Ideology-Detection

Mass media is an important apparatus for shaping public opinion. Any bias prevalent in the policy discourse in mass media can thus have a significant impact on how people understand policies. In this study, we aim to understand the ideological bias existing in Indian mass media, in terms of the coverage it provides to the statements given by influential people on key economic and technological policies. We adopt a hierarchical methodology using Recursive Neural Networks to model the semantics of these statements. Our results show that the Indian media is ideologically biased, typically covering pro-policy statements much more than anti-policy statements and favoring a technology deterministic viewpoint more than the other side of the discourse. Alongside estimating the ideological differences existing among these influential people or entities on these policies, we are also able to make a generic policy-level classifier that can push the media towards self-regulation in terms of achieving diversity in the policy discourse.