The sad fate of small Facebook audiences: India
Team Members
Anushikha Chaudhuri
Laura Dea Vamper
Anabella Villanueva
Contents
1. Introduction
This research project is a subset of a bigger experimental project to analyze the content exposed to small audiences on Facebook against the backdrop of the general elections that are going to take place in 2024. Besides Mexico, the United States, Algeria, Croatia and North Macedonia, our team has researched the political advertisement environment of India with its general elections taking place around April/May 2024. In this massive pool of Ads with few consumers; misinformation, scams, and information operations are likely to be rampant. Thus, we aim to find out what type of content these small audiences are exposed to in India and how Ads are operated through various propaganda.
Due to India’s rich diversity and many political parties, we have narrowed our research down to focusing on the two major political parties that have the most influence across the country, namely the BJP (Bharatiya Janata Party) and Congress (Indian National Congress). BJP has been in power in India since 2014, succeeding the Congress party, which dominated Indian politics for the preceding 20 years.
We conducted an audit of various Ads sourced from the Meta Ads Library, enabling us to assess the political transparency of election campaigns. Our analysis aimed to determine if these campaign messages exhibited bias or a preference for a specific political party. Additionally, we investigated the Ads for any signs of suspicious activities, such as scams and examined potential connections to religious activities which may be used as propaganda.
Our focus is on a small-scale audience ranging from 1,000 to 500,000 individuals. India's global standing as one of the most populous countries, boasting an estimated population of 1.435 billion people, positions it as the world's most populated nation. This demographic magnitude is particularly significant when considering that India hosts the largest election processes globally (worldometers, 2023; Haidar, 2023). Moreover, the sizes of small audiences can exhibit significant variation due to the country's population density.
2. Research Questions
- With our project, we aim to answer the following questions:
- What type of materials are “small audiences” exposed to on Facebook in the light of general elections 2024 in India?
- What is the cumulative size of all the small audiences exposed to problematic content?
- What is the level of transparency these pages showcase through their personal information section on Facebook?
- What level of engagement is achieved by the problematic content shown to small audiences?
- In the context of India, given the right-wing alignment of the ruling party, how do political parties leverage religion as a tool to influence voting outcomes in elections?
3. Methodology and initial datasets
The initial datasets were sourced from two categories within the Meta Ads Library: the "Search Ads" search bar and the "download information and track spending" category.
In the former dataset, 3780 advertisements were retrieved using filters such as Country (India) and Ad category ("Social Issues, Elections, or Politics") with keywords "BJP" and "Congress." We filtered the amount spent on the Ads based on the local Indian currency below 100 rupees (approximately 1 euro) and we narrowed it down to 1293 Ads with impressions ranging from 1,000 to 500,000.
For a detailed examination of states with either no or substantial opposition to the ruling party BJP (Vaishnav, 2023), we utilized the latter search category, downloading a total of 2,257,844 Ads in total. Our focus was on Mizoram and Telangana with 334 and 638 Ads respectively, as these two states fit the ‘no or substantial opposition’ criteria most.
In order to investigate the issue of potentially spreading misinformation amongst small Facebook audiences, we used the Meta Ads Library to find advertisements about BJP and Congress by filtering them by location (“India”) and advertisement category (“Issues, elections or politics”), then typing in the party names separately. After initially getting tens of thousands of results, we used the filters section to search for posts that were 1) active ads, 2) posted between 01.01.2023. and 01.10.2024., and 3) their estimated audience size was equal to or below 500 thousand people (which is to be considered a small audience size in the case of India).
Upon downloading the datasets in csv format, we first analyzed the results’ authors one by one. We examined their posts to find out what kind of content they promote which may be biased politically and/or religiously and also considered the comments to measure audience engagement. Moreover, looking at Page Transparency within the ‘About’ section, we looked at History to see if the pages’ names were changed over time, which may indicate scam activity. Furthermore, we reviewed the ‘People who Manage this Page’ section, to see if there are any page administrators outside of India, which again, may indicate suspicious, potential scam activity.
Lastly, we accessed the Meta Ads Library again, to look at the pages’ latest advertisements, then clicked on See Ad Details, and mainly inspected the ‘About the disclaimer’ and ‘About the advertiser’ sections, to detect any unusual traits, such as fake email addresses, accounts and websites. We were able to find multiple suspicious, scam-like accounts by following this methodology.
4. Findings
After looking at the profiles from our dataset, we discovered that small audiences on Facebook are predominantly exposed to content centered around religion, voting slogans and politicians' visits to places of worship. An interesting find was that Ads with a budget of less than 100 rupees sometimes had impressions exceeding 300 thousand, probably due to the big size of ‘small audiences’ in India.
