Clicks And Editorial Decisions: How Does Popularity Shape Online News Coverage?
University of Toulouse 1 – Toulouse School of Economics (TSE)
University of Pennsylvania – The Wharton School
Does the popularity of news stories influence the way editors allocate resources to them, and if so, how? With news available online, editors have the ability to track popularity or reader demand (i.e., the number of clicks) for individual stories. Using a unique online news dataset from a large Indian English daily newspaper, we provide evidence that editors indeed expand coverage of stories which receive more clicks. To establish a causal link between clicks and coverage, we use a novel instrumental variables strategy exploiting rainfall and power outages as exogenous shocks to reader access to online news. Further, we find that the newspaper responds asymmetrically across different types of news stories by giving additional coverage based only on the higher clicks received by ‘hard’ news stories. We provide evidence for ‘hard’ news crowding out ‘soft’ news and not vice-versa. Moreover, we relate our results to firm strategy and the inability of firms to handle ‘big data’.
Clicks And Editorial Decisions: How Does Popularity Shape Online News Coverage? – Introduction
On the 7th of January 2015, a terrorist attack on the satirical newspaper Charlie Hebdo resulted in the deaths of 12 people in Paris. In the days that followed, the event dominated news headlines around the world. On the same day, reports also began to emerge about the slaughter of thousands of Nigerians at the hands of Boko Haram. In contrast to the Charlie Hebdo headlines, this story had little follow-up. Disinterest or a lack of demand on the part of readers was a major factor cited to explain the dearth of media attention on the Boko Haram massacre.
In this paper we analyze the role of a media outlet’s need to maximize demand or readership, to increase advertising revenue, on editorial decisions in the context of online news. In particular, we examine whether there are more follow up articles on topics which get a larger number of clicks initially. Estimating the impact of popularity on editorial coverage decisions is difficult for a number of reasons. First, some stories are simply seen as more interesting than others. These stories will attract a larger number of clicks, and will likely be followed by additional articles covering the same, interesting, event. This makes it difficult to disentangle cause and effect: are editors choosing articles they know readers will like, or are both readers and editors responding to fundamental story characteristics? We address this issue by instrumenting for clicks using shocks to page views which are unrelated to the characteristics of stories published that day. These shocks come from two different sources: days with rain and power outages. On rainy days, readers are constrained in their activities and may be more likely to choose to read the news; power shortages, on the other hand, limit individuals’ ability to use electronic devices or connect to the internet. Second, testing the relationship between clicks and editorial coverage requires data at the article level which is hard to come by since it is used by newspapers to price advertisements. To this end, we acquire propietary information on clicks at the level of a URL from a leading English language Indian national daily newspaper.
Documenting whether decisions of news editors systematically cater to reader interest is important for a variety reasons. First, there is overwhelming evidence that news affects voting (Gentzkow et al. (2011)), judicial outcomes (Lim, Snyder and Stromberg (2015)), policy (Eisensee and Stromberg (2007)) and financial decisions (Fang and Peress(2009)). Hence, we are adding another link to this chain by analyzing how many clicks generate a certain amount of coverage which in turn has an impact on different outcomes of interest. Second, Best (2009) shows theoretically that selective coverage of stories due to their ‘sensational appeal’ by a newspaper can distort the beliefs of readers and make them demand a sub-optimal amount of public spending on policy issues than if they were perfectly informed. Hence, clicks (or popularity) based coverage can have potentially adverse effects on the beliefs and subsequently, the actions of the readers3. Relatedly, identifying whether news stories are chosen because of demand side (or supply side) reasons is important for the design of media regulation, as highlighted by Prat and Stromberg (2013). The Federal Communications Commission (FCC), for example, have ignored demand side incentives while proposing regulations for media ownership assuming a positive correlation between ownership of the media outlet and the viewpoint expressed by it. If editors systematically follow up on stories based on their popularity or sensational appeal then there could be additional regulatory implications.
See full PDF below.