The data involves about 38 million URLs related to civic discourse that was shared on Facebook between Jan’ 2017 and July 2019. The data mostly showed results about the opinion of users on shared links to be fake or hate speech and if they were opened or liked. Plus, Facebook is also providing researchers with the demographic details of the users.
Facebook also announced in 2018 that it will give scholars access to the shared-link data, but its own experts suggested that it will compromise the user’s privacy.
In order to handle this matter, Facebook developed a method that hide the name of the users showing them as anonymous individuals in the data set. This method was extremely useful and many researchers are looking forward to conducting studies over it to understand how politically charged news spread on social media.
As soon as Facebook decided to give out the data to researchers, Garry King from Harvard University and Nathaniel Persily from Stanford University formed a non-profit entity named Social Science One.
The Social Science One hosted the data on its website and vet requests to access it. Within sometimes, Social Science One was able to collect about $11 million as a charity for the researchers who want to research on the data, so Social Science Research Council started to give out grants.
Joshua Tucker (a professor of politics and Russian studies at New York University) received about $50,000 in the grant from which he was able to conduct his research on the data set. Tucker found that the elderly people i.e. people older than 65 years were more likely to share misinformation on Facebook as compared to the ones in their 20s.
The research included traditional surveys to collect information on the user online the behavior for those who agreed upon it. Tucker wanted to expand the research by obtaining results from Reddit and Twitter and linking it to the nonpublic user data of Facebook, but it was not available.
Now, the researchers need to learn how to apply traditional methods to analyze the data. King along with Georgina Evans, a graduate student taught how to perform linear regression on differentially private data sets.
Scientists need to make sure that their results are correct so that the community can mold themselves into the new privacy approach. The available data can do wonders if used correctly, so that actions can be taken to control the spread of misinformation on the platform.
Photo: FB
Read next: Facebook to Go Ahead With Massive Satellite Launch Plan
