Grouped counts of news articles mentioning European politicians in Russian-language media (2025)

This dataset contains cluster-level data used to estimate intra-class correlation coefficients (ICCs) for the salience of twelve European politicians in Russian-language media during 2025. The data underpin the analysis of within-story homogeneity in entity salience reported in the accompanying paper.For each politician, the dataset includes a selected sample of dates with sufficiently rich contextual structure -operationalised as days with a large number of detected contextual clusters - to permit estimation of between-cluster variance. Each row corresponds to a single contextual cluster on a given politician-date combination and records cluster-level counts used as inputs to a logistic random-intercept model.Columns:politician: Name of the politician.report_date: Calendar date for which the counts were recorded (YYYY-MM-DD).contextual_cluster: Unique identifier of the contextual cluster (story-level grouping of documents) detected on that date.m_g: Number of documents in the cluster in which the politician is mentioned.n_g: Total number of documents in the cluster (cluster size).