follow CCP

Recent blog entries
popular papers

What Is the "Science of Science Communication"?

Climate-Science Communication and the Measurement Problem

Ideology, Motivated Cognition, and Cognitive Reflection: An Experimental Study

'Ideology' or 'Situation Sense'? An Experimental Investigation of Motivated Reasoning and Professional Judgment

A Risky Science Communication Environment for Vaccines

Motivated Numeracy and Enlightened Self-Government

Ideology, Motivated Cognition, and Cognitive Reflection: An Experimental Study

Making Climate Science Communication Evidence-based—All the Way Down 

Neutral Principles, Motivated Cognition, and Some Problems for Constitutional Law 

Cultural Cognition of Scientific Consensus

The Tragedy of the Risk-Perception Commons: Science Literacy and Climate Change

"They Saw a Protest": Cognitive Illiberalism and the Speech-Conduct Distinction 

Geoengineering and the Science Communication Environment: a Cross-Cultural Experiment

Fixing the Communications Failure

Why We Are Poles Apart on Climate Change

The Cognitively Illiberal State 

Who Fears the HPV Vaccine, Who Doesn't, and Why? An Experimental Study

Cultural Cognition of the Risks and Benefits of Nanotechnology

Whose Eyes Are You Going to Believe? An Empirical Examination of Scott v. Harris

Cultural Cognition and Public Policy

Culture, Cognition, and Consent: Who Perceives What, and Why, in "Acquaintance Rape" Cases

Culture and Identity-Protective Cognition: Explaining the White Male Effect

Fear of Democracy: A Cultural Evaluation of Sunstein on Risk

Cultural Cognition as a Conception of the Cultural Theory of Risk

« Fracking freaks me out | Main | If you think GM food & vaccine risk perceptions have any connection to the "climate change risk perception" family, think again »

Who fears what & why? Trust but verify!

Patrick Moore, aka "@EcoSenseNow," posed this question to me:


Probably Patrick & a friend were involved in a discussion about whether those who are (aren't) concerned about climate change are the "same" people who are (aren't) concerned about nuclear power and GM food risks.

A discussion/argument like that is pretty interesting, if you think about it.

We all know that risk perceptions tend to come in intriguing packages -- intriguing b/c the correlations between the factual understandings they comprise are more plausibly explained by the common cultural meanings they express than by any empirical premises they share. 

E.g., imagine you were to say to me, "Gee, I wonder whether crime rates can be expected to up or instead to go down if one of the 40 or so states that now automatically issue a permit to carry a concealed handgun to any adult w/o a criminal record or a history of mental illness enacted a ban on venturing out of the house with a loaded pistol tucked unobtrusively in one's coat pocket?"

If I answered, "Well, I'm not sure, but I do have some valid evidence that human activity has caused the temperature of the earth to increase in recent decades--surely you can deduce the answer from that," you'd think either I was being facetious or I was an idiot (maybe both; they can occur together--I don't know whether they are correlated).

But if I were to run up to you all excited & say, "hey, look--I found a correlation between believing that the temperature of the earth has not increased as a result of human causes in recent decades and believing that banning concealed handguns would cause crime to increase," you'd probably say, "So? Only a truly clueless dolt wouldn't have expected that."

You'd say that -- & be right, as the inset graphic, which correlates responses to the "industrial strength risk perception measure" as applied to "private ownership of guns" and "global warming," illustrates -- b/c "everyone knows" (they can just see) that our society is densely populated with "types of people" who form packages of related empirical beliefs in which the reality & consequences of human-caused climate change are inversely correlated with beliefs about the dangers posed by private ownership of handguns in the U.S.

The "types" are ones who share certain kinds of commitments relating to how society and other types of collective enterprises should be organized.  We can all see our social world is like that but because we can't directly observe people's "types" (they & the dispositions they impart are “latent variables”), we come up with observable indicators, like cultural worldviews" &/or "political ideologies" & various demographic characteristics, that we can combine into valid scales or classifying instruments of one sort or another. We can use those to satisfy our curiosity about the nature of the types & the dynamics that generate the puzzling pattern of empirical beliefs that they form on certain types of disputed risk issues.

