Right Back, Left Forward: How news biases played out during the Adani-Hindenburg Episode
Quantitative analysis shows that Indian right-wing media expressed strong opinions and defensiveness but ignored Hindenburg’s 88 questions. Centrists and business outlets prioritized facts over opinion. The left-wing made it about the ruling establishment but raised stronger points.
On January 25th, 2023, Hindenburg Research sent shockwaves through Indian news media when it accused the Adani Group, a prominent Indian conglomerate, of engaging in various financial irregularities and environmental violations. This report caused the Adani Group’s share prices to plummet and sparked controversy across India.
However, the Adani Group vehemently denied the allegations and dismissed them as “blatantly erroneous.”
Although the Indian media covered the Hindenburg Research report extensively, the coverage was not uniform across various media outlets. Most importantly, very few media organizations provided in-depth factual reporting on the 88 questions raised in the report and instead chose to take sides.
At a high level, the right-wing media outlets were more opinionated and defensive. They frequently brought up “the world against India” and “an attack on India’s growth,” whereas centrist media outlets reported more facts and published statements by both sides. Even the left wing had a decent number of points despite going on the offensive against the ruling establishment.
To thoroughly analyze the right-wing media’s reporting style, we analyzed over 500 English-language articles from over thirty media organizations from January 25th to March 16th. Using pre-trained models, we gained a deep understanding of the reports’ sentiment and emotional flow from start to finish. We also identified clusters of media houses, analyzed word clouds and lexical dispersion, and employed other types of analysis.
After combining all the results of our analysis, we arrived at conclusions supported by a contextual basis. It’s important to note that this was a more complex and generic sentimental analysis, as it required a more nuanced and in-depth approach.
A quantitative analysis of the right-wing reporting
What is clear is the right-wing media wasn’t ready for a fight from the onset of the Hindenburg allegations. In contrast, the Left and Centrists took to it more doggedly.
Clubbing The Republic, SwarajyaMag, and OpIndia as right-wing media outlets–given their regular orientation–during the analysis period (from 25 Jan to 16 Mar), the three contributed only 20-25 percent of the total number of articles written despite the newsworthiness of the Hindenburg story.
This lack of reporting stands in stark contrast to topics that put the government and groups like Adani, who are allies of the government, in a positive light.
To gauge the sentiments of each block throughout an entire article, we split the pieces into 400 to 500-word blocks. We identified nine different sentimental flows across all article blocks, which were then used as a parameter to decipher the flow of all articles. These flows are as follows:
- Negative-Negative
- Negative-Neutral
- Negative-Positive
- Neutral-Negative
- Neutral-Neutral
- Neutral-Positive
- Positive-Negative
- Positive-Neutral
- Positive-Positive
To analyze specific outliers or patterns, we used the baseline sentiment analysis of the story, which can tend to be negative due to the allegations against Adani. We used pre-trained models designed for financial reporting and stories to ensure objective and context-based analysis.
Our analysis of right-wing publications revealed that they tend to follow certain patterns. Specifically, the negativity quotient at the story’s beginning is lower than the average baseline, the middle section is more neutral than average, and the ending typically has a higher neutrality or positivity factor.
In contrast, left-wing publications exhibit different patterns. These stories begin with a high degree of negativity, are less neutral in the middle, and have less positivity at the end.
Understanding the context and form of reporting is crucial in grasping the presented narrative. And the beginnings and endings of a story can offer insight into its direction.
For example, consider this story by SwarajyaMag. It rightly starts by acknowledging the seriousness of the allegations against Adani. But the story becomes biased in its reporting soon after.
The last two sections of the story are devoted to raising concerns about the accuracy of foreign entities’ claims against large corporations. It casts doubt on the timing and validity of Hindenburg’s report without delving into the actual allegations in any way.
Multiple studies and citations analyze the ending of a news story and tie it up to an inherent bias in the author’s intent. How a news story concludes significantly impacts the reader’s perception of the story. For instance, a literature review on “Automated identification of media bias in news articles” asserts that “Journalists create the spin of an article on all textual levels, e.g., by supporting a quote with an explanation (phrase level), highlighting certain parts of the event (paragraph level), or even by concluding the article with a statement that frames all previously presented information differently (article level). The order in which facts are presented to the reader influences what is perceived (e.g., some readers might only read the headline and lead paragraph) and how readers rate the importance of reported information”. In the above context, the story ends with raising concerns about Hindenburg’s research and the timing.
We see a clear trendline when we average out the Left and right-wing numbers. Reports on the Left start off more negative, less neutral, and end with less positivity. However, the ones on the Right start less negatively, stay neutral, and end more positively.
