Srs Bsns: How A Recent Pew Research Study Provides Evidence Of Twitter’s Bubble

The concept of online echo chambers, or filter bubbles, is a well-known phenomenon on the internet, particularly when it comes to social media. Twitter is no different than the rest when it comes to being susceptible to this, but a recent study gives us some startling insight into just how much of a problem this occurrence is on the platform, especially in terms of politics.

In mid-October, the Pew Research Center put out a poll, titled “Differences in How Democrats and Republicans Behave on Twitter,” conducted between November 2019 and September 2020. Their findings proved that a small minority of users on the platform created the vast majority of tweets from U.S. adults, with over two-thirds of these identifying as Democrats. Before we dive too deeply into this, let’s first explain what exactly a “bubble” is, and why this supports the concept.

The Twitter Bubble:"Just 10% of [Twitter] users produced 92% of all tweets from U.S. adults since last November, and 69% of these highly prolific users identify as Democrats or Democratic-leaning independents."https://t.co/ZWnzdzK51D

— Glenn Greenwald (@ggreenwald) October 23, 2020

Usually the result of web service algorithms that personalize search results and newsfeeds tailored to what the user likes to see, a filter bubble is the phenomenon that traps users into online echo chambers filled with like-minded individuals who share their beliefs. These social media bubbles are typically based on the user's web history, geolocation, likes, interests and other personal information. While this might not be a problem in every case, bubbles have been largely criticized over the years for preventing users from being exposed to opposing viewpoints and creating online echo chambers, which can even result in instances of political extremism.

As we enter the peak of America’s 2020 presidential election season, anyone who’s spent time online lately knows that political discussions have reached an all-time high. That being said, what insight can we glean from this new Pew’s study, and what does it tell us about the pervasiveness of Twitter’s bubble — especially in the context of what many are denoting as the “most important election of our time.”

On the whole, the poll’s findings show that Twitter behaviors of U.S. adults largely depend on how they identify their political affiliation. According to the analysis, while most Americans post only rarely on Twitter, a small percentage of highly active users, mostly Democrats, produce a vast majority of tweets: just 10 percent of users produced a staggering 92 percent of all tweets from U.S. adults since last November, and 69 percent of these highly prolific users identify as Democrats or Democratic-leaning independents.

(Small share of highly active Twitter users produce the bulk of tweets from U.S. adults.)

To explain this phenomenon and the factors that contribute to it, the Center stated that Twitter contains a larger number of Democrats than Republicans, but also that the 10 percent most active Democrats typically produce almost twice as many tweets in a month as the 10 percent most active Republicans (157 vs. 79).

The differences between both parties who use Twitter and those who don’t are also notable. Republican Twitter users and non-users, 60 percent of the former and 62 percent of the latter, both described themselves as “very or somewhat conservative,” with 60 percent of Democrats on the platform describing themselves as liberal-leaning (24 percent as “very liberal”), compared to non-users at 43 percent (and 12 percent “very liberal”).

(Democrats on Twitter more likely to identify as liberal than Democrats who are not Twitter users.)

Aside from purely the amount they tweet, both Democrats and Republicans on Twitter are also much more likely to follow accounts from one particular party than the other, such as Barack Obama with 42 percent of Democrats but 12 percent of Republicans, and Donald Trump with 35 percent of Republicans but 13 percent of Democrats. The only major crossovers between the two political groups were a small number of the most-followed accounts on Twitter, mostly popular celebrities or entertainers, with similar shares of U.S. adults belonging to each.

The study also found that Republicans and Democrats are both more likely to follow members of their own party, as well as candidates, journalists or other Twitter accounts who share their ideologies and affiliations — meaning they don’t often get opposing views or information.

Another interesting discovery showed that ages and education differed between the two, as well as non-Twitter users. Pew stated that Democrats and Republicans on the site, including independents, tended to be younger and more highly educated than Democrats or Republicans who didn’t use Twitter.

Perhaps the most interesting bit of this study is that, regardless of part affiliation, the vast majority of all Twitter users tweet rarely – about once per month on average – but a small segment of the site produces the bulk of all tweets. Between Dems and Reps, Democrats are certainly the more active users on the platform, posting more on average than their counterpart. They also have more followers on average (and the amount they follow themselves), but the Center also stated that most adults on Twitter had relatively few followers across the board.

(Twitter behavior dominated by a small group of highly active U.S. adults.)

While the average Twitter user is relatively inactive, posting infrequently with fewer followers, the top 10 percent of the most active tweeters produced a staggering 92 percent of all tweets from U.S. adults, and 69 percent of those are Democrats, with just 26 percent Republicans. Of these most prolific tweeters, those who identified themselves as Dems produced nearly twice as many tweets per month as their top Republican counterparts (157 vs. 79 on average). Additionally, between November 2019 and September 2020 when the study was conducted, the average Democrat in this group tweeted over 1,600 times compared to around 800 for the median Republican.

Taking into account the various findings of this whole study, it’s easy to see how the existence of echo chambers or filter bubbles is quite prevalent among both parties or their relevant ideologies. While this information is especially important and insightful leading up to this year’s heated election, it’s also not exactly new.

Back in 2017, the Proceedings of the National Academy of Sciences (PNAS) published a study on this phenomenon and analyzed over half a million tweets to prove that “morally outraged” tweets tended to be widely shared and engaged with by members of their own political spheres, but rarely broke outside those bubbles. So while Republicans or Democrats expressed their emotionally charged opinions on issues, these takes rarely fell on ears that weren’t already the same as theirs.

(Network graph of moral contagion shaded by political ideology, showing how tweets from either side rarely reach the other.)

Despite this information simply reinforcing the existence of social media bubbles and echo chambers, the Center’s study offers the most current insight into this internet phenomenon as we head into the final days of the U.S. presidential election. Twitter itself has attempted to remedy some of this by rolling out new changes to the platform in recent years, such as its “topics” feature in 2019, but other studies have also shown that bursting these bubbles can sometimes have a negative impact.

In early 2018, a team from Duke University, New York University and Princeton got over a thousand Republicans and Democrats on Twitter to follow bots that retweeted posts from their opposing party’s politicians, pundits and journalists. What they discovered by polling the participants before and after the study was that exposure to conflicting ideas or views actually made people more entrenched in their existing beliefs.

The study is just one example of this and only a relatively small amount of users, but still, perhaps it supports the idea that social media is simply not the right medium for bringing people of different political ideologies together in the first place.