follow CCP

Recent blog entries
popular papers

Science Curiosity and Political Information Processing

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

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

« Weekend update: Does marching around in costumes help overcome cultural polarization? Comparative data might help answer this Q | Main | ICT eats RAT & CAT for breakfast: More (and more data on) religiosity, political predispositions, and "anti-science" »

Weekend update: Decisive proof of "conservative distrust in science"? You tell me...

This worked pretty well before, so why not try again?

Recently I posted some data on right-left political outlooks, religion, and positions on disputed science issues and asked you, the 14.33 billion readers of this blog, to say what inferences the data support.  The responses, including one submitted by Chris Mooney in his Washington Post Wonkblog, were really interesting, and led to an informative set of exchanges, which continued when I finally added my own assessment.  I'm pretty confident the discussion wouldn't have been as enlightening had I offered my own views of the data in the original post.

Well, here's some more data! 

These are all from the General Social Survey data set, 1973-2012.  The question are ...

Do these data support inferences on the cause of public controversy over science issues such as climate change?  In particular, how do they bear on the commonplace claim that such controversy originates in a growing "distrust"of or hostility toward science associated with right-of-center political outlooks?

Well, you tell me!  I'll say nothing, nothing for at least 48 hours (±6 hrs).


Okay, I said I would hold back for a bit on my own conclusions, but here are a few reactions to these very data from others, just to prime the pump of reasoned engagement with evidence:




 Hey--once again: prize to commentator who offers "best" interpretation (as determined by expert consensus survey)!

PrintView Printer Friendly Version

EmailEmail Article to Friend

Reader Comments (23)

Could probably plot the % of liberal/democrat scientists over time to current high of over 90% and realize that the spin that comes with their "science" is what republicans do not like.

November 22, 2014 | Unregistered CommenterBill

Should add that republican trust science but do not trust left-liberal scientists!

November 22, 2014 | Unregistered Commenterbill

@Bill-- so does that mean you agree w/ the representative commentators? Sounds like it, actually.

November 22, 2014 | Registered CommenterDan Kahan

Spinning my words - The terms "believe in Science " and "anti-science" are spin (deliberate attempt to belittle or slander). I do not expect you to be able to "see" or acknowledge your own blind spots but somewhere in the past 10 -20 years scientists quit debating the actual science started trying to dehumanize the other side. Pity the poor Republicans - they are just not able to grasp the problem or the science because of their mental issues. I will add a sarc tag to that for the non self aware.

November 22, 2014 | Unregistered CommenterBill

Dan -

First some housekeeping:

==> " If I observe the evidence doesn't actually support an answer to a disputed question, then it's a bit odd to be told by those who cite it that the evidence in question "misses the point..."

I don't think you "miss the point" (and that's also not exactly what I said). I don't know that there's a "the" point. And as far as the range of points go, I think you're directly on target for a lot of them.

What I am getting at is that I think that there's real evidence for a trend in this country in the interrelationship between political ideology and "science" as a broad and vague term, and so then I'm not convinced that it is meaningful to say that Americans in general are uniformly "pro-science" - which again I think is a very broad and vague term. It's important to consider that if we talk about "science" we are talking about some mixture of scientific institutions, scientists, scientific method, and scientific product. And I would prefer to use less/more trust in science as opposed to "anti-" or "pro-science."

==> "So tell me what it is about Gauchat that "gets to the point"?"

(Again, I didn't write what you put into quotes - and I don't think the distinction between that paraphrase of what I wrote and what I actually wrote is merely academic).

What I"m referring to can be summed up in this excerpt from Gauchat w/r/t his findings about a trend in how respondents answered questions about their levels of trust in science:

One possible interpretation, supported by a growing number of studies, is that social factors such as race/ethnicity, income, religiosity, social capital, and political identifications are at least as important as knowledge and education in predicting trust in science (Gauchat
2008, 2010; Sturgis and Allum 2004; Yearly 2005).

And from an interpretation of his findings:

Gauchat's findings may also reflect clashes between supporters of conservative ideology and particular areas of science. The rise of the Christian right, which rejects evolutionary theory and opposes research on embryonic stem cells, may have been particularly important – Gauchet found a similar decline in trust in science among frequent churchgoers. More recently, conservatives have come to fear that global warming will provide an excuse for "big government" to restrict their personal freedom through environmental regulation.

