Well, things are going slowly in the kitchen, so here’s another “vaccine risk perception” appetizer — on the house

Okay, so my goal was to get a big (N = 2000) study that combines public opinion and experimental analysis of vaccine risk perceptions done by today.

I wanted to do that mainly so the evidence would be out there at the same time as my Perspective piece today in Science, which uses the HPV vaccine disaster and empirically uniformed risk communication about public attitudes toward childhood vaccines to draw attention to the need for a more systematic policy of “science communication environment protection,” both in government and in relevant professional and civic institutions.

But it’s easier to be in a magazine than to run one, which requires among other things meeting all kinds of deadlines etc.

I’m not going to meet mine for getting “the report” out.  I want to fine tune some things (including estimates made with survey weights that I’ve now fine tuned more precisely).  Maybe it won’t matter but I’d rather feel 100% comfortable before calling peopel’s attention to something that I hope can help them make decisions of consequence.

But I’m okay giving you a bit more to chew on — more “conventional wisdom” that has zero evidence and when examined turns out to be untrue (like the idea that there is some connection between positions on climate change & evolution & concern about vaccines).

Know how people say, “belief that vaccines cause autism is for the left what climate denial is for the right …” blah blah? I guess that’s based on a poll– of Robert Kennedy, Jr.

Here’s evidence from a nationally representative sample of 900 ordinary people. It’s a cool lowess plot that shows how political outlooks shape differences in perceived vaccine risk perceptions.

The y-axis uses the industrial strength risk perception measure for vaccines, global warming, guns, and marijuana legalization, and the x-axis is a continuous right-left ideology measure formed by aggregating party affiliation and liberal-conservative ideology.

Gee, becoming progressively more liberal doesn’t make people think childhood vaccines are more risky.

Actually, people become more concerned as they become more conservative.

But the effect is genuinely tiny —  as you can see by holding it up to comparison w/ other politically contested risks as a benchmark.

You can’t figure out the practical significance of variation by looking at a correlation coefficient or a complicated structural equation model. You have to know what sort of variance is being explained/modeled

Here it’s the difference between thinking it is genuinely asinine to worry about vaccines and thinking that it’s just really really dumb.

And to complement yesterday’s data, here is a look at how perceptions of the balance of vaccine risks and benefits (y-axis!) relate to science comprehension (measured with a pretty powerful composite scale that fortified the NSF’s science indicator battery with an extended “Cognitive Reflection Test” battery) and also to religiosity (again, a highly reliable composite scale, here comprising church attendance, “importance of God,” and frequency of prayer):

Well?  There are relationships– the balance tips a tad toward benefit as science comprehension increases and toward risk a tad as religiosity does.

But again, these are small effects, in statistical terms, and irrelevant ones in practical ones.  Those at both ends of both spectra are concentrated toward the “benefit greater than risk” end of the measure.

It’s not enough to explain variance; one has to know what the difference is that is being explained.

Actually, though, the religiosity & science comp relationship is more interesting than this picture lets on. It turns out that these two interact. So even though it looks like science comprehension has no effect, it does– but it depends on how religious one is!  Sound familiar?  Same thing as in climate change, where the impact of science comprehension turns on whether one has a cultural predisposition toward crediting or dismissing environmental risks.

Except not really:

This figure plots the interaction in relation to a composite scale that combines a bunch of indicia into a (very reliable!) scale that measures perception of the value of universal vaccination as a public health measure.  That scale is normalized — the units are standard deviations.  Same thing with the “science comprehension measure.”

So basically, we are talking about a shift of about 1/4 of standard deviation in every standard deviation of difference in standard deviation in science comprehension.

Hey– I could put three “***” next to the coefficient that measures the interaction b/c it is really really significant. But only in a “statistical” sense, not a practical one.

Unlike people who are below average in religiosity, people who are above average in religiosity don’t become even more enamored of vaccines as they become more science comprehending.  But everyone in this story loves vaccines– the mean on the scale reflects things like 75% of people agreeing with the statement that “I am confident in the judgment of the public health officials who are responsible for identifying generally recommended childhood vaccinations.”

Yeah but only super confident–why not super duper, like people who are below average in religiosity and above in science comprehension?

So maybe you see where this is going?

But actually, the report is not “all about nothing.”

The something has to do with what happens when you stick in people’s faces information that tells them that “anti-vaccine,” “climate change skepticism,” “denial of evolution” are all of a piece in some massive assault on science in our socciety….

So more on that. Tomorrow. I think!

Leave a Comment