News agencies report news from different viewpoints and with different writing styles. In this study, we propose a method to extract the characteristic descriptions, especially about people and organizations, written by news agencies. To extract the characteristic descriptions of a given person or organization by a news agency, we analyze co-occurrence relationships among words which appear in the same sentence based on their SVO roles. Then, we extract a description which is often used by the news agency but not commonly used by the others. We carried out an experiment by using the articles from there major Japanese newspaper agencies. The experimental results shows that our method can elucidate the different features of each agency.