This study presents a framework of discourse truth for analyzing how truth is constructed in news discourse. Using the UAM CorpusTool, this framework is applied to a corpus of 160 news articles, totaling 142,894 words, collected from the websites of Cable News Network (CNN) and Fox News. The analysis highlights three key dimensions of discourse truth: authenticity, neutrality, and authority. Authenticity is primarily constructed through evidentiality (including direct evidence, hearsay, and inference) and quotations (direct or indirect). Neutrality is achieved through third-party attribution (specified or unspecified sources) and low modality (lexical or grammatical). Authority is established through identity credentials (e.g., names, titles, and institutional affiliations) and personalization strategies (i.e., personalizing the experience of subjects, journalists, or audiences). Both CNN and Fox News emphasize authenticity, neutrality, and authority, focusing primarily on the first two aspects, although they differ in their use of specific discourse devices. These findings provide novel insights into how the truth of news is constructed and understood.