How Social Media Experiences Changed from March 2023 through May 2024
With new data through May 2024, we find that Reddit is increasingly informing users, while X isn’t across 5 survey waves conducted between 2023 & 2024
This is a post where Senior Advisor Matt Motyl analyzes and summarizes data from the University of Southern California Marshall School’s Neely Center Ethics and Technology Indices. The project builds on industry precedent to create independent verifiable, cross-platform metrics of user experiences, which we leverage in discussions with companies, the press, and policy makers. You can read more about the methodology of the project here and find a full list of analyses here.
Social media and technology companies had an eventful couple of years in 2023 and 2024. They laid off tens of thousands of employees, a disproportionate number of whom specialized in protecting users from harmful experiences. Predictably, hate speech and propaganda on X (formerly Twitter) increased, as did Russian disinformation on Facebook and Instagram. Journalists and watchdog groups also discovered that Facebook and Instagram’s recommendation systems helped to connect and promote vast networks of pedophiles consuming and selling child sexual abuse material. Consequently, executives of these companies were called to testify in front of the United States Senate regarding concerns that they were failing to protect young people who use their platforms (for an in-depth review, see this post-mortem co-authored by yours truly on the Integrity Institute blog). Additionally, regulatory bodies in Australia and the European Union launched investigations into these platforms’ treatment of harmful content. Brazil even went so far as to block access to X (Twitter) in Brazil for its failure to crackdown on disinformation. Given these developments over the past 1.5 years, it is worth examining how users’ experiences on these platforms may be changing.
Thanks to USC Marshall’s Neely Center for Ethical Leadership and Decision Making, the Psychology of Technology Institute, and the John S. and James L. Knight Foundation we were able to start our longitudinal panel survey of US adults’ experiences on social media in early 2023. Now, we are able to examine how experiences across these platforms have changed over the course of a tumultuous year in social technology. In today’s report, I compare usage of and experiences on these platforms and services from our initial survey in March-May 2023, and expand upon an earlier report where I examined changes from November 2023-February 2024 to integrate an additional survey wave that ran through May of 2024.
Usage of Online Services
Overall usage of social apps and platforms generally remained stable among US adults. Just two of the 16 largest platforms included in the Neely Social Media Index survey showed significant declines in their monthly users. X (Twitter) lost more than 17% of its US adult monthly users, most of whom left the platform in the middle of 2023. It appears that the platform has since stabilized with around 15.2% of US adults reporting using it in the past month (relative to 18.4% in early 2023). YouTube is the only other large platform in our survey showing a significant decrease in monthly US adult users in this time frame, with a 5.1% relative drop in users from the March-May 2023 survey period through the February-May 2024 period. In raw terms, monthly YouTube use among US adults dropped 3.4%. On the other side, Reddit is the only platform that showed a marginally significant increase in the percentage of US adults using the platform. No other platform showed significant growth or shrinkage.
The overall rankings of the top 8 most widely used platforms remained the same. Two platform agnostic modes of social communication -- Email and Text Messaging continued their reigns as the top 2 most widely used services. Facebook, YouTube, Instagram, Facetime, TikTok, and Snapchat all maintained their relative popularity. WhatsApp showed the biggest climb in the ranked popularity, jumping 3 spots, followed by Reddit jumping 2 spots, and Pinterest climbing 1 spot. X (Twitter) dropped 3 spots, followed by LinkedIn, Online Gaming, and Nextdoor, which all dropped 1 spot.
We also asked our survey panel about Mastodon and Threads, two newer social media platforms that received some popular press coverage. Fewer than 1% of US adults used Mastodon, and we removed that from the list of platforms included in our survey. Threads, despite becoming the most rapidly downloaded app ever, struggled to retain users. In August 2023, we found that 4.6% of US adults had used the app in the previous 28 days, but that number was cut in half and has hovered around 2-3% in each of the survey waves since then. Recently, the number of downloads of Threads has accelerated while downloads of X (Twitter) have continued to decline. External data reveals that Bluesky is showing substantial growth, gaining more than 13 million users since it first opened to the public in February 2024. If we are able to obtain sufficient funding to continue this survey, we will add Bluesky to the list of platforms that we monitor.
Next, I report changes in user experiences assessed by 4 main top-level questions that respondents who use each platform were asked in the two survey waves:
In the past [28 days/4 weeks], have you personally witnessed or experienced something that affected you negatively on [service used]?
In the past [28 days/4 weeks], have you witnessed or experienced content that you would consider bad for the world on [services used]? (examples could include content that is misleading, hateful, or unnecessarily divisive)?
In the past [28 days/4 weeks], have you experienced a meaningful connection with others on [services used]? (examples could include exchanging emotional support or bonding over shared experiences)
In the past [28 days/4 weeks], have you learned something that was useful or that helped you understand something important on [services used]?
