Location: PSH (Professor Stuart Hall Building) - 326,
Goldsmiths, University of London, Building 2
Campus Map Contributor: Kelli Burns, University of South Florida, United States
Background: Perhaps no other campaign has reached greater success in terms of participation, donations, social media chatter, and attention in popular culture than the ALS Ice Bucket Challenge. This campaign was truly a grassroots effort, fueled by friends of an ALS victim and then spread on social media by millions of participants ranging from average citizens to celebrities. Significant coverage by traditional media and participation by personalities on television further accelerated awareness of the campaign and the cause. On Twitter, many celebrities also shared their involvement in the campaign and spurred conversation and engagement among other Twitter users.
Objective: This study sought to understand the impact of influencers in spreading discussion of the ALS Ice Bucket Challenge campaign on Twitter. This study examined the top accounts for retweets and mentions; the correlation between the most followed, most retweeted, and most mentioned users; and the network structure of the conversation. Finally, the conversation was examined to see whether it was "on message."
Methods: A historical data grant from Texifter provided access to almost 550,000 tweets and enterprise access to DiscoverText, a “cloud-based, collaborative text analytics solution.” Tweets were selected based on the criteria of inclusion of the hashtags of #alsicebucketchallenge or #icebucketchallenge and use of the English language during the period of August 18-22, 2014. Of the millions of English-language tweets during this timeframe using these hashtags, 15 percent of the tweets were randomly selected for inclusion in the sample, resulting in a sample size of 545,563 tweets.
Results: Analysis of the more than 500,000 tweets determined the top influencers in terms of mentions and retweets and also explored correlation between the most followed users, the most retweeted, and the most mentioned, finding a moderate correlation between the most retweeted and mentioned. T2G 0.3 for Python was used to extract all edges from tweets with multiple mentions from a sample of 1,000 tweets, resulting in 1,457 edges to be analyzed in NodeXL. Plotting the data using the Fruchterman-Reingold algorithm resulted in large nodes (representing the vertex in-degree metric) for certain celebrities and close relationships for others. The vertices were grouped by cluster using the Clauset-Newman-Moore algorithm and further examined. This study also modeled a larger sample of the dataset using Gephi and D3. Finally, although ALS was mentioned in a larger percentage of tweets, variations of the word donate were mentioned much less frequently.
Future Work:
The implications of this study would be relevant to other activist Twitter campaigns that mobilize celebrity influencers.
