follow CCP

Recent blog entries
« Aken's worldview | Main | Go local, and bring empirical toolkit: Presentation to US Global Change Research Program »
Friday
Aug242012

Motivated consequentialist reasoning

Nice paper by Liu & Ditto just published (advance on-line) in Social Psychology and Personality Science ("What Dilemma? Moral Evaluation Shapes Factual Belief," doi: 10.1177/1948550612456045).  It presents a series of studies-- from variants of the "trolley problem" to ones involving evidence on stem cell research--supporting the hypothesis that people will conform their assessments of an action or policy's consequences to their appraisals of its intrinsic moral worth.

As Liu & Ditto acknowledge, their findings are in keeping with those of other researchers who have been studying the influence of culturally or ideologically motivated cognition.  The design of their studies, however, was specifically geared to detecting how readily disposed their subjects were to resort to consequentialist justifications for nonconsequentialist positions. In one cool experiment, e.g., they found that exposure to compelling nonconsequenitialist arguments generated changes in the perceived deterrent efficacy of capital punishment! 

This feature of their paper enables the motivated-reasoning position to square off directly against two other important positoins in contemporary moral psychology.

The first, associated most conspicuously with Jonathan Haidt, is that ideological or partisan conflicts over policy reflect a fundamental difference in "liberal" and "conservative" moral styles. Conservatives, Haidt argues, focus on nonconsequentialist evaluations of "purity" or "sanctity," whereas liberals focus on "harm."

But as Liu & Ditto note, conservatives, every bit if not more than liberals (more on that in a second!) adopt a default utilitarian perspective. What divides contemporary American who identify as "liberals" and "conservatives" is not the normative authority of Mill's "harm" principle. It's a set of disputed factual claims that about whether forms of behavior symbolically associated with one or the other's cultural style causes  harms of the sort that any Millian Liberal would agree warrant legal redress.  

That people are impelled to impute harm to behavior that denigrates their cultural norms is, of course, the nerve of Mary Douglas's work, in particular Purity and Danger. I very much agree with Douglas's view. Indeed, I think the view that public policy debate can be characterized as one between philosophical Liberals and Antiliberals -- i.e., between those who believe that law should be confined to promotion of secular ends and those who believe that law is also a proper instrument for propogating a moral orthodoxy -- is one only those who spend far too much time in university moral philosophy seminars are likely to form. 

The second position with which Liu & Ditto join issue is the dual process theory of moral psychology. I view Josh Greene as the leading exponent of this perspective. Greene is a subtle thinker; like Haidt, he is both a first-rate philosopher and an amazing psychologist, But he has not been shy about equating nonconsequentialist (or "deontological") reasoning with emotion-driven, unconscious "system 1" (in Kahneman's terms) reasoning and consequentialism with conscious, reflective "system 2."

I don't buy it. Indeed, cultural cognition -- the tendency of people to fit their assessments of risk and related facts to their group values -- is all about the distorting force that motivated reasoning exerts over consequentialist judgments.  Greene depeicts "deontological" reasoning as a form of confabulation. But precisely because consequentiaist frameworks so often rest on contentious behavioral conjectures and contested forms of empirical proof, they furnish a notoriously pliable set of resources for those who feel impelled to reason, as opposed to intuit, their way out of policy conclusions they find ideologically noncongenial. 

If anything, it seems like those who are adept at system 2 reasoning will be more vulnerable to motivated cognition. They will be better than those who are less reflective, more intuitive, in manipulating the various bendable empirical bits and pieces out of which utilitarian argument tend to be formed. This was the premise of our Nature Climate Change study, which presented evidence that greater science comprehension magnifies cultural cognition.

But like any other proposition (or any proposition worth discussing), the claim that consequentialist reasoning is more hospitable to motivated cognition than other sorts is open to empirical testing. I count Liu & Ditto's studies as evidence in support of that conclusion.

Now, there is one other issue to discuss.

As I said, Liu & Ditto find that conservatives, as well as liberals, resort to consequentialist reasoning. Conservatives don't naturally frame their position in nonconsequentialist terms, much less confine themselves to such justifications. Indeed, in one of the studies they feature in their paper, Liu & Ditto observe "the tendency to perceive morally distasteful acts as also being practically disadvantageous was significantly more pronounced ... for political conservatives." 

So this raises the perennial (for me, in this blog; I am getting treatment, but still can't shake my obsession) issue of the "asymmetry thesis"-- the claim (ably advanced in Chris Mooney's Republican Brain) that motivated consequentialist reasoning is more characteristic of conservatives than liberals.  Is the Liu & Ditto paper evidence in "favor" of the asymmetry thesis?