Negative Experiences that Personally Affected Users
First, we look at one of the two negative experiences users may have on social media and communication services included in our survey -- whether they personally witnessed or experienced something that affected them negatively in the 28 days prior to completing the survey. The rate of negative experiences showed a slight numerical decrease across all platforms except LinkedIn, but none of these decreases exceeded the margin of error. As the table below shows, X (Twitter), Facebook, and Nextdoor are tied for being the platforms with the highest rates of users reporting negative experiences in our survey. LinkedIn was the only platform showing a numerical, but not statistically significant, increase in the rate of negative user experiences reported.
On the other end of the spectrum, Pinterest and Facetime remained the platforms with the lowest rates of users reporting negative personal experiences. Neither of these platforms showed a significant decrease over time, but that is more due to the fact that their baseline rates of negative experiences were so low. In fact, fewer than 2.4% of their users reported a negative experience a year ago, and that number has fallen to around 1% for both platforms. The 0.8% and 1.2% decreases year over year for Facetime and Pinterest, respectively, marks 38.1 and 50% decrease relative to their baseline. Those are huge relative decreases, but the baseline rates were not different from a rate of 0% (i.e., both platforms prevalence estimates are smaller than the margin of error at all time points).
Experiences that are Bad for the World
Next, we look at the second of our two negative experiences users may have on social media and communication services -- whether they witnessed or experienced content that they would consider bad for the world. As the table below shows, X (Twitter) remained at the top of the list with the highest rate of users reporting experiences that they consider bad for the world of any platform in our survey. X (Twitter) did show some marginal improvement in the August-September 2023 data, which coincided with media coverage criticizing the company for not addressing anti semitism or bigotry following the Center for Countering Digital Hate and the Anti-Defamation League’s reports showing that hate speech was spiking on the platform. In the months that followed, experiences with content users perceived to be bad for the world on X (Twitter) trended upward.
On the other end of the spectrum, Discord, Pinterest, Facetime, and LinkedIn all had rates lower than 4.4% of their users experiencing content on those platforms that they perceived to be bad for the world. Discord did show a 51.7% increase in this type of experience relative to their baseline, but the raw change of 1.5% is a fraction of the margin of error meaning that this change may be statistical noise.
Three platforms--Facebook, YouTube, and Email--showed significant reductions in experiences with content that users deem bad for the world. No other platforms showed a significant change.
Learning Something Useful or Important
Next, we shift to looking at one of the two positive experiences users may experience on social media that are included in the Neely Social Media Index survey -- whether users learned something useful or important on these services and platforms in the previous 28 days. As seen in the table below, YouTube held onto the top spot with 53.6% of its users reported learning something that was useful or that helped them understand something important in the last 28 days. YouTube mostly held steady over the first 7 months of the longitudinal survey, but then showed a significant increase between November 2023 and February 2024, before increasing slightly in the most recent wave. Pinterest held steady across the duration of the survey, but after Reddit showed significant improvements it now shares the second spot for the highest rate of users reporting learning something useful or important on the platform in the previous 28 days.
On the other end of the spectrum, Snapchat and Online Gaming retained their positions as the two services with the lowest rates of users feeling that they learned something useful or important. Neither of those services showed a systematic trend over time; instead, their rates alternated between numerical increases followed by decreases that fell within the margin of error.
There was a tendency for users across many of the platforms to be less likely to say that they learned something useful or important in September and October than they were between March and September. This was most pronounced for TikTok and LinkedIn whose users reporting learning something useful or important dropped 66 and 46%, respectively, from the first two survey waves to the September-October survey wave. Yet, both of these platforms exhibited a rebound effect in the subsequent survey wave, and held steady through the most recent wave.
Additionally, there was an overall tendency for users reporting that they felt informed across platforms in the November 2023-February 2024 survey wave, which could imply exogenous factors like seasonality or specific world events (like the holiday season) may explain some of the changes above and beyond platform-specific features.
Over the course of the full year, 7 of the 16 most widely-used platforms showed significant increases in the percentage of their users who stated that they recently learned something useful or important. Reddit showed the greatest improvement, improving 16.4% between March 2024 and May 2024, which translates to a 58.2% increased likelihood of its users reporting learning something useful since our baseline measurement. Nextdoor and Discord showed the next most improvement of just over a 10% increase over the year. Given both of these platforms’ initial low rates of useful learning experiences, their relative improvement since then is remarkable. Nextdoor users’ likelihood of reporting learning something useful increased 97.3%, followed by Discord users’ likelihood of reporting learning something useful increased 75.6%.
Email, text messaging, and Instagram also improved significantly over the year. No platform exhibited a significant decrease in learning something useful or important, but X (Twitter) was the only platform showing a numerical decrease (though this decrease was less than the margin of error).
Meaningful Connections with Others
Next, we look at the other of the two positive user experiences people may have on social media and communication services that we included in our survey -- whether or not they experienced a meaningful connection with others on the service within the past 28 days. As seen in the table below, the two services with the highest rates of their users reporting a meaningful connection with others -- Facetime and Text Messaging -- retained the top two spots, with the majority of users of both services reporting having these experiences. On the other end of the spectrum, Pinterest retains its position as the platform with the lowest rate of users experiencing a meaningful connection with others on the platform.