We can all readily think of indicators of the sorts of “types” whose perceptions of the risks of climate change & guns are likely to be highly convergent, e.g.

Those risks are "politicized" in right-left terms, so we could use "right-left" political outlooks to specify the "types" & do a pretty decent job (a walk or bunt single; hey, it’s spring training!).

We could do even better (stand-up double) if we used the cultural cognition "worldview" scales -- & if we tossed in race & gender as additional indicators (say, by including appropriate cross-product interaction variables in a regression model), we'd be hitting a homerun!

But here’s another interesting thing that Patrick’s query—and the argument I’m guessing was the motivation for his posing it: our perceptions of the packages and the types aren’t always shared, or even when widely held, aren’t always right.

Not that surprising, actually, when you remember that the types can’t be directly observed. It helps too to realize that the source of our apprehension of these matters—the packages, the types—is based on a form of sampling rife with potential biases.  The “data,” as it were, that inform our perceptions are always skewed by the partiality of our social interactions, which reflect our propensity to engage with those who share our outlooks and interests. 

That sort of “selection bias” is a perfectly normal thing; only a lunatic would try to “stratify” his her social interaction to assure “representativeness” in his or her personal observations of how risk perceptions are distributed across types of persons (I suppose one could try applying population weights to all of one's interactions, although that would be time consuming & a nuisance).

But it does mean that we’ll inevitably disagree with our associates now & again—and even when we don’t disagree, all be wrong—about who fears what & why.

E.g., many people think that concern over childhood immunizations is part of one or another risk-perception package held by one or another recognizable “type” of person. 

Some picture them as  part of the package characteristic of the global-warming concerned, nuclear-power fearing tribe of “egalitarians, [who] oppose . . . big corporations and their products.”

When others grope at this particular elephant, they report feeling the “the conservative don’t-tread-on-me crowd that distrusts all government recommendations”—i.e., the same “type” that is skeptical of climate-change and nuclear-power risks.

Well, one or the other could have been right, but it turns out that they are both just plain wrong.

As the CCP report on Vaccine Risk Perceptions and Ad Hoc Risk Communication documents, all the recognizable “types”—whether defined in political or cultural terms—support universal childhood immunization.

The perception that vaccines cause autism is not part of the same risk-perception package as global warming: climate-change skeptics and climate-change believers both overwhelmingly perceive the risks of childhood immunizations to be low and the benefits of them to be high.

The misunderstandings about who is afraid of vaccines and why reflects selection bias in an echo chamber, reinforced by the reciprocal recriminations and expressions of contempt that pervade climate change discourse and that fill members of each with the motivation to see those on the other as harboring all sorts of noxious beliefs and being the source of myriad social ills. (Is this a new thing? Nope.)

So … back to Patrick’s question!

It’s not news—it’s a staple of the public study of risk perceptions and the cultural theory of risk in particular—that perceptions of climate-change and nuclear-power risks are part of a common “package” and are associated with distinctive types.

So my guess is that either Patrick or his friend (the one he was having an argument with; nothing inherently unfriendly about disagreeing!) was taking the position that GM-food risk perceptions was part of that same package as climate & nuclear ones.

Actually, the view that GM foods are “politically polarizing” is a common one.  “Unreasoning, anti-science” stances toward GM foods, according to this view, are for “liberals” what “unreasoning, anti-science” stances toward climate are for “conservatives.

But this is the toxic echo chamber once again.

As the 17.5 billion regular followers of this blog know (welcome, btw, to new readers!), GM foods get a big collective “enh,” at least in the view of the general public.  Most people have never really heard of GM foods, and happily consume humungous helpings of them at every meal.

Advocacy groups of a leftish orientation have been trying to generate concern—trying, moreover, by resort to exactly the “us-vs-them” incitement that is poisoning our science communication environment—but remarkably have been getting absolutely nowhere.

Here in the U.S.; matters are different in Europe. Why there but not here?! These things are truly mysterious—and if you don’t see that, you get a failing grade on the basic curiosity & imagination aptitude test.

