The Political Attitudes of Social Media Users
We examine the political identities, extremity, and polarization of users of specific social media platforms.
In the past 2 months, we’ve introduced the Neely Social Media Index and reported our findings for social media experiences that users perceive as bad for the world, experiences that negatively affect users personally, which platforms people use to learn useful and important information, and where people experience the most meaningful connections with others online. This week, we pivot to the political attitudes of users of social media platforms.
The majority of Americans use at least one form of social media, which makes it a valuable tool for political candidates and groups to reach potential donors, supporters, and voters. Political actors’ reliance on social media has grown increasingly with each passing election, and effective social media activity can boost donations, garner incredible free media coverage, and very precisely-catered ads to small groups of voters who are likely to be more receptive to certain appeals. Given this increased role of social media in elections, it is not surprising that politicians fear bias against their beliefs, and that the majority of Americans, regardless of their political leanings, believe that social media censors congenial political viewpoints and is biased in favor of uncongenial viewpoints.
Beyond the left-right political arguments on and/or about social media’s potential biases, numerous academics, civil society organizations, and journalists raise concerns about some of the potential effects social media may have on democratic societies. One concern often raised is whether social media divides and polarizes societies. Some have argued that newer technologies create filter bubbles, where people increasingly see news that confirms their preferred beliefs about the world and shields them from seeing news that might challenge their preferences. Others have argued that social media creates echo chambers that allow people to interact with ideologically homogeneous others, which can fuel hostility towards and misperceptions of people who hold different political beliefs, and increase polarization. Recent studies funded by Meta, the parent company of Facebook, Instagram, and WhatsApp, found that short-term experimental algorithm changes caused only very small increases in polarization, or did not cause detectable increases in polarization, even as they did lead to increases in the distribution of what many would consider polarizing content. Reviews of the broader evidence suggest that platform effects on polarization are a longer term phenomenon that affects publisher incentives in ways that may not be detectable in short term studies that change individual user experiences. Debates continue as to social media’s role in political discourse.
Therefore, today’s post seeks to add some context to these political elements of social media. We ask three main research questions here:
What are Americans’ political leanings?
How do platforms differ in how liberal, moderate, conservative, Democratic, Republican, and extreme their users are?
Which platforms’ users are most and least polarized? Which platforms’ users are most tolerant and intolerant of politically similar and dissimilar others?
Americans’ Political Leanings & Social Media Preferences
To understand the politics of different social media platforms, we first need to understand the politics of Americans. Historically, especially since 1991, more Americans identify with the Democratic Party than with the Republican Party. The current sample reflects this pattern, whereby 29.2% of Americans are registered Democrats and 26.9% of Americans are registered Republicans.
Many registered voters who are not members of a political party do tend to favor one party over the other. Despite some regional variability, these independent voters who lean towards one party or another tend to be evenly split between the Democratic and Republican parties.The current sample also reflects this pattern observed in other nationally representative sample, where roughly 4% of unaffiliated voters lean towards the Democratic Party and another 4% of unaffiliated voters lean towards the Republican Party. The overall estimated number of Independent voters is lower than some external estimates, but this may partially reflect the fact that we are in the midst of a presidential primary campaign and that the number of people identifying as independent tends to decrease in election years.
Party affiliation and leaning increasingly shed light on how liberal, moderate, or conservative Americans may be, but there is substantial variability in the political orientations of people in the major parties. To capture people’s political orientations more directly, political scientists ask people where they place themselves on a continuous spectrum from “Very Liberal” to “Very Conservative.” This approach takes an extraordinarily complex network of beliefs and simplifies them into a single dimension, but admittedly brushes over some important differences. Nonetheless, this is predictive of party affiliation, attitudes, and voting. Historically, more people identify as conservative than liberal, but the plurality of people tend to be moderate. This pattern holds true in the current sample, too.
Does social media use differ by people’s politics?
In our survey, we presented respondents with a list of 19 popular social media and communication services, ranging from Facebook and TikTok to email and text messaging, and asked them to indicate which they used. In the full sample, 50% of people use at least 6 different social media and communication platforms. Democrats and liberals tend to use about 1 more social media platform than do Republicans, Independents, conservatives, and moderates. These results are consistent with past surveys showing that Democrats and political liberals tend to use more social media platforms than Republicans and conservatives (and, this survey from Pew, too).
