Abstract

First, we would like to thank this reviewer [1] for their insightful comments on our paper [2]. Although endogeneity may be an issue of concern for these types of framings, our methodology builds on numerous studies (now cited in the revised paper) that—after controlling for many correlated variables—show how the emergence of online echo chambers is partly due to contagion dynamics, partly due to homophily, partly due to influence effects, and is not simply explained by one single mechanism (eg, political ideology alone). Nevertheless, our strategy has been proven effective to separate network structure from information spread dynamics. In the revised manuscript, we explained the various assumptions of the model, some potential limitations related to endogeneity, and referred to work illustrating the robustness of the adopted approach.
The reviewer is absolutely correct in that the real-world political ideology distribution may not match the one on Twitter. In fact, in the revised manuscript, we now refer to various studies that confirmed the same skewed online ideology distribution we observed in our study of Twitter. Since the data we observed is heavily left skewed, we used binning to facilitate comparison between left-and right-partisan users. This approach is consistent with prior work, which we now cite in the revised paper. We should note that as our study is restricted to Twitter, any insights we gleaned should only be assumed to be applicable to this platform—an important limitation that we now underscore in the revised manuscript, which, however, we do not think takes away from the importance of our work given the prominence of Twitter …

Date

August 5, 2021

Authors

Julie Jiang, Xiang Ren, Emilio Ferrara

Journal

JMIRx Med

Volume

2

Issue

3

Pages

e32266

Publisher

JMIR Publications Inc., Toronto, Canada