follow CCP

Recent blog entries
popular papers

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

Ideology, Motivated Cognition, and Cognitive Reflection: An Experimental Study

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

« More on Hameed's "Pakistani Dr" -- "explaining contradictory beliefs" begs the question | Main | "Anyone who doesn't agree must be a Marxist!" Plus "bans," "decibans," & Turing & Good on "evidentiary weight" »

What “bodycams” can and can’t be expected to do. . . plus coolest study of the year

I definitely favor police “bodycams” as a means of promoting  greater police accountability to the public and greater public confidence in their police.

But there’s a pretty straightforward reason why bodycams won’t prove to be a silver bullet in the effort to subdue societal conflict over excessive police force: perceptions of who did what to whom in such disputes are among the class of factual beliefs influenced by cultural cognition.

When it comes to the impact of cultural cognition, there’s nothing special about brute sense impressions.

Indeed, the foundational study of motivated reasoning—of which cultural cognition is one form—involved distortion of visual perception.  Described in Hastorf & Cantril's 1954 paper, “They Saw a Game,” the experiment showed that students from rival colleges formed opposing perceptions of disputed officiating calls featured in a film of a football game between their schools.  The students' stake in experiencing solidarity with their classmates, researchers concluded, had unconsciously influenced what they saw when viewing the film.

Whether the police can be trusted to refrain from abusing their authority turns on a host of disputed facts symbolically identified with membership in important cultural groups.  Accordingly, the stake that individuals have in experiencing and expressing solidarity with those groups can likewise be expected to unconsciously shape what they see when they view filmed depictions of violent police-citizen interactions.

People who remember the divided reactions to the Rodney King video probably have a sense of that—although, in fact, when people are experiencing this sort of cognitive dynamic, they tend to notice its impact only on those with whom they disagree and not on themselves (Robinson,  Keltner, Ward,  & Ross 1995).

But there is also experimental evidence corroborating the impact of cultural cognition on visual perceptions of behavior in police-citizen confrontations.

These include two CCP studies:  They Saw a Protest (Kahan, Hoffman,  Braman,  Evans, &  Rachlinski 2012) which involved a film of police and political protestors who were variously characterized as demonstrating against abortion rights or against the military’s “don’t ask, don’t tell policy”; and Whose Eyes Are You Going to Believe (Kahan, Hoffman & Braman 2009), which involved film shot from inside a police cruiser that deliberately rammed that of a fleeing suspect.

In both instances, the studies found that what the subjects reported observing --protestors blocking access to a building or people shamed into avoiding entry; a driver veering wildly into lanes of oncoming traffic or police "taking out" a motorist for defying their will -- depended on the subjects' cultural identities. 

But the coolest study on motivated reasoning and perceptions of police force was featured in an article that just came out,  Justice is not blind: Visual attention exaggerates effects of group identification on legal punishment (Granot, Balcetis,  Schneider, & Tyler 2014).

Indeed, GBST  is for me the run-away winner in the contest for “coolest study of the year.”

Actually, GBST reported the results of two related  studies. In one, the researchers correlated perceptions of a violent citizen-police encounter with subjects’ moral predispositions toward the police generally.

In the other, the researchers correlated the subjects’ group membership with perceptions of the behavior of two brawling private citizens, who were identified variously as belonging either to the subjects’ group or to a rival one.

The super cool part of the study was that the researchers used an eye-tracking instrument to assess the predicted influence of motivated reasoning on the perceptions of the subjects.

Collected without the subjects’ awareness, the eye-tracking data showed that subjects fixed their attention disproportionately on the actor they were motivated to see as the wrongdoer—e.g., the police officer in the case of subjects predisposed to distrust the police in study 1, or the citizen identified as an “out-group” member in study 2.


Before reading this study, I would have assumed the effect of cultural cogntion was generated in the process of recollection: that people were fitting bits and pieces of recalled images onto narrative templates featuring police force and the like (cf. Penningon & Hastie 1991, 1992)

But GBST's findings suggest the dynamic that generates opposing perceptions in these cases commences much earlier, before the subjects even take in the visual images.  

