Week 7: Echo Chamber Index

Hello everyone, and welcome back to my blog! This week, I finally combined the values I have been calculating over the past three weeks: network isolation, topic homogeneity, and sentiment analysis. Using different weight values for each aspect, I built scores for each community on both X and Reddit

How I chose the weights

Out of the three parts, I chose to weight network isolation the most with a value of 0.45. This is because echo chambers are primarily created by people only interacting with the same group of people. If people are mainly talking in one small group and rarely see outside perspectives, then an echo chamber can exist there. 

The second component, topic homogeneity, has a weight of 0.35. This is mainly to see if everyone in a community is talking about the same ideas. Even in a very connected network, there can only be an echo chamber if the same ideas are constantly being mentioned. This makes topic homogeneity important for finding echo chambers, but not as important as network isolation.

Lastly, sentiment is more of a supporting aspect. This measures if everyone feels the same way about the same topics. Even though sentiment can make an echo chamber stronger, it doesn’t prove that there is an echo chamber on its own. Echo chambers can exist even without everyone feeling the same way about the topic. 

X Communities

Political Campaigns and Elections: 0.8616
International Relations and Strategy: 0.7896
National Security and Defense: 0.6650
Education and Teaching: 0.7301
Research and Scientific Development: 0.7007
Economic Policy and Finance: 0.7209
Social Issues and Race Relations: 0.6572
Media, Communication, and Public Discourse: 0.7351

X shows consistently high echo chamber index scores, with all scores being between about 0.65 and 0.86. The strongest community overall is Political Campaigns and Elections, which means it has the highest probability of really being an echo chamber. As I analyzed my results, this outcome made the most sense, since political content tends to be the most polarizing subject that people discuss. . 

What is also interesting is that less political categories, like Education and Teaching, still showed strong echo chamber values. This could mean that X as a platform could reinforce clustering, which could lead to more communities acting like echo chambers. 

Reddit Communities

Private Prisons: 0.4917
Tax Policy: 0.5273
Healthcare Policy: 0.4085
Donald Trump Policies: 0.4722
Economic Inequality: 0.4068
Iran Sanctions: 0.4143
Religious Freedom: 0.4040
Middle East Policy: 0.3917
Abortion Access: 0.3948

In contrast, Reddit communities has much lower echo chamber scores, ranging from about 0.39 to 0.53. Even the highest scoring topic, Tax Policy, is still much lower than the top communities on X. Most communities had scores around the low 0.40 range, meaning that these communities were rather weaker echo chambers overall.

Reddit discussions could be less isolated and more open to varied perspectives. Even for very polarizing topics like Abortion Access or Middle East Policy, the scores were still relatively low. One reason for this could be Reddit’s structure, where users can talk in shared subreddits rather than personal user networks, allowing for more interaction between various viewpoints. 

Overall, there is a contrast between both platforms. X has consistently strong echo chamber communities, while Reddit seems to be more moderate. This shows the importance of how the way a platform is built can influence how information is spread. 

In the following weeks, I will work to evaluate my echo chamber index values through sensitivity analysis, where I will make small changes to values in the index to see how much of a difference they make. After, I will finally begin to decide what threshold for the index I will use to identify if a community really is an echo chamber or not.

Thank you for reading, and I will see you all next week!

Harish