Experimenting with Choice Architectures to Enhance News Diversification

The next speaker in this ICA 2024 conference session is Nicolas Mattis, whose interest is in news diversification. This builds on notions of democratically (rather than merely economically) motivated news recommender design, the purported links between news diversity and democratically desirable effects, and emerging experiments with various news diversification metrics.

Any kind of digital news environment can be thought of as a choice architecture: they filter, rank, and present news stories to their user. News diversification facilitates engagement with diverse news, and seeks to facilitate normatively desirable outcomes; this involves the adjustment of such choice architectures. Democratic theory can provide a framework for developing the metrics that test the effects of such interventions.

The project conducted an experiment with some 700 participants in the Netherlands to explore the effect of choice architecture in the context of stories on a proposed new transgender law in the country; people were asked to explore available articles in an experimental news app and then vote on the issue. Articles used different levels of activating language, affective language, and alternative voices to simulate diversity of coverage, and they were presented in differently structured newsfeeds.

Overall, however, such news diversification had very limited effects; this is likely to be also a result of the short-term experimental setting. Also, positive and negative effects might have cancelled each other out. News diversification might also have resulted in lower perceptions of article quality. But it remains to be seen what effects there could be in the longer term – this cannot be measured in experimental work alone.