Dirty data: missingness, contradictions and their usefulness for historians

Abstract In the twenty-first century, quantitative history seems to have made a comeback. And with it, the phrase ‘cleaning data’ has become commonplace. Data seems to be considered dirty when it includes blanks or gaps (missing information), weirdness (exceptions, unusual phrasings, things written in the margins of administrative forms, etc.) and contradictions (in the same source, as in the case of traces of deletion, and especially across sources). This article advocates for keeping these features of the data and for taking them seriously, because they can be key to source criticism as well as to substantive interpretation.