Selective and deceptive citation in the construction of dueling consensuses

Five files are attached associated with the publication "Selective and deceptive citation in the construction of dueling consensuses."

The first, Analyze_Data.py, is Python code necessary for replicating this work. Some packages may need to be downloaded, which you can find in the second file, requirements.txt.

The third file consists of annotation data for this paper. It is an Excel file with several sheets. The first, "Twitter_URLs," consists of the Twitter URLs we analyzed in our paper, the number of times they were shared in our dataset, shorthand names we assigned to these URLs, the number of outgoing citations we annotated from these sources, and whether we considered them to be explaining the science behind the effectiveness of masks. "Twitter_Edges" contains our annotations of citations on Twitter. They are directed from our shorthand names for Twitter URLs to reference titles for the citations. Data on these reference titles can be found in the "Twitter_Sources" document. The Date columns marks the most recent publication date we annotated for this source. The sheet "Twitter_URLs_50Shares" is an extended list of all URLs in our dataset shared at least 50 times. This data is not analyzed in the paper, and is provided for reference. The sheet "CitingPolarity" provides our annotations of sources we considered to be generally skeptical of masks' effectiveness ("Anti") and those that had either a positive or neutral stance. The "Citation_Context" page provides example context sentences for the 5 papers were reference in Table 1. Finally, the "WOS_Edges" sheet provides citation annotation for sources we retrieved from the Web of Science. See our paper for additional details.

The zipped file WOS_Raw.zip contains citations collected as November 18, 2021 from the Web of Science. It is required to run the Analzye_Data.py script. Additional details can be found in our paper. Finally, the "Beers_ConsensusNetworks_VisualizationWorkbook.gephi" file provides visualization data for the network visualization in Figure 1. It can be opened in the network program Gephi.

Additional Notes: In the "Twitter_URLs" datasheet, the URL on Row 157 (https://wwwnc.cdc.gov/eid/article/26/5/19-0994_article) was often erroneously unwound into the error page found on Row 6 (https://wwwnc.cdc.gov/eid/404.html?aspxerrorpath=/eid/article/26/5/19-0994_article). Both reference the same article, linked to in Row 157.