- July
Posted By : Adrian Chan
No Twitter in Meme-tracking and News Cycle Research

Interesting research was reported recently in the New York Times about the relationship between blogs and mass media. The research, out of Cornell University, focused on the news cycles observed in mass media and in the blogosphere. News cycles were measured by the distribution of a “meme.” Memes were defined by the researchers as quotes — full or partial. The research observed a tight 2.5 hour echo, memes peaking rapidly first in the mass media and then in the blogosphere shortly thereafter.

The research pretty much stands for itself, and while the researchers claim to have found a faster peak and decay in the lifecycle of memes than may have been expected, there don’t seem to be many surprises in their overall findings. Having defined a news cycle meme as a quote, it’s not surprising that quote-able news peaks in a meme like fashion. After all quotes are quote-able and are easily tracked (the research was limited to recognizable variants on the original quote: hence it excluded other kinds of discourse). Quotes can be repeated with limited analysis and context: a quote speaks for itself. Quotes are a common thread in news, insofar as quotes are the juicy bits of what our politicians and celebrities have to say.

It is no surprise that blogs would pick up on these memes soon after their appearance in mass media. Many blogs serve as news sources themselves. And news blogs need news to blog about: news is no different in mass and social media. Nor should the news cycle be much different from one medium to the other.

That said, there might be other cycles unique to social media that would be different. Unfortunately the research doesn’t cover these. Research didn’t directly address conversational (realtime) social media, most importantly Twitter. Twitter poses some challenges to media research: posts of 140 characters lose context, are reworded, shortened, and otherwise corrupted in ways that make them difficult to relate reliably to source quotes and memes. I think we could comfortably assume that twitter echoes the news cycle in ways like the blogosphere, although faster, and often preceding the mass media. The appearance of twitter-sourced stories in mass media, then through to blogs, has been covered already (e.g. earthquakes, Iran protests).

There are a few reasons social media would make an interesting distinct study, were it possible to reliably constrain research. Social media are more than news media, and that they are frequently driven by talk, interaction, or conversation in the form of tweets, comments, and status updates. Were it possible to conduct the research, it would be interesting to know:

  • Is there a long tail distribution of information in conversational (realtime) social media?
  • Does the distribution of information in conversational social media tell us something about relationships of credibility, influence, trust, authority, intimacy, etc and how they facilitate the distribution of information?
  • Are cycles of information distribution in conversational social media more “organic”: subject perhaps to daily rhythms of users and their habits and routines of use?
  • Does the “imitation” of information cited by researchers as one of two key ingredients function differently in conversational social media? Specifically, can it be determined whether or not imitation reflects social motives: retweeting for attention; retweeting for association; retweeting to get attention; tweeting for influence; tweeting for social inclusion; and so on.
  • After a news quote decays, is there long tail pickup in social media that reflects depth of interest? Can the commenting depth (not addressed by the research but often used by analytics tools) expose a kind of media authority more participatory than mass media, and credible for insight, commentary, analysis and not just news.
  • Is there a social graph ingredient in the distribution of news in conversational social media that is not explained by the echoing of news stories but which might offer valuable insight into lines of influence and which could render social relations of different kinds? Motivated not just to report the news, but to associate oneself, identify with a person, event, or to help tell a story, conversationalists can show us who they talk to, and about what.