Another interesting find was that verified (blue-ticked) pages of politicians often lack transparency regarding their email IDs. We found that most of these pages are outsourced to different marketing agencies or individuals, but personal information and email IDs are seldom revealed by the politicians. In the Northern and Western Indian states where the Bharatiya Janata Party (BJP) faces little to no opposition, scam-oriented ad campaigns are scarce. Instead, there is a notable prevalence of religious propaganda on pages catering to small audiences. A significant observation is that the majority of advertisements exhibit an extremist approach to 'Hinduism.' The audiences' engagement further highlights this trend, with some comments often featuring religious slogans such as 'Jai Shree Ram' (Glory to Lord Rama), indicating a strong religious alignment. However, there is no general trend in this, since audience alignments vary with different pages.
Further investigation into states like Mizoram and Telangana, characterized by either no or substantial opposition to the ruling BJP, reveals that some pages operate from different countries with minimal personal details disclosed. Moreover, the advertisements in Telangana and Mizoram from our dataset remain largely neutral with voting slogans and politicians visiting different villages and places of worship. This auditing investigation helped us discover that BJP and Congress politicians have slightly different types of approaches when it comes to advertising their campaigns on Facebook with the former focusing on heavy religious propaganda and the latter discrediting BJP or making neutral posts.
It is imperative to mention that these are just preliminary findings and might not represent the true picture of the pages that we have retrieved from the Ads Library. Starting in 2019, Meta has taken steps to curb fake profiles and released advertising policies applicable to both Facebook and Instagram. This has also been implemented in countries like the United States, the United Kingdom and Brazil. Facebook product manager Sarah Schiff and Shivnath Thukral, the company's public policy director in India, expressed their commitment to establishing a new standard of transparency and authenticity for political advertising. In a statement, they emphasized that significant changes are being made ahead of India's general elections, emphasizing the platform's dedication to enhancing the integrity of political advertising on their platforms (Iyengar, 2019). These measures have been taken to prevent an outcome similar to the 2014 India general election, where deceptive social media tactics were used to influence its citizens. During 2014, major political parties in India took advantage and used techniques to gain fake likes, comments, shares and reactions. Nowadays, social media is used to demonstrate how big the influence is of these politicians, and although it does not directly help acquire real power, social media has become a medium to rise up in the ranks (Wong and Petersen, 2023).
5. Discussion
India possesses remarkable data sources for political and social research, notably through decennial censuses, renowned surveys like the Indian Sample Surveys/Indian National Election Studies and comprehensive data collected by government agencies. However, scholars face challenges in utilizing these sources for statistical analysis due to difficulties in accessing and merging the data. Many records are maintained at the state or district level, and older data, including census information before 1991, were published in book format. Historical election results are often in PDFs, and public data in India may be inconsistently formatted. The rise of the Internet, increased storage capacity and a growing emphasis on transparency are changing this landscape, with more government agencies and private organizations making substantial data available online, offering a valuable resource for researchers studying Indian politics and society (Jensenius and Verniers, 2017).
As shown by our findings, political Ads on Facebook seem to be transparent and are most likely run by the political parties themselves. Many electoral candidates can be found on Facebook and whilst they all run their political party on social media, not all of them use the same methods or have the same target audience. The Meta Ads Library and its transparency center allowed us to check each advertiser profile, giving us information on their activity, target group, ad serving and disclaimers on profiles not following guidelines.
How active a political party is on social media or whether they use misinformation and fake news can endanger or change political outcomes. One of the most notorious examples of this was the presidential election in the United States in 2016, as Donald Trump's digital media director stated: “Facebook and Twitter were the reason we won this thing", even Trump himself attributed his success to his social media account (Müller, Schwarz and Fujiwara, 2020). The political environment in India's social media has numerous ad impressions, the lowest amount alone had impressions exceeding 300 thousand. The amount spent on the service ad is 100 rupees, which is approximately 1 euro. This indicates that India's Facebook has influence and has helped to reach and engage with a larger number of voters in a short amount of time and thus it is successful at polarizing politics (Adhana and Saxena, 2018). The Meta Ads Library provided an overview of the most active candidates, and it was apparent that these political parties are really active in rural areas. They primarily target small audiences, which most likely indicates that demographic target groups who are at a disadvantage, are being persuaded by these political parties through emotional connection and they make voters feel more directly involved with the campaign through posts, comments, and likes. Unlike traditional media, Facebook allows politicians to actively engage with potential voters and followers by exposing them to their campaign strategies. Facebook sharing can ultimately create a personal connection through behind-the-scenes views of their campaign (Narasimhamurthy, 2014). One of those behind-the-scenes connections strategies is the use of religious beliefs during their campaign. There was a common denominator on the candidate profiles we discovered through the Meta Ads library: religion was used as a medium to approach and connect with voters. Essentially their campaign strategy was the documentation of candidates visiting temples and practicing religious traditions, which were later scattered through Facebook.