The identity-protective impressions people form originate in a kind of biased sampling: by training their attention on the actor who they have the greatest stake in identifying as the wrongdoer, people are--without giving it a conscious thought, of course--prospecting in that portion of the visual landscape most likely to contain veins of data that fit their preconceptions.

Sadly, the benefit of gaining this remarkable insight into the workings of motivated cognition comes at the cost of intensified despair over the prospects for resolving societal conflicts over the appropriateness of the use of violent force by the police.

These disputes look like ones that could be resolved if we only had more information about the facts.  Hence the proposal that the police wear bodycams.

But this understanding has things backwards: the cultural conflict that this policy is meant to dispel will in fact shape what people see when they watch the bodycam videos.

Thus, the full value of the bodycam video policy—which I think can be considerable—will actually depend on our dispelling the antagonistic meanings that make police-citizen encounters a focal point for cultural conflict.

But in fact, that’s part of why I support the bodycam policy. 

The policy involves a significant commitment on the part of society to monitor police, and on the part of the police themselves to make their conduct amenable to monitoring.

Accepting that obligation itself conveys a signal, to the citizens who have the most reason to doubt it, that society and the police themselves are dedicated to assuring that the police will use force appropriately—to protect rather than violate the rights of the members of the community they serve.

More than this gesture will be needed, of course, to create the conditions of reciprocal cooperation and trust necessary to vanquish the distorting influence of cultural cognition on perceptions of violent confrontations between police and individual citizens.

But it’s a good start.


Granot, Y., Balcetis, E., Schneider, K. E., & Tyler, T. R. (2014). Justice is not blind: Visual attention exaggerates effects of group identification on legal punishment. Journal of Experimental Psychology: General, 143(6), 2196-2208. doi: 10.1037/a0037893 

Pennington, N., & Hastie, R. (1991). A Cognitive Theory of Juror Decision Making: The Story Model. Cardozo L. Rev., 13, 519-557.

Pennington, N., & Hastie, R. (1992). Explaining the Evidence: Tests of the Story Model for Juror Decision Making. Journal of Personality and Social Psychology, 62(2), 189-206.

Hastorf, A. H., & Cantril, H. (1954). They saw a game: A case study. The Journal of Abnormal and Social Psychology, 49(1), 129-134. doi: 10.1037/h0057880

PrintView Printer Friendly Version

EmailEmail Article to Friend

Reader Comments (12)

Interesting stuff.

==> "Before reading this study, I would have assumed the effect of cultural cogntion was generated in the process of recollection...But GBST's findings suggest the dynamic that generates opposing perceptions in these cases commences much earlier, before the subjects even take in the visual images."

Not sure why you would have made that assumption? ...seems to me like it is consistent with essentially (unconsciously) deciding which data you are going to collect or scrutinize -- in order to confirm your biases.

December 26, 2014 | Unregistered CommenterJoshua


Yes, it is analogous to biased search. Not surprisingly (!) this is a case of EIOOYKTA. It would have been just as "obvious" if they had found that the motivated reasoning effect was mediated entirely by biased reconstruction of events. but in fact I'd say the "obvious" result they corroborated was less obviously obvious-- it surprises me that people would have somewhere in their minds the "knowledge" that they'd find the preconception-confirming information in a particular place in the field of vision.

I (they) might also be overreading the results. It could be spurious correlation: the inclination to fixate on expected wrongdoer *&* to reach the conclusion that the expected wrongdoer is blameworthy are both caused by same thing-- a cultural predisposition -- w/o the fixation actually playing any mediating role in between the predisposition and the blaming.

December 26, 2014 | Registered CommenterDan Kahan

==> "it surprises me that people would have somewhere in their minds the "knowledge" that they'd find the preconception-confirming information in a particular place in the field of vision. "

Yeah. That is interesting. I mean if someone has a "motivation" when they're watching the video, they would utilize a strategy that is most likely to conform with their motivation. But why, then, would that play out in a particular pattern (disproportionately watching the "wrongdoer?")