Sure. In fact, in one of their studies Liu & Ditto present a statistical analysis that shows that subjects' tendency to adopt empirical positions supportive of their intrinsic moral assessments increased as subjects became more conservative. As I've noted before, proponents of the "asymmetry thesis" usually don't try to assess whether any differernces observed in the force of motivated reasoning across the ideological spectrum (or cultural spectra) is statistically, much less practically, significant. Liu & Ditto did make such an assessment.

But does that mean the asymmetry thesis is "true" after all?

It's a mistake (a sadly common one) to view scientific studies as "proving" or "disproving" claims in some binary fashion. Valid studies supply evidence that gives us more reason than we otherwise would have had to credit one hypothesis relative to some alternative one. If one wants to form a provisional judgment--and all judgments must always be viewed as provisional if one is taking a scientific attitude toward empirical proof--then one has to aggregate all the available pieces of evidence, assigning to each the weight it is due in light of how much more consistent it is one with hypothesis than another.

There's just much more valid & compelling evidence in support of the "symmetry" thesis -- that ideologically motivated reasoning is uniform, for all practical purposes, across ideologies--than there is in support of the "asymmetry" position. I myself don't view the Liu and Ditto finding of "asymmetry" as a reason to substantially revise my view of the likelihood that that position is correct.

Indeed, I don't think Liu and Ditto themselves view their results as particularly strong proof in favor of the asymmetry thesis. They note that the "associations between moral and factual beliefs" they observed--on issues like the death penalty, promotion of condoms to fight STDs,  stem cell research, and forceful interrogations--" were stronger for conservatives but "still significant for ... political liberals." "[W]hile our political psychology results can be taken as consistent with the body of work associating conservatism with heuristic and motivated thinking," they conclude, "it is important to also note the modest size of these interaction effects and that significant moral-factual coordination was found across the political spectrum." 

The paper is not a "show stopper" on the "asymmetry" question. On the contrary, it is, in this respect like the others, something much better than that: a pertinent, informative, and indeed elegant addition to an ongoing scholarly conversation.

 

PrintView Printer Friendly Version

EmailEmail Article to Friend

Reader Comments (12)

Hi Dan. Thanks for the blog.

I'm a little confused about your analysis of this study and how it relates to Haidt's Moral Foundations theory. You talk about how the Liu and Ditto study indicates that people, including conservatives, use consequentialist reasoning to justify positions formed on nonconsequentialist grounds. And the way I read it, you think this goes against Haidt's position that conservatives use nonconsequentialist values like purity and sanctity, since in these experiments they used consequentialist justifications as much or more than liberals.

But it would seem to me that the study as you've described it fits well with Haidt's theory. If conservatives are more receptive to a position based on values of sanctity and purity, they will then be motivated to justify that position. At that point they can use consequentialist justifications. But that doesn't mean that the initial disposition towards the the belief didn't derive from the nonconsequentialist values.

Perhaps that would help describe why the tendency was more pronounced in conservatives. If their moral foundations tend to be based on more nonconsequentialist values, then they may end up resorting to motivated consequentialist reasoning more often. In other words, it's not that conservatives are more susceptible to motivated reasoning, but it's that their starting positions are based on values that naturally tend towards requiring more consequentialist reasoning which provides more opportunity for that reasoning to be affected by cultural motivation.

August 24, 2012 | Unregistered CommenterJason Hahn

Hi, Jason. I can certainly see why you are confused. I presented the point in a confusingly compact way.

I see the Liu & Ditto study as supportive of a model in which (a)consciously speaking, everyone is naive or default welfarist utilitarian, but (b) their perceptions of the factual inputs in the utility function are unconsciously influenced by their commitments to one or another set of non-consequentialist cultural norms. Thus, to the extent that we are talking about ordinary members of the public, "conservatives" are just as attentive to "harm" as "liberals" when they engage in conscious moral assessment; & liberals' *perceptions* of harm are as motivated by commitment to a culturally partisan set of non-consequentialist social norms as are the "harm" perceptions of conservatives.

If this is right, then the most straightforward reading of Haidt -- liberals subscribe to a moral theory that focuses on "harm" prevention, and conservatives on one that focuses on "purity" -- seems wrong. They both are focused on harm prevention (or like welfarist objectives); and they both tend to impute harm to behavior that is "impure" in the sense of demeaning of or threatening to nonconsequentialist norms.

