Policing Trump’s (or anybody’s) tweets: the problem is context

Donald Trump tweeted that mail-in ballots in California would lead to fraud. Twitter added a content warning on his tweets. If you think this is the start of Twitter fact-checking tweets, you’re wrong. It can’t. The problem is context.

Let’s review what actually happened. Here’s what Trump tweeted:

Twitter did something it rarely does: included a content warning on the tweet. The content warning looks like this:

If you click on it the warning, it goes to this page, with the warning below:

Trump makes unsubstantiated claim that mail-in ballots will lead to voter fraud

On Tuesday, President Trump made a series of claims about potential voter fraud after California Governor Gavin Newsom announced an effort to expand mail-in voting in California during the COVID-19 pandemic. These claims are unsubstantiated, according to CNN, Washington Post and others. Experts say mail-in ballots are very rarely linked to voter fraud.

What you need to know

– Trump falsely claimed that mail-in ballots would lead to “a Rigged Election.” However, fact-checkers say there is no evidence that mail-in ballots are linked to voter fraud.
– Trump falsely claimed that California will send mail-in ballots to “anyone living in the state, no matter who they are or how they got there.” In fact, only registered voters will receive ballots.
– Five states already vote entirely by mail and all states offer some form of mail-in absentee voting, according to NBC News.

Trump threatened to close down platforms like Twitter if they continued to show liberal bias.

Media reports about Twitter’s action were quick to label what the platform did as “fact-checking” the president, and liberals and other opponents of the President hoped it would lead to more warning labels on Trump tweets. That won’t happen. Let’s take apart what happened here, and why fact-checking on Twitter is and will always be a problem.

Twitter’s “fact-checks” apply only to election-related tweets

Twitter shuts down accounts and removes tweets that violate its policies. This includes harassing and threatening people, posting nudes without consent, and posting excessively violent content.

It does not include lying.

Trump has lied hundreds of times on Twitter. He’s misrepresented who pays tariffs. He’s quoted inaccurate poll numbers. He lies repeatedly about easily verifiable facts. Just in the last week, he retweeted doctored pictures of Nancy Pelosi.

Twitter doesn’t care.

It is impossible to police lies on Twitter. I could tweet that I have a PhD (I don’t), that my son is an olympic swimmer (he isn’t), and that 24 million Americans died in the Vietnam (they didn’t). Lying on Twitter is not a violation of terms of service, whether you are the leader of a nation or a blogger with no interest in reality.

So why did Twitter act on the Trump tweets about the election? Because, as as a Twitter spokesperson said, the tweets “contain potentially misleading information about voting processes and have been labeled to provide additional context.”

Lying about the election is a violation of Twitter’s terms of service. If I tweeted that the date of the presidential election is November 10 (it isn’t, it’s November 3) and that the polling place in your district was the Mormon Temple (it isn’t), then Twitter could kick me off and probably would.

So why didn’t they kick off Trump? Well, certainly, that would be a controversial move. Trump has violated their terms of service by threatening people, for example, and they haven’t shut his account for that.

But let’s actually take apart Trump’s statement on Twitter and evaluate it for truth.

  • “There is NO WAY (ZERO!) that Mail-In Ballots will be anything less than substantially fraudulent.” This is a prediction. Based on past history, it is unlikely. But like all predictions, there is no way to verify its truth. If I predict that the Cleveland Browns will win the 2021 Super Bowl, I’m almost certainly wrong . . . but you can’t know until (or if) the Super Bowl is played.
  • Mail boxes will be robbed, ballots will be forged & even illegally printed out & fraudulently signed.” Also a prediction. In fact, this is likely true. If one person robs a mailbox, forges a ballot, or fraudulently signs one, this could be true. That’s not massive voter fraud, but can you prove this won’t ever happen at all?
  • “The Governor of California is sending Ballots to millions of people, anyone………living in the state, no matter who they are or how they got there, will get one.” This is almost completely false. The governor is sending ballots to registered voters, not everybody in the state. But there will inevitably be errors, resulting in some ballots sent to people who registered falsely. And once they get to an address, someone else could open them and attempt to vote, even though they are not a registered voter. So there may be a tiny particle of truth here amid the unsubstantiated falsehood.
  • “That will be followed up with professionals telling all of these people, many of whom have never even thought of voting before, how, and for whom, to vote.” Politicians will tell people whom to vote for. Election officials, such as the secretary of state of California, will tell people how to vote, but not whom to vote for. This statement is confusing, but not inaccurate, depending on your definition of “professionals.”
  • “This will be a Rigged Election.” What is the definition of “Rigged Election?” If one fraudulent vote is cast, is the election “rigged?” This is too vague to evaluate for truth.

So this statement, while misleading, isn’t a clear-cut falsehood, unlike the many actual verifiably false statements that the president has made.

Twitter’s action was not to delete the tweet or kick Trump off Twitter, it was to indicate that the tweet was problematic and include a link to more information. Trump-haters will argue that is not enough. Trump backers will argue that it is outrageous.

