Scientific Twitter: The flow of paleontological communication across a topic network

Abstract

The field of paleontology, which is based principally on observations of the natural world, includes an active community that is engaged across multiple social media platforms, consisting of museums, academic researchers, amateur fossil collectors, paleontological artists, and commercial fossil dealers. As such, it represents an ideal environment for examining the people, interactions, and flow of scientific information. Using interactions involving the four most popular Twitter hashtags for paleontology, this embedded mixed methods study defined the members of this social world and investigated how they influenced and controlled the flow of information, as well as how their expression of scientific practice was related to their identity. Results provide further evidence for the diversity of people and practice involved in this domain of science and indicate that the magnitude and breadth of the public's impact may be larger than previously projected. Certain types of messages were shown to be effective for different segments of the community, but news posts, essentially media outlet stories, were ineffective for generating any form of engagement. This study adds to our understanding of the important scientific contribution being made by members of the public as they interact with professional scientists and educators as peers in an open social media platform that supports a diverse and active community.

Fig 1. Quantities of members by classifications.

Members are depicted in this diagram at the Categorical and Type levels of the PIT as members can be classified into such tiers regardless of Structure. At the Categorical level, Public made up the majority (61.6%, n = 2,199), followed by Scientists (24.4%, n = 872), Education and Outreach (12.3%, n = 440), and Commercial (1.6%, n = 58). At the Type level, Public-Interested Party made up the majority (53.6%, n = 1,915) whereas Scientist-Paleontologists made up just over 10 percent (11.1% n = 396).

Fig 2. Network map following cluster analysis.

Clusters are partitioned and labeled as groups (e.g., G1, G2, etc.). Members within groups are indicated by nodes, which are proportional in size to their degree of control (i.e. betweenness) and colorized by Category. Interactions are indicated as grey lines between nodes.

Fig 3. The flow of messages across groups within the social world.

Members within groups are proportional in size to their degree of control (i.e. betweenness) and colorized by Category (Fig 2). New messages (i.e. tweets) are colorized by type.

Fig 4. The flow of messages between groups by type.

The link between the nodes represents the flow of messages between groups calculated using InfoMap. The node sizes are proportional to the number of members within each group.