Haidt for sure sees a connection between his account of liberal vs. conservative moral styles and motivated reasoning. But I don't think he's clear about what the connection is -- or what about what it is that is *motivating* liberals once one recognizes that their "harm" perceptions are cognitively derivative of something else -- something that is presumably just another brand of "purity."

This is what I think Mary Douglas would say. That doesn't make it right, certainly. But maybe when I put it this way, what I'm saying is clearer?

I actually think that Haidt has, over the course of his career, migrated away from a psychological theory of morality that featured motivated reasoning toward a philosophical theory of morality that roots valuation in personality traits. I find the older, "emotional-tail-wagging-rational-dog" Haidt more congenial. (I suspect that Liu & Ditto feel the same way; take a look at how they alternately connect their findings to, and contrast them with, various of Haidt's works. But maybe we'll get really really lucky & they -- or one of them -- will chime in!)

--Dan

August 24, 2012 | Registered CommenterDan Kahan

Basing this more on his TED talk than anything since -- perhaps that was back in his dog-wagging days? -- my understanding is that Haidt thinks liberals have ~2 fundamental moral values (harm-avoidance and fairness, which I think can encompass equality) and conservatives have ~6 fundamental moral values (liberty, loyalty, obedience, purity, AND harm-avoidance and fairness). Both groups care about harm. So when a member of either group feels a strong moral tug stemming from one of his other values, he'll seek to reconcile all his of values, including the consequentialist harm-avoidance value, with that moral intuition. If I'm correct that that's Haidt's understanding, is it really at odds with Liu & Ditto's findings?

August 25, 2012 | Unregistered CommenterMW

@MW: Take concealed weapons & climate change. How can one be motivated by "harm avoidance" or "fairness" to see particular facts on those issues when the "harm avoiding" & distributionally "fair" positions themselves turn on perceptions of fact subject to motivated cognition? Something *else* must be motivating "liberals" to see the facts that make "harm avoidance" & "fairness" cohere with the "liberal" position on those issues.

Or how about this: tell me how on your understanding of Haidt = Liu & Ditto you'd expalin "They Saw a Protest"? I am inclined to tell you how "liberals" & "conservatives" formed opposing percreptions of "harm" (of a sort any Millian would agree can be legally regulated) in that experiment based on an account like the one that Luker gives of their opposing cultural visions of virtue & gender roles. The "liberal" one in her case cannot be reduced to "fiarness" or "harm" --rather, that cultural vision is what makes "liberals" *see* harm (& unfairness; tell me who is "against fairness"?) in the experiment conditional on what sort of protest they think they are watching. So let's try to specify what those visions are & come up w/ valid psychometric indicators of the disposition to subscribe to them. And then let's see whether it turns out that they are in fact exactly the same cultural visions that are making "liberals" & "conservatives" see opposing "harm" & "fairness" facts on climate & guns. & HPV. & geoengineering. & the death penalty. & -- cetera.

August 25, 2012 | Unregistered Commenterdmk38

Dan,

I think I understand what you're saying. But my understanding of Haidt is different. I always took his moral foundations to refer to the intuitive moral beliefs we have. In other words, people come to hold a certain set of beliefs based on their moral foundations (and the moral foundations of their in-group). Separately from that, they use motivated reasoning to mold the facts into supporting their intuitive belief. Certain positions will "feel right" to a conservative more than a liberal (and vice versa) because each values something different. Only then will either side be roughly equally likely to find consequentialist justifications for the position they were predisposed towards.

In your reply to me you said, "'conservatives' are just as attentive to 'harm' as 'liberals' when they engage in conscious moral assessment." But my understanding of Haidt was that his moral foundations apply to the steps before that, the subconscious moral assessments.

Granted, I'm just some random layman who is only familiar with a sliver of the research and discussion done on these subjects and I haven't even read the Righteous Mind yet, so I might be way off. But to me this explanation makes more sense and fits in with what little research I've heard about. I guess he would have to clarify to be sure.

August 25, 2012 | Unregistered CommenterJason Hahn

Maybe I'm reading (watching?) too much into Haidt, but I think the values that he sets out pretty logically entail the opposing cultural visions that have taken shape. Valuing "fairness" without valuing obedience to authority (this is really hierarchy, no?), loyalty, and purity, leads to valuing equality for all sexes, races, and social classes (egalitarianism, right?). If you value fairness but also value purity and obedience to authority, then you might think it is fair for a woman who has premarital sex to suffer negative consequences, because she's defied other values (a simplification, but you get the idea). I don't think that's a stretch to go from those two sets of primary values to these two cultural visions, do you? In other words, I think the effect of these primary values on specific views on issues like gun control, etc., is mediated by cultural commitments. Obedience/loyalty/purity --> hierarchy --> pro-concealed weapons/pro-life/etc.

