Description

The rapid and wide dissemination of disaster information through tweeting and retweeting has made Twitter an important communication channel during disasters. However, due to its 140-character limit, a tweet could be considered uninformative during disasters. Based on Shannon and Weaver’s communication theory, we use the measure of entropy to quantify the extent to which a tweet is considered informative. We theorize that as a tweet’s entropy increases, its informativeness decreases, and importantly, so too does the probability of retweeting. To assess tweets’ entropy, we use topic modeling to discover topics in tweets. Using tweets collected during the 2013 Colorado floods, we empirically examine the relationship between tweets’ entropy and retweet frequency. We take this investigation one step further by examining the interaction effect of the number of URLs on that relationship. Our empirical results demonstrate the negative effect of entropy on retweet frequency. In addition, the effect of the number of URLs depends on entropy. Our findings suggest that tweets’ entropy is an important factor in explaining tweets’ retweeting mechanism during disasters and enhance the understanding of the relationship between short-length tweets and information dissemination. As a result, our study contributes to IS research on the role of Twitter in emergency communication.

Share

COinS
 
Aug 10th, 12:00 AM

How Do Multiple Topics in Terse Tweets Affect Retweeting? Evidence from the 2013 Colorado Floods

The rapid and wide dissemination of disaster information through tweeting and retweeting has made Twitter an important communication channel during disasters. However, due to its 140-character limit, a tweet could be considered uninformative during disasters. Based on Shannon and Weaver’s communication theory, we use the measure of entropy to quantify the extent to which a tweet is considered informative. We theorize that as a tweet’s entropy increases, its informativeness decreases, and importantly, so too does the probability of retweeting. To assess tweets’ entropy, we use topic modeling to discover topics in tweets. Using tweets collected during the 2013 Colorado floods, we empirically examine the relationship between tweets’ entropy and retweet frequency. We take this investigation one step further by examining the interaction effect of the number of URLs on that relationship. Our empirical results demonstrate the negative effect of entropy on retweet frequency. In addition, the effect of the number of URLs depends on entropy. Our findings suggest that tweets’ entropy is an important factor in explaining tweets’ retweeting mechanism during disasters and enhance the understanding of the relationship between short-length tweets and information dissemination. As a result, our study contributes to IS research on the role of Twitter in emergency communication.