ABSTRACT Based on the theories of media frames and their application to discourse analysis, the objective of this study is to identify the discursive strategies of American newspapers in the coverage of the 2016 primary elections. A large corpus of more than 3,000 articles published in six national papers between January and March 2016 is analysed using the textual statistical tools Iramuteq and TXM. Several semi-automated operations, based on lexical co-occurrence (notably hierarchical descendant classification and factorial analysis) enable a re-construction of semantic cues, revealing ideological postures, argumentative strategies and thematic disparities between the different newspapers. In the paper, we address the subjective interpretation schemes inherent to the way each media outlet presents the different candidates depending on their communication contract, with a final focus on the candidate Donald Trump.

中文翻译：

---

2016 年美国总统初选：媒体报道的定量和定性方法

摘要 基于媒体框架理论及其在话语分析中的应用，本研究的目的是确定美国报纸在 2016 年初选报道中的话语策略。使用文本统计工具 Iramuteq 和 TXM 分析了在 2016 年 1 月至 2016 年 3 月期间发表在六篇国家论文中的 3,000 多篇文章的大型语料库。几种基于词汇共现的半自动操作（特别是层次后代分类和因子分析）能够重建语义线索，揭示不同报纸之间的意识形态姿态、辩论策略和主题差异。在论文中，