Also, the values also serve as the permissible modes of argument for each of the two cultural groups. Who is against "fairness"? Nobody, of course, and nobody's against harm avoidance, which is why those are the two ways people couch their arguments in intercultural public discourse. Conservatives can cite the other values as well when they're talking to each other -- "X didn't obey his superior and that's bad" -- but not when they're talking to liberals -- "well, was there some good reason for X to disobey his superior? If so, who cares?"

I'll admit the individualist/communitarian distinction is a little tougher to pull out of Haidt's proposed value split. I mean, it's not tougher if you believe that only conservatives value liberty. But that's pretty clearly wrong. Again, I don't know that his six core values are a perfect or complete set, but I do think they make some sense, and they could easily bring about the grid (if not group) cultural split.

August 25, 2012 | Unregistered CommenterMW

@Jason Hahn: Rigth, I agree, feelings, iintuitions & then arguments, which are essentially confabulations. That's what Haidt says in emotional/rational dog & tail. But what makes me uneasy in the "liberal vs. conservative feelings" model, is the liberal "feelings" become far too much like the values that liberal confabulations are made of-- "harm" in particular. We won't find the motivating disposition of liberals on the pages of On Liberty (even if we find a lot of good stuff there). It's not "danger" feeling vs. "purity" feeling; it's purity vs. danger period -- purity feeling aleph vs. purity feeling bet, both of which insist the other is dangerous. So in the interest of explaining, predicting, and prescribing, let's get straight about what sort of little Jacobin ghost is haunting liberals & not just focus on whatever vestigial cultural zealot is haunting conservatives (and really only haunting them, too; most "conservatives," like most liberals, are too busy playing with their ipads to make th effort to impose a moral orthodoxy on anyone. thank goodness for capitalism).

@ MW: You are helping but not enough. Why do I want to say (a la Douglas) it's purity -> danger all the way down? You do see *if* it's purity->danger all the way down, then the idea that the egalitarians believe in the "harm" principle doesn't work? that it also doesn't work to say "oh-- conservatives; they are old fashioned types who value 'purity' & who therefore get upset about harmless wrongdoing like xxxxx." If you think liberals don't know about purity, see how one feels the next time you smoke wi/ 25 meters of him or her. You think egalitarians value people who don't obey their moral orthodoxy? Go watch how much mileage Michael Moore gets by exciting real disgust & contempt for people who have cultural style different from those of people who like his movies. Douglas & Wildavsky call the egalitarians "sectarians" in Risk & Culture. I'm not quoting scripture. But the point is, they aren't raving lunatics. It has to be possible to conjure what they are saying & why it's compelling even if it's incomplete or wrong (really just to see why it's incomplete & wrong). Don't hold a mirror up to the egalitarians if you want to describe them. They are not "sectarians," of course, b/c that grossly overstates their level of ambition & their attention to the public sphere (they are playing w/ their ipads). But you can't explain them either if you don't start w/ their ghosts.

August 26, 2012 | Registered CommenterDan Kahan

Hey Dan,

So what (~) likelihood ratio would you ascribe to this study for the hypothesis that the asymmetry thesis does not exist? And how can we be sure that you aren't using your prior to influence that assessment?

August 27, 2012 | Unregistered CommenterNick

Let me first thank Dan for blogging about our paper on his site. His work and thinking have been very influential to us. It is great to have such a thoughtful person working on a topic I see as so central to our current political clusterf**k -- while I have to add how frustrating it is to start working on an exciting idea only to find that Kahan already had it and is already publishing a terribly clever study supporting it!

Here is my brief take on the comments above. First, in the broadest sense I see our work as perfectly consistent with Haidt's ideas about moral intuitionism. We are top-down moral thinkers -- we have the intuitions then recruit supporting justifications (including factual beliefs) in bass-ackward fashion. The implication is that people don't use principles or principled approaches like deontology or consequential to generate moral conclusions. This assumption, which runs deep in both most moral philosophy and moral psychology (especially trolleyology) assumes bottom-up moral thinking. Instead, Brittany and I think that people construct moral beliefs from the top-down -- and thus once they start the justification process (which btw people likely only do if they are pressed to do it) they will recruit any cognition that seems to support their moral intuitions whether that is a broad deontological principle or consequential justifications for their moral positions. It is not an either-or thing but rather an "all of the above" approach.