It is also important to consider frequency of use. If Democrats and liberals use a wider breadth of social media than Republicans and conservatives do, the implications would be very different if one side uses those platforms less frequently than the other. In the plots below, we actually see that Independents and moderates are generally more likely to report using social media and communication services multiple times per day than are their partisan and more ideological counterparts. There are some exceptions to this pattern, where Democrats and liberals tend to be more likely to use Reddit and Discord more than Independents and moderates, and both Democrats/liberals and Republicans/conservatives are more likely to use Twitter multiple times per day than are independents and moderates.
The fact that Democrats and liberals use about 1 more platform, on average, than Republicans and conservatives do, may lead to a relatively higher percentage of users on social media platforms being Democratic or liberal than one would predict if platforms were a perfect reflection of the political breakdown observed in America. This disparity may be most pronounced on Reddit and Discord, due to more frequent use of those platforms by left-leaning Americans. Next, I compare the political composition of the social media platforms included in our survey.
Social Media Platforms’ Users’ Politics
When looking at the political affiliations of users of each of the platforms, we see that the only platform with more Republican than Democratic users is Facebook, and this numerical Republican advantage on Facebook is slight -- just 1.5%, which is half of the margin of error (3.02%). Twelve (75%) of the sixteen most popular social media and communication platforms had more Democrats than Republicans. Just 3 (18.75%) of the 16 platforms had the same percentage of Democrats and Republicans. The two platforms with the highest percentage of Democratic users, Discord and Reddit, have more than twice as many Democratic as Republican users (54% vs. 23%, and 50% vs. 23%, respectively).
Additionally, most platforms have more Democratic and more Republican users than would be expected if the user bases were reflections of the national estimates of those parties. The only service where Democrats were less represented than expected based on the national average is online gaming, and the only platforms where Republicans were less represented than expected based on the national average are Discord and Reddit.
The mostly higher-than-average percentage of Democrats and Republicans on social media platforms might imply social media is more polarized than the United States, overall. Yet, party affiliation is crude and fails to differentiate extremity from political leanings. Therefore, looking at the political orientation of users by platforms should be informative of how polarized users may be.
Even though there are more Democratic users on most platforms, there are actually more conservatives than liberals on most platforms. This is due to more Democrats being more moderate in the political orientation than Republicans are. Again, Facebook has the highest percentage of conservative (33%) vs. liberal (24%) users, and WhatsApp also has a relatively large percentage of conservative (36%) vs. liberal (29%) users. On the other extreme, Discord and Reddit had the largest percentage difference in the liberal (43% and 44%, respectively) vs. conservative (15% and 16%, respectively) users.
The plurality of users on most platforms is politically moderate, with people using online gaming being the most politically moderate. Just 2 (12.5%; Twitter & Reddit) of the 16 platforms have more liberal than moderate users, and just 1 (6.25%; WhatsApp) platform has more conservative than moderate users. This moderacy is reflective of the national average.
Another way to examine political extremism is by looking at how far from the midpoint of the 101-point political orientation scale, and estimating the number of people who are within 15 points of either the “Extremely Liberal” or “Extremely Conservative” scale anchors. In the plot below, I show the percentage of users on each platform who selected these extreme political orientations. NextDoor has the fewest politically extreme users (24.4%), with Email, Facebook, Text Messaging, Online Gaming, and YouTube also having fewer extreme users than the national average (27.68%). All other platforms have more politically extreme users than expected if the platforms reflected the national average. Reddit and Discord have the highest percentage of politically extreme users.
Together, users of social media platforms tend to be more Democratic and moderate than Republican and ideological. Importantly, though, our conclusions are limited to platforms we included in our survey. Therefore, we do not have apples to apples comparison data for alternative social media platforms like Truth Social, Gab, or Parler. A recent survey from Pew shows that 66% percent of users on these sites are Republican, or tend to vote for Republicans, suggesting that social media, as a whole, isn’t necessarily more appealing to one party or ideology over another.
In terms of extremism, 63% (10 out of 16) of platforms have more politically extreme users than the United States overall. These observations tell us about the identities of the users of these platforms, but do not speak directly towards attitudes toward and tolerance of people with similar vs. dissimilar politics. Additionally, these results cannot determine whether platforms are causing changes in political attitudes and extremism. In the next section, I examine the feelings users on different platforms have towards liberals and conservatives, and how comfortable these users are being friends with liberals and conservatives.
Social Media & Political Polarization
Political polarization is measured in many different ways. Some measure it by taking the difference in attitudes towards specific policies and infer polarization based on the size of that difference. Polarization, measured this way, is typically not very great, as there tends to be more agreement than disagreement on policy issues (except on certain hot-button issues, like abortion). Another type of polarization, often referred to as affective polarization, is based on how people feel about politically similar others relative to politically dissimilar others. This type of polarization is significantly higher than issue-based polarization, and has been growing in recent decades in the United States. Additionally, greater affective polarization corresponds with greater prejudice against the political outgroup.
