Since I’m getting ready to return from a trip to Europe, I thought it would be a good time to mention the work that CCP has been doing to investigate “cross-cultural cultural cognition.”
In our research, we use two scales—“Hierarchy-egalitarianism” (HE) and “Individualism-communitarianism” (IC)—to measure the “worldviews” featured in Douglas & Wildavsky’s (CTR). HE and IC (in the form of factor scores extracted from a collection of attitudinal items) are used as predictors to test various hypotheses about how group predispositions influence perceptions of risk and related facts.
“Cross-cultural cultural cognition,” as I’m using this term, involves applying the same methods to non-U.S. samples. In this first of two posts, I’ll describe some of the key theoretical/conceptual issues involved in cross-cultural cultural cognition. In the second, I’ll show some results for studies involving test subjects in the UK and Australia.
Part 1: What generalizes and what doesn’t
The point of “cross-cultural” study of cultural cognition, of course, is to identify the extent to which the dynamics we observe in our studies generalize across societies. But to avoid confusion, it’s necessary to frame the “generalizability” question in reasonably fine-grained terms. The approach we are using to engage in cross-cultural study of risk perceptions addresses generalizability separately with respect to three elements of the cultural cognition framework: (1) motivating dispositions, (2) disposition indicators, and (3) culture-risk mappings.
A. Motivating dispositions
“Motivating dispositions” refer to the group affinities that orient individuals’ perceptions of risk. In the cultural cognition framework, these dispositions are the CTR worldviews that we measure with the HE and IC scales. The dispositions are described as “motivating” because they are what orient the various modes of cognition that unconsciously link cultural worldviews to perceptions of risk and related beliefs.
Cross-cultural cultural cognition—at least as I’m using the concept here—posits that the dispositions featured in CTR do generalize across societies. In other words, we should expect the worldviews of every society’s members to vary systematically along cross-cutting HE and IC dimensions that everywhere reflect the same orientations toward social institutions.
This is a strong claim. HE and IC are simultaneously distinctive and spare. One could easily imagine that in a particular society, individuals’ preferences and expectations wouldn’t meaningfully vary along one or the other of these two dimensions; that is, one might think that particular societies would be relatively homogenous with respect to either HE or IC. In addition, one might imagine that the members of at least some societies might vary along worldview dimensions that can’t be reduced to either of these two.
But rather than get worked into a state of philosophical agitation about whether HE and IC generalize, I would treat the claim that they do as a hypothesis, and cross-cultural cultural cognition as an empirical test of it. If attempts to construct universal HE and IC measures go nowhere, then the claim that these dispositions generalize will be of philosophical interest only. If, in contrast, a project of this sort does contribute materially to explanation, prediction, and prescription across diverse societies, then no philosophical objection to universal motivating dispositions will be sufficient to refute it.
Nevertheless, my motivation for hypothesizing the universality of the HE and IC dispositions is not really that I think that claim is true. The value of the hypothesis is its contribution to systematizing empirical research. Tests of the hypothesis will likely prove successful and thus generate instructive models of risk variance in many societies. In others, it will probably fail, while still yielding insight into what is likely to work better and why.
B. Disposition Indicators
In our research, we use a latent variable modeling strategy to measure the motivating dispositions associated with Douglas’s group-grid framework. A latent variable is one that doesn’t admit of direct observation or measurement; it is measured indirectly by aggregating measures of indicators—observable variables that correlate with the latent variable.
That’s exactly what the items that make up our HE and IC scales are—reliable and valid latent- variable indicators. Responses to them covary in patterns that are consistent with their being measures of two unobserved attitudinal orientations, which themselves cohere with other things (from other attitudes to demographic characteristics to preferences and behaviors of one sort or another) that one would expect people who hold the worldviews formed by the intersection of HE and IC to display.
Should we expect the indicators of the HE and IC dispositions to generalize across societies? I certainly wouldn’t.
