Politician engagement with media, and its impact on news propagation

Dibyendu Mishra and Joyojeet Pal

(suggested citation: Mishra, D. and Pal, J. (2020) Politician engagement with media, and its impact on news propagation. https://joyojeet.people.si.umich.edu/polarity-of-indian-news-sources-among-politicians/ )

In this piece we examine the impact of politicians’ engagement with news stories from specific news sources using retweets as a metric for news propagation.

We plotted all the retweets from 2000 most followed politicians in India in 2019 to understand which news sources they used, as well as what is the effect on that news item once it is retweeted by a politician. Our results are as visualized in this graph. An earlier version of this with the basic description of the data and the outcomes is here.

Here is a visualization of our results

Media houses retweeted by politicians arranged by political spectrum and size of retweets — Figure 1: Media sources retweeted by Indian politicians arranged by political spectrum and number of retweets

Bubble location on x-axis is a metric for polarization
1. The farther a bubble to the right (+ x-axis), the proportionately more the engagement of BJP politicians
2. The farther a bubble is to the left (- x-axis) the proportionately more the engagement of non–BJP politicians
3. If the bubbe is on an extreme end — +1 or -1, then it is only engaged by a single party
Bubble location on y-axis indicates the number of engagements by politicians with the source. We use a log scale
Bubble size is a metric of propagation: the larger the bubble, the higher the median retweets of the message when a politician engages with it
Bubble color indicates relative propagation influence of the two parties
1. Each bubble is two bubbles – one orange and one blue, both with the same midpoint
  1. Orange bubble is the number of retweets when a BJP politician retweets that media house
  2. Blue bubble is the number of retweets when a non-BJP politician retweets that media house
2. If the blue bubble is larger, that source gets more retweeted when non-BJP politicians retweet that source
3. If the orange bubble is larger, that source gets more retweeted when BJP politicians retweet it
4. If the bubble has a single color, it either means that there is no difference in the relative retweets from members of the two parties (eg TimesOfIndia, EconomicTimes, JagranNews) or means that only one party ever retweets that source (eg-TheCaravanIndia: non-BJP, DDNewsHindi: BJP)

Key results:

Sources that are the most engaged by politicians (see Figure 2 below)
1. ANI (dominant source, more engagement than next four sources combined)
2. TimesNow
3. AajTak
4. NDTV
Sources that get relatively more engaged when a politician retweets (the bigger a bubble)
1. ANI
2. ZeeNewsHindi
3. BBCHindi
4. OpIndia
5. NDTVIndia
6. ttIndia (The Telegraph)
7. Interpreting the above finding is very important. it suggests that these sources may be seen as inherently political in nature. For instance, when @timesofindia gets retweeted by a politician, it gets much less throughput than @theWire_in or @OpIndia, even though they both get relatively the same number of retweets
Sources that do better when a BJP politician retweets them
1. ANI
2. ABPNews
3. Republic etc (wherever orange is the outer concentric circle)
Sources which get more retweeted when a non-BJP politician retweets an item
1. BBCHindi
2. TheWire_in
3. NDTVIndia etc (wherever blue is the outer concentric circle)
Hindi language news gets more retweeted for the same source in the case of NDTV — ie @ndtvindia is much more retweeted during engagements with politicians than

Figure 2: Media sources cited by Indian politicians - orange is number of BJP engagements, blue is number of non-BJP engagements — Figure 2: Media sources cited by Indian politicians – orange is number of BJP engagements, blue is number of non-BJP engagements

Methodology

On this graph, we have plotted major news sources based on how much their news stories get retweeted by party. We took the top 1000 highly followed BJP politicians and top 1000 non-BJP politicians and pulled all their tweets(~5.1 M) since Jan 2019. We then record retweeting of news sources from the Twitter handles of 212 news sources. We find about 64,000 instances of such retweets. Of these we select 35 twitter handles who were retweeted 52,000 times by these politicians, the overwhelming majority of represented engagements by politicians. A part of this selection included discarding those news handles who were retweeted less than 200 times. The logic for plotting retweets, as opposed to likes, replies, mentions or retweets with mentions (quotes) is that a straight retweet is the closest approximation of one account wanting to propagate the words of another.

The x-axis is the polarity – thus a news source at the extreme right of the graph (score +1) was only ever tweeted by BJP politicians, whereas a news source at the extreme left (score -1) of the graph has only ever been tweeted out by non-BJP politicians. A news source that appears around the middle, on ‘0’ on the x-axis was equally used by the BJP and non-BJP sources. The y-axis is a measure of the frequency of use of the news source – thus eOrganiser or scroll_in are not used very often, whereas ANI at the top of the graph with a high score on the y-axis is a frequently used source.
The size of the bubbles indicate the median of the retweets received by the source tweets. The colors indicate the party of the politicians who retweeted these tweets. For example, ANI has a larger orange bubble which means that the tweets that were retweeted by BJP politicians received a higher retweet on median in comparison to the tweets that were retweeted by members of other parties.