The 2024 U.S. Election in Social Media: Changes in How Americans Used and Experienced 17 Platforms

Experiences across platforms remained relatively stable, suggesting that platform norms and designs may matter more than the time period. Negative experiences attributed to politics increased.

Introduction

Continuing the pattern of previous elections, the 2024 U.S. election generated a great deal of social media discourse, misleading information, and polarizing content. Some key events included: the unexpected withdrawal of former President Joe Biden from the race, the viral ascent of Vice President Harris among Gen-Z users, two failed assassination attempts on President Trump, all amplified through content algorithms. While this comes as no surprise with social media’s increasing presence in politics, this election also comes as the first amidst our AI technology boom. Even more broadly, the rise of TikTok and short-form content in recent years has presented a novel challenge, allowing for rapid dissemination of political information.

Recognizing the stakes, many major social media platforms took measures in updating election-based policies in months leading up to November 5th. For instance, X (formerly Twitter) stands out as a platform for accessing political information, with 59% of users reporting that they use it to keep up with politics. X rolled back content moderation and misleading information labeling, which may have affected user experiences during the election. For example, researchers found that 74% of the 283 misleading X posts they analyzed on the 2024 election were not correctly labeled; these false claims accumulated 2.2 billion views.

Many other platforms also reduced previous restrictions on fact-checking, advertisements, etc. established during the 2020 election, citing concern of restricting political speech and influencing election results. Some experts have suggested that these decisions may have contributed to more distrust, hostility, and even political violence amongst social media users. Thus, a critical question emerges: How did Americans’ social media experiences during this U.S. election cycle change across different social media platforms?

We seek to understand how millions of users were engaging with and learning from the political content on these platforms from their own experiences. Using data from the Neely Social Media Index, a longitudinal survey of a representative sample of Americans’ online experiences, we examine how users report experiencing both harmful and beneficial outcomes across platforms in the months before and after the election.

Method

We asked respondents the following question for each platform:

1. In the past [28 days/4 weeks], which of the following online services have you used?

For each of the platforms that the respondent selects they have used, we then asked the following top-level four experience questions:

2. In the past [28 days/4 weeks], have you personally witnessed or experienced something that affected you negatively on [service used]?

   1. If Yes: What was the impact of your negative experience(s) with [services used]?

   2. If Yes: Did your experience(s) on [services used] relate to any of these topics?

3. In the past [28 days/4 weeks], have you witnessed or experienced content that you would consider bad for the world on [service used]?

   1. If Yes: What negative impact do you feel your experience(s) with [services used] could have on the world?

   2. If Yes: Did your experience(s) on [services used] relate to any of these topics?

4. In the past [28 days/4 weeks], have you learned something that was useful or that helped you understand something important on [service used]?

5. In the past [28 days/4 weeks], have you experienced a meaningful connection with others on [service used]?

Responses were weighted to construct a nationally representative sample. We assessed responses in the two most recent waves of the Neely Social Media Index survey. The survey wave conducted pre-election (February 5 - May 19, 2024) was designated as the baseline period. During the election period (October 14, 2024 - January 4, 2025), we initially randomized 50% of participants to receive the survey during the pre-election period and 50% during the post-election period. Final group assignment was based on the actual date of survey completion, categorizing participants into during- or post-election groups accordingly. Since we asked participants to recall their social media use in the past 28 days in the survey, we assigned participants who responded before Dec 3, 2024 as the “During Election” group, and who responded on or after Dec 3, 2024 as the “Post Election” group. We note that this self-reporting may not yield a clean distinction between during- and post-election periods, but because people are generally more sensitive to more recent events, our estimations of potential effects are likely 1) reflective of differences between periods and 2) conservative in that they include some amount of “noise” due to some mixing of time periods. We report the resulting prevalences with sample size below. We bolded changes that exceeded the margin of error and thus are considered statistically significant. The highlighted color green reflects statistically significant increases over time and red reflects statistically significant decreases over time.

Platform usage rises on several platforms, including Facebook, TikTok, and Reddit, post-election

Note: Neely Social Media Index nationally representative survey conducted between February 5 - May 19, 2024 (2,913 respondents) and October 14, 2024 - January 4, 2025 (2,666 respondents). We excluded platforms with less than 100 users. Reported data represents weighted prevalence of U.S. Adults who used each platform during each period as percentages. The weighted margin of error for all use data is +/-2.7%. Bold values indicate changes that exceed the margin of error and thus are considered statistically significant. Sample sizes (n) are shown for each time period.