Digital platforms are being used as a battleground for democracy and propaganda. Both BJP and Congress have different tactics to gain new votes and one of these strategies to gain the favor of possible new voters is the use of misinformation, fake news and the circulation of negative news discrediting the opposite political party. Most BJP candidates are led by Hindu nationalists agenda groups who are using this as a form to compete against the Congress party, comments such as “If the BJP is here, your children will be safe. Hindus will be safe”, ‘’Hindu men were murdered by Muslims”, ‘’Hindu girls are being groomed by Muslim men to join the Islamic State’’, etc. (Shih, 2023). Hate speech and disinformation has grown increasingly and marginalizing religious minorities in the country is part of BJP’s Hindu nationalist agenda in a few states. In an interview with BJP staffers and allies, they stated that circulated posts aimed to exploit the fears of India’s Hindu citizens through
WhatsApp groups with the help of 150,000 social media workers (Shih, 2023). Contrary to
WhatsApp, Facebook content and profiles seem to have a more transparent political propaganda with verified accounts, which could be due to Facebook increasing data transparency for political advertising and is making sure its users know more about political ads. Furthermore, politicians have been required to verify their identity and location (Odutayo and Abedayo, 2019; Iyengar, 2019). Although most of the Facebook profiles we discovered were transparent, it is important to mention that a few of these profiles were managed by multiple individuals around the world. Overall BJP’s audience engagement surpasses the Congress Party, with the latter having less active audiences.
6. Conclusions
Building on the observations made by Fuller and Weizman (2021, p.159), the challenges of controlling information flow, especially in the era of leaks and social media openness prompt authorities to adopt a specific tactic for concealing exposed secrets. This tactic involves flooding the public space with an abundance of varied data, effectively concealing the original signal beneath layers of unrelated information. In our efforts to bring transparency to this dynamic, we have utilized Open Source Investigations (OSINV) by going through the Meta Ads Library to disclose some of this data. Nevertheless, the expansive political landscape of pages operating on Facebook, often under different names in case of India, suggests that more in-depth research in this domain is warranted. Understanding who is influenced and how, through advertisements on these platforms remains a focal point for future investigation.
7. References
Adhana, D. K., & Saxena, M. (2018). Role of Social Media in the Changing Face of Indian Politics: A Study with Special Reference to Facebook. International Journal of Research and Analytical Reviews (IJRAR), 6(1), 935-951.
https://www.researchgate.net/profile/Deepak-Adhana/publication/342162878_R OLE_OF_SOCIAL_MEDIA_IN_THE_CHANGING_FACE_OF_INDIAN_POLITICS_A_STUD Y_WITH_SPECIAL_REFERENCE_TO_FACEBOOK/links/5ee62247299bf1faac55c83a/R OLE-OF-SOCIAL-MEDIA-IN-THE-CHANGING-FACE-OF-INDIAN-POLITICS-A-STUDY- WITH-SPECIAL-REFERENCE-TO-FACEBOOK.pdf.
Fuller, M., & Weizman, E. (2021). Investigative aesthetics: Conflicts and commons in the politics of truth. Verso Books.
Iyengar, R. (2019, February 7). Facebook starts labeling political ads in India as election looms. CNN Business.
https://edition.cnn.com/2019/02/07/tech/facebook-india-political-advertising/ind ex.html
Jensenius, F. R., & Verniers, G. (2017). Studying Indian politics with large-scale data: Indian election data 1961–today. Studies in Indian Politics, 5(2), 269-275.
Müller, K., Schwarz, C., & Fujiwara, T. (2020, October 30). How Twitter affected the 2016 presidential election. CEPR.
https://cepr.org/voxeu/columns/how-twitter-affected-2016-presidential-election
Narasimhamurthy, N. (2014). Use and Rise of Social media as Election Campaign medium in India. CORE.
https://core.ac.uk/display/72802902?utm_source=pdf&utm_medium=banner&ut m_campaign=pdf-decoration-v1.
Shih, G. (2023, October 31). Inside the vast digital campaign by Hindu nationalists to inflame India. Washington Post.
https://www.washingtonpost.com/world/2023/09/26/hindu-nationalist-social-me dia-hate-campaign/
Odutayo, D., and Abedayo, B. (2019, January 21). Facebook not taking foreign political ads ahead of Nigeria’s general election. CNN.
https://edition.cnn.com/2019/01/21/africa/facebook-political-ads-nigeria-election /index.html.
Vaishnav, M. (2023, December 7). Decoding India’s 2024 election contest. Carnegie Endowment for International Peace.
https://carnegieendowment.org/2023/12/07/decoding-india-s-2024-election-cont est-pub-91178.
Worldometer. 2023. Countries in the world by population (2024). N.d.
https://www.worldometers.info/world-population/population-by-country/.
Wong, J. C., & Ellis-Petersen, H. (2021d, April 15). Facebook planned to remove fake accounts in India – until it realized a BJP politician was involved. The Guardian.
https://www.theguardian.com/technology/2021/apr/15/facebook-india-bjp-fake-a ccounts#:~:text=Thehttps://www.theguardian.com/technology/2021/apr/15/face book-india-bjp-fake-accounts#:~:text=The%20Facebook%20loophole-,Facebook% 20planned%20to%20remove%20fake%20accounts%20in%20India%20%E2%80% 93%20until%20it,a%20BJP%20politician%20was%20involved&text=Facebook%20 allowed%20a%20network%20of,being%20alerted%20to%20the%20problem.