Abstract
For the analysis of two-way contingency table with large number of rows and columns, correspondence analysis and clustering of rows and/or columns based on chi-square distance are frequently used in practice. In this short article, it is shown that the ratio of chi-square statistics of the original contingency table and a reduced table after clustering can be interpreted as the contribution ratio of the reduced table. In this note, the relationship between the chi-square statistic and correspondence analysis is effectively used.