Pure speculation, but maybe it's because there's an inherent trust in the "rightdoer? If I accept that the "rightdoer"' was right, then it would be logical for me to search for what the "wrongdoer" did wrong. I"m not trying to confirm that the "rightdoer" did right - because I'm already assuming that is the case.

With that in mind, I would guess that it wouldn't be that hard to manipulate the experiment to alter the field of vision choice: If it were a situation where I felt it that confirming my assumption of the rightness of the "rightdoer" were more important than finding something incorrect about the "wrongdoer's" narrative.

Let's say we found out that there were a video that showed the whole Michael Brown incident. I'd say that someone inclined to confirm the Brown as wrongfully shot narrative would be primarily interested in watching Brown in the video to see if he had his hands raised in surrender when he was shot. The issue of Wilson's action becomes less important: There is no question as to whether he shot Brown. I don't confirm my bias by watching Wilson (of course, there are details...such as whether Wilson reached out and grabbed Brown, whether Wilson shot at Brown when had his back turned while fleeing... that might incline that person to watch the "wrongdoer....just trying to use an example here).

==> " It could be spurious correlation: you are inclined to fixate on expected wrongdoer *&* to reach the conclusion that the expected wrongdoer is blameworthy by the same thing-- a cultural predisposition -- w/o the former being a mediating cause of the latter"

Well, isn't that obvious? :-)

We're not adjusting our field of vision because we're trying to find out the truth. We're adjusting our field of vision to confirm our bias. Cultural predisposition "motivates" our choices of what to watch just as it motivates us to use information to confirm our bias.

December 26, 2014 | Unregistered CommenterJoshua

Agree; one could do a study in which one makes someone observe the portion of the video someone with opposing predispostions focused on & see what happens ....

One of the points here is that there really is no stopping point when one asks, “What ‘s the mechanism for ...?”

If “...” is polarization over facts in police-citizen confrontations, the answer, a study like They Saw a Protest or Whose Eyes suggests, is “cultural cognition” or related species of motivated reasoning.

But then the question becomes, “what ‘s the mechanism for how cultural cognition generates opposing perceptions of images in a filmed confrontation between police and private citizens?”

“People w/ opposing predispostions focus their attention on the portion of the visual field in which they are most likely to yield predisposition-confirming conclusions,” says GBST.

“But what ‘s the mechanism that explains why they focus their visual attention in that way?!” ... etc.

We shouldn ‘t go on forever. But we certainly go as far as necessary to satisfy our interest in explanation/prediction/prescription; and also our curiosity ...

Another reason to think the “reconstruction” explanation rather than the “visual-field selection” one would be the the answer to the 2d-level mechanism question here is that we know visual perception depends on massive virtual or imputed information. We don ‘t have the cognitive “space” to map on to our brains all the visual information that makes up our image of, say, the scene outside the window of our office. So we sample -- taking a bit of information and imputing the rest, generally as necessary to make use of the information we are taking in that way. It is easy to show that this strategy can be confounded; that is the point of Simon ‘s basketball-passing gorilla study: you don ‘t see the gorilla (or wouldn ‘t if I hadn ‘t told you to watch for him/her) b/c there is no way you ‘d impute an image with a gorilla in it as you watch a basketball game. Similarly, people see the political protest; they will impute most of the information based on what they expect to see. People w/ different predispositions will have differnt expectations -- templates, essentially, that reflect their expectations about how people with the relevant commitments and identities behave. As a result they will see different things. That account wouldn ‘t predict that the difference in what people w/ opposing predispositoins see will be mediated by differences in “visual field selection.”

But maybe that ‘s wrong!

Or maybe we need another study to sort out a possible confound in GBST -- the spurious correlation explanation. Maybe they have an explanation already that deals with that.