Excerpts from Meme-tracking and the Dynamics of the News Cycle

Tracking new topics, ideas, and memes across the Web has been an issue of considerable interest. Recent work has developed methods for tracking topic shifts over long time scales, as well as abrupt spikes in the appearance of particular named entities. However, these approaches are less well suited to the identification of content that spreads widely and then fades over time scales on the order of days the time scale at which we perceive news and events.
As our principal domain of study, we show how such a meme-tracking approach can provide a coherent representation of the news cyclethe daily rhythms in the news media that have long been the subject of qualitative interpretation but have never been captured accurately enough to permit actual quantitative analysis. We tracked 1.6 million mainstream media sites and blogs over a period of three months with the total of 90 million articles and we find a set of novel and persistent temporal patterns in the news cycle. In particular, we observe a typical lag of 2.5 hours between the peaks of attention to a phrase in the news media and in blogs respectively, with divergent behavior around the overall peak and a heartbeat-like pattern in the handoff between news and blogs.
First, the set of distinctive phrases shows significant diversity over short periods of time, even as the broader vocabulary remains relatively stable. As a result, they can be used to dissect a general topic into a large collection of threads or memes that vary from day to day. Second, such distinctive phrases are abundant, and therefore are rich enough to act as tracers for a large collection of memes; we therefore do not have to restrict attention to the much smaller collection of memes that happen to be associated with the appearance and disappearance of a single named entity.
From an algorithmic point of view, we consider these distinctive phrases to act as the analogue of genetic signatures for different memes. And like genetic signatures, we find that while they remain recognizable as they appear in text over time, they also undergo significant mutation.
Outside of computer science, the interplay between technology, the news media, and the political process has been a focus of considerable research interest for much of the past century [6, 22]. This research tradition has included work by sociologists, communication scholars, and media theorists, usually at qualitative level exploring the political and economic contexts in which news is produced [19], its effect on public opinion , and its ability to facilitate either polarization or consensus [15].
We perform this analysis both at a global level understanding the temporal variation as a wholeand at a local level identifying recurring patterns in the growth and decay of a meme around its period of peak intensity.
We also show how the temporal patterns we observe arise naturally from a simple mathematical model in which news sources imitate each others decisions about what to cover, but subject to recency effects penalizing older content.
Among the fastest sources we find a number of popular political blogs; this measure thus suggests a way of identifying sites that are regularly far ahead of the bulk of media attention to a topic.
Some of the key research issues here have been the identification of topics over time [5, 11, 16], the evolving practices of bloggers [25, 26], the cascading adoption of stories [3, 14, 20, 23], and the ideological divisions in the blogosphere [2, 12, 13].
Our goal is to produce phrase clusters, which are collections of phrases deemed to be close textual variants of one another. We will do this by building a phrase graph where each phrase is represented by a node and directed edges connect related phrases. Then we partition this graph in such a way that its components will be the phrase clusters.
Notice how the plot captures the dynamics of the presidential campaign coverage at a very fine resolution. Spikes and the phrases pinpoint the exact events and moments that triggered large amounts of attention.
To begin with, there are interesting potential analogies to natural systems that contain dynamics similar to what one sees in the news cycle. For example, one could imagine the news cycle as a kind of species interaction within an ecosystem [18], where threads play the role of species competing for resources (in this case media attention, which is constant over time), and selectively reproducing (by occupying future articles and posts).
We argue that in formulating a model for the news cycle, there are two minimal ingredients that should be taken into account. The first is that different sources imitate one another, so that once a thread experiences significant volume, it is likely to persist and grow through adoption by others. The second, counteracting the first, is that threads are governed by strong recency effects, in which new threads are favored to older ones.
When there is only a recency effect but no imitation (so the probability of choosing thread j is proportional only to (t – tj) for some function ), we see that no thread ever achieves significant volume, since each is crowded out by newer ones.
When there is only imitation but no recency effect, (so the probability of choosing thread j is proportional only to f(nj) for some function f), then a single thread becomes dominant essentially forever: there are no recency effects to drive it away, although its dominance shrinks over time simply because the total number of competing threads is increasing.
In general, one would expect the overall volume of a thread to be very low initially; then as the mass media begins joining in the volume would rise; and then as it percolates to blogs and other media it would slowly decay. However, it seems that the behavior tends to be quite different from this. First, notice that in Figure 7 the rise and drop in volume is surprisingly symmetric around the peak, which suggests little or no evidence for a quick build-up followed by a slow decay.
The majority of phrases first appear in news media and then diffuses to blogs where it is then discussed for longer time. However, there are also phrases that propagate in the opposite way, percolating in the blogosphere until they are picked up the news media. Such cases are very important as they show the importance of independent media.

Authors: Jure Leskovec Lars Backstrom Jon Kleinberg

New York Times article about the news cycle research
Study Measures the Chatter of the News Cycle


Leave a Reply