This data accompanies our article entitle 'Can we trace the social affiliation of rooks (Corvus frugilegus) through their vocal signature?'. The data is required to run the analysis code, hosted at https://gitlab.com/kimartin/rook_vocal_signature. Included are the audio, annotations, and various intermediate results used for the article. The audio and annotations are stored in the dataset.zip archive, in the following format: five folders of audio .flac files and corresponding .txt files (each pair of files has the same name, just different extensions). Each folder contains recordings of a different colony / recording site (three around Strasbourg, France: Strasbourg_Captive, Contades, and CNRS; two around Cambridge, UK: Cambridge Captive and Cambridge Wild). The recordings span mostly the 2022 breeding season from around March to May, except for the Strasbourg Captive colony where recordings were taken in 2020 and 2021. All recordings were taken in naturalistic conditions (i.e. with background noises and other rooks vocalising). Annotations correspond to clean instances of the 'nest call', a particular call produced only by brooding females within the nest. Inside the dataset.zip archive is the tableau_vocs.tsv file, which aggregates the various .txt files for ease of use with the code. The intermediate_results.zip archive stores the data used for the statistical analysis, along with the R scripts needed for the statistical analysis and for creating the figures. All scripts should run out-of-the-box with the results stored in the same folder and reproduce the same results as the article, however up-to-date versions can still be found on the GitLab repository if needed.