So I'm not suggesting that there's a "the point" here, but that included in the multiple influences in how Americans respond to questions about their level of "trust in science" are religiosity and political identification. And more to the point, I have an expectation of such for the following reasons:

There is, to some degree, an inherent tension between religion and science.

There is more specifically, a direct conflict between the "consensus" of scientists and some pretty important tenets of specific religiou faiths such as: whether evolution is the product of divine intervention, how the universe was created, whether homosexuality is a "sin" or the product of biology, etc. Other questions, such as whether stem cell research is a sin or whether it is a sin to produce and distribute birth control pills aren't exactly a direct conflict between specific tenets of faith and science, whether the bible is the word of god and should be interpreted literally, but they do drive a wedge between people who hold particular religious beliefs and the work product of some scientists

There has been, over the past 5 decades? or so, a concerted effort on the part of conservative thought leaders to exploit a deliberate politicization of religious believes for the purpose of partisan expediency.

There is a pretty widespread and apparent effect of those efforts - for example, a fairly common ideological theme that I hear from some conservatives that the country is deteriorating because of godless liberals, and godless academic liberal scientists (which means practically any academic scientists), and godless liberals at the CDC or the NIH or NASA or virtually any government-related scientific organization.

So IMO, it isn't that you miss the point and Gauchat gets it. It's that you are making some important observations that are helpful in uncovering problems with the arguments that "conservatives are anti-science" - but IMO, some of the conclusions that you are drawing don't explain a phenomenon that I think I am observing. Yes, my observations are anecdotal. But I know that I"m not dreaming up the often expressed antipathy from some conservatives towards scientific institutions and many scientists, and I think that the antipathy so easily stimulated from many conservatives towards the CDC from the Ebola scare, for example, is real and meaningful for some reason. Is there some trend in that antipathy? I don't know - but it does seem like there is one. What does that trend mean? I don't know. But I"m dubious about an analysis of trends in the interrelationship between ideology and "trust in science" in this country if it doesn't explain my observations.

Sorry if much of that was repetitious from things I've said before.


===> "Do these data support inferences on the cause of public controversy over science issues such as climate change?"

I think that climate change is a special case...I suspect that it is really hard to generalize about the cause of public controversy over science across different issues. I suspect that the "cause" w/r/t climate change is probably quite different than controversy w/r/t whether conducting medical research using stem cells is a sin. It could be that religious identity is more "causal" for the views of conservatives, generally, about stem cell research while political ideology is more "causal" for views w/r/t climate change. In that sense, "hostility" towards science might be more applicable for views on stem cell research than climate change.

==> "In particular, how do they bear on the commonplace claim that such controversy originates in a growing "distrust"of or hostility toward science associated with right-of-center political outlooks."

I"m not sure how this could ever be answered meaningfully without first establishing agreement among discussants, very specifically, what "hostility" towards science means. For example, does it mean something completely different than what might be reflected in answers to questions related to trends in hostility towards our existing societal scientific institutions? If someone were displaying hostility toward science, would the necessarily mean throwing out their smartphones and GPS', refusing to seek guidance from doctors who rely on medical research to inform their practice, etc.?

==> "In particular, how do they bear on the commonplace claim that such controversy originates in a growing "distrust"of or hostility toward science associated with right-of-center political outlooks?"

I think that this question is too broad to be really meaningful. Even getting past the ambiguity of what "science" means there - perhaps there are differences w/r/t distrust or hostility towards science (whatever that means) among different groups of conservatives. For example, how about highly religious conservatives? Or how about conservatives with a strongly libertarian bent?

November 22, 2014 | Unregistered CommenterJoshua

Dan -

I think you can probably just skip my previous (you were probably going to do so anyway after reading only the first couple of sentences!!!) and still get to my point by simply answering whether this statement:

I do not expect you to be able to "see" or acknowledge your own blind spots but somewhere in the past 10 -20 years scientists quit debating the actual science started trying to dehumanize the other side.


1) reflective of a "distrust of science" to any degree (with an understanding that there is obviously some level of distinction between "science" and "scientists") and,
2) a kind of statement that we would likely see increasing in frequency over the past 5 decades.