Throughout the 2024 U.S. election period, most social media platforms observed relatively stable usage by American adults, although several platforms experienced notable changes. Some platforms saw an increase in usage: Facebook experienced the largest jump of 5.7 added percentage points from during- to post-election levels; similarly, Email and TikTok both increased by 4.4 percentage points post-election, followed by Nextdoor, Online Gaming, and Reddit at 3.9, 3.3, and 2.9 percentage points, respectively.

On the other hand, platforms that saw a decrease in user prevalence post-election include FaceTime, whose usage declined by 3.2 percentage points from during- to post-election, and Snapchat, whose usage declined by 2.8 percentage points. Pinterest observed the only significant change in U.S. adult users from our baseline wave to election wave, dropping 3.6 percentage points.

Negative personal experiences decline post-election; Politics dominates negative experiences on all platforms and increased during the election on many platforms.

Note: Neely Social Media Index nationally representative survey conducted between February 5 - May 19, 2024 (2,913 respondents) and October 14, 2024 - January 4, 2025 (2,666 respondents). % of U.S. adults on each platform witnessed or experienced something that affected them negatively during each period.

During the 2024 U.S. election period, many individuals experienced harassment and hostility related to their political identities among other negative personal experiences on social media. Research has demonstrated that these election-related attacks can disincentivize political participation, increase polarization, and ultimately undermine democratic discourse. In our sample, we observed that the majority of user-reported personal negative experiences were attributed to political reasons, with users reporting being annoyed, angry, worried, or more distrustful of people/institutions.

Several platforms showed significant post-election decreases in negative personal experience rates. Dating Apps exhibited the largest decrease from 12.9 percentage points from 22.8% during- to 9.9% post-election. Facebook rates declined 7.1 percentage points from 20.7% during- to 13.6% post-election. Snapchat rates fell completely from 3.4% to 0.0%. Threads was the only platform showing a change in negative personal experiences from baseline to election wave, rising from 3.1% to 12.4%. We do note the small sample sizes for Dating Apps, Snapchat, and Threads in these results.

Nextdoor consistently yielded the highest negative personal experience rates across all periods (20.8% baseline, 17.1% election wave). Twitter/X and Facebook also remained among the top platforms for negative experiences.

Figure 1. During the election, politics was increasingly cited as leading to negative personal experiences across all platforms, ranging from 54.8% to 81.2% in the baseline period and 60.6% to 81.9% in the election period.

Breakdown of topics of negative personal experiences reported by users on each platform. Note: Percentages are calculated by dividing the number of people reporting the topic on each platform by the number of people reporting the experience on each platform. Source: Neely Social Media Index surveys of 2,913 U.S. adults conducted Winter/Spring 2024 and 2,666 U.S. adults conducted Fall 2024/Spring 2025.

Figure 2. Annoyance, anger, and worry were the most common reactions to negative personal experiences, with only 5.0% to 14.7% of respondents reporting no effect.

Breakdown of impacts of negative personal experiences reported by users on each platform. Note: Percentages are calculated by dividing the number of people reporting the impact on each platform by the number of people reporting the experience on each platform. Source: Neely Social Media Index survey of 2,666 U.S. adults conducted Fall 2024/Spring 2025.

We chose to examine the topical breakdown of negative personal experiences as reported for six of the largest, most public-facing platforms during the 2024 election (Figure 1). We saw that across Facebook, YouTube, Instagram, TikTok, Twitter/X, and Reddit, the relative rankings of topics contributing to negative personal experiences remained largely the same in both baseline and election waves. During the baseline period, political reasons dominated negative experiences, accounting for 83.3% of reports on Twitter/X, 59.6% on Reddit, 59.4% on YouTube, 58.8% on Facebook, 56.3% on TikTok, and 20.% on Instagram. The prevalence of politics-related negative experiences decreased slightly on Twitter/X to 81.9% but increased significantly overall during the election period, reaching a heightened 72.6% on YouTube, 70.4% on Facebook, 65.8% on Reddit, 60.6% on TikTok, and 59.8% on Instagram. Across platforms, the percentage of negative experiences attributed to politics increased significantly, from 61.1% to 69.7%, with significant increases on Facebook, YouTube, and Instagram.

