First, a note on sources. Although I originally did searches about article frequency on Google News some years ago I can’t seem to find this Google feature anymore. Instead, the above graphs are generated from searches entirely on Factiva (which requires a subscription). Using Factiva, I did a series of searches. For example, I searched from Jan. 1, 2012 to Dec. 31, 2012 in the Hindu for the words “Supreme Court”. It came back with 6110 hits where there were articles in the Hindu that included these two words. I did the same for “Parliament” generating 4438 hits and then finally for “Manmohan Singh” retrieving 3309 hits. I repeated this for every year between 1998 and 2014 in both the Hindu and in the Times of India. To create the Prime Minister results, for the years 1998-2003 I searched “Vajpayee” for 2004-2013 “Manmohan Singh” and for 2014 “Narendra Modi”.
Now there are clear flaws to the above methodology, which I adopted mostly to save time as I just wanted a gist of these relationships. I’ll point out the two most blatant. (1) In election years, where there was a transition in power, I should have divided the year between the two governments and tracked who was Prime Minister accordingly. (2) The searches involving Parliament and Prime Minister are under-inclusive. I didn’t want to use “Prime Minister” as a search term because it returns articles that reference the British or Japanese Prime Minister, but not the Indian, thus over-inflating the result. However, there are certainly a number of articles that reference the Indian Prime Minister’s Office without naming the Prime Minister. These, and similar situations, will be missed in the above results. Also, the returns for Parliament would be higher if I included searches for “Lok Sabha” and “Rajya Sabha”. Particularly in election years the reference to “Lok Sabha” spikes well above references to Parliament, while in other years it’s below. Many articles reference both the words “Lok Sabha” and “Parliament” or “Rajya Sabha” and “Parliament”. One can avoid this duplication effect with appropriate multiple searches, but I decided for this exercise it just took too much time. The point of the above graphs is not to say this branch of government of that branch of government is covered most, but instead track trends in relative coverage over time.
One thing I find reassuring in terms of methodology is that both graphs broadly tell a similar story. From 1998 to about 2003 the Supreme Court was the least covered of the three search terms in both the Hindu and the Times of India. From 2003 to 2013 it becomes the most covered. In 2014, in both graphs “Narendra Modi” becomes the most written about.
Now what, if anything, of use can we glean from these graphs. Well, at the very least it indicates media interest in the different search terms. It’s more questionable though if such data is any indication of the power or relative activity of Parliament, the Prime Minister, or the Supreme Court. For example, there could be a lot of coverage of Parliament because of a corruption scandal that is discrediting the institution or because it is passing a series of laws and programs that is giving it new popularity. It is interesting to note though that the relative rise of coverage about the Supreme Court corresponded with Manmohan Singh being Prime Minister – a PM who was not particularly adept with the media and who had to vie with the power of Sonia Gandhi, who exercised control mostly out of the limelight. The increased coverage of the Prime Minister in the media may correspond with an increased ability of the Prime Minister to shape a media narrative in their favor when it comes to future institutional clashes, but then again it may not.
For now, I will take the safe academic answer to the meaning of this data, which is that we won’t really know what it is a good predictor of until we have more data. Factiva does not go back further historically for these publications, and while 16 years is not an unsubstantial amount of time, it really does not allow us to read too much into the data with much certainty although it should not stop us from hypothesizing.
Join the discussion