If the answers to both those questions are "yes," then why do you think #2 is the case?

November 22, 2014 | Unregistered CommenterJoshua

Should have been more specific:

2) a kind of statement that we would likely see increasing in frequency from a certain subsets of conservatives (i.e., libertarians and the religious right) in comparison to any trend in frequency of hearing that type of comment from other types of conservatives, moderates, and liberals over the past 5 decades?

And Bill - fyi, your own individual political ideology and/or religiosity wouldn't be particularly important for answering my questions.

November 22, 2014 | Unregistered CommenterJoshua


Do you think I only read first couple of sentences generally? :(

this is direct quote from your comment (more than a couple of sentences in!) that was basis of question that you are addressing now & saying reflects unfair reading:

I know that you have some questions about some aspects of Gauchet's analyses - but I think that some of his work supports an argument that jibes with my anecdotal take on the interface between politics and "trust in science," and that I think that you're kind of missing ...

You still haven't answered my question-- what exactly do you see in Gauchat that supports you & that I'm "missing"? (To prove I read: I reject the notion that citing Gauchat's own citiations of his or others works in bkrd or lit review engages his evidence; and you can see from this post what I think of relying on journalists' and other commentators' interpretations of studies it is unlikely they have actually read & for sure haven't critically engaged.) This woul indeed be a good place to finally do so, certainly.

I answered your latest question already in answer to @Bill.

But my point to *you* is that you should recognize the risk that the impressions you collect in this manner are based on biased sampling; you should test your impressions with evidence to see if they are right; & to be sure you aren't crediting evidence b/c it matches your impressions, you should actually look at it closely, & when others who don't see why it supports you say so, explain why it does

Do you think the data in this post corroborate your impressions?

November 22, 2014 | Registered CommenterDan Kahan


I get your point.

But here's a chance to actually address evidence. There is evidence in this post. Your suggestion that concservatives have reason to distrust scientists seems to concede tht the evidence presented in this post supports claims like "conservatives lose faith in science," "fewer than ever before believe in...," not to mention @Joshua's report of his impressions.

Do you mean to concede that?

November 22, 2014 | Registered CommenterDan Kahan

If I disapprove of the communication methods of someone on an ethical basis, then their technical arguments about science lose weight. There should be a complete distinction between science and scientists. Science by definition is not owned by any group and is never complete or perfectly understood and there is always more understanding to be discovered over every new hill.

It is pure slander to say that one group or another does not "believe in science".
What people mistrust is scientists who spin the science for political or financial advantage ( see any number of medical research scandals in recent years.)
The data evidence above shows an increasing distrust of scientists over time which coincidentally happens to correlate with an increase in politically active liberal scientists. It is pure politics to say that conservatives increasingly distrust science when they really are distrusting activists.

November 22, 2014 | Unregistered Commenterbill


Try harder.

November 23, 2014 | Registered CommenterDan Kahan

Consider the difference between "Science" and "What scientists say".

"On the one hand, as scientists we are ethically bound to the scientific method, in effect promising to tell the truth, the whole truth, and nothing but — which means that we must include all the doubts, the caveats, the ifs, ands, and buts. On the other hand, we are not just scientists but human beings as well. And like most people we'd like to see the world a better place, which in this context translates into our working to reduce the risk of potentially disastrous climatic change. To do that we need to get some broadbased support, to capture the public's imagination. That, of course, entails getting loads of media coverage. So we have to offer up scary scenarios, make simplified, dramatic statements, and make little mention of any doubts we might have. This 'double ethical bind' we frequently find ourselves in cannot be solved by any formula. Each of us has to decide what the right balance is between being effective and being honest. I hope that means being both."

If somebody read the above, and decided that they had less confidence in "what scientists say" as a result, would that reflect less confidence in "Science"?

Suppose they argue that Science was about being entirely honest, and that's what they trust, but if they discover there are some scientists who are striking some sort of "balance" and therefore they trust them less, is that being anti-science? Or just anti-scientist? Or could you argue that people who strike any sort of "balance" are not actually scientists, and therefore the survey participants shouldn't have been counting those?

Remember, this is a survey about "the people running these institutions..." It would be like arguing from poll results showing a loss of confidence in the President to justify headlines that the public have lost confidence in constitutional government. Some Presidents do seem to think "L'Etat, c'est moi," but it's not, really.

