This dataset contains tabular files with information about the usage of "actually" in contemporary British speech. We draw on two spoken corpora: (i) The demographically sampled part of the Spoken BNC1994 (Crowdy 1995) and (ii) the Spoken BNC2014 (Love et al. 2017). For both corpora, we list the usage rate observed for each speaker (total number of words produced, number of actually tokens, normalized frequency of actually expressed as per million words), along with information about the sex and age of the informant. In total, the dataset includes n = 1,408 speakers (Spoken BNC1994DS) and n = 668 speakers (Spoken BNC2014). For each corpus, we offer data tables with additional speaker meta-data. For a subset of the Spoken BNC1994DS (speakers with available information on gender and age; n = 886 speakers; n = 2,688 tokens), we also report on the position of actually in the clause (initial, medial, final), which was annotated manually. Related publication: Sönning, Lukas & Manfred Krug. 2022. Comparing study designs and down-sampling strategies in corpus analysis: The importance of speaker metadata in the BNCs of 1994 and 2014. In Ole Schützler & Julia Schlüter (eds.), Data and methods in corpus linguistics: Comparative approaches, 127-159. Cambridge: Cambridge University Press. https://doi.org/10.1017/9781108589314.006