The Wartime Propaganda Puzzle

This paper presents the efforts of the Media War Matching project to combine the puzzle pieces of Second World War propaganda collections, namely audio recordings of radio broadcasts, transcripts of radio broadcasts and newspapers. To give researchers a richer and broader picture of propaganda dynamics in the Nazi-occupied Netherlands, the project published digitised radio transcripts in the CLARIAH Media Suite, an online platform for humanities research, so that they could be analyzed both in isolation and in combination with other wartime media collections. In this paper, we first introduce the wartime propaganda landscape, as well as the Media War project, which combined quantitative and qualitative approaches to analyzing propaganda in pro- and anti-Nazi Dutch language media. Subsequently, we discuss the Media War Matching project, which emanated from the challenges and opportunities that the Media War research team encountered during their work. Then, we relate how we published a specific set of radio transcripts in the Media Suite so that researchers could optimally search and browse them, including the development of an algorithm to automatically date the individual pages. Next, we explain how we attempted to link the audio of radio broadcasts to the relevant transcripts, using Automatic Speech Recognition (ASR). Finally, we demonstrate how the publication of propaganda collections in the CLARIAH Media Suite, despite the limitations of the source material and the lack of success of linking, still enables the identification and comparison of quantitative patterns that can act as a starting point for further qualitative analysis of narratives. In conclusion, this paper reflects on the lessons learned and aims to inspire others to digitise, publish and link media collections.