But besides that, I'd ask about error bars, sampling bias, post-processing (e.g. smoothing), changing definitions (are we talking about the same institutions 40 years ago and today? The same definitions of left/right?), and natural background variation. The level for "a great deal" varies between about 40% and 55% for the left, and between about 35% and 55% for the right, which is not a lot of difference, and the sample period is too short to be able to say this is the full range of variation. If these things wobble around randomly anyway, driven by a multitude of factors and perceptions, then maybe it's just coincidence that the data end point coincides with a dip.

I thought it interesting that the trend line for "liberal" was different to the one for "democrat". That suggests a shift in the population distribution of political opinion, which could cause changes in the correlation even if nobody actually changed their opinion on the people running the scientific community at all.


Although I will admit that I do like the headline "Conservatives lose faith in science over last 40 years". That's not a conclusion justified by the data presented here, but it might be the explanation. "Faith" is exactly the right word for what conservatives are complaining about, and faith ought to have nothing to do with science. Nullius in Verba.

November 23, 2014 | Unregistered CommenterNiV


I don't think having CIs helps much, or in any case are necessary to figure out whether the data support the inferences being drawn.

But here you go.

For sure the "differences" in probability that someone right of center would pick "great deal" changed "significantly," as did the difference in probability that right & left of center subjects would, over this very very very long time period (I don't see how you coull want more; the trend is real--issue is what sort of inference it supports).

This exercise is about thinking .P-values are usually a substitute for thought--people say, "Oh -- the result is significant! That proves it!" Proves *what*?

November 23, 2014 | Registered CommenterDan Kahan

"I don't think having CIs helps much, or in any case are necessary to figure out whether the data support the inferences being drawn."

In this case, no. As Bill has already said, the inferences drawn are obviously not supported because they're talking about different things. One is discussing the people running the scientific community, the other is discussing science. It's so obviously a different topic that I find it hard to believe that's really what you're asking.

I'm guessing you're looking specifically for another problem with these inferences, besides the obvious one. If so, you'll have to give me a bit more of a clue about what you're thinking of.

But generally, I always say that data without error estimates is meaningless. Are those percentages +/-1%, or +/-10%, or +/-50%? If you don't know, then what could you possibly conclude, even if the data was relevant to the question being asked? I have to say, though, they're wider than I was expecting.

"I don't see how you coull want more; the trend is real--issue is what sort of inference it supports"

Who says there's a trend? There's a change, certainly, but the concept of a 'trend' as determined by the OLS trend line calculation assumes that the outcome is a linear function of time plus identical, independent Gaussian errors. What makes you think so? And what if it's not?

An alternative model is trendless autocorrelated noise - that the value just wiggles slowly around some level, and the apparent 'trends' are spurious. There's no evidence here that the trend is 'real'. All you can say is that the level has changed.

"people say, "Oh -- the result is significant! That proves it!" Proves *what*?"

Proves that if whatever assumptions the CIs are based on are true, then something unlikely just happened which may be worth paying attention to. I agree with you, it's only ever one small part of the story.

But what I was thinking was that an absence of p-values is often an invitation not to think about the evidence, to just accept the conclusion.
Wouldn't you agree that can be a bigger problem?

November 23, 2014 | Unregistered CommenterNiV


I think p-values are lame. But I accept your point -- I think 0.95 CI or SE or something that enables visualization of precision of measure is usually appropriate.

But there are certainly situations where knowing precision of measure is pretty unimportant -- like where there is a huge N or where it's obvious from sign & size of effect that the data do/don't support the hypothesis -- & where CIs can just clutter things.

As for trend, just look at the raw data (the ones in the squares). You tell me what sort of hypothesis you have, & then we can fiture out how to model to test that.

