While the number of studies about online content warning labels are on the rise, the graphical design of these labels has been understudied. To begin filling this gap, we perform a 13-arm between-subjects experiment (N=1,288) across two graphic elements of warning labels: intrusiveness and color. We find that highly intrusive labels that require click-through to view content are significantly less effective at decreasing trust in false information than lesser intrusive designs-conditional on those consumers clicking through the highly intrusive label to actively engage with the content. However, highly intrusive warning labels are still better than having no warning label. We find no evidence that color impacts the recall or effectiveness of content warning labels. These results should be taken in concert with the limitations of the study, which are discussed in depth in the paper.
Liban et al. (Sat,) studied this question.