when there is no a-priori hypothesis at hand. It is also a descriptive technique. Correspondence analysis is used to simplify complex data avoiding extreme simplification and provides an exhaustive analysis of data at hand. Correspondence analysis allows multivariate comparison of data. In order to conduct such analysis, data has to be presented as a rectangular matrix with non-negative entries. Tables typically include categorical variables. Correspondence analysis is most effective when the matrix of data at hand is large enough and simple statistical or graphical analysis cannot provide insight into data structure or properties. Since the purpose of a typical correspondence analysis is to represent entries in a data table of frequencies in terms of distances between rows or columns, data that is typically used for correspondence analysis contains homogenous variables. Finally, correspondence analysis is most effective when the structure of the data matrix is either unknown or cannot be interpreted. The two web-sites that were used in this paper were Unesco.org and Statsoft.com. The latter provides a detailed account of the application of CA, while the former provides a basic overview of the distinguishing features of the technique. ReferencesStatSoft (n.d.) How To Analyze Simple Two-Way and Multi-Way Table, Correspondence Analysis. Available at: http://www.statsoft.com/textbook/correspondence-analysis/ Unesco (n.d.) Correspondence Analysis. Available at: …