Source-linked AI summary
Rising tides or rising stars?: Dynamics of shared attention on Twitter during media events
Yu-Ru Lin, Brian Keegan, Drew Margolin, David Lazer
TL;DR
The paper asks whether shared attention during media events changes collective Twitter behavior beyond other bursts of activity and whether the change is population-wide or elite-driven. Using eight 2012 U.S. political events and approximately 200,000 politically engaged users, it compares event behavior with news and typical periods. The results indicate increased activity concentrated around elite users, hashtags, and tweets, while the authors note limits to generalizability.
Problem
The paper asks whether media events produce behavioral patterns distinct from breaking news and typical time, and whether changes reflect broad population shifts or elite-user behavior.
Method
The study compares communication patterns across eight 2012 U.S. political events using approximately 200,000 politically engaged Twitter users.
Results
Media events increased activity but primarily concentrated attention on elite users, hashtags, and tweets rather than distributing it broadly.
Takeaways & Limitations
The findings suggest that shared-attention events replace existing social dynamics with collective attention focused on rising stars.
Takeaways & Limitations
The findings are limited by eight politically related events observed over roughly six weeks and a sampling strategy that may oversample debate-active users.
Abstract
from arXiv · showhide
"Media events" such as political debates generate conditions of shared attention as many users simultaneously tune in with the dual screens of broadcast and social media to view and participate. Are collective patterns of user behavior under conditions of shared attention distinct from other "bursts" of activity like breaking news events? Using data from a population of approximately 200,000 politically-active Twitter users, we compare features of their behavior during eight major events during the 2012 U.S. presidential election to examine (1) the impact of "media events" have on patterns of social media use compared to "typical" time and (2) whether changes during media events are attributable to changes in behavior across the entire population or an artifact of changes in elite users' behavior. Our findings suggest that while this population became more active during media events, this additional activity reflects concentrated attention to a handful of users, hashtags, and tweets. Our work is the first study on distinguishing patterns of large-scale social behavior under condition of uncertainty and shared attention, suggesting new ways of mining information from social media to support collective sensemaking following major events.
Introduction
The paper asks whether shared attention during media events changes Twitter behavior beyond ordinary bursts and whether those changes reflect population-wide shifts or elite users’ behavior. It proposes studying media-event-driven behavioral change using politically engaged users and comparisons across event types.
- Motivation: Twitter supports information sharing during both breaking news and scheduled media events, increasingly through dual-screen participation.Users can react to live broadcasts while simultaneously communicating on social media.
- Motivation: Because Twitter activity is concentrated around a few stars, media events might reproduce elite-centered attention rather than broaden participation.The paper frames this as a tension between concentrated attention and wider participation.
- Research questions: The paper examines whether users behave differently during media events than during breaking news or normal time.It focuses on changes in hashtags, mentions, and retweets, not only total activity.
- Research questions: It also asks whether behavioral differences arise across the population or from changes among elite users.This distinguishes broad behavioral change from an artifact of elite-user activity.
- Research approach: Using eight 2012 U.S. political events and approximately 200,000 politically engaged users, the study compares media events with news events and typical time.The design examines rates and concentration of communication patterns, connectivity, and responsiveness.
Background
The background distinguishes ceremonial, scheduled media events from uncertain breaking-news events and asks whether dual screening changes behavior across politically active Twitter users. It situates the question within research on collective attention and information spreading.
- Media events: Media events are scheduled, ceremonial occasions that interrupt routines and invite large audiences into shared, real-time viewing.Their defining features include advance planning, media prominence, and collective experience.
- News events: News events can be large and covered live, but their unplanned nature leaves them without the same ceremonial and enthralled-audience features.Disasters and scandals are presented as examples where information and implications are uncertain.
- Shared attention: Social media back channels support rapid information sharing, broad participation, and collective sensemaking during major events.These processes occur as audiences respond to narratives across television and mobile-device screens.
- Dual screening: Dual screening connects political information consumption with social-media activity during live debates.The passage reports that 11% of live debate watchers used social media while watching television coverage.
- Related work: Prior research links collective attention and information propagation to finite attention, novelty, early reactions, forgetting, and retweeting.The paper notes that retweets label content and signal community membership.
- Research gap: The paper addresses whether dual screening produces behavioral changes across a politically active Twitter population and how those changes appear across population levels.It notes that prior work often assumes steady-state operation rather than exogenous shocks affecting many users.
Media event-driven behavioral change
The paper theorizes that shared attention during media events can displace fragmented everyday contexts and reorganize collective Twitter behavior. It contrasts this possibility with the alternative that activity simply increases while elite concentration remains unchanged.
