Karen Cyphers: Few want government regulating online speech; Meta’s move to TX may decrease distrust

The dynamics of American social media shifted dramatically last week when Meta CEO Mark Zuckerberg announced sweeping changes to content moderation and fact-checking policies across Meta’s platforms — Facebook, Instagram and Threads.

And he didn’t stop there.

Zuckerberg followed this with a surprise appearance on “The Joe Rogan Experience” podcast, where he doubled down and expanded on many of the points he had alluded to in his original statement. Most notably, he revealed how the Joe Biden administration had exerted significant influence over the nature of content that Meta would approve, suppress, or fully deplatform.

According to Zuckerberg, this wasn’t a matter of subtle persuasion — it was outright interference. He described heated phone calls filled with yelling and demands. These weren’t limited to issues of national political importance; they also extended to petty matters, like memes the president or his staff simply didn’t like. If Zuckerberg’s account is accurate, it raises serious concerns about government overreach — concerns that should trouble not just opponents of the Joe Biden administration but also its supporters who value free speech.

Sachs Media regularly surveys Florida voters on political and cultural issues; we went into the field immediately after Zuckerberg’s appearances to see how Floridians felt about these changes.

This survey, conducted Jan. 9-12, 2025, shows that the issue isn’t abstract or theoretical.

Our research shows that nearly half of Floridians who use social media (45%) say a platform has restricted or removed their content at some point. This experience spans political affiliations. Even more striking, 6 in 10 (59%) report witnessing someone else’s content being unfairly censored or removed.

Given how this issue affects so many people, cuts to the heart of First Amendment rights, and shapes how news is shared, Sachs Media has been tracking public opinion on the politics of content moderation for several years. We’ve focused particularly on how Americans feel about government involvement in online speech. Shifts in public opinion are typically gradual, but over the past 14 months, we’ve seen a dramatic shift — a rare “sea change” in how people view the expression of opinions online.

Following Zuckerberg’s announcements and podcast appearance, we surveyed 800 Florida voters. To measure how opinions have evolved, we asked a mix of new questions and repeated others verbatim from November 2023.

The following provides a detailed analysis, starting with the timeliest insights.

Half of Floridians (51%) reported being aware of Zuckerberg’s announcement before encountering it in this survey. Among those familiar with the news, a majority (57%) expressed general approval of the changes Zuckerberg is implementing, while 31% disapproved and 12% were neutral. Support for the changes was significantly higher among Republicans (88%) and nonpartisan voters (61%), compared to just 26% of Democrats. The new policies also appear to influence how users will behave: 25% of Republicans and 23% of nonpartisans said they’re now somewhat more likely to use Facebook/Meta. In comparison, 33% of Democrats indicated they are less likely to do so.

Asking people whether a policy change will influence their behavior is one thing. Testing that influence is another. Survey experiments allow us to measure whether specific interventions lead to statistically significant differences in responses.

One of the most intriguing hypotheses stemming from Zuckerberg’s announcement involves Meta’s decision to relocate its content moderation team from California to Texas — a move explicitly designed to build user trust by signaling that moderation decisions are fair and free from political bias.

In other words, Zuckerberg’s team appears to believe — and may have even conducted their own research to confirm — that hiring content moderators in a more conservative state like Texas could help address concerns about political bias in content moderation.

But will establishing a Texas-based moderation headquarters actually change public perception?

To test this, we designed an experiment featuring a fictional social media company. Respondents were told that the company had hired content moderators and fact-checkers to review potentially harmful content and misinformation. All participants were shown an image of the (fake) company’s headquarters, but they were randomly split into two groups: one with an image of Fort Worth, Texas, and the other with an image of Palo Alto, California.

Respondents were then asked how much they trusted the company’s content moderators to make fair decisions. Overall trust levels remained relatively stable across the test groups, but distrust was noticeably reduced among Republicans and nonpartisan voters (NPAs) when the supposed moderators were located in a more conservative state.

Specifically, 74% of Republicans didn’t trust the California-based moderators, compared to 60% who saw the “Texas” label — a 23% reduction in distrust. This aligns closely with the 25% of Republicans who earlier indicated they’d be more likely to use Facebook under similar changes. Among NPAs, distrust dropped from 55% in the “California” group to 43% in the “Texas” group.

Among Democrats, the pattern was reversed. Trust was higher among those who saw the “California” label (44%) compared to those who saw the “Texas” label (33%). However, among these same Democrats, there were no statistically significant differences in distrust levels between Texas (25%) and California(30%).