Our scales work for members of the U.S. population because they capture reasonably well certain words that contemporary Americans use to express their commitments. But that’s just a matter of historical happenstance. Those same statements (e.g., “[i]t seems like the criminals and welfare cheats get all the breaks, while the average citizen picks up the tab”) might not even make sense to, much less divide people with opposing cultural outlooks in, Sweden or Brazil. If so, scales formed by aggregation of responses to those items would be neither reliable nor valid.
That wouldn’t necessarily mean, though, that there aren’t hierarchical individualists, hierarchical communitarians, egalitarian individualists, and egalitarian communitarians in those countries. It would mean only that if there are, measuring their dispositions would require alternative indicators—such as attitudinal items the wordings of which capture how Swedes or Brazilians with those outlooks express their commitments.
I’ll say more about that—and in particular about how one can determine whether society-specific indicators are measuring the same dispositions across societies—in the next post. But for now, it is enough to say that it’s just a mistake to think the cross-cultural study of cultural cognition demands not only that the motivating dispositions associated with Douglas’s group-grid scheme be universal but also that the indicators used to measure them be uniform across societies.
C. Cultural mappings of risk perception
In my view, there’s no reason to expect the mappings of risk perceptions onto worldviews to generalize across societies either. Like the items used to form the HE and IC scales, what risks mean in relation to group-grid worldviews will likely be a matter of contingent historical circumstances and thus vary across place and over time.
Take gun risks, for example. The “gun debate” in American society is one over competing risk claims: the assertions that widespread gun possession increases the incidence of gun accidents and crime, on the one hand; versus the argument that gun control undermines the ability of law-abiding citizens to protect themselves from violent precaution, on the other. Relying on CTR, Donald “Shotgun Braman” and I have conjectured that egalitarian communitarians would be motivated to worry more about the risks associated with too few restrictions on guns, and hierarchical individualists the risks associated with too many, and our studies support that hypothesis.
Some commentators, including Mary Douglas, have expressed puzzlement over this finding. They asserted that hierarchists should support restriction of private gun possession in line with their general commitment to social regimentation and control of individuals.
This expectation, we replied, overlooks the distinctive history of guns in the U.S.: their association with Southern honor norms; their use in settlement of the western frontier; their role in enabling resistance to Reconstruction in the 19th Century and to civil rights in the 20th. Against this background, aversion to guns conveys a recognizable egalitarian style, and enthusiasm for them (particularly among white males) a recognizable hierarchical one. But those meanings are specific to the U.S..—and thus suggest nothing about how gun risk perceptions will map onto group-grid in some other society having an entirely different historical experience with guns.
Again, it is a mistake to think that CTR, to be meaningfully cross-cultural, demands that who fears what and why generalize across societies. It requires only that the diversity of risk perceptions that people form across societies or within particular ones of them over time all be meaningfully connected to the motivating dispositions featured by group-grid.
Or at least that seems to me like the most plausible and profitable conjecture to pursue by empirical testing.
Indeed, the prospect of identifying cross-cultural divergences in how risks map onto the HE and IC worldviews is what excites me most about extending our methods to non-U.S. samples.
Within any society, the fraction of risk issues that provoke cultural conflict relative to the ones that could but don’t is always small. The primary mission of the science of science communication, in my view, is to understand the forces that divert this small set of issues from the pathways of collective-knowledge transmission that usually guide diverse citizens to the best available understanding of how the world operates.
Ideal for acquiring such knowledge would be a rich cross-cultural data set that links uniform risk-perception predictors—the cultural disposition scales derived from society-specific indicators—to distinctive patterns of variance across societies. With such data, researchers could formulate and test hypotheses about what happened in one society but not in another to cause same putative risk to become a source of cultural contestation.
On the basis of what such study revealed, we’d then be in a position to systematize our knowledge of how to design procedures that hold the precipitants of such conflict in check or counteract them when preemptive interventions have failed.