This is perfectly consistent with a moral intuitionist position, so consistent with Haidt. The issues are a little murkier when you get into moral intuitionism 2.0 -- Moral Foundations Theory. I think Dan's characterizations of it are fair, but it has never been clear in the theory whether the foundations are really seen as deontological principles all -- or some deontological and others more consequentialist, etc. I think of them as points of "sacredness" and therefore more like deontological principles. Most centrally, to me they are moral buttons that if pushed will motivate thinking and post hoc justifying and the two sides just have different buttons that bug them. Liberals are more likely to be offended by acts that harm people (especially underdogs) or violate equality-based fairness whereas conservatives' buttons are violations of in-group, authority, purity, and perhaps liberty.

A paper that may be helpful in thinking about this is Kurt Gray's recent target article in Psychological Inquiry on harm as the "essence" of morality. Much of Dan's thinking on how harm is created from things like purity violations is just what Kurt wants to say -- that all morality, once you distill it down, is about harm. Lots of interesting stuff in that issue that people here would find interesting I think -- to complicated to work thru in any detail here.

So, to finish up. I think our work is very consistent with moral intuitionism generally though esoteric quibbles about whether it is consistent with MFT are possible, and generally challenges Josh Greene's ideas about firm distinctions between deontological and consequentialist thought. I agree with Dan that consequentialism is fertile ground for motivated reasoning -- people in our studies essentially are doing motivated cost-benefit analyses which I bet are extremely common in general. I want to make people think again about whether cost-benefit analyses should be held up as a normative standard -- they are as likely to be a product of affective and intuitive forces as deontological principles.

August 27, 2012 | Unregistered CommenterPete Ditto

@Nick: I don't know who you are but I know you are obviously a very evil, diabolical person (you might even be my cat, except she can't spell as well as you can).
You are right to ask these questions. The answer to the first (what LR for the study in connection with the hypothesis that "asymmetry thesis is false") the answer is 0.75. The answer to the second -- "how can we be sure" my priors have not contaminated my LR-- is, of course you can't be *sure*; so *you* should adjust the LR *you* assign down to 0.66 to compensate for this possibility.
Now, you asked how "we" -- i.e., you -- should know whether my priors contaminated by assessment of the LR associated with L&D's study (or that part of it relating to the asymmetry thesis). I'd like to know how *I'd* know, b/c it bothers me to think that could be going on. I can't be sure either, of course. But what I do is give an account of why I see things the way I do & see what people like you have to say. I think I can recognize thoughtful people reacting in a reflective way, whether or not they agree with me. Accordingly, if I saw such people either telling me they suspect I was succumbing to confirmation bias or reacting in such a way suggestive of that, I think (hope) I'd notice *that* & give such evidence the weight that it is due.
Still not foolproof. But the one thing I am sure of is that this is the sort of approach that it makes sense to take.
What do you think?

August 28, 2012 | Unregistered Commenterdmk38

I didn’t mean to be snarky with my questions – I respect you and your work very much. And sorry for my sloppy wording: I meant “we” as in 'we the readers', and “sure” in the colloquial sense, not referring to literal certainty (p = 1.0).

I have followed with enthusiasm your case against Mooney’s asymmetry hypothesis, and I find your analyses compelling. One thing from this post did bother me, though. You acknowledge Liu & Ditto’s findings do support the asymmetry thesis, yet you state, without much explanation, that you “don't view the Liu and Ditto finding of "asymmetry" as a reason to substantially revise my view of the likelihood that that position is correct.”

Is this because you have (excellent) polynomial analyses showing the opposite? If so, then it seems to me that your prior is potentially influencing the weight you ascribe to the LR for the Liu & Ditto study. You ask how you would know if this is occurring. One way to think about it is that your LR for the Liu & Ditto study as it relates to the asymmetry hypothesis should be ~ equal to the LR from a person who is completely ignorant (in an E.T. Jaynes sense) about the Cultural Cognition findings that bear on the hypothesis. It is, of course, silly to think this way, and certainly no reader of this blog would be in this position, but such ignorance would provide an ‘unbiased’ estimate of the LR associated with the study. [note that is amendable to empirical testing.]

You may have simply have been stating that your prior on the asymmetry hypothesis is so low that the LR for this study does not change your posterior very much. That is perfectly coherent but I would still be interested in what’s happening to your LR (even if its effect on the posterior is trivial).

Does this make sense? I hope I’m not being too evil.

Submitted respectfully.

August 28, 2012 | Unregistered CommenterNick

@Nick-- makes sense! you are not so evil; I was just pulling your tail, kid!

August 28, 2012 | Registered CommenterDan Kahan

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):
Post:
 
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>