Cultural Cognition: Theory, Sources, Methods, Results

In my last post, I wrote about the “white male status anxiety effect,” which describes how cultural cognition generates demographic variation in risk perception. I now want to fill out the background on cultural cognition, describing briefly (okay, “briefly” is a pathetic lie; sorry) the theory behind it, its sources, the methods we’ve used to investigate it so far, some key empirical findings, and some follow-up studies now in the works.

As we use the term, “cultural cognition” refers to the impact that certain key group commitments have in shaping individuals’ factual beliefs about law and public policy. The core idea, which can be called the cultural cognition thesis, is that culture is prior to fact on charged policy issues. Normatively, culture might be prior to fact in the sense that cultural values determine what significance individuals attach to the consequences of environmental regulation, gun control, the death penalty, and the like. But more importantly, culture is cognitively prior to facts in the sense that cultural values shape what individuals believe the consequences of such policies to be.

The cultural cognition thesis is based on an amalgam of sources. The primary inspiration (intellectual and also emotional) comes from the so-called “cultural theory of risk perception.” Famously (and polemically) articulated by anthropologist Mary Douglas and political scientist Aaron Wildavsky, this theory posits that individuals selectively credit claims of environmental and technological risk in a way that reflects and reinforces their cultural “worldviews.” Because they are dedicated to the autonomy of markets and other private orderings, “individualists,” according to Douglas and Wildavsky, tend to dismiss claims that commerce is environmentally dangerous. So do “hierarchists,” who see claims of environmental and technological risk as implicit attacks on the prerogatives of social and governmental elites. In contrast, “egalitarians,” particularly ones of a solidaristic or communitarian bent, naturally credit such claims, the public acceptance of which would license regulation of commercial activities that are symbolic of individual self-interest and productive of economic and social disparities.

The cultural cognition thesis also rests on various bodies of work in social psychology. Research in this discipline identifies a plausible set of mechanisms through which culture could be expected to have the impact on risk perceptions (and other policy-related beliefs) asserted by Douglas and Wildavsky. These include cognitive dissonance avoidance, which might incline individuals to resolve contested empirical claims in a manner compatible with their cultural identities; biased assimilation, which would induce them to credit or discredit factual information in a manner supportive of their prior, culturally grounded views; affect, which plays a critical role in risk perception and which is likely shaped by cultural influences; and various in-group and out-group dynamics, such as naïve realism, reactive devaluation, and group polarization, which would motivate individuals to trust those who share their cultural allegiances and distrust those who do not when cultural groups disagree about risk.

We have designed and carried out a study to test the cultural cognition thesis. With funding from the National Science Foundation, we conducted a National Risk and Culture Survey. Our national sample comprised 1800 persons selected by random-digit dial. Subjects participated in interviews of approximately 30 minutes, in which we collected various types of information pertinent to assessing the impact of cultural orientations and other influences on perceptions of various types or risk.

To measure subjects’ “cultural orientations,” we developed reliable attitudinal scales based on Douglas’s “group-grid” typology. In this scheme, a “high group” worldview supports a communitarian society, in which collective needs trump individual ones. A “low group” worldview, in contrast, coheres with an individualist social order, in which persons are expected to secure their own needs without collective assistance and without collective interference. A “high grid” worldview favors a hierarchical society, in which resources, opportunities, duties, rights, political offices and the like are distributed on the basis of conspicuous and largely fixed social characteristics--such as gender, race, class, lineage. A “low grid” worldview favors an egalitarian society, in which such characteristics are denied significance.

The results, many of which are featured on our Cultural Cognition Project site and in papers that can be downloaded there, strongly supported the cultural cognition thesis. We were able, for example, to corroborateDouglas and Wildavsky’s own conjectures about culture and environmental risk perception. As individuals become progressively more hierarchical and individualistic, they worry less and less about risks associated with global warming, nuclear-power generation, and pollution generally. Indeed, cultural orientations more strongly predict this (and other) risk perceptions more powerfully than any other individual characteristic, including education level, income, gender, race, political ideology, and personality type.

