Adding background images associated with positive emotions in places where discussions happen online.

Decrease uncivil behavior.

Park and Singh (2022) conducted an online experiment simulating discussions around an existing news article from The Washington Post describing US immigration where participants had to comment on the article and reply to each other. Participants were sorted into three experimental conditions. In the first, the news article was presented in a blog with a gray background, while in the second and third, the background was a positive image associated with positive emotions (in color in the second condition; and greyscale in the third). Significantly, participants only interacted with others in the same condition and only one image was tested. Among other things, they found that using an image associated with positive emotions decreased participants' perception of incivility (effect size: p = .038, eta-squared = .09) and the usage of uncivil expressions, as noted by a single coder (p < .001, eta-squared = .08).

Design choices like this one can be easily implemented, and thus, even if the effects are small, they could be broadly applied.

The findings of the key study backing up this intervention are based on a small sample (N=105) of participants, and each unit in each of the treatment conditions is non-independent (since they influenced each other). Even though the study was peer-reviewed, this limitation bars the “Convincing” grade to this intervention.