The takeaway? Meta’s move to Texas may not immediately build trust, but it appears to have a tangible effect on reducing high levels of distrust among those who are already skeptical. At the same time, it doesn’t seem to significantly alter the views of those who already hold greater baseline trust.

We also tested whether voters agreed with several key points from Zuckerberg’s announcement by presenting them as slightly edited, unattributed statements.

Here’s how the overall agreement ranked, from most to least:

— 77% agree: “Social media restrictions on topics that are the subject of frequent political discourse and debate are unfair. It’s not right that things can be said on TV or the floor of Congress, but not on our platforms.”

— 74% agree: “What started as an effort to be more inclusive has gone too far, such as on topics like immigration or gender, and censorship has wrongly been used to shut down opinions and shut out people with different ideas.”

— 69% agree: “Professional ‘fact-checkers’ hired by social media companies have been too politically biased, destroying more trust than they’ve built in their moderation efforts.

— 68% agree: “Europe has an ever-increasing number of censorship laws; Latin American countries have secret courts that can order social media companies to take content down; and China censors many apps from even working in their country. It’s important for the U.S. to work with countries around the world to reduce censorship.”

— 60% agree: “Whether well-intended or simply political, the Biden administration put pressure on social media companies to censor content.”

— 58% agree: “Over the past four years, social media companies have censored out too much content that should be allowed in mainstream discourse.”

— 56% agree: “The incoming Trump administration is determined to restore the expression of free speech.”

While agreement with these statements varies significantly across political parties, even a majority of Democrats align with the first set of statements — those that are more general and less focused on events of the past four years.

For instance, among Democrats:

— 65% agree that social media should allow content related to topics debated on TV or the floor of Congress.

— 56% agree that what began as an effort to be more inclusive has gone too far.

— 57% agree that it’s important for the U.S. to collaborate with other countries to reduce global censorship.

— 54% agree that professional fact-checkers have destroyed more trust than they’ve built.

However, Democrats are far less likely to agree regarding statements about alleged interference by the Biden administration. Only 29% of Democrats believe the administration pressured social media companies and just 38% agree that social media platforms have censored too much content over the past four years. Unsurprisingly, only 19% of Democrats believe the incoming Trump administration is determined to restore free speech.

Regardless of their feelings about Zuckerberg’s announcement or the differences between Presidents Biden and Trump, a plurality of Floridians (42%) believe censorship should occur only in extreme cases, such as incitement to violence. This perspective is shared relatively evenly across party lines. However, opinions on other approaches to content moderation reveal stark divisions. While 29% of Floridians think there should be no censorship at all, 28% believe social media companies should actively moderate to prevent misinformation. Party differences are particularly pronounced: 42% of Republicans favor no censorship, compared to just 18% of Democrats. Meanwhile, 43% of Democrats support active moderation, with only 14% of Republicans feeling the same.

To provide context for these findings, we revisited a series of questions we first asked Florida voters in November 2023. When asked who should have the most control over what is shown on social media platforms, 30% of voters in 2023 preferred government-related agencies — whether state, federal or international. That number has since dropped by nearly half, to 17%.

At the same time, the preference for individual users to have control has increased from 41% in 2023 to 51% today, while support for private social media companies having the most control has risen modestly, from 30% to 32%.

Additionally, more voters now believe courts should protect online speech, even when the content in question is distasteful or upsetting. This sentiment has grown from 61% in 2023 to 68% today. The remainder of respondents believe courts should focus on protecting the rights of private companies to decide what types of speech to allow on their platforms.

Finally, support for increased government control of online speech has also dropped significantly, from 47% in 2023 to 30% today. This decline is evident across all major political groups, including Democrats, nonpartisans and Republicans.

These findings suggest a growing preference among Floridians for individual control over online content and diminished trust in both government and corporate entities to moderate speech fairly.

Whether it’s a Babylon Bee satire piece flagged and removed despite being clearly labeled as humor … real images incorrectly flagged as fake … theories that COVID-19’s origins were prematurely suppressed but then later being recognized as plausible by the scientific community … or inconsistent treatment of public figures, Americans appear increasingly eager to reclaim the ability to judge content credibility for themselves or at the very least, to avoid the government taking on that role for them.

___

Karen Cyphers, Ph.D., is a partner and director of research at Sachs Media, a teaching faculty member at Florida State University, and runs the Decyphered Substack. You can reach her at [email protected].