Data Sheet 1_Toxicity levels in Spanish political communication on Twitter/X: a comparative analysis of major political parties, 2015–2023.xlsx

IntroductionSocial media, especially Twitter (now X), have long since become integral to the communication strategies of political parties, providing a direct and agile medium for interaction with the public and dissemination of messages. But this digital omnipresence also poses worrying challenges, such as the polarisation of political discourse. Several studies have explored the communication strategies of political parties on social media, as well as the toxic language and hate present in such communication. However, most of the work has looked at specific parties and in electoral contexts.
MethodsThe present study focuses on examining in depth the levels of toxicity in political discourse on Twitter, particularly on the accounts of the 10 major Spanish political parties, from 2015 to 2023. Computational methods and the Google Perspective API have been used to identify the levels of toxicity, severe toxicity, insult, profanity, threat, and identity attack present in the contents published on these accounts during that period (N = 265,122).
ResultsAmong the main findings, a generalized temporary increase in the presence of all toxicity indicators since 2015 is highlighted. Comparatively, a significant difference is perceived between most of the parties and Vox, the party that scores highest in virtually all indicators of toxicity. Furthermore, higher levels of toxicity appear to generate greater engagement in terms of interaction.
DiscussionThese findings confirm trends noted earlier at the international level, such as the leading role of the populist far-right in the propagation of toxic discourses and the generalised rise of toxicity and polarization in political debate.

引言（Introduction）
社交媒体，尤其是推特（Twitter，现为X平台），早已成为政党传播策略中不可或缺的组成部分，为政党与公众互动、传播政治讯息提供了直接且灵活的渠道。但这种数字化全覆盖也带来了令人担忧的挑战，例如政治话语的极化现象。已有多项研究探讨了政党在社交媒体上的传播策略，以及此类传播中存在的有毒语言与仇恨言论。然而，此类研究大多聚焦于特定政党以及选举语境下的情况。
研究方法（Methods）
本研究旨在深入分析推特（Twitter，现为X平台）上的政治话语毒性水平，尤其聚焦2015年至2023年间西班牙10个主要政党的官方账号。本研究采用计算方法与谷歌视角API（Google Perspective API），识别该时段内这些账号发布内容中的毒性、严重毒性、侮辱性内容、亵渎性内容、威胁性内容及身份攻击内容，共计样本量N=265,122。
研究结果（Results）
主要研究结果显示，自2015年起，所有毒性指标的出现频次均呈现普遍的阶段性上升趋势。对比来看，多数政党与沃克斯党（Vox）之间存在显著差异——沃克斯党在几乎所有毒性指标上的得分均为最高。此外，毒性水平越高，相关内容的互动参与度似乎也越高。
讨论（Discussion）
本研究结果印证了此前国际层面已观测到的趋势，例如民粹主义极右翼在有毒话语传播中的主导作用，以及政治辩论中毒性与极化现象的普遍上升。