Generalized trust in science: nukes vs. climate
Tuesday, May 30, 2017 at 8:50AM
Dan Kahan

Here’s another helping of “trust in science” data, this time compliments of Pew Research Center.

Yesterday's data“Yesterday”tm  I posted some data from the General Social Survey that showed that liberals but not conservatives become more inclined to worry about climate change as their trust in science increases. That’s a pattern that goes against one popular narrative, which attributes climate skepticism to an anti-science disposition on the part of right-leaning individuals.

But at least one commentator was understandably dissatisfied with the GSS outcome variable, which solicits on a five-point scale respondents’ assessments of how dangerous they  “think that a rise in the world's temperature caused by the `greenhouse effect', is for the environment.” As the commenter pointed out, this item presupposes that climate change is occurring, a proposition rejected by around 25% of the U.S. population.

Well, the Pew data (which, like the GSS, can download for free) contains a more conventional “belief in climate change” measure in which respondents can indicate whether there believe there is “solid evidence that the average temperature on earth has been getting warmer over the past few decades,” and if so, whether that change is “mostly because of human activity such as burning fossil fuels or mostly because of natural patterns in the earth’s environment?”

When one regresses on belief in human caused climate change on respondents’ political outlooks, their score on a trust in science scale, and the interaction of those variables, one sees pretty much the same pattern as in the GSS data.  That is, it looks like more left-leaning respondents are influenced by their level of trust in science but more right-leaning ones aren’t (in fact, the difference in how much the trust scale affected respondents conditional on their political outlooks was “significant” at only p = 0.10, but if are interested in what the analysis adds to the weight of the evidence, we shouldn’t get too hung up on that; maybe “tomorrow”tm I’ll elaborate on that).

That we got the same answer from two different data sets, which used different measures, should make us more inclined to reject a generalized “science trust deficit” theory of why conservatives are more climate skeptical. (Conservatives might be more “distrustful” of climate scientists, but that response would best be viewed as just an indicator, not an explanation, of the latent disposition toward skepticism on climate change ([Poortinga & Pidgeon 2005]).

Another commenter wondered what would happen if we substituted nuclear power for climate change in the GSS item, and in particular how liberals would respond.

Well, the GSS dataset contains such an item. It asks respondents to indicate, on the same 5-point scale, how dangerous they regard “nuclear power stations” to be for the environment.

Here at least, we see the “science trust deficit” doing what it is often advertised as doing—namely, predicting less concern among individuals as their trust level (measured as it was “yesterday”tm) increases.  It does so, moreover, for both liberals and conservatives:


What to make of this?  Well, you can tell me.

But I would say that, for myself, I was a bit surprised.  I was expecting to discover  that high levels of trust in science in general have no impact on disputed applications of particular forms of decision-relevant science.  Because more trust was associated with less concern for both liberals and conservatives, I have less confidence than I did that a generalized trust measure is of little use.

BTW, you can find the 7 items I used to form the Pew “trust in science” scale here.  The dataset had more candidate trust items in it, but this combination of items displayed the highest reliability score (α = 0.69).


Poortinga, W. & Pidgeon, N.F. Trust in Risk Regulation: Cause or Consequence of the Acceptability of GM Food? Risk Analysis 25, 199-209 (2005).




Article originally appeared on cultural cognition project (
See website for complete article licensing information.