This question was on an online exam given recently to more than 3,000 American high school students. I think it exposes a threat to both our education system and our national security.

“Below is a screenshot of a Facebook page from October 2016. . . . You can watch the video from the post here: (Note, there is no audio.)” The grainy video showed election workers surreptitiously stuffing ballots into bins. The video said this happened during a Democratic primary election in 2016. Across the screen it said: “Have you ever noticed that the ONLY people caught committing voter fraud are Democrats?”

Students were asked: “Does the Facebook post provide strong evidence of voter fraud during the 2016 Democratic primary election?” They could check the video for evidence and search the Internet for relevant background.

The correct answer was no. The clips actually showed a polling place in Russia, which the students could have verified online. Yet 52 percent of that large and representative sample of American youths — children of the Internet Age — said yes. A quarter of them rejected the video but could not say why. About 9 percent gave a relevant explanation for their doubts, such as lack of context. Only 3 out of 3,119 respondents found the BBC news story that exposed the Russian fraud.

This is only part of a distressing study, “Students’ Civic Online Reasoning: A National Portrait,” released by Stanford University researchers Joel Breakstone, Mark Smith and Sam Wineburg. Results from other questions — on the validity of websites on gun control and on global warming, of a tweet on National Rifle Association member opinions and differentiation of advertisements and stories on Slate — were similarly disappointing.

In another survey, the researchers found 60 percent of a group of Stanford freshmen thought the socially conservative 500-member American College of Pediatricians was a more reliable source than the mainstream 64,000-member American Academy of Pediatrics.

The study concluded that these failures to catch falsehoods in social media are not from lack of intelligence but from lack of training. When asked to navigate unknown websites, professional fact-checkers acted very differently when the researchers compared them with Stanford freshmen and university professors at four different institutions given the same tasks.

The fact-checkers left the site they were supposed to examine quickly and “opened new browser tabs to search for information about the trustworthiness of the source,” the study said.

“In contrast,” it said, “the students, as well as the academics, typically read vertically, examining the original site’s prose, references, About page, and top-level domain (e.g., .com vs. .org) — features that are all easy to manipulate.”

I asked Wineburg, Margaret Jacks professor of education at Stanford, if the many high school students who thought the fake primary election video was legitimate might have been too lazy or unmotivated to take the exam seriously. After all, it wasn’t affecting grades on their report cards. Wineburg said he didn’t think so. He said the students were volunteers and rated themselves highly as Internet researchers, on average 7.31 points out of 10 in that skill.

“In most cases, unless it is directly pointed out to them,” Wineburg said, “students have no idea how lacking they are in the evaluation of digital material. . . . You’ll find very few civics/government classes in this country that are explicitly teaching students how to become reliably informed citizens using the medium of a screen.”

The study revealed that school lessons on Internet use commonly don’t tackle real sources. Instead, students are given multiple-choice questions or true/false items, and not asked to investigate actual online material.

How about creating a ninth-grade social studies lesson in online sleuthing required of every student? A month of examining fake videos and attractive but inadequate Web pages, and being forced to see their errors, might have an effect.

Break the class into teams. Make it a competition. There are an infinite number of fraudulent pages on the Web, ready for detection by teenagers whose energetic skepticism, in that case, could improve their grade.