In any case, nothing is ever “finally” resolved. We just keep gathering pieces of evidence, assigning them their proper “deciban” levels & updating forever & ever. Unless one is a Marxist.

December 26, 2014 | Registered CommenterDan Kahan

I have seen that Gorilla video before. Just fantastic! The brain is fucking awesome.

==> "That account wouldn ‘t predict that the difference in what people w/ opposing predispositoins see will be mediated by differences in “visual field selection.”"

I'm still not getting the scaffolding of your logic. If visual field selection would aid me in the process of confirming bias, why wouldn't I utilize that tool?

I'm wondering if part of it isn't related to risk avoidance.

Some thoughts;

If I want to think that a given (generic) incident is evidence of mistreatment of minorities by police....

I'd rather watch the cop. If I watch the cop and I see him being the "wrongdoer," I can confirm my bias. If I watch the cop and I fail to see him be a "wrongdoer" I can still believe that the kid that was shot was innocent - the clip just failed to capture meaningful evidence. Either way, my bias remains intact.

If I watch the kid who was shot: If I fail to see him be the "wrongdoer" my bias remains intact. If I see him be the "wrongdoer" I'm in the uncomfortable situation of confronting my bias.


If I want to think that a given (generic) incident is evidence of a minority kid acting in such a way that he "got what he deserved.....

If I watch the kid being a "wrongdoer" my bias is confirmed. If I watch the kid but fail to see him doing something justifying being shot, I can just think that the video simply failed to include the meaningful evidence where his actions justified the shooting. Either way, my bias remains intact.

If I watch the cop and fail to see him doing something wrong, my bias is confirmed. If I watch the cop and I see him being the "wrongdoer," I'm in the uncomfortable situation of confronting my bias.


Not a perfect example, for sure. I'm sure that we could play with it to make it not really hold up - for example, how would I conclude that the video of the kid missed important information if it followed the kid the whole time. But my larger point is that it would makes sense to me that we'd have a reflexive algorithm for hedging our bets (minimizing the risk of confronting our bias).

Of course, no matter what I see, I can take a biased approach to analyzing the evidence I'm given. I don't have to confront my bias if I simply discount any evidence that contradicts the conclusion I'm predisposed to make. But I think that there may be a risk of having to work harder, contingent on odds related to visual field selection.


December 29, 2014 | Unregistered CommenterJoshua


Not goofy.

But it does start to strain the priniple that the mechanisms for selective attention of this sort can't be consciously selected; one can't "choose" what to believe so we need selection mechanisms that will unconsciously tend toward styles of reasoning that promote the stake a person has in forming identity-congruent beliefs (an important Elster point).

IN general, too, the sorts of information processing dispositoins that could have this effect might not be uniquely determined; or if it is the case that one is more likely to work than another, it miight not be obvious why it is.

So just to say, "it woudl be useful if..." is not necessarily a compelling reason to expect things to be that way.

My naive take was that "imputed visual data" was most likely candidate for how motivted reasoning generates differences in visual perception.

But the GBST evidence is reason to think otherwise (I'm still not sure I *believe* that the "selective visual prospecting" account is right but I give the GBST finding a likelihood ratio of 100 or if you prefer a weight of 20 decibans in favor of that position.

December 29, 2014 | Registered CommenterDan Kahan

===> "So just to say, "it woudl be useful if..." is not necessarily a compelling reason to expect things to be that way."

Sure. I'm not saying that I have a "compelling" answer.

I'm just thinking aloud - that there might be some unconscious calculus going on = I have more to lose by watching the rightdoer than by watching the wrongdoer.

It reminds me, somehow, of when I'm watching a sports event and rooting for one team heavily. I'm going to embarrass myself here, but sometimes I'll start feeling that by watching the game on TV, I can affect the outcome (I know plenty of sports fans who have the same feeling, or similarly do something like wear the same underwear as long as the wining streak keeps going, etc.). Interestingly, I never think that by watching I will have a positive influence on the game, and sometimes there is another pattern that shows up...