If people can't figure out from just looking at the raw data what's going on here, then they are much more likely to be fooled that be protected from being fooled by analyst who applies a model

November 24, 2014 | Unregistered Commenterdmk38

Haven't read all comments carefully, but since it doesn't look like anyone else is taking the low(est?) hanging fruit: the data don't support the inference that conflicts originate in conservative distrust of science -- or even in conservative distrust of scientists. First, the difference between percent of liberals who trust scientists and the percent of conservatives who trust scientists (not to mention democrats and republicans, who are barely distinguishable) is pretty small and not close to the magnitude of the division over climate change. (Gallup's easy to pull up.) And only a tiny portion of conservatives hardly trusts scientists at all. So distrust in scientists could only be driving a little bit of division, if any. Second, there's no inflection point or other great change preceding conflict over climate change specifically (or any recent science-related cultural battle, from what I can see). I should be better versed in this, but climate change became divisive starting around the late 90s, right? Conservative opinion of scientists is pretty flat around then. So that's a weak origin story.

In sum, if I thought division in trust of scientists were driving conflict over climate change, I'd expect to see a large division between these groups, comparable to the division over climate change, starting around the time of the politicization of climate change or a bit before. And we don't see that. (Tell me if my "comparable size" expectation or others are ill-conceived!)

I do think there may be something to (a more modest version of) the very last comment in your examples, that some conservatives have started to resent certain scientists (see also, Bill's and NiV's comments). And it's not implausible that the trend in recent years relates to distrust of some climate scientists. It's only a small trend, but it's somewhat intriguing.

November 24, 2014 | Unregistered CommenterMW

"You tell me what sort of hypothesis you have, & then we can figure out how to model to test that."

The problem is that arbitrarily complicated hypotheses can be generated to fit any data set arbitrarily closely. You need an exogenous reason for picking a particular form of model, and I don't have one.

But for the sake of an example, let's pick ARIMA(1,1,0).

November 24, 2014 | Unregistered CommenterNiV


By the way, I'm intrigued - what's with the numbers on the x axis? What are the tick marks supposed to represent? And why?


Good point. Could the opinions have been diluted by all the other good scientists that conservatives still do have trust in?

November 24, 2014 | Unregistered CommenterNiV

Last figure first: The trust that Libs and Cons have placed in Big Capitalists changes fairly similarly with time.

In contrast, first figure shows that the trust that Libs and Cons have placed in Big Scientists has changed differently since sometime between 1980 and 1990. It's hard to say because these are just regression splines, but Libs seem to trust science more beginning sometime between 1980 and 1990, while Cons continue to increasingly mistrust science.

These data alone suggest 1) the interpretation that climate change is just the latest fight in a long series of increasingly divisive fights over the role of science in policy, and 2) combined with the unaccountably large partisan split over climate change, the hypothesis that Americans by and large trust science, yet mistrust different sectors of science differently by identification.

The middle two graphs with the regression lines are difficult to interpret. I have a request: denote the %age of category X that trust scientists a great deal by Trust_X. Could you please plot (Trust_Lib - Trust_Con) against (Trust_Dem - Trust_Rep)? I'm looking for an excursion from linearity as time proceeds, which may indicate when an issue that typically divides people only by worldview suddenly becomes more (or less) partisan.

November 24, 2014 | Unregistered Commenterdypoon


Y axis is sequence order of surveys. The survey was admnistered 27 times between 1974 & 2012. There were periods in which it was administered annually but it is now every-other-yr. So data are plotted in relation to "number in sequence 1-27." I've used year labels for reference.

November 24, 2014 | Registered CommenterDan Kahan

"Do these data support inferences on the cause of public controversy over science issues such as climate change? In particular, how do they bear on the commonplace claim that such controversy originates in a growing "distrust"of or hostility toward science associated with right-of-center political outlooks?"

No - I do not think these data support that inference because both sets of data infer a reduction in trust of the people running these institutions. The reduction of trust in executives follows from scandals like Enron, dotcom, etc and the reduction in trust of scientific community also can come from the many medical, fake peer review, etc. scandals as well as distasteful political activism.

November 24, 2014 | Unregistered Commenterbill


Am trying to get the request. You want 2 Figs, one for difference in % selecting great deal for lib & con & other difference in % for Dem & Repub, both over time. Is tht right?

You also want the labels to refer to "trust"? If so, why shouldn't the label correspond to the actual language of the item?

November 25, 2014 | Registered CommenterDan Kahan


Discussion moved here.

MW is winner (she selected Turing machine robo-dog; last one in stock...)

November 25, 2014 | Registered CommenterDan Kahan
Comments for this entry have been disabled. Additional comments may not be added to this entry at this time.