Facebook only shares data with researchers, does not finance or direct conclusion of research findings

Contrary to popular perception, Facebook hasn’t given any research grant to 60 academics from 30 institutions across 11 countries that are studying the impact of social media on elections

Contrary to popular perception, Facebook hasn’t given any research grant to 60 academics from 30 institutions across 11 countries that are studying the impact of social media on elections as part of an effort to prevent manipulation of social platforms. The researchers were selected under a review process by the Social Science Research Council and the independent group Social Science One.

A Facebook spokesperson told Firstpost that the social networking giant’s role was limited to sharing of information with independent researchers.

‘’Our primary role is to facilitate the sharing of privacy-protected data with researchers. We had no role in the selection of the individuals or their projects. Also, we won't have any role in directing the findings or conclusions of the research,” said the spokesperson.

Prof Vinay K Nangia, professor emeritus, Netaji Subhas University of Technology, Delhi, put the significance of the research initiative in context.

Facebook only shares data with researchers, does not finance or direct conclusion of research findings

Image: Reuters

“The objectives of the research won’t be clearly articulated in the media reports because of lack of access. Be that as it may, there are several unanswered questions. Who has written the proposal? And who has received it? Who is the biggest financier? Those who have invested the most will have the last word. Also, who is the end user of this research?” Nangia raises some pointed questions.

He wonders about the Principal Investigator (PI) of the research. The PI is a person who will know the research’s grand design.

Usually, these international researches are given to some key foundation in a reputed university and a professor from that institution is chosen as the PI. The PI, in turn, chooses professors from some universities and divides the work functionally or based on geographies. The work can be further distributed to other varsities for field research.

Explaining Facebook’s approach, the spokesperson said, “So far, there are 12 separate research projects and each has its own PI and a project description that can be accessed on the Social Science Research Council website. Facebook is not directly or indirectly funding any of these foundations. Instead, the foundations are providing financial support, through the Social Science Research Council, directly to the Universities and researchers.”

The Council is sourcing its funds from NGOs such as John and Laura Arnold Foundation, Democracy Fund, the William and Flora Hewlett Foundation, John S and James L Knight Foundation, Charles Koch Foundation, Omidyar Network, Sloan Foundation and Children’s Investment Fund Foundation.

Prof MP Gupta, who heads the Department of Management Studies (DMS) at IIT-Delhi, said, “Nowadays, social media is very powerful media. If you analyse the past elections in the US and India, there has been an extensive use of social media. If you look at the kind of messages that are being exchanged through WhatsApp, it appears that each political party has an army of bright people who are working day and night to create content to suit their outfits and discredit their rivals. It’s hard to distinguish the origin and authenticity of the content.”

On the methodology of the research, Prof Arpan K Kar, associate professor, DMS, IIT-Delhi, cited his similar research. “Last year, we did a similar study on US elections for Twitter. Typically, we took the profile of users but protected their identities.

Prof Kar also shared the Twitter Analytics Model as below:

Descriptive Analytics Content Analysis
Tweet statistics (Tweet, Reply, RT) Sentiment analysis
User statistics (ie number of users) Polarity analysis
URL analysis Emotion analysis
Hashtags analysis Topic modelling
mention analysis Lexical diversity
Word cloud (most frequent words)
Reach metric

 

Network analysis Space-time analysis
Friend-follower networks Time-trend analysis
Network layout Time series comparisons
Network diameter Geo-spatial analysis
Centrality analysis Geo-location analysis
Cluster detection Topic evolution
Information flow networks

Source: The Untold Story of USA Presidential Elections in 2016 -Insights from Twitter Analytics; Purva Grover, Arpan Kumar Kar, Yogesh K. Dwivedi, and Marijn Janssen

On the question of privacy of data, the Facebook spokesperson agreed with Prof Kar.

“The URL data set will be aggregated and will remain anonymous to prevent researchers from identifying any individual Facebook users. You can read more about the extensive steps we’ve taken to ensure people’s privacy as part of this effort here,” said the Facebook spokesperson.

Tech2 is now on WhatsApp. For all the buzz on the latest tech and science, sign up for our WhatsApp services. Just go to Tech2.com/Whatsapp and hit the Subscribe button.

Loading...





also see

science