The Effect of Hiding Dislikes on the Use of YouTube's Like and Dislike Features

Soon-Gyo Jung, Joni Salminen, Bernard J. Jansen

Abstract: Using data from a major international news organization, we investigate the effect of hiding the count of dislikes from YouTube viewers on the propensity to use the video like/dislike features. We compare one entire month of videos before (n = 478) and after (n = 394) YouTube began hiding the dislikes counts. Collectively, these videos had received 450,200 likes and 41,892 dislikes. To account for content variability, we analyze the likes/dislikes by sentiment class (positive, neutral, negative). Results of chi-square testing show that while both likes and dislikes decreased after the hiding, dislikes decreased substantially more. We repeat the analysis with four other YouTube news channels in various languages (Arabic, English, French, Spanish) and one non-news organization, with similar results in all but one case. Findings from these multiple organizations suggest that YouTube hiding the number of dislikes from viewers has altered the user-platform interactions for the like/dislike features. Therefore, comparing the like/dislike metrics before and after the removal would give invalid insights into users’ reactions to content on YouTube.