In both the lead up to and the immediate aftermath of the US presidential election, President Donald Trump made claims of voter fraud and a rigged election, using all channels available to him, including Twitter. Despite the apparent lack of evidence for these accusations, they have arguably influenced the beliefs of millions of Americans.

Twitter has been a primary means by which the president has sought to set the agenda. Since he first took office, many people have speculated that some of Trump’s tweets were deployed to distract from negative media coverage. For example, when the press reported on the US$25m Trump University settlement, he tweeted about the Hamilton play controversy. When COVID-19 failed to “just go away” but instead took a stranglehold on the US, he tweeted about the “OBAMAGATE!” conspiracy theory.

At least some of these distractions seem to have worked. For example, our previous research showed how there was far greater public and media interest in the Hamilton controversy than the Trump University settlement. But the evidence had been anecdotal – until now.

Our new research presents the first empirical evidence that Trump’s tweets systematically divert attention away from topics that are potentially harmful to him. Perhaps even more importantly, we found that this diversion works and suppresses subsequent coverage of potentially harmful news stories.

We asked two questions: is potentially harmful media coverage followed by increased diversionary Twitter activity by Trump? And does such diversion reduce subsequent media coverage of that topic?

To test the hypotheses, we focused on the content of the New York Times (NYT) and ABC World News Tonight (ABC) headlines and all of the approximately 5,000 Trump tweets during his first two years in office. We chose the Mueller investigation into potential collusion with Russia as a harmful topic. We then selected a set of keywords – “jobs,” “China,” and “immigration” – that we assumed would be Trump’s go-to topics at the time, based on a systematic content analysis of his campaign materials and major talking points.

The team hypothesized that the more the NYT and ABC reported on the Mueller investigation, the more Trump’s tweets would mention jobs, China, and immigration, which – if the diversion were successful – would then be followed by less coverage of the Mueller investigation by NYT and ABC the following day. The logic is illustrated in the graphic below.