Neural topic modeling reveals German television’s climate change coverage

Abstract The implementation and acceptance of climate policies depend on public perceptions of climate change. The media play a crucial role in informing the public discourse. While previous research has predominantly focused on written news, television remains the primary source of information globally. Here, we present an algorithm based on natural language processing techniques for identifying climate change coverage from subtitles of the leading German television news program, Tagesschau . Combining a dictionary approach with neural topic modeling, we classify the topics of over 28,000 news items (2015–2023). Our results show that climate change accounts for 4% of the total coverage, surpassed, for example, by sports coverage (9%). Acute crises, such as COVID-19, are covered more frequently and positioned more prominently. 80% of climate change coverage reports on climate policy, while only 10% covers climate impacts, like weather extremes. The latter tend to be covered in later news slots, indicating lower news value.