Morality, Empathy, And The Value Of Theory

Let’s solve a problem together: I have some raw ingredients that I would