Suppose the teams are tied late in the game. I'm more likely to watch my team when it is on offense and then turn the TV off when it is on defense (and then turn it on again when my team goes on offense again). That strategy is based on some instinctive feeling that if my watching the game has a negative influence, there's more "risk" involved if I'm watching on defense; if the other team scores (because I'm watching) the negative effect is clear.

On the other hand, if my negative influence of watching affects my team when it is on offense, causing it not to score, it isn't as clearly damaging. In a tie game, my team not scoring isn't as clearly a negative outcome as the other team scoring.

Again - it is an unconscious "decision" to avoid the risk/disappointment of seeing what I don't want to see.

December 29, 2014 | Unregistered CommenterJoshua

I guess I should add, in case anyone wonders, I don't actually believe that I'll affect the outcome of a game by watching. But it couldn't hurt my team's chances if I turn off the TV while my team is on defense, now could it? :-)

December 29, 2014 | Unregistered CommenterJoshua


I think Pascal had some theory like that

& if your favorite sport was double-slit photon racing, then yes, watching could affect the outcome.

I think if the GBST mechanism is genuine (and I'm open to that certainly), then likely it is the case that people are more likely to find materials suited for constructing identity-protective blame attribuitions by watching the party who they are predisposed to see as wrongdoer. If that's true, then, I think they'd be as likely unconsciously to gravitate toward that way of positioning themselves to take in information in that situation as they would be to gravitate unconsciously toward any other form of information processing that conduces to the stake they have in identity protection

December 29, 2014 | Registered CommenterDan Kahan

I would expect that the motivation in selecting what parts of a scene to watch (you can't watch everything simultaneously!) would not be based on any sort of identity-protective choice, but on what hypotheses were being considered and where they differed the most.

Say you have two hypotheses (there may be more, but just to keep things simple for explanation). They both make predictions about what will be observed, and the probabilities. Some of these predictions will be the same for both hypotheses - gravity will continue to operate, there will be no random gorillas wandering the scene, etc. - and some will differ. Those predictions that are either equally likely or equally unlikely convey little information about the hypotheses - the likelihood ratio is 1 and the number of decibans is zero. Those parts of the picture where the hypotheses differ the most in their predictions convey the most information - the log likelihood ratio is very large (whether positive or negative).

If you can only look at a limited portion of the picture, it makes sense to look at the parts that convey the most information about the question you are being or expect to be asked. This will depend both on what hypotheses you are considering, and what predictions you make from them (i.e. the statistical model being used), both of which may differ between participants of differing political perspective.

You really need to try to elicit what hypotheses subjects are considering in these experiments, but I'd speculate that in a "blame" scenario where you're being asked what "caused" the subsequent events, they would be looking to the one they believed most likely to cause trouble; to see if they do. The alternate hypotheses will be hypotheses about this actor.

But that's just speculation. Try asking people why they do what they do. You don't necessarily have to believe it, but it's useful data even so.

December 30, 2014 | Unregistered CommenterNiV

Thinking about it further, it would also depend on your prior probabilities for the hypotheses. If there are ten or twenty hypotheses, but only three or four with high prior probabilities, and the information-rich regions of the scene differ between them, with no one element giving good information for all of them, then it makes sense to look at the regions that distinguish the most likely hypotheses. It's not the information as such, but the expected information (i.e. probability weighted) that one is attempting to maximise.

Thus, even with the same set of hypotheses and likelihoods, people can still differ if they have different prior expectations.

December 30, 2014 | Unregistered CommenterNiV

Dan -

Don't know if you'll catch this.. but what's playing out in South Carolina right now made me think of this thread...

I think of how outcomes of the shooting were headed down such a different path before/after the video came out. I also have to think that this type of incident getting so much publicity has to have some impact on public attitudes towards the police and police attitudes towards use of deadly force (not that I'm predicting what shape that impact will take).

April 9, 2015 | Unregistered CommenterJoshua

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>