There was a general numerical tendency for a higher percentage of users of most platforms to report experiencing a meaningful connection in the most recent survey window. In our last report, we noted that the increase could have been due to the holidays potentially driving up the rate of meaningful connections across the board. This may still be the case, but it seems less likely because the rates of meaningful connections generally stayed at the elevated rates through May 2024.
Over the course of the five survey waves, 4 of the 16 most widely-used platforms or services showed significant changes. Facetime’s 10.7% improvement marks the largest increase, but LinkedIn’s 6.6%, Snapchat’s 6.5%, and Text Messaging’s 4% improvements followed closely behind. Some of the services with smaller user bases reported larger numerical gains (e.g., Discord), but due to the smaller number of users the margin of error is larger which raises the magnitude of improvement needed to be confident that the improvement is reliable. Nonetheless, Discord, Pinterest, Online Gaming, and WhatsApp did show numerical improvements.
How do these platforms rank based on the rate of their users with each of these experiences?
Examining the percent change in experiences on platforms is instructive as it tells us whether user experiences within each of the platforms are improving or worsening over time. The downside of this approach is that it fails to display how these platforms stack up to each other accounting for their changes across these different dimensions. To address this, I created the following heat map that shows the ranking of each service across the four experiences we measured in our survey. To allow for easier comparison of each service’s current ranking to their ranking from one year ago, I included their previous rank in parentheses next to their current ranking. I also color-coded each tile so that better scores (i.e., more positive experiences and fewer negative experiences) are shaded in green and worse scores (i.e., fewer positive experiences and more negative experiences) are shaded in yellow.
There are four main take-home points from these rankings:
Traditional social media platforms (e.g., X/Twitter, Facebook, Reddit, TikTok) continue to tend to have more bad experiences than the more direct communication services (e.g., Facetime, text messaging) and the more focused / niche social media platforms (e.g., LinkedIn, Pinterest).
These services are relatively stable in their rankings over the first and last waves relative to each other. Even when one platform is improving in a user experience over time, they tend not to improve sufficiently to overtake the platforms above them in the rankings. The main exceptions to this are LinkedIn and Instagram, which both dropped several spots on both negative experience dimensions. NextDoor also leapfrogged several platforms in the learning something useful dimension, while Facebook and X (Twitter) both dropped several spots.
Pinterest has overtaken the top spot for the fewest negative experiences, and held strong near the top of the list for the highest rates of learning experiences. Given that Pinterest does not prioritize user to user communication, it makes sense that they have the lowest rate of users reporting meaningful connections with each other on the platform.
X (Twitter) has the highest rate of negative experiences, and has held that position across all 5 waves of the survey. X (Twitter) also showed the biggest drop in the rate of users reporting learning something useful or important, falling from 7th to 13th out of 16.
Summary
The Neely Social Media Index survey is reflecting other publicly available data sources in showing that the percentage of US adults using some social media and communication services are decreasing, as is the case with X (Twitter) and YouTube, or increasing, as seems to be the case with Reddit. Other platforms, like Facebook and Snapchat, have also trended downwards in the percentage of US adults using them in each subsequent wave. Neither of these downward trends currently exceed the margin of error, but if the trends persist, they will.
In our first annual report examining changes in user experiences on social platforms and services over 12 months, we see more significant improvements than significant regressions. While we cannot isolate the specific cause(s) of these improvements, it is reasonable to presume that advertiser, legal, and regulatory pressure are part of the story. Specifically, numerous advertisers pulled their advertising campaigns from social platforms where ads were being displayed adjacent to child sexual abuse material or pro-Nazi content, 45 different US states sued Meta for not doing enough to protect children’s well-being, and regulatory bodies in the US, EU, and Australia are seeking large financial penalties for failures to comply with regulations, counter disinformation, and remove illegal content. Some of these improvements may also be due to these companies realizing that promoting better user experiences makes good business sense and tends to improve long-term growth, as internal research from Meta described in Jeff Horwitz’s Broken Code demonstrated and as our Senior Advisor and Integrity Institute fellow (me, Matt Motyl) argued in Tech Policy Press.
We are now in the field with a sixth wave, conducted over the course of the 2024 election and are in development for a teen version of the index, to be launched by the end of the year. Since starting this project, we’ve presented our data and methods to numerous regulatory authorities, including the European Commission, the UK’s OfCom, Australia’s eSafety group, members of the United States Senate, and the United States’ Surgeon General. We’ve also had discussions with numerous companies about their results and how it maps to their product work. Finally, we have several academic projects ongoing that aim to publish formal longitudinal analyses of these data in academic journals. Momentum is growing for comparative user experience measurement. We are hopeful that more and more companies and jurisdictions will continue to track their longitudinal impact on users’ experiences and are glad to be part of this growing practice.