Internal Seminar: Dong Ook Choi (Sangmyung University)

Title : News Selection by News Aggregators and Incentives for Newspapers to Invest in Journalism: The Case of South Korea

Abstract:
We study news selection by two prominent Korean news aggregators. For this purpose, we collect all online news articles of 52 national newspapers from South Korea for the year 2015, along with all the articles posted on the homepages of the two aggregators, whose joint share of news traffic was around 80%. We apply text analysis to the collected news articles to group them by topic through a clustering process and measure their degree of copying from preceding news content within each topic cluster. Using this data, we calculate the originality rate of each article, along with other quality factors such as article length, the presence of bylines, and whether the title includes breaking news or exclusive label. We first study selection of news topics (i.e., news clusters) by the two aggregators and find that the probability of selection increases with the size of the cluster, the number of newspapers which contribute articles, the average originality of articles and the average length of articles. We also find that on average, the 10.8th article within a specific cluster is the first to be selected by the aggregators. We then study selection of news articles by the aggregators and find that articles with higher originality, longer length, and exclusiveness are more likely to be selected. Surprisingly, we find that the rank of an article in terms of the time order it is posted is unrelated to its selection probability. These results suggest that there is little premium for being the first to break a news story, which seriously undermines Korean newspapers’ incentive to invest in journalism. We provide an explanation based on incomplete information about importance of topics and short-termism.

Room A406.