Academic Paper Shows How Badly The Mainstream Media Misled You About Section 230
We've had to publish many, many articles highlighting just how badly the mainstream media has misrepresented Section 230, with two of the worst culprits being the NY Times and the Wall Street Journal. Professor Eric Goldman now points us to an incredible 200 page masters thesis by a journalism student at UNC named Kathryn Alexandria Johnson, who did an analysis entirely about how badly both the NYT and the WSJ flubbed their reporting on Section 230.
The paper is actually more than just that, though. It includes a really useful description of Section 230 itself, along with its history, and some of the often confused nuances around the law. Johnson clearly did her homework here, and it actually is one of the best summaries of the issues around 230 I've seen. The paper is worth reading for just that section (the first half of the paper) alone.
But then we get to the analysis. Johnson notes that the Times and the Journal are basically the most powerful agenda setting" newspapers in the US, so how they cover issues like Section 230 can have a huge impact on actual policy. And they failed. Badly.
The thesis explores the data in multiple ways, but one chart stands out: when talking about the impact of 230, both newspapers almost always frame the law as having a negative impact. They almost never describe it as having a positive impact.
That is, out of 116 articles in the NY Times that talk about the impact of Section 230, 107 described it negatively. Another six gave a combination of negative and positive, and only two (two!) described the impact positively. For the WSJ, it's basically the same story: 88 articles discussing the impact of Section 230, 80 of them purely negative. Another four with a combination of negative and positive, and just three describing the law's impact positively. That means, grand total, 91.7% of the articles in these two agenda-setting newspapers described the law's impact as negative, with another 4.9% describing both negative and positive impacts, and just 2.5% describing the impact positively.
That's pretty amazing. Now, some may argue that if you truly believe that the impact of Section 230 is negative, then these two publications are only being accurate in their descriptions. But, for those of us who have studied Section 230, and understand its broadly positive aspects, the whole thing seems crazy.
I've had many people argue over the years that the big newspapers like the Times and the Journal have an institutional interest in trashing social media and the internet, because it takes away from their gatekeeping powers. And I've always brushed that aside as an exaggeration. But the numbers here are pretty damn stark.
The paper also explores how these newspapers sought to frame Section 230, and found that they did a very poor job explaining how it has multiple functions, often choosing to focus on one framing - rather than a more accurate framing of how Section 230 is structured to encourage multiple things. It protects websites from being held liable as a publisher of third party content, which encourages more websites to allow for more speech, and it protects them from content moderation decisions creating liability, enabling them to cultivate their communities in the way they see fit. Understanding both of these is kind of important to understanding Section 230, but it appears that these papers rarely gave a complete description. Also, perhaps oddly (or perhaps because they're just super confused themselves), they often used the publisher framing, even though they were really talking about the content moderation function - which may very well be why so many others, including politicians, are so confused about 230.
As previously discussed, the majority of definitions including only the publisher" frame.Interestingly, despite a majority of definitions referencing only platforms' protection fromliability for the content posted by third-parties (59.5%), a large majority of articles were focusedon the societal impacts of censorship and deplatforming. Such issues most closely map to the content moderation" frame. And despite many of the articles' focus on censorship anddeplatforming, very few articles included definitions with only the content moderation" frame.
For the purposes of creating the most informed electorate, the most helpful definitions arethose that present both of Section 230's functions. These articles were coded as Both" whendiscussed above. Only a third of the definitions of Section 230 included both the publisher andcontent moderation frame, indicating a weakness in journalists' reporting on this issue. Coveragein The Wall Street Journal more frequently defined Section 230 in terms of both publisheractivity and content moderation activity than The New York Times, but coverage in The WallStreet Journal still mentioned both legal frames less than half the time. Journalists could improvecoverage could be improved by including definitions that explain both legal frames associatedwith Section 230, regardless of the focus of the article.
Then there's the question of how often these two famed newspapers just flat out got things wrong about Section 230. The data may be lower than you might expect, as Johnson found it happened 16.2% of the time, but that's still kind of astounding. This is a fundamental issue that has gotten a ton of attention and to still get it wrong in about one out of every six articles is indefensible.
It is interesting, though, to note that the WSJ misrepresented the law at nearly double the rate of the NY Times. Again, people have pointed out that Rupert Murdoch, who owns the WSJ, has more or less declared war on the entire internet, and noted that could impact the coverage of things like Section 230. I always assumed that would be a stretch, but the data here is, once again, noteworthy.
As Johnson notes in her paper, many of the misrepresentations were not necessarily outright falsehoods (though there were some of those), but rather statements lacking enough important context or requiring clarification."
Then there's this:
Every misrepresentation identified in the entire sample could be credited to an unattributed source. Therefore, journalists themselves were the source of each misrepresentation. This finding suggests that either journalists themselves do not fully understand the nuance of how Section 230 is applied or that journalists do understand how Section 230 functions but are not accurately conveying that knowledge to the reader
For what it's worth, it may also be the fault of the editors, rather than the journalists. I am familiar with at least one situation in which a major newspaper misrepresented Section 230, and the journalist later explained to me that they had fought for the correct representation, but their editor insisted on running a misleading one.
Johnson's paper also highlights how these misrepresentations can lead to further misunderstanding of Section 230.
Understanding that the First Amendment, and not Section 230, enables platforms tomoderate content is important to social understanding regarding how platforms would function ifSection 230 was reformed or repealed. Without the portion of Section 230 that precludespublisher liability, platforms would still be able to remove content, that for example violatedtheir community standards; however, platforms would be less likely to do so because they wouldonce again, be liable for any unlawful content that they did not remove.
Johnson also, correctly, summarizes what would actually happen with the removal of Section 230: there would be fewer places to speak online.
In fact, Australia's high court recently ruled that news media outlets are to be treated aspublishers" of the unlawful content that is posted in comments sections on social media. Inresponse, news media outlets began disabling their comments sections due to their inability toconstantly moderate all comments. Removing the comments section was the easiest way toprotect themselves from legal liability. This anecdote suggests that if Section 230 was changedand platforms were treated as publishers of third-party content, platforms would beginrestricting users' ability to post on their sites-severely stifling the ability of the public to sharecontent and ideas online. Limiting the public's ability to communicate online has negativeimplication for self-governance beyond just debate and discussion regarding Section 230. Theinternet provides a forum for citizens to ask questions, seek answers, and engage in debate aboutimportant policy issues. As a vast democratic forum[ ]" the internet has democratizedspeech by lowering the barrier of entry for individuals to speak, be heard, and engage in debatesabout important issues facing society. In this way, Section 230 creates a causality dilemma.Section 230 is necessary to create the speech environment online that is required for individualsto debate and discuss issues related to Section 230.
Johnson's paper also highlights how many stories about 230 inaccurately refer to it as a safe harbor" rather than an immunity." As it notes, this is an important distinction. DMCA 512 is a safe harbor, and in order to make use of it, you need to meet a bunch of qualifications. This is why there is a long history of case law involving extensive litigation about a bunch of different factors to determine if a site qualifies for the DMCA safe harbor or if it loses" the safe harbor. But 230 is an immunity, which is different. You can't lose an immunity. You don't have to take any steps to get the immunity. And one of the biggest misconceptions about 230 is that sites can take some sort of action that loses them the protections. That's not true, but when news organizations report on it as a safe harbor, they support that misconception.
There's much, much more in the paper, but it's quite an excellent thesis, incredibly detailed, including getting a lot of very nuanced and complex topics correct that (as the paper itself shows) journalists often get very, very wrong. And it also adds clear data to the discussion. Just an all around excellent piece of scholarship.