A survey on information pollution in common data-bases of Islamic Azad University of Tehran Science & Research Branch in retrieving library information from the viewpoint of postgraduate students

The main purpose of this study is determining the information pollution in data bases shared by the Islamic Azad University, Tehran Science and Research Branch in library Information Retrieval from the viewpoint of postgraduate students and is determining the average information pollution in content of abstracts, bibliographic information, keywords and contamination caused by writing. The methodology in this research is documental and analytical survey. Two societies of research include five database: Emerald, Pro Quest, Science Direct, ISI Web of knowledge, Scopus and graduate students of Library and Information Science, Islamic Azad University, Tehran Science and Research branch. For this study two different questionnaires were used. The findings showed between the mean score of pollution Factors of bibliographic data to separate databases are significant differences. So that Proquest in this category is in the first place. It also was determined. the mean score of pollution factors in the abstract of five databases, is no significant difference database. But ISI Web of knowledge Had little more pollution than others. Comparison of pollution in the keyword information shows the mean scores between pollution factors to separate data bases are no significant different, so that Emerald has more pollution. It also was determined between the mean score pollution factor caused by writing data exists significant difference. So that in this group Science Direct is in the first place. Although these five bases are studied in the humanities, but the results showed each database does not meet all user needs entirely. To reduce pollution using standard thesaurus, fix any spelling mistakes and inappropriate information with competent judgment, reduce the contradictory information, stained bias and biased information, misleading and unfounded is recommended.