- Competing explanations: Media events might raise activity volumes without changing the underlying rates or distribution of mentions, hashtags, and retweets.This is the paper’s baseline alternative to behavioral change.
- Theory: The paper argues that media-event contexts interact with dual screening to produce media event-driven behavioral change.The proposed change concerns both social context and repertoires of technology use.
- Shared attention: Shared attention creates a common focal activity that displaces fragmented social contexts and organizes new social relations.Without such events, users’ activities remain distributed across places, groups, and interests.
- Shared attention: The theory describes media events as shifting users from relatively disordered behavior toward a highly ordered state centered on one shared activity.Users become reflexively aware that others are exposed to the same information.
- Liminal space: Social media can create a liminal space where shared interpretation and improvisation around televised events alter prevailing communication structures.The paper uses debate references such as “big bird,” “binders,” and “bayonets” as examples.
- Predictions: The paper expects shared attention to change retweet, mention, and hashtag behavior and asks whether increased attention benefits elites or spreads across users.Its competing possibilities are rising stars versus broader participation by less frequently amplified users.
Research design
The study treats eight 2012 U.S. political events as natural experiments, comparing event peaks with typical pre-debate periods in a panel of politically engaged Twitter users. It measures peak activity and relational network structure across users, hashtags, and tweets.
- Events: The dataset contains eight events from approximately six weeks in late August through mid-October 2012, including conventions, debates, and news events.The events span different expected levels of shared attention.
- Baseline: The design compares event peaks with activity peaks four days before each debate during periods without major media or news events.The pre-event windows reduce systematic bias from weekday variation and provide a conservative baseline.
- Sample: The researchers track a defined panel of politically engaged Twitter users rather than randomly sampling the Twitter garden hose.This computational focus group supports longitudinal comparison but limits generalizability.
- Measurement windows: For each event, analyses focus on the one-hour window containing the peak cumulative activity within a 48–96-hour event window.The study summarizes users, tweets, retweets, mentions, and hashtags observed at peak activity.
- Validation: The political relevance ratio measures the fraction of peak-time tweets containing candidate names, handles, or event terms.It validates distinctions among the political events.
- Network measures: The study constructs networks from hashtags, mentions, replies, and retweets to measure centrality distributions for users, hashtags, and tweets.Retweet user in-degree aggregates how often all of a user’s tweets were retweeted.
Results
During media events, Twitter users became more active, but communication shifted toward topical broadcasting, retweeting, and concentrated attention around a small number of users, hashtags, and tweets.
- Changes in communication: Tweet volumes during debates were three to four times greater than during other event types, while hashtag use nearly doubled.
- Changes in communication: A 40% decline in mentions and replies made communication less interpersonal and more declarative during media events.
- Changes in communication: Retweeting increased 20% during conventions and debates compared with typical and news events.
- Changes in concentration: Media events increased concentration in hashtags, mentions, replies, and retweets rather than distributing attention across more users.
- Changes in concentration: The top 25% of users’ tweets accounted for approximately 75% of all retweet activity during media events.
- Changes in concentration: Users became more active overall, but the additional activity predominantly benefited a handful of users, hashtags, and tweets.
Discussion
Media events increase activity, but the added participation primarily concentrates attention on elite users, hashtags, and tweets rather than broadly changing user behavior. The study is limited by its small, politically focused event sample and potentially debate-biased sampling.
- The findings support a theory that shared attention and intense activity displace normal social foci, producing more highly ordered behavioral states.
- Media events made users more active, but the additional activity primarily benefited a handful of users, hashtags, and tweets.This pattern supports a concentration of collective attention rather than an evenly distributed increase across the system.
- References to hashtags, user mentions, and retweets became significantly more centralized during media events without correspondingly large changes in average user behavior.
- The newfound attention was concentrated among users with the largest audiences rather than distributed across users with different follower counts.
- The analysis is limited by eight politically focused events over six weeks, potentially debate-biased sampling, and aggregated traces that omit tweet content and user motivations.
Conclusions
The paper explains why media-event bursts differ from other bursts by analyzing both activity levels and network structure. Its findings show that increased attention is driven more by elite “rising stars” than by broadly distributed “rising tides,” with implications for collective sensemaking.
- The paper develops a theory explaining why some bursts of social-media activity generate different behavioral patterns than others.
- Analyzing activity levels and network structures reveals that media-event activity is driven more by elite “rising stars” than broadly distributed “rising tides.”
- These findings have implications for collective action under uncertainty and suggest ways to mine social media for collective sensemaking after major events.
- Twitter can bring otherwise segmented audiences together in a third space to participate in consequential events.