Computer Analysis of Facts and Comments Separating Standard in Ukrainian Online Media

Abstract

Today, there is a need to measure the quality of media content, in particular, compliance with professional standards. In Ukraine this task is performed manually. In this study, we offer an auxiliary algorithm of computer analysis for the study of adherence to the journalistic standard of separating facts from comments in Ukrainian online mass media. It is assumed that based on the analysis of the verbs in the references after the citations it is possible to assess compliance or violation of this standard. A number of programs written in Python allow to download and process media publications from archives. On the material of 7,111 journalistic texts from all-Ukrainian and regional media sites analysis several examples of compliance and violation of separating facts from comments professional standard were found. It was established that all-Ukrainian mass media can more often violate the standard of separating facts from subjective opinions, statements and comments of experts than regional ones (49.3% vs. 37.2%), although this observation can be explained by a significant difference in the number of materials published by the media daily. At the same time, the following evaluative and emotionally expressive verbs were found in the majority of mass media: "declare", "mark", emphasize" (it was necessary to replace the neutral "say"), "underline", "remind", "assure", "notice" (used by journalists to evaluate what someone said). Common "neutral" vocabulary was also recorded: "inform", "say", "add", "tell", "write". This method of media research can be implemented not only to check verbs with emotional, solemn, expressive, stylistic or evaluative content, but also to analyze other parts of languages (nouns, adjectives, adverbs) or popular phrases or phrases that can also be used by journalists to provide the subjective meaning of what someone said about a certain fact or event. Therefore, the introduction of elements of computer analysis, such as the one proposed in our paper, into media monitoring can help experts see those texts with some markers of a possible standards violation.