Accurate, Fast, But Not Always Cheap: Evaluating “Crowdcoding” as an Alternative Approach to Analyze Social Media Data

Crowdcoding, a method that outsources "coding" tasks to numerous people on the internet, has emerged as a popular approach for annotating texts and visuals. However, the performance of this approach for analyzing social media data in the context of journalism and mass communication research has not been systematically assessed. This study evaluated the validity and efficiency of crowdcoding based on the analysis of 4,000 tweets about the 2016 U.S. presidential election. The results show that compared with the traditional quantitative content analysis, crowdcoding yielded comparably valid results and was superior in efficiency, but was more expensive under most circumstances.