I’d like to see Twitter do more to police the truth, but this is about as far as they will go.

The problem of context

On Facebook, there is an attempt to police false articles. If you post a link to one, and fact-checkers have determined that it is false, Facebook will mark it as inaccurate.

The problem with tweets that don’t link to anything is context. In 280 characters, context is lacking. You can say things that are mostly false or highly misleading, but aren’t directly contradicted by facts.

For example, suppose that reported COVID-19 cases in Texas decrease on a Thursday, and on Friday I tweet “COVID-19 on the decline in Texas!” Is that accurate? In a broader context, cases in Texas may be going up most days, but they did go down on one day. Should that tweet be labelled as false?

A whole news article is subject to fact checking (and news articles are held to a higher standard — misleading and unverified information is not appropriate). But tweets are hard to police. You can see why Twitter does not want to get into the business of fact-checking tweets for accuracy.

This is why Twitter is such a rampant cesspool of disinformation and lies, and why it is so easy for trolls to spread rumors. And it is also why the president is able to twist the platform to support his world view and share inaccurate information on it.

We count on institutions like tech platforms, media sites, and government officials to abide by their own codes of conduct. Doing so would reduce the amount of inaccurate garbage in the world, and could restore the idea of truth to the common discourse.

Unfortunately, many of those institutions and the people in them are no longer dedicated to truth.

I look forward to the day we can fix this. And I wish I knew how to do it.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.


  1. I want places like Twitter and Facebook and the like to post without comment or censure whatever anyone posts. Maybe they ought to remove posts that are not posted by the account-holder, that sounds reasonable.

    I would like folks that post to be held responsible for the accuracy of their non-opinion posts. That is done with comments from others or iggy.

    I do not find traditional new outlets to be trustworthy. This is by design. Their goal is to get attention and fast, so sensationalism is key. Journalists are not familiar, much less expert, in almost anything that they report, so anything with any science or critical thinking is suspect. Journalists and their outlets are biased. All of these do not matter who or what, it is. This, like the politics of today, is not new.

    I do not believe the problem is solvable, but one could temper the problems by being slower, not worrying about attention or influence, be conscious about biases, and by checking with experts before publication (and correcting mistakes and clarifying positions.)

    However, these all, in practice, lead to less (money, attention, influence, readers, etc.). So, why do those?

  2. I’m rather confused by the comment by Norman above.

    Let me try:

    The element of truth in Norman’s comment (if I understand it correctly) that I zero in on, is that it all comes down to the consumer of the information, whatever the source.

    GIGO = Garbage In, Garbage Out – If there is a demand for Garbage, there will surely be someone to produce it, since money flows to them.

    The AI of Social Media platforms are merely a form of mirror – algorithms driving toward the ultimate in individual personalization (predicting what each one wants to consume, based on past consumption patterns).

    More time online, more clicks, more views = more money. Journalism, Social Media – they all are largely funded by advertising.

    UGC = User Generated Content – Individual Social Media account holders participate in another way, as they attempt to provide content that also appeals to as large a group as they can, to gain visibility, status, and/or fame.

    Given the short form of Social Media, there is incentive to grab attention with the sensational. Hence the circle of GIGO.

    Bottom line: If we cannot police our own consumption, then how can we ever be happy with a third party attempting to do it for us?

    Norman, did I get it right?

    1. BM: That’s good. What confused you about my comment?

      I do not see the problem as a new problem…I remember reading the big papers (WashPost, NYT, WSJ) daily in the 80s/90s and regularly seeing so many “truth” issues. Social Media is just the venue du jour.

      Russ Roberts had a great webinar for Foundation for Teaching Economics today on data and science. Truth is scarce in the social sciences (and to a lesser extent, the hard sciences). “It’s complicated.”

  3. While I agree with the assessment here about context, as it gives nuance to specific claims, I think that whole focus by Twitter is an error. It is missing the big picture.

    The strategy being employed is “Flooding the Zone with Crap”. It actually doesn’t matter what is true. It is meant to confuse and exhaust.

    It creates a high level skepticism, such that large portions of the population don’t know who or what to believe anymore. When that happens, it only becomes a game of signalling.

    The more outrageous and galling the claim, the more signal value.

    If Twitter is serious about the problem, then they have to decide what they want to do about users of their platform employing this strategy.

    Same with Journalism / the News Media – do they continue down every rabbit hole to provide nuance, context and clarity, or do they recognize they are being set up to exhaust themselves down rabbit holes and exhaust their audience too.

    They’ve also done themselves a disservice by over-reacting or making out-sized claims of their own in response – feeding into that skepticism so desired by their target.

    This Flood the Zone strategy works because of the shape of our media consumption and incentives built within that (see my comment above).

    I’d like to think that Twitter (and other platforms) and the News Media would figure out a way to deal with it, but I fear that the incentives for any one participant would mean they will continue to let this play out.