Twitter/X comprised the largest concentration of politics-related user-reported negative experiences, reinforcing its role as a space for charged political discourse. The rate of politics-related negative experiences increased across all platforms except Twitter/X from baseline to election periods. This could be due to a ceiling effect, as most negative experiences on Twitter were already about politics before the election.

During the election wave, among user-reported impacts of negative experiences (Figure 2), annoyance was the most common (38.0% to 66.7%), followed by anger (38.8% to 54.6%) and worry (31.0% to 49.5%). Trust-related effects were also notable, with 25.5% to 43.6% of users across platforms experiencing content they felt reduced their trust in people and 19.4% to 39.9% experiencing content they felt reduced their trust in institutions, both of which were highest on Twitter/X. More severe self-reported impacts included psychological harm (24.6% to 34.1%), feeling unsafe (6.5% to 21.7%), and self-censorship effects (6.6% to 25.9%). Only 5.0% to 14.7% of users reported “no effect,” therefore indicating that negative social media experiences yield real-world consequences. These impacts range from emotional harm to reduction of civic trust and reduced willingness to engage in discourse and self-expression.

Twitter/X and Facebook lead in political content users find “Bad for the World”. Reports dropped on several platforms post-election.

Note: Neely Social Media Index nationally representative survey conducted between February 5 - May 19, 2024 (2,913 respondents) and October 14, 2024 - January 4, 2025 (2,666 respondents). % of U.S. adults on each platform witnessing or experiencing content they consider to be bad for the world during each period.

Political misleading information, false claims regarding election integrity, and malicious rhetoric are all examples of what may be harmful, or “bad for the world” content users encountered during the 2024 election cycle. Similar to negative personal experiences, the real-world impacts of such content can be profound and range from reduced democratic trust to increased political violence.

From during- to post-election, several platforms showed significant declines in exposure to content perceived to be “bad for the world”. Facebook demonstrated the highest prevalence and a prominent decrease with a 7.5 percentage point drop from during- (26.8%) to post-election (19.4%) levels. Nextdoor followed with an 8.5 percentage point decrease (18.8% to 10.3%), followed by Threads’ 4.9 percentage point decrease (16.0% to 11.2%). Dating Apps showed an unusual pattern with a 3.1 percentage point increase from during- to post-election (8.1% to 11.2%). Discord declined 4.0 percentage points from baseline to election period, yielding the only significant wave-level change.

We found that Twitter/X maintained the highest rates of users reporting encountering content considered bad for the world across all time periods, increasing from 27.8% during the baseline wave to 33.0% during the election wave. Facebook again followed, increasing from 22.7% in user harmful content experiences to 25.0% during the election period. Conversely, more communication-focused platforms like FaceTime (1.9% to 2.2%) and LinkedIn (0.7% to 2.6%) maintained the lowest rates throughout all periods.

Figure 3. Political “bad for the world” content increased across most platforms from baseline to election period compared to other categories that observed overall decline.

Breakdown of topics of “bad for the world” content reported by users on each platform. Note: Percentages are calculated by dividing the number of people reporting the topic on each platform by the number of people reporting the experience on each platform. Source: Neely Social Media Index surveys of 2,913 U.S. adults conducted Winter/Spring 2024 and 2,666 U.S. adults conducted Fall 2024/Spring 2025.

Figure 4. Misleading content was the top concern of “bad for the world” content, with Twitter/X showing highest rates across all impact types (61.5%-81.3%).

Breakdown of impacts of viewing “bad for the world” content reported by users on each platform. Note: Percentages are calculated by dividing the number of people reporting the impact on each platform by the number of people reporting the experience on each platform. Source: Neely Social Media Index survey of 2,666 U.S. adults conducted Fall 2024/Spring 2025.

Figure 5. Politics-related negative personal experiences and harmful content exposure across social media platforms increased from baseline to election waves. Reddit observes anomalies in reduced institutional trust and concerns for misleading content.

Comparison of key election-related variables on each platform during baseline and election periods. Note: Percentages are calculated by dividing the number of people reporting the topic or impact on each platform by the number of people reporting the experience on each platform.

Source: Neely Social Media Index nationally representative survey conducted between February 5 - May 19, 2024 (2,913 respondents) and October 14, 2024 - January 4, 2025 (2,666 respondents).