Before looking at platform-specific feelings towards liberals, let’s look at feelings towards liberals in the general US population. 42.7% of US adults are favorable and 31.5% of US adults are neither favorable nor unfavorable towards liberals. Just 23.7% of the US adult population holds negative feelings towards liberals.
When we look at favorability by users of the 16 most popular platforms included in our survey, we see that user bases on all platforms are more favorable than unfavorable towards liberals. The majority of users on 5 (31.25%) of the 16 platforms are favorable towards liberals, and no platforms have a majority of users who are unfavorable towards liberals. No platform had even a quarter of its users expressing they were unfavorable towards liberals. Reddit, Whatsapp, and NextDoor had the lowest percentages of users expressing unfavorability towards liberals. This demonstrates people have relatively positive, or at least neutral, feelings towards liberals on social media.
But, what are their feelings towards conservatives?
Again, we’ll review the estimated favorability towards conservatives in the general US adult population. Here, we see that the plurality of US adults are favorable towards conservatives, and that this pattern is similar to favorability towards liberals. Specifically, American adults are more favorable towards liberals and conservatives than unfavorable towards them.
When we look at favorability by users of the 16 most popular platforms included in our survey, we see that all platforms are more mixed in their feelings towards conservatives. No platform has a majority of users that are favorable, neutral, or unfavorable towards conservatives.Yet, 62.5% (10/16) of platforms have users that are more favorable towards conservatives than observed in the general population. Discord, Reddit, and Twitter users were more likely to be unfavorable towards conservatives. This demonstrates people have relatively positive, or at least neutral, feelings towards conservatives on social media, though these attitudes are more mixed than the attitudes towards liberals across platforms.
Another method sometimes used to assess prejudice against social groups relies on asking people how comfortable they are being friends with members of a social group. If people are less comfortable being friends with people in one social group than another, then they harbor a prejudice against people in the disfavored group. The plurality (44.8%) of US adults state that they are equally comfortable being friends with liberals and conservatives. Slightly more US adults are more comfortable being friends with liberals (29.8%) than with conservatives (23.3%).
The general pattern applies to most social media platforms, where the plurality of users on all platforms besides Reddit, Twitter, and Discord, report being equally comfortable being friends with liberals and conservatives. Reddit, Twitter, and Discord have a higher percentage of users who are more comfortable being friends with liberals than conservatives.
In Summary
While there are many debates about political biases of platforms, our data can only speak to the political affiliations, attitudes, and prejudices of users on these platforms. If platforms were to have an overwhelming majority of users subscribing to one ideology or belonging to one party, it may make users of the minority party feel uncomfortable on a given platform. However, no platforms exhibited an overwhelming majority of liberals or conservatives, Democrats or Republicans. Certainly, most platforms (except Facebook) have user bases that are somewhat more liberal than conservative, but none have user bases with a large majority adhering to a particular ideology or party. Reddit and Discord’s user bases each stood out as substantially more liberal than conservative, more unfavorable towards conservatives than liberals, and more politically extreme than the general US adult population.
In summary, most platforms have more liberal and Democratic users than conservative and Republican users. Additionally, a larger number of platforms exhibit less prejudice towards liberals than towards conservatives. These findings could explain why some conservatives feel that social media platforms are biased against them, as they may be more likely to encounter liberals and anti-conservative prejudice than their liberal counterparts. While possible, data from some of the social media platforms that conservative content gets more views and is amplified by ranking algorithms, would imply it is less likely that conservatives are encountering too many liberals or liberal ideas.
It is entirely possible that social media platforms themselves (as opposed to their users) are biased against (or towards) certain political groups, but our survey of users cannot speak to biases baked into recommendation algorithms that decide what content to include in people’s feeds. Our longitudinal approach combined with occasional deep dives will provide greater insight into whether any of these platforms are causing increased polarization or extremism, or if it is merely the case that some platforms seem to have more polarized and extreme users than others.
Note: This survey was conducted prior to Twitter changing its name to X, and before Threads was launched. In subsequent waves of this survey, we have updated the name Twitter to X, and are including Threads as a new platform.
For additional analyses and plots, check out this Google doc.
So y'all are always gonna be fine with conservative self-reports where they never call themselves extreme, they just think women who have abortions should be thrown in the same dungeon as Fauci for making them experience the tyranny of a mask, huh?