This is part 2 of the (or an) answer to the question: “Why cultural worldviews rather than liberal-conservative or Democrat-Republican?” in our studies of risk perception & science communication.
In the last post, I connected our work to Aaron Wildavsky’s surmise that Mary Douglas’s two-dimensional worldview scheme would explain more mass beliefs more coherently than any one-dimensional right-left measure. (BTW, I don’t think our work has “proven” Wildavsky was “right”; in fact, I think that way of talking reflects a mistaken, or in any case an unappealing understanding of the point of identifying the sources of public contestation over risk, something I’ll address in the final installment of this series of posts.)
Part 2: Motivated system 2 reasoning
I ended that post with the observation that the cultural cognition worldview scales tend to do a better job in explaining conflict among individuals who are low in political sophistication. In this post, I want to suggest that cultural worldviews are also likely to shed more light on conflict among individuals who are high in technical-reasoning proficiency—or what Kahneman refers to as “system 2” reasoning.
In Kahneman’s version of the dual process theory, “System 2” is the label for deliberate, methodical, , algorithmic types of thinking, and “System 1” the label for largely rapid, unconscious, heuristic-driven types. (Before Kahneman, a prominent view in social psychology called these “systematic” and “heuristic” processing, respectively.) Kahneman implies that cognitive biases are associated with system 1, and are constrained by system 2—or not, depending on how disposed and able people are to think in a rigorous, analytical manner.
Our work (consistent with—indeed, guided and informed by—the earlier dual process work) suggests otherwise. We have examined how cultural cognition interacts with numeracy, a form of technical reasoning associated with system 2. What we have found (so far; work is ongoing) is that individuals who are high in numeracy are more culturally polarized than those who are low in numeracy.
To us, this shows that those who are more adept at System 2 reasoning have a unique ability— if not a unique disposition—to search out and construe technical information in biased patterns that are congenial to their values. In effect, this is “motivated system 2 reasoning.” It is as much a form of “bias” as any mechanism of cultural cognition that operates through system 1 processes (although whether it makes sense to think of either system 1 or system 2 mechanisms of cultural cognition as “biases” is itself a complicated matter that depends on what we understand people to be trying to maximize and on how we ourselves feel about that).
It’s not clear to me that political-party identity or liberal-conservative ideology can account for motivated system 2 reasoning. Indeed, as I discussed in connection with John Bullock’s interesting work, the juxtaposition of partisan identity with measures of reasoning style like “need for cognition” seems to produce results that are simply unclear (although intriguingly so).
“Need for cognition” & other quality-of-reasoning measures that rely on self-reporting might be less helpful here than ones that rely on objective or performance-based assessments. Numeracy is one of those.
In some new analyses of data collected by the Cultural Cognition Project, I looked at how CRT measures (a subcomponent of our numeracy scale) relates to the cultural worldview measures. I found that Hierarchy and Individualism were both correlated with CRT— but that they had opposite signs— positive in the case of Hierarchy, negative in the case of Individualism.
I also found that a scale that reliably combined Republican party affiliation/conservative ideology (α = 0.75) was correlated with CRT in the positive direction. This is probably not the association one would expect, btw, if one subscribes to the “asymmetry” thesis, which sees political conflict over risk and related facts as linked to reasoning deficiencies unique to conservative thought.
And the package of correlations doesn’t bode well for any one-dimension left-right measure as a foundation for explaining risk perception & science communication. For if System 2 reasoning does have special significance for the sort of conflict that we see over climate change, nuclear power, etc., then a one-dimensional measure that merges Hierarchy & Individualism into a generic “conservativism” will be insensitive to the potentially divergent relationships these dispositions have with the system 2 reasoning style.
Enough! (for now anyway)