We also examined the topical breakdown of content users considered “bad for the world” across the same six major platforms as we did for negative personal experiences (Figure 3), with similar findings. Politics once again contributed to the most harmful content reports across all platforms in both time periods, accounting for 86.4% of bad content reports on Twitter/X, 79.2% on Facebook, 71.5% on TikTok, 70.7% on Reddit, 68.3% on YouTube, and 63.9% on Instagram during the baseline period. This prevalence mostly increased significantly during the election period on many platforms, reaching 91.7% on Twitter/X, 82.5% on YouTube, 82.2% on Facebook, 77.5% on TikTok, 71.4% on Instagram, and 69.6% on Reddit. Twitter/X maintained the highest concentration of political content users found “bad for the world.”

Among user-reported potential impacts of harmful content, misleading information emerged as the primary concern) (Figure 4), with rates ranging from 68.2% on Instagram to 81.3% on Twitter/X. Hate, fear, and anger ranked second (58.7% to 76.6%), with Twitter/X again showing the highest rates. Political polarization also posed a major concern, particularly on Twitter/X (77.0%) and Facebook (65.5%), as was risk of violence (42.6% to 61.5%). Only 1.1% to 9.1% of users reported that harmful content had “no effect.”

Drawing across negative personal experience and harmful content metrics, we visualize the key political variables across baseline and election periods (Figure 5). These variables were chosen because, together, they track key election-period concerns: prevalence of political content, erosion of institutional trust, risk of misleading information, and polarization. We find that negative experiences relating to political content exposure increased virtually uniformly during the election period. Political polarization effects increased during the U.S. election cycle on most platforms, and were especially pronounced on Twitter/X. Reports of reduced trust in institutions were also substantial. Interestingly, Reddit observed decreases in content related to reduced trust in institutions and misleading information from the baseline to the election wave; it is also the only pseudo-anonymous platform that focuses on community interaction and moderation methods.

Private communication platforms yield more meaningful connections than public social media. Several platforms showed reduced meaningful connections post-election.

Note: Neely Social Media Index nationally representative survey conducted between February 5 - May 19, 2024 (2,913 respondents) and October 14, 2024 - January 4, 2025 (2,666 respondents). % of U.S. adults on each platform who experienced a meaningful connection with others during each period.

Different social media platforms demonstrated varying capacities in fostering meaningful personal connections during the 2024 U.S. election cycle. Unsurprisingly, communication-centered platforms saw the highest prevalence of user-reported meaningful connections; Facetime exhibited the largest increase of 12.6 percentage points from 49.2% during- to 61.8% post-election. However, all other significant changes we found involved decreases, with LinkedIn experiencing the largest decline of 7.0 percentage points from 21.9% to 14.9%, while Instagram declined 4.5 percentage points from 26.5% to 22.0%. WhatsApp, Instagram, Pinterest, Facebook, and Dating Apps all showed moderate decreases of 3.6 to 4.9 percentage points, although WhatsApp maintained strong overall connection rates at 54.2% during the election wave.

Private communication platforms WhatsApp, Facetime (52.0% during the election wave), and Text Messaging (51.8%), Facetime (52.0%) yielded the highest meaningful connection percentages while traditional, public-facing social media platforms showed more modest rates: Facebook at 35.3%, Instagram at 25.4%, and Twitter/X at only 9.0%. Personal communication tools therefore may yield genuine connections even during politically charged periods.

YouTube and Reddit top platforms for learning useful information, Twitter/X shows election-period spike

Note: Note: Neely Social Media Index nationally representative survey conducted between February 5 - May 19, 2024 (2,913 respondents) and October 14, 2024 - January 4, 2025 (2,666 respondents). % of U.S. adults on each platform who learned something that was useful or that helped them understand something important during each period.

Perhaps most critically, social media platforms serve as information and misleading information sources during election periods; their effectiveness in broadcasting useful or important content, however, varies significantly across platforms. Interestingly, for many countries outside the U.S., social media can be seen as a democratic good and promoter of healthy discourse. In our sample of U.S. adults, we found that many top platforms with users reporting learning something useful involved video-based content or community discussion.

Twitter/X yielded the largest increase in user-reported useful learning experiences of 9.1 percentage points from baseline to election period (17.0% to 26.1%), likely reflecting heightened information seeking as a more politics-focused platform. On Reddit, useful learning increased 6.1 percentage points post-election (46.1% to 52.2%); WhatsApp increased 9.9 percentage points (14.7% to 20.6%); LinkedIn decreased the most at 12.9 percentage points post-election from 26.3% to 13.4%.