We also confirmed our own hypotheses about gun risks. The gun-control debate is naturally framed as one between competing risk claims: control proponents argue that too little control will increase gun violence and accidents, while control opponents argue that too much control will render innocent persons unable to defend themselves from violent criminals. We hypothesized that which of these risks individuals found more important would turn on their cultural orientation. Hierarchs and individualists, we surmised, would worry more about being rendered defenseless, because of the importance of guns in enabling hierarchical social roles (hunter, protector, father) and symbolizing individualistic virtues (courage, self-reliance, prowess). Egalitarians and solidarists, in contrast, would worry more about gun violence, because of the association of guns with patriarchy and racism, and with distrust and indifference to the well-being of strangers. And sure enough, the data show that cultural orientations explain gun-risk perceptions better than any other factor, including one’s gender, one’s race, one’s region of residence, one’s community type, one’s political ideology, one’s personality type, and so forth.

While hierarchists and individualists agree about guns and the environment, but that’s not always so. Hierarchists worry, for example, about the societal dangers of drug distribution and promiscuous sex, and the individual dangers of marijuana smoking; individualists don’t. Likewise egalitarians and individualists aren’t worried about the personal risk of obtaining an abortion, or contracting AIDS from surgery; hierarchists are. These patterns, too, conform to the logic of the worldviews in question. Hierarchists morally disapprove of behavior that defies conventional norms, and thus naturally believe that deviant behavior is dangerous. Egalitarians morally disapprove of norms that rigidly stratify people, and individualists disapprove of norms that constrain individual choice generally, so these types naturally believe that deviant behavior is benign.

Our study of cultural cognition is ongoing. The rationale for using survey methods was that the distribution of risk perceptions could itself supply evidence of the cultural cognition thesis. Because there is no reason to think that persons of one outlook have more access to information about risk or are better able to process it than persons of any other, the power of cultural orientations to explain the distribution of beliefs across persons (and the clustering of them across seemingly unrelated risks) furnishes strong proof that these group commitments are indeed integral to the formation of risk perceptions and related beliefs.

But this is an admittedly indirect form of proof. Although the results of the National Risk and Culture Survey suggest that cultural cognition exists, they don’t tell which of the various mechanisms that might explain cultural cognition is responsible for it and to what extent. We are now designing and carrying out laboratory experiments to help us determine that. (Feel free to ask me about the experiments, but I’ve gone on quite enough for now!) The results of these experiments, we hope, will not only strengthen the case cultural cognition as a positive matter, but also enable the design of risk-communication strategies and other policy interventions that help overcome the barriers that cultural cognition can pose to the formulation of sound, politically acceptable public policies.

Comments

[frank cross]

I remember when the way you could tell the difference between a liberal and a conservative was not any policy position but whether they believed the police or the Black Panthers fired first.

How much of this might be conscious? I think we naively like to think that people seek the truth and consequently, when people's understanding of the truth differs and even is dubious, we ascribe that to cognitive features such as those you've discussed. My casual observation, though, suggests that some of this is simply prioritizing other values (ideological, religious, etc.) over descriptive truth. It's not too uncommon, I think, for a person to essentially acknowledge losing a factual argument but respond with something like "I don't care." I believe what I believe. That sort of attitude seems to go beyond cognitive dissonance to a believe that the descriptive truth itself isn't that important but just a tool used to advance other goals.

[Dan Kahan]