Here are some data to illustrate that point and to answer Patrick’s question.

First, look at “packages”: 

Here gun-possession, nuclear, GM-foods, and childhood-vaccine risk perceptions are plotted in relation to climate change risk perceptions (the plotted lines reflect locally weighted regression -- they are "truer" to the raw data than a lnear regression line, reflecting the correlation coefficient I've also reported for each, would be).

Yes, GM food risk perceptions are correlated with global warming ones.  But the effect is very modest. It’s nothing like correlation between guns and climate change or nuclear and climate change.  You’ll find plenty of people—ones without two heads and who don’t think contrails are a government plot—who think climate change is a joke but GM foods a serious threat, and vice versa.

It’s really not part of the “climate change risk perception family.” 

How about in terms of “type”?

Enlarging a bit on some data that I’ve reported before, here are various risk perceptions plotted in relation to conventional left-right political views (measured with a composite scale that combines responses to party-identification and liberal-conservative ideology items):

Pretty clear, I think, that GM foods is just not a left-right issue.

As regular readers know, I’ve also examined GM food risks in relation to other types of “type” indicators, including the cultural cognition worldview scales and “interpretive community” scales derived from environmental risk perceptions.  It just doesn’t connect in a practically meaningful way.

So what to say?

Well, for one thing, there’s certainly no reason for embarrassment in finding out that things aren’t exactly as one conjectured on these matters.

As I said,  “risk packages”—because they reflect unobservable or “latent” dispositions, and because we are constrained to rely on partial and skewed impressions when we observe them—definitely have fuzzy peripheries.

In addition, the packages breed dynamics of misinformation, including the echo chamber effect and strategic behavior by deliberate science-communication environment polluters.

Under these circumstances, we should all adopt a stance of conscious provisionally toward our impressions here.  We shouldn’t “disbelieve” what our senses tell us, but we should expect evidence now & again that we have misperceived—and indeed, seek out such evidence before making decisions of consequence that turn on whether our perceptions are correct.

In the words of a famous scholar of risk perception—I can’t remember his name; early sign of senility?nah, couldn't be!—said (in some other language, but this is rough translation), “trust but verify!”

Maybe it’s just me, but I actually love it when evidence bites me in the ass on something like this.

Not just because I want to be sure the beliefs I hold are free of error, although of course I do feel that way.

But because every time the evidence surprises me I experience anew the sense of wonder at this phenomenon.

What is going on here?!  Why are there packages? Who are the “types”?

Why do some “risks”but not others become entangled in conflicts between diverse groups—all of which are amply stocked with individuals who are high in science comprehension and all of which have intact practices for transmitting what’s collective known to their members?

I really want to know the answers—and I know that I still just don’t!

“Tomorrow,” in fact, I’ll show you something that  is definitely freaking me out

PrintView Printer Friendly Version

EmailEmail Article to Friend

Reader Comments (5)

This may be more relevant to "tomorrow"'s post (so I can wait til then to elaborate as needed) but I've been thinking more about the concept of agency (or perceived agency) on the part of an individual as they assess risk... and whether the collective sense of agency toward avoiding negative outcomes influences how polarized a topic becomes in the minds of people of different cultural worldviews. In your experience- what topics are the most polarized? Recalling past diagrams you've shown, I'm thinking climate change, guns, nuclear power, water pollution? Other biggies? Does fracking land somewhere similar to water pollution? Is it more severe in polarization than mere "water pollution?" Just curious.

March 10, 2014 | Unregistered CommenterJen

The plots are interesting. Could you show us error bars on the fits? It is hard for me to look at the plots and have a feeling of how many other plots decently fit the given data. Since I don't know the error bars, it is hard to know how many conclusions statistically fit the data.

March 11, 2014 | Unregistered CommenterEric Fairfield

"Probably Patrick & a friend were involved in a discussion about whether those who are (aren't) concerned about climate change are the "same" people who are (aren't) concerned about nuclear power and GM food risks."

Or they could have been having a discussion about whether anti-GMO fanatics are overwhelming lefties. Or they could have been having a discussion about whether environmentalists are overwhelmingly anti-GMOers. Or they could have been having a discussion about whether anti-GMOers are overwhelmingly environmentalists.

