No, that's not my point. I don't know what real thing(s) make up the factor that factor analysis spits out and we label Openness, but my point is that if openness, novelty, new ideas, etc, really were universally valued, then there could be no such factor because everyone would recognize the Openness-loaded questions (they're transparent) and answer maximally. The fact that an Openness factor emerges out of pretty transparent questions shows that in the general population, there are a lot of people who will freely tell you on a questionnaire that they aren't keen on new ideas, who don't feel the need to seek out novelty, who won't be particularly creative on anything, and are fine with all that.

Let me rephrase it a little more subtly.

Suppose someone said, 'In general, I like to think I support free speech. I support letting Nazis march in Skokie, I agree with many of the ACLU's lawsuits, I'm horrified by what happened to Brendan Eich, etc. And yet, doesn't everyone think they support free speech? Can you imagine ever saying that free speech is a bad thing and unpopular groups should not be allowed to speak? I sure can't. So maybe I'm not nearly as liberal and tolerant as I think, and I'm merely engaged in self-congratulations on how wonderful I am!'

Sounds plausible, right? Just ask the neo-reactionaries how far free speech goes, right?

Except many, if not most, Americans hold views on free speech that would horrify you and me. (We don't even need to go outside the West to make this point.) For example, the General Social Survey asks Americans about free speech; using 2000+ data (so hardly historic), Razib Khan does some analysis. He's interested in splitting by intelligence so there's not an overall position unless I want to do more work extracting weighted averages than a comment is worth, but just scroll down to the first table and look at the various breakdowns:

• 2-26% of Americans thinks gay people should not be allowed to speak
• 3-36% of Americans think that atheists like me (& 79.2% of LW) are better seen than heard
• 7-49% think militarists should shut up
• 16-48% of Americans disagree with free speech for racists
• 16-75% claim Muslims should not be permitted to preach against America

If we were to poll these questions on LW, I would be surprised if any category had >10%, and 49% (or 75%...) would be completely beyond the pale. Yet over these categories, probably a majority of the American populace holds at least one of these illiberal positions.

(Nor is free speech an exception; I have seen in a number of places that the majority of Americans often hold views I find cruel, disgusting, and despicable.) Surprising as it may seem, a lot of core values on LW are not universal even in the American populace. So yes, it absolutely is possible to be more open-minded than most people; more supportive of free speech than most people - in fact, it's not even hard. (Overestimating is a sin, but so is underestimating.)

What this example shows is that it's not as easy to apply social desirability bias as it may seem. It is socially desirable in the bubbles that LWers inhabit to support free speech, and so we might expect the true level of support for free speech to be lower than that expressed verbally, but these bubbles still filter for people with unusual support for free speech and given the general population survey numbers, we expect the true LW support to still be much higher than the general population support (even though it's hard to realize just how divergent we are because of the bubbles).