Using impression data to improve models of online social influence

Rui Liu,Kevin T Greene,Mihovil Mandic,Ruibo Liu,Soroush Vosoughi,V S Subrahmanian,Benjamin A Valentino

doi:10.1038/s41598-021-96021-3

Abstract

Influence, the ability to change the beliefs and behaviors of others, is the main currency on social media. Extant studies of influence on social media, however, are limited by publicly available data that record expressions (active engagement of users with content, such as likes and comments), but neglect impressions (exposure to content, such as views) and lack “ground truth” measures of influence. To overcome these limitations, we implemented a social media simulation using an original, web-based micro-blogging platform. We propose three influence models, leveraging expressions and impressions to create a more complete picture of social influence. We demonstrate that impressions are much more important drivers of influence than expressions, and our models accurately identify the most influential accounts in our simulation. Impressions data also allow us to better understand important social media dynamics, including the emergence of small numbers of influential accounts and the formation of opinion echo chambers.

Highlights

Influence, the ability to change the beliefs and behaviors of others, is the main currency on social media
Our regressions show that when the impressions of a particular account on the platform increase by one standard deviation, the expected number of times users cite the account as influential more than doubles. These results suggest that the single biggest determinant of influence on social media platforms is impressions, a factor that has hardly been studied in the growing online social influence (OSI) literature
To measure how an observer was influenced to nominate a given user we developed two kinds of networks that “traceback” the activities of observers, direct influence networks, and full influence networks

Summary

Introduction

The ability to change the beliefs and behaviors of others, is the main currency on social media. This has led to an explosion of research on online social influence from computer science6–9, physics[10,11], and the social sciences[12,13] Almost all these studies have important limitations: (1) due to their reliance on publicly available data from social media platforms, extant models of influence rely exclusively on expression data, that is, data about the active engagement of users with content. In addition to commonly used expression data, the software platform captured impression data, which consists of the posts that users viewed, but did not actively engage by liking, re-posting, commenting, or following Users knew they were participating in a simulation, most agreed that the experience faithfully represented social media use in the real world. Further details regarding the simulation are described in the methods section

Objectives

Methods

Results

Conclusion