I have had that discussion a number of times. For example, here's one where I had a convo about the following comment:

The curious thing is that climate sceptics generally are not anti-GMO and not anti-vaxxers. Greens tend to be believers in CAGW, dangers of GMOs and vaccines.

The comment thread:

See the comments from J M, JH, and Tom Scharf.

Now your data do not single out the demographic of "environmentalist" as a particular variable of interest, but do you think that I am wrong that it would be unlikely that issues such as climate change and nuclear energy would have such a clear signal in your data, and GMOs would show absolutely no signal, if anti-GMOism was as characteristic of, and dominated by, environmentalists as my interlocutors claim?

And Dan -

"Advocacy groups of a leftish orientation have been trying to generate concern—trying, moreover, by resort to exactly the “us-vs-them” incitement that is poisoning our science communication environment—but remarkably have been getting absolutely nowhere."

I think that is very one-sided. It isn't only "groups of a leftish orientation" that are trying to portray GMOs as a polarizing issue - it's people like Keith Kloor, it's "conservatives" of the sort who were commenting in that thread, and the stances reflects how many "conservatives" approach many environmental issues - such as the use of DDT, the banning of CFCs, etc. I don't think that it is particularly productive for you to speak of that as a characteristic of "groups of leftish orientation" w/o grounding that in the larger context. In fact, I think that doing could only increase the polarization - with "lefties" getting defensive and "righties" feeling vindicated for their attempts at incitement.

March 11, 2014 | Unregistered CommenterJoshua


As I'm sure you vividly recall, if one constructs an "environmental risk concern" scale out of concerns about global warming, nuclear power, toxic waste disposal, and air pollution (all of which are very much part of one happy or terrified or pissed off or whatever family), then it does correlate positively w/ concern about GM food risk concerns. But the correlation isn't nearly as big as the correlation between the perceptions that the scale comprises; one wouldn't want to treat GM food risks as an "indicator" of the latent disposition measured by the scale. But yes, you'd find more GM food concerned people among those who score highest on that measure.

On Keith, I still am inclined to disagree. If he inferred, as many do, that being left of center correlates meaningfully w/ anxiety of GMO from the fact that anti-GMO activists tend to be "left" in orientation, then he'd have been wrong. I don't know if that's what he's saying, but I don't think it matters in the conversation we've been having about his GMO writings.

My impression (show me if I'm wrong) is that he is angry at activists who are trying to glue GM foods to climate, nuclear etc. I think he is very right both to see those activists as left in their orientation and to be mad at them. If there are those on the right trying to ideologically "brand" positions on the issues, they are #scicomm environment polluters too.

But trying to fight polluters is not the same thing as trying to antagonize or provoke or stigmatize the group the polluters are associated with (& generally are betraying by abusing the trust placed in them by those who rely on them for information).

I don't think that Kloor's writings are goading ordinary left-leaning folks into seeing GM food risks as one of "their tribe's" issue or that there is any potential it could. He's saying that it's a mistake for advocates to be trying to create tribes here.

What am I missing?

March 11, 2014 | Registered CommenterDan Kahan

@ Eric--

Here the issue isn't precision of estimates but directions of signs, so I don't think it adds any info to put in error bars or CIs & in fact just clutters things (I know you are not one of them, but some people think leaving out confidence intervals when one plots a line is like blaspheming -- they see statistics as a set of religious injunctions to be piously observed rather than a set of tools for helping to guide, orient, discipline inference).

The error bars or CIs will be pretty small b/c the sample size is so big (you can see the p values are 0.00x...) But here you go, by way of illustration.

As for "fit," the OLS "linear" regression lines are necessarily "best fitting" unless one thinks a nonlinear model would be better, something one could get a sense of by looking at the lowess lines for the first set of graphs. I think the lowess lines help to show that for vacs & GM foods, across the vast expanse of x, there's nothing interesting happening at all in y (especially for vacs).

March 11, 2014 | Registered CommenterDan Kahan

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>