How do tweeters feel about scientific misinformation: an infoveillance sentiment analysis of tweets on retraction notices and retracted papers

The traditional retraction mechanism's failure to eradicate the retracted papers' continued effects urges for more control and monitoring systems to warn against low-quality and flawed papers. To investigate the potential of Twitter in reflecting social attitudes about retracted papers, this study analyzed the sentiments expressed in the tweets about the papers and contrasted them against two benchmarks: the retraction notes and their tweets respectively serving as authorities’ voices and their social resonance. Using a sentiment analysis method, the study examined a collection of Scopus-indexed retracted papers, their retraction notices, and their tweets. The opinions expressed in the texts were mined using the SentiStrength. The findings revealed a high rate of untweetedness for the retracted papers (91.54%) and retraction notes (90.72%). However, the paper tweets mostly contained texts and were not limited to URLs, except for a low percentage (2.78%). While the retraction notices were mostly negative, followed by neutral polarity, the note and paper tweets were dominated by neutrality followed by negativity. Nevertheless, the paper tweets were more negative either in the pre-, or post-retraction phases. Moreover, negative tweets were comparatively more retweeted than positive and neutral polarities. The research findings implied tweet potentials in increasing the visibility of and awareness about low-quality and erroneous papers, even before being disclosed by official authorities, provided that more users are actively involved in the discussions on the platform. The potential can be regarded as a kind of monitoring applied by social users who feel responsible and show sensitivity towards the quality of science, though they may be scarce in number and selectively react to some papers.