This so-called article is a piece of crap.
It purports to provide the results of a study and conflates the numbers in the study with society as a whole in ignorant ways.
For example, second paragraph:
Just ask college students. A fifth of undergrads now say it’s acceptable to use physical force to silence a speaker who makes “offensive and hurtful statements.”
A fifth of undergrads? No. A fifth of the 1500 undergrad students they surveyed. That's 300 or so.
Villasenor conducted a nationwide survey of 1,500 undergraduate students at four-year colleges.
Nationwide? There are far more than 1,500 four-year colleges (for those of you not American, the word includes universities). How were the colleges chosen? How were the students chosen? How many were chosen at each university? How many overall were from the same discipline? There's no way to know. We don't even know if he chose accredited schools, or those pay-for-a-degree places. Did they ask at Ivy League schools, the majority of whose students come from well-off families? Did they ask at places like City College of New York, where the tuition is much lower and people who are there are from a variety of backgrounds, not wealthy? Ag and tech colleges, out in the countryside, or only urban colleges?
Further down it says the margin of error is 2-6 percent, "depending on the group." Oh, really? Which group is 2% and which is 6%? We aren't told. It appears we are to be grateful that a margin of error was even mentioned.
The whole thing is supposed to be about undergrads' understanding of First Amendment-protected free speech. Since we are not told the exact wording of the questions asked, it's impossible to know if the responses were appropriate to them, or if the questions were leading the students to a specific response.
And then there's this:
Let’s say a public university hosts a “very controversial speaker,” one “known for making offensive and hurtful statements.” Would it be acceptable for a student group to disrupt the speech “by loudly and repeatedly shouting so that the audience cannot hear the speaker”?
Astonishingly, half said that snuffing out upsetting speech — rather than, presumably, rebutting or even ignoring it — would be appropriate. Democrats were more likely than Republicans to find this response acceptable (62 percent to 39 percent), and men were more likely than women (57 percent to 47 percent). Even so, sizable shares of all groups agreed.
It gets even worse.
Respondents were also asked if it would be acceptable for a student group to use violence to prevent that same controversial speaker from talking. Here, 19 percent said yes....
Let's look more closely, ignoring the editorializing sentence for the moment. Half of who? Half of 1500 people is 750 people, scattered across the US. And then again -- 19% of who? Everyone? Women? Men? Democrats? Republicans? We aren't told.
Meanwhile, the entire other side of this survey is ignored. By stressing the minority and ignoring the majority, the minority's views are inflated and made more important. Let me turn this around for you: more than 80% of undergrads say that violence is not acceptable in dealing with an unwanted speaker. Try turning around all the other numbers, and the story falls apart. Instead of "students" substitute "students surveyed", and it also falls to pieces. Who cares what 1500 people out of 200 million think? If we don't know why those 1500 were specifically chosen, why should we care?
I have worked with surveys, written surveys, conducted and analyzed surveys. It is possible to have a statistically perfect survey with 1500 people surveyed, but only if the respondents are very carefully selected to avoid bias. There is no way to tell if that was done with the evidence given in this story. For all we know, those respondents could have been selected from the same departments or majors at all the colleges. The colleges could have been technical schools or enormous state universities or religion-affiliated schools. There is no way to know. Why does this matter? Liberal arts, political science and pre-law students are more likely to have read about the First Amendment than optics majors or engineers, for instance. I'm not saying the optics majors or engineers would be more conservative or liberal -- but they are less likely to have discussed free speech in a class. Improper choice of respondents can provide very slanted results -- for example, the survey that said Dewey would win over Truman was conducted by telephone, and the calls went to houses on the corners of two streets; this meant that people who were wealthier (because corner houses pay higher taxes, based on road frontage) were questioned, while their less wealthy neighbors (who voted for Truman) were ignored.
Also, by not including any context relative to current events, there is no way to know if the small percentage who thought violence was acceptable was the same as during the Vietnam War, for instance, or Desert Storm. I guarantee you, it was not the same percentage as during the Revolutionary War, when those who spoke against any prevailing view to an audience who disagreed would have been lucky to have been ridden out of town on a rail, if not tarred and feathered. (Feel free to do the research if you wish; be sure you have a strong stomach for the details of what happens when boiling tar is applied to skin.)
What it all comes down to is this: this story is written poorly by someone who does not understand how statistics should be used, and was not properly edited. It was published in order to scare people, although the publisher may not have realized its propaganda value. By not including the whole story, and by allowing editorializing in the middle of it, it slants the results.
This would not have been published during the time when Kay Graham was publisher. Editor Ben Bradlee would not have let this story pass. He would have told the reporter to rewrite it, clean it up, and get more depth into it.
And the reason I am writing this is that this is not the only paper that misleads with statistics, and you need to be aware of this, and of what to look for when someone is quoting a study, badly, misleadingly, in a way that bids fair to be used for propaganda. Be cautious and critical when you see numbers and statistics, and look for whether the writing is made personal/editorialized. It matters.