It’s important to note that the headlines of news stories can significantly influence readers’ interpretations by setting the tone or inviting confirmation bias. Even if the story’s content may differ considerably in information or sentiment, the headline prejudices the reader.
The actual content of the article notwithstanding, sharing of headlines, especially on social media, further influences secondary and tertiary coverage. Regarding story titles, right-wing publications tended to be more neutral and positive, while left-wing publications were more negative and neutral.
Another aspect of our analysis involved identifying and combining the emotions expressed in each paragraph with the overall sentiment of the piece. We categorized each paragraph into “anger,” “fear,” “joy,” or “sadness.”
Most media outlets exhibited a “Neutral+Anger” combination as the predominant “Sentiment+Emotion” pairing. Publications such as Scroll, The Hindu, and The Print had higher levels of “Negative+Anger” even though the overall average for all publications was lower.
On the other hand, right-wing and neutral publications scored higher in “Positive+Anger,” “Positive+Fear,” and “Positive+Joy.”
Neutral publications like Business Today, CNBCTV18, Mint, and The Economic Times often ended their stories positively or neutrally, and the average sentimental flows in their stories can put them into the bucket of unbiased reporters of the Hindenburg report.
Word Cloud Analysis
After normalization, we can see three different dimensions regarding using words and mentioning topics and entities. Notably, upon the release of the Hindenburg report, Adani’s official response was to invoke nationalism and quote the “Jallianwala Bagh massacre.”
Additionally, Adani’s defense referred to India’s financial growth and how this was an attempt to denigrate India’s economic growth and superiority. While it could be possible, none of the reports went ahead of merely quoting Adani’s response to the eighty-eight questions raised by the Hindenburg.
The third dimension was George Soros’s statements on the Adani allegations and Adani’s alleged proximity to Prime Minister Narendra Modi, and the BJP.
Amid all this, factual and objective reporting was lost in the right-wing’s reporting.
The word cloud from SwarajyaMag, Republic, OpIndia, India Today, and Hindustan Times, shows words like “Hindenburg,” “Congress,” “Rahul Gandhi,” “Indian,” “Stock market,” “Stock manipulation,” etc., playing a significant role.
Rather than focusing on the allegations’ authenticity, this shooting of the messenger is in line with our overall analysis of stories that lean towards the right wing where the reporting gets less factual and more rhetorical.
Moreover, the objectivity and factuality in right-wing reporting considerably reduced as days passed since the day of the release of the Hindenburg report.
More opinion pieces started to crop up. Also, while trying to play defensive, the right wing attempted to distance the Adani enterprise from being associated with Narendra Modi. Considering the uncertainty surrounding the allegations against Adani, this might have been a sensible strategy.
On analyzing the centrist or neutrally reporting media houses, we see a stark difference in reporting the Adani issue. The word cloud shows a high objectivity and factuality quotient in their reporting with terms like “FPO,” “Assets,” “promoters,” “Power,” “Investor,” “Transmission,” etc., with negligible mention of any political party or leaders.
The terms are all related to the allegations and the market trends upon the report’s release, and most of them are financial.
This neutral word cloud was from business media outlets like CNBCTV18.com, Times of India, Financial Times, Mint, The Hindu Business Line, Outlook India, and The Economic Times.
When analyzing left-wing publications, we found that while they naturally mentioned the government and Prime Minister Modi quite often, they also included a good amount of objective and factual reporting in their stories. However, the negativity factor was higher compared to previous analyses. The left-wing stories also mentioned Adani’s alleged excesses in Bangladesh and Sri Lanka, bringing up past allegations.
We also conducted an analysis of word pairings in the stories, specifically looking at words that preceded and followed the word “Modi”. In right-wing stories, we noticed that most preceding words were “Prime Minister”, while the sentences that followed were justifying and defending his actions related to the Adani case. On the other hand, the left wing hardly referred to him as the Prime Minister, and the sentences that followed mostly linked him to the Adani fiasco. In contrast, centrist publications focused on highly factual reporting when connecting the word “Modi” to the contextual subject in focus.
Finally, we conducted a lexical dispersion analysis to identify the occurrences of keywords throughout the length of a news article, from start to finish.
The right-wing media showed a clear pattern of excessive jingoism and rhetoric, with frequent use of the term “India” at the beginning of the articles. They also made numerous mentions of the opposition party, “Congress.” In contrast, the centrist media used more objective language and avoided the excessive use of such terms.
Overall, the right-wing media had lower levels of objectivity and factuality than the Left and centrist media outlets. Their articles were also approximately 40% shorter in length compared to other media publications. After normalization, it became clear that the right-wing media had more opinionated phrases than other media outlets.
Balaji Sri Raj contributed to this report.
Author
