Table 1.
Content analysis coding categories for messages from Boston Marathon Bombing.
Table 2.
GLM negative binomial model using source, style and theme variables predicting number of per-tweet retweets during the Boston Marathon Bombing.
Table 3.
GLM negative binomial fixed effects predicting number of per-tweet retweets during the Boston Marathon Bombing.
Fig 1.
Thematic content of official messages is strongly related to expected retweet rates.
Solid bars indicate the estimated retweet rate multiplier associated with the presence of each content type; error bars depict 95% confidence intervals. Tweets containing content related to advisories, hazard impact, or emotional, judgmental, or evaluative statements were on average retweeted at 2–3.5 times the rate of messages without such content. “Thank you” messages, by contrast, were retweeted at just under 50% the base rate.