Across all periods, YouTube consistently maintained the highest learning rates (53.6% baseline, 50.9% election wave), followed by Reddit (44.6% baseline, 47.8% election wave), Pinterest (40.2% baseline, 41.6% election wave), LinkedIn (25.4% baseline, 23.3% election wave), and Text Messaging (21.8% baseline, 21.7% election wave). Thus, different platforms and their functions may inherently promote or restrict learning outcomes.

Challenges in Monitoring Social Media Experience Around Public Events Using National Surveys

In our study, to capture the social media use experience before, during, and after election, we designed the study as follows:

1) participants were randomized to receive the survey either pre- or post-election,

2) final group assignment was based on the actual date of survey completion.

3) given the recall nature of self-reported social media experience, to align reported experiences with exposure to the election, we assigned participants who completed the survey before December 3, 2024 to the “During Election” group (based on a 28-day recall period), and those who completed it on or after December 3 to the “Post Election” group.

We note several limitations in the current study design.

• Noncompliance with random assignment

Although participants were initially randomized to pre- and post-election survey waves, actual group assignment was determined by survey completion date. This introduces noncompliance with random assignment, which may have led to systematic differences between groups (e.g., people who complete surveys earlier may differ in important ways from those who respond later).

• No true control group

For events such as elections, the “treatment” is universally affecting the entire U.S. population. As a result, everyone is exposed to it to some degree, making it impossible to construct a “never-treated” control group. This limits the possibility of estimating causal effects using traditional experimental or quasi-experimental logic.

In other words, our goal here is not to quantify the effect of “election” itself, but instead to monitor the social media user experience around that period. The results may signal how a variety of events during the election period (e.g., misleading information spikes, major platform policy changes) may influence user experience.

• Ambiguity in social media experience recall window

We relied on a 28-day recall window, but participants’ exposure to the election could vary widely even within that window. For example, someone responding on December 1 might mostly recall October experiences, while someone responding on December 3 might mostly recall post-election content. This introduces temporal ambiguity in what the “pre” and “post” groups actually represent. The distinction is not clean, with respondents often being sensitive to more recent effects. The noise introduced by this likely leads us to underestimate any potential effects.

The recall window issue could be hard to resolve in large-scale national surveys. If more resources are available, we could invite participants via social media platforms to monitor the temporal experience or shorten the recall window.

In Summary

Through descriptive analysis of data from the two most recent Neely Social Media Index surveys, we report significant changes in both negative and beneficial user experiences around the 2024 U.S. election, with significant rises in political content that either affected users negatively or were perceived to be “bad for the world”. Twitter/X emerged overwhelmingly as the platform where users reported the most politics-related negative experiences (~82%), while Facebook, YouTube, and Instagram all saw ~10% related increases. The vast majority of users experiencing negative content also reported real personal and societal impacts—only 5-15% reported “no effect.” These impacts involved anger, hate, misleading information, polarization, erosion of trust, reduced self-expression, and increased risk of violence among others. In terms of general social media use, we saw that post-election, platforms such as Facebook, TikTok, and Reddit experienced significant usage increases among American adults, while usage of FaceTime and Snapchat decreased.

While we cannot draw causal relationships between the 2024 election specifically and these outcomes, our findings align with broader expert discourse around social media as critical social democratic infrastructures. Furthermore, they should be understood in the context of changing election-related policies and technologies across major platforms between 2020 and 2024: Meta removed political ad bans; AI deepfakes falsely portrayed candidates; sentiments generally shift towards Republican criticism of content moderation as “censorship.”

We should also consider aspects of social media platforms that contribute to successes. Uniquely, Reddit saw decreases in reports of misleading information and content that reduced institutional trust from baseline to election periods. FaceTime saw a substantial increase in users reporting forming meaningful connections during- to post-election, and Reddit, WhatsApp, and Twitter/X saw significant increases in useful learning. Generally, platform experiences across platforms remained relatively stable, suggesting that the norms and designs of platforms have more of an effect on user experiences than the time period. The only consistent change was a greater percentage of negative experiences being reported as relating to politics. The question remains: what platform designs and features can meaningfully shape healthy online election environments? Given the consistent results across platforms, the answers may include serving logged off users via search, reduced reliance on engagement based feeds, private 1:1 interactions, and the existence of negative feedback mechanisms.