%0 Conference Paper %1 10.1145/3345645.3351107 %A Sales da Costa Melo, Allan %A Balby Marinho, Leandro %A Alonso Veloso, Adriano %B Proceedings of the 5th International Workshop on Social Media World Sensors %C New York, NY, USA %D 2019 %I Association for Computing Machinery %K kilimod %P 5–6 %R 10.1145/3345645.3351107 %T Media Bias Characterization in Brazilian Presidential Elections %U https://doi.org/10.1145/3345645.3351107 %X News media bias is commonly associated with framing information so as to influence readers judgments. It is not rare to find different news outlets reporting the same events under different perspectives with the intention to deliberately influence the reader. For example, making one side's ideological perspective look better than another. This may be an indication of a well known cognitive bias, the framing effect, which states that people may change their judgment based on how the information is presented (or framed). According to a 2017's survey from the Knight Foundation and Gallup, Americans believe that 62% of the news they consume is biased 1. Still according to the survey, there is a sharp divergence of bias perception across Republicans and Democrats regarding news organizations. This implies that the perception of bias may be affected by whether one agrees (or not) with the ideological leaning (when present) of the news source. How to expose such biases in an automatic fashion from textual content only? One way to do that is by comparing different news outlets on the same stories and look for divergences. In this talk, we present an investigation on news media bias in the context of Brazilian presidential elections by comparing four popular news outlets during three consecutive election years (2010, 2014, and 2018). We analyse the textual content of news stories in search for three kinds of bias: coverage, association, and subjective language. Coverage bias is related to differences in mention rates of candidates and parties. Association bias 2 occurs when, for example, one candidate is associated with a negative concept while another not. Subjective bias 3, has to do with wording that attempts to influence the readers by appealing to emotion, stereotypes, or persuasive language. We perform a thorough analysis on a large scale news data set where several such biases are exposed. %@ 9781450369039

@inproceedings{10.1145/3345645.3351107, abstract = {News media bias is commonly associated with framing information so as to influence readers judgments. It is not rare to find different news outlets reporting the same events under different perspectives with the intention to deliberately influence the reader. For example, making one side's ideological perspective look better than another. This may be an indication of a well known cognitive bias, the framing effect, which states that people may change their judgment based on how the information is presented (or framed). According to a 2017's survey from the Knight Foundation and Gallup, Americans believe that 62% of the news they consume is biased [1]. Still according to the survey, there is a sharp divergence of bias perception across Republicans and Democrats regarding news organizations. This implies that the perception of bias may be affected by whether one agrees (or not) with the ideological leaning (when present) of the news source. How to expose such biases in an automatic fashion from textual content only? One way to do that is by comparing different news outlets on the same stories and look for divergences. In this talk, we present an investigation on news media bias in the context of Brazilian presidential elections by comparing four popular news outlets during three consecutive election years (2010, 2014, and 2018). We analyse the textual content of news stories in search for three kinds of bias: coverage, association, and subjective language. Coverage bias is related to differences in mention rates of candidates and parties. Association bias [2] occurs when, for example, one candidate is associated with a negative concept while another not. Subjective bias [3], has to do with wording that attempts to influence the readers by appealing to emotion, stereotypes, or persuasive language. We perform a thorough analysis on a large scale news data set where several such biases are exposed.}, added-at = {2021-04-15T16:07:22.000+0200}, address = {New York, NY, USA}, author = {Sales da Costa Melo, Allan and Balby Marinho, Leandro and Alonso Veloso, Adriano}, biburl = {https://www.bibsonomy.org/bibtex/2c0b98fc61303ba213dc66dfb76c3ed77/albinzehe}, booktitle = {Proceedings of the 5th International Workshop on Social Media World Sensors}, doi = {10.1145/3345645.3351107}, interhash = {a8ccb7e93f8d61c683191a7ffac664b4}, intrahash = {c0b98fc61303ba213dc66dfb76c3ed77}, isbn = {9781450369039}, keywords = {kilimod}, location = {Hof, Germany}, numpages = {2}, pages = {5–6}, publisher = {Association for Computing Machinery}, series = {SIdEWayS'19}, timestamp = {2021-04-15T16:07:22.000+0200}, title = {Media Bias Characterization in Brazilian Presidential Elections}, url = {https://doi.org/10.1145/3345645.3351107}, year = 2019 }