President Trump has been nothing if not consistent in his approach to the investigation by special counsel Robert S. Mueller III. From the outset, he's repeated a few key phrases meant to cement his position in the minds of Americans. No collusion. No obstruction. And, at times, no votes affected by Russia.

This last point has a very specific purpose. In the days after the election, as vote totals in the upper Midwest were finalized, it became obvious just how narrow his victory was. In Michigan, Pennsylvania and Wisconsin he won by a combined 78,000 votes while losing the popular vote by millions nationally. His campaign transitioned from warning about voter fraud to defending against allegations that the vote had been tainted. When media outlets began reporting on Russia’s efforts to aid his candidacy, he was quick to dismiss the idea that Russian actors had, as some theorized, gone into state voting systems and affected the tally. He wanted his victory to be his, not Russia’s.

The release of a redacted version of Mueller's final report on his investigation into Trump's campaign and possible coordination with Russia's interference efforts included a brief section addressing that specific claim. Mueller's team, it states, didn't actually investigate whether Russian attempts to access voting systems, which occurred, were successful. (Outside of the Mueller investigation, there hasn't been much evidence that it was.)

One of several tweets from the president addressing the report's release included a mention of vote manipulation.

It’s true, definitionally, that anything Russia did in 2016 occurred while Barack Obama was president. It’s not true that he did nothing about it, though. He pushed for a sweeping statement addressing the Russian effort, but couldn’t get agreement from Republicans in Congress to do so. His Homeland Security director and director of national intelligence did release a report specifically warning state elections officers against the threat of being hacked by Russia. Unfortunately for its impact, it was released on Oct. 7, 2016, right before The Washington Post published the “Access Hollywood” tape and WikiLeaks began releasing emails stolen from Hillary Clinton’s campaign chairman, John Podesta.

Which brings us to the last, critical part of Trump's tweet from Thursday evening.

"Most importantly, the vote was not affected,” he writes. What Mueller showed is that this claim is entirely wrong.

Here's what we know about the Kremlin's playbook for creating division in the U.S. (Jenny Starrs/The Washington Post)

It may be — and appears to be — true that Russia didn’t manipulate actual voting results, changing a county’s pro-Clinton votes into pro-Trump ones. Perhaps the Oct. 7 warning played a role in that success. But Russia’s efforts absolutely affected the vote, as they were intended to — and as Trump and his campaign hoped they would.

How? Well, those Podesta emails for one. Mueller's indictments against a number of alleged Russian intelligence agents details how they accessed the Democratic National Committee network and Podesta's email account (among other targets) and transferred the data to WikiLeaks. There, the stolen material became a central component of the election coverage for much of the last month.

In October 2017, we made this graphic, showing how often WikiLeaks was mentioned on cable news, compared to former FBI director James B. Comey (whose last-minute announcement about Clinton’s email server certainly affected the results) and to “Access Hollywood.”

All of that conversation during the course of that month certainly had some effect.

That’s a point that Mueller makes explicitly. When Attorney General William P. Barr released a brief letter last month providing an overview of what he said Mueller’s report contained, he quoted Mueller as writing “the investigation did not establish that members of the Trump campaign conspired or coordinated with the Russian government in its election interference activities.” The full quote, though, says something slightly different.

“Although the investigation established that the Russian government perceived it would benefit from a Trump presidency and worked to secure that outcome,” the report states, “and that the campaign expected it would benefit electorally from information stolen and released through Russian efforts, the investigation did not establish that members of the Trump campaign conspired or coordinated with the Russian government in its election interference activities.”

The campaign “expected it would benefit electorally from information stolen and released through Russian efforts,” it says. The campaign knew that WikiLeaks was releasing material damaging to Clinton, prompting Trump to praise the organization frequently that October. (It was also known by then that Russia had a hand in the hacking of the Democratic Party material.) It — and Trump — tried to get more information about what WikiLeaks had, dispatching the candidate’s longtime ally, Roger Stone, to try to learn more. Why? Because they believed it would help their campaign.

There’s certainly an indication that such a belief would be warranted. In late October, Suffolk University and USA Today asked Americans whether emails stolen from Podesta that included excerpts from speeches which might contradict Clinton’s policies would affect their vote. Ten percent said that WikiLeaks releases would make them more likely to support Clinton. Thirty-seven percent said they’d be less likely to.

That was one part of one of the two major WikiLeaks releases that year. The other came in July, aimed at sowing dissension among Democrats by boosting the idea that Sen. Bernie Sanders (I-Vt.) was treated unfairly in the election. Assuming that Russia didn’t affect the 2016 election means assuming that no part of the July release of material stolen from the Democratic Party and no part of the steady trickle of releases in October had any impact on any voter’s decision. That no voter, say, opposed Clinton because they thought the DNC had tried to position her over Sanders — a belief that became tightly tied to WikiLeaks releases.

Clearly, the WikiLeaks releases had more than zero influence.

How much influence it had is hard to ascertain. FiveThirtyEight tried in December 2016. The closest it came was to determine that voters who made up their minds in October preferred Trump by a 14-point margin according to exit polls. This was the period when interest in WikiLeaks was near its peak.

It may well have not been enough to change the results of the election, but it’s simply indefensible to say that Russia didn’t affect any votes. If WikiLeaks affected the vote and if, as Mueller’s evidence shows, it got that material from Russia, then Russia affected the vote. The claim is particularly indefensible coming from Trump, who celebrated WikiLeaks publicly precisely because he knew it would help his candidacy. He knew it and Mueller knew it.

It was, after all, one of the key reasons that Russia was interfering in the election in the first place.