A big data analysis of the adoption of quoting encouragement policy on Twitter during the 2020 U.S. presidential election

Abstract

This research holds significance for the fields of social media and communication studies through its comprehensive evaluation of Twitter’s quoting encouragement policy enacted during the 2020 U.S. presidential election. In addressing a notable gap in the literature, this study introduces a framework that assesses both the quantitative and qualitative effects of specific platform-wide policy interventions, an aspect lacking in existing research. Employing a big data approach, the analysis includes 304 million tweets from a randomly sampled cohort of 86,334 users, using a systematic framework to examine pre-, within-, and post-intervals aligned with the policy timeline. Methodologically, SARIMAX models and linear regression are applied to the time series data on tweet types within each interval, offering an examination of temporal trends. Additionally, the study characterizes short-term and long-term adopters of the policy using text and sentiment analyses on quote tweets. Results show a significant retweeting decrease and modest quoting increase during the policy, followed by a swift retweeting resurgence and quoting decline post-policy. Users with fewer connections or higher activity levels adopt quoting more. Emerging quoters prefer shorter, positive quote texts. These findings hold implications for social media policymaking, providing evidence for refining existing policies and shaping effective interventions.