On 29 January 2017, the YourNewsWire fake news site published an article asserting that an NPR study had determined that 25 million fraudulent votes had been cast for Hillary Clinton in the 2016 president election:
NPR: 25 Million Votes For Clinton ‘Completely Fake’ — She Lost Popular Vote
A study published by NPR reveals that over 25 million Hillary Clinton votes were completely fraudulent, meaning that the Democratic candidate actually lost the popular vote by a huge margin.
Were the claim correct, it would mean more than a third of Clinton's near 67 million popular vote tally consisted of votes were "completely fake." There was, of course, no truth to that claim.
YourNewsWirecited a 27 January 2017 InfoWars.com article, which more accurately held:
“A report by the Pew Center on the States finds that more than 1.8 million dead people are currently registered to vote, and 24 million registrations are either invalid or inaccurate,” NPR reported in 2012, which is ironic given how NPR is heavily controlled by Democrats.
"... The Pew study found that almost 3 million people are registered to vote in more than one state,” NPR added.
The purported "NPR study" determining that 25 million fraudulent votes were cast for Hillary Clinton in November 2016 was not compiled by NPR, had nothing to do with the 2016 election, and did not declare that 25 million votes had been determined to be fraudulent (in any election).
The referenced "NPR study" was in fact a 14 February 2012 NPR article about a Pew Research Center report issued in 2012. That Pew report observed that "approximately 24 million voter registrations in the United States are no longer valid or are significantly inaccurate," but the NPR article specifically noted that "there's little evidence that this has led to widespread voter fraud":
A new report by the Pew Center on the States finds that more than 1.8 million dead people are currently registered to vote. And 24 million registrations are either invalid or inaccurate.
There's little evidence that this has led to widespread voter fraud, but it has raised concerns that the system is vulnerable.
Election officials say one problem is that Americans move around a lot. And when they do, they seldom alert the local election office that they've left ... or if a person dies in-state — there's often a delay before election officials are alerted. It's also not always clear that the individual on the death certificate is the same one who's registered to vote. Election officials still have to do a lot more digging to avoid accidentally taking someone off the rolls who is very much alive.
Washington Secretary of State Sam Reed says it's amazing how many times his state has come across names on the voter rolls that appear to be the same person, but turn out not to be.
"We've even had cases, in very small counties, people [with the] same name and same birth dates," added Reed.
He said that has led to inaccurate reports that "dead" people are voting. He admits there have been a few cases in his state where widows or widowers have cast ballots for former spouses, but he said such fraud is very rare.
The continued inclusion of inaccurate or outdated registrations on voter rolls is a widespread problem that continues to plague many states and local voter boards, largely because of difficulties in matching up persons who have died or moved with the registration records they left behind. However, the existence of inaccurate or outdated registrations does not constitute voter fraud per se, nor is it evidence or proof that voter fraud is taking (or has taken) place. Voter fraud only occurs when someone uses an invalid registration to illegally vote &mdash by, for example, assuming another person's identity, voting multiple times, or voting in states in which they are no longer resident. The Pew report did not track or measure any such fraudulent voting activity.
The Pew Research Center report did not, as YourNewsWire falsely declared, state that 25 million fraudulent votes were cast in 2016 or in 2012 (the year the report was published) or in any other year, for Hillary Clinton or for any other candidate. The report antedated the 2016 election by more than four years, dealt only with voter registrations (not actual votes), and could not possibly be construed as documenting that 25 million fraudulent votes were cast for Hillary Clinton in 2016.