On whether the culture effect on risk perception is rationalization of facts or candid indifference to fact, I think the answer for some issue is "both." That is, people as a result of cultural cognition genuinely believe facts congenial to their values ("gun control deters gun homicides"; "the dealth penalty doesn't deter murder") and also believe their values justify many of those same policy positions regardless of the facts ("private gun ownership is an affront to democracy whether or not gun ownership leads to crime"; "executing someone is uncivilized, regardless of the impact of capital punishment on the homicide rate"). Indeed, we have evidence that this is so on gun control. http://research.yale.edu/culturalcognition/index.php?option=content&task=view&id=104
However, I can tell you from experience w/ focus groups, that although people will eventually say, "even if the facts were x, I would still support/oppose this policy on moral grounds" it takes a lot of coaxing; it's an unnatural way for most people to think!
In addition, there probably are *some* issues where people honestly believe that the facts are decisive, and that they would change their mind on the issue if the facts could be so shown to be otherwise, without realizing that the only facts they are capable of being persuaded of are ones congenial to their cultural values. Individualists, for example, don't believe that global warming poses a significant environmental threat; I doubt they would say, "well, even if there is catastrophic consequences, commerce and industry are sufficiently vital to the good life that it's worth it...."
But for both types of issues, I think the policy prescription is the same: devise policies that appeal to the values of culturally diverse persons. Those policies will be appealing as ends in themselves, and will quiet culturally grounded status anxieties that interfere with acceptance of facts material to societal well-being.

[Eileen Braman]

As I have been reading these posts the ideas you put forth sound a lot like an extension or at least close relative of Social Dominance Theory (and Orientation) put forth by Jim Sidanius in social psych at UCLA and his co-author Felicia Pratto at Connecticut -- Could you distinguish between the two for me and maybe talk a bit about how your measure of hierarchy differs from their SDO measures -- what distinguishes between the two empirically and theoretically? Are they highly correlated or distinct? Oh and thanks for sharing -- this is a very interesting application to a domain I had not previously considered.

[Dan Kahan]

Hi, Eileen. Very ineresting point/question--one to which we have in fact given some thought!

For one thing, SDO purports to measure a personality disposition toward hierarchy, while our worldview measures, which derive from Douglas’s anthropological work, purports to describe a set of shared attitudes that reflect and reinforce a way of life. You can get some sense of that by comparing our hierarchy-egalitarian items (there’s a link to our scales in my post) with SDO items. Whereas the latter focus generally on preferences for group domination (e.g., “Sometimes other groups must be kept in their place”; “It’s probably a good thing that certain groups are at the top and other groups at the bottom”) many of the former relate in a more nuanced way to practices in a household or community that has certain conventional forms of role differentiation and stratification (e.g., “A lot of problems in our society today come from the decline in the traditional
family, where the man works and the woman stays home”; “Parents should encourage young boys to be more sensitive and less ‘rough and tough’ ”). As closely related as “personality dispositions” and “worldviews” (in this sense) might be, moreover, the latter seems to offer a more complete explanation for something like a set of beliefs about risk. That is, once I understand how “hierarchy” has been instantiated in various practices within American society, I can offer a cogent explanation (or at least a hypothesis worth testing) for why hierarchists would believe what they do about matters as diverse global warming, gun control, and promiscuous sex. SDO might or might not predict these things, but it doesn’t readily explain them--and it doesn’t purport to.

In addition, our worldview scheme, which as I indicated is rooted in Douglas’s group-grid scheme, contains not just a hierarchy-egalitarian measure, but also an individualism-solidarism measure. SDO doesn’t have a corresponding dimension. We’ve found that how much a person values role differentiation is largely independent of how much he or she values communal provisioning for individual needs. Accordingly, having the two dimensions allows us to explain patterns of risk perception that would evade a scale that had only hierarchy in it or that blurred that with collectivism.

All that said, it would disappoint me--indeed, it would make me worry--if there weren’t some correlation between our hierarchy-egalitarian measure and the SDO scale. One would expect the sorts of people we describe as hierarchs to have some taste for group dominance. Indeed, to validate our scales, we are collecting data (outside of our national study) on the relationship between how someone scores on our measures and how they score on things like SDO, the F-Scale, the “Big 5” personality test and the like. We aren’t done yet, but preliminary results suggest that there is indeed the sorts of correlations one would expect to see, given what we understand ours scales to be measure. But needless to say, the correspondence isn’t perfect. And that’s a good thing, from our point of view, too!

[eileen braman]

I think so too!(And note that as you do that analysis you may want to look at the great pains Sidanus and Pratto took to distinguish their measure from those other scales you mention to show it is related but distinct).

Thanks for the clarification. I look forward to learning more about this project this week and in future iterations of empirical work. Best, Eileen