It’s no secret that politically active Twitter users punch well above their weight. The most prolific political tweeters provide politicians and members of the press with an instantaneous and, therefore, valuable feedback mechanism. In exchange, they are granted disproportionate influence over the national political agenda. Study after study has shown how much clout Twitter wields over the national political environment. Books have been written about the ways in which “Twitter is changing newsroom cultures and practices” and about the “concept of social capital and agenda setting.” But fewer efforts have been made to both qualify and quantify precisely what “Twitter,” in this context, is. Who are these powerful tweeters? What are their politics, and how do they behave online? The Pew Research Center has shed some light on who is driving the national dialogue, and it’s not a pretty picture.

From June 2018 to June 2019, Pew studied more than 1 million tweets produced by a select group of adults who constitute a representative sample of the 22 percent of American adults with active public Twitter accounts. To determine what constitutes a political tweet, Pew limited its focus to online messages that involved discussions around national political issues or parties, ideological groups and institutions, or explicitly political behaviors like voting or directly engaging with a politician (though that engagement doesn’t need to be explicitly political in nature). Given that criteria, Pew found that “prolific political tweeters” make up just 6 percent of the population of active users on the website, but they generate 20 percent of all tweets and 73 percent of tweets with political themes. All told, while only 13 percent of the tweets Pew studied could be considered political in nature, a staggering 97 percent of those were composed by just 10 percent of active users.

So, who makes up this select group of influential Americans? It may not surprise you to learn that they are wildly unrepresentative of the broader population of politically active Americans. You may be equally unperplexed to discover that these movers and shakers are overwhelmingly on the left.

From mid-November to mid-December 2018, in the middle of this study period, Pew found that a full two-thirds of Twitter users with public accounts disapproved of Donald Trump’s performance in office while just 30 percent approved. By contrast, the Real Clear Politics average of job approval surveys from this same period found Trump’s disapproval rating fluctuating between 53 and 54 percent of surveyed adults and voters. Twitter users who disapprove of President Trump generated 80 percent of all studied tweets and 72 percent of political tweets during the timeframe Pew observed. Comparatively, those who strongly approve of the president produced just 11 percent of all tweets but a full quarter of political tweets. The “center”—e.g., Americans with more nuanced views of this presidency and the president—are all but unrepresented.

Sixty-seven percent of users consistently composing tweets with political themes identify as Democrats or as leaning toward the Democratic Party. Twenty-eight percent lean toward or identify with the GOP. America’s “prolific political tweeters” are also disproportionately older than the population. Those over 50 crafted 29 percent of tweets but 73 percent of political tweets. Those age 65 and over produce just 10 percent of tweets but a full third of political commentary. This predominantly Democratic population is far more likely to self-select into bubbles composed of ideologically like-minded groups and they are more likely than their Republican counterparts to believe that the news that they encounter on social media is accurate and reliable.

Now, it’s important to note that this study does not capture the themes and conversations around social and cultural issues that are often mistaken for politics. A self-sustaining media ecosystem is devoted to chronicling every latest outrage that is supposedly reflective of a broader socio-economic malady so universal and culturally ingrained that it is impervious to legislative remedy. More often than not, that outrage is illustrated by only a handful of tweets. Pew’s study narrows its focus to politics properly understood and is therefore more valuable than an anecdotal examination of how click-starved media outlets manufacture trends. It demonstrates how Twitter’s hostile takeover of the national agenda is being conducted by one of the most atypical and parochial subsets of the American public imaginable, and it’s having profoundly deleterious effects on the national dialogue.

George Mason University Economics Professor Tyler Cowen astutely observed, in February, that Twitter is accelerating an already noticeable trend toward weaker political institutions and stronger political personalities. Left unchecked, coherent political platforms will be subsumed into the daily, even hourly, dictates of the social-media news cycle. Concepts and programs that are easily expressed will supplant more complex matters. “Technocratic dreams will fade,” he wrote, “and fiery rhetoric and identity politics will rule the day.”

Well, sure. But the most interesting finding in Pew’s survey is how many millions of Americans, including millions of Twitter users, opt out of this competition. More remarkable still is how woefully underserved they are by the political class and the press. It’s true (and obvious) that social media is a self-reinforcing, homogenized feedback loop that encourages predictable codes of behavior, but that does not make it less relevant. Far from it. Politicians and media provide the platform with status and ensure that events that occur online have real-world consequences. The question is, should they? Pew’s study seems to confirm rather definitively that the answer is no.