Replication Data for: Parties’ ideological cores and peripheries: Examining how parties balance adaptation and continuity in their manifestos

This dataset accompanies the article “Parties’ Ideological Cores and Peripheries: Examining How Parties Balance Adaptation and Continuity in Their Manifestos”, published in the British Journal of Political Science. It provides all replication data required to reproduce the empirical analyses presented in the paper. The dataset contains party-manifesto data for 137 Western European parties across 17 countries (1946–2022), derived from the Manifesto Project Corpus. Each manifesto is split into sections representing a party’s ideological core and policy periphery, following the party-family–based operationalisation introduced in the study. For every quasi-sentence, the dataset includes predicted policy-category probabilities generated using manifestoberta—a multilingual XLM-RoBERTa language model fine-tuned on Manifesto Project annotations (https://manifestoproject.wzb.eu/information/documents/manifestoberta). These predictions are aggregated to compute economic and cultural left–right positions, ideological extremity, and inter-election positional change separately for core and peripheral manifesto sections. This archive enables full reproduction of all statistical analyses, figures, and robustness checks reported in the article. It is intended to support transparency, facilitate reuse in future research, and provide a foundation for studying party competition, programmatic change, and the strategic distinction between core and peripheral issue domains. Instructions for how to use the scripts and datasets, including a full replication workflow, are available in the accompanying GitHub repository: https://github.com/FabianHabersack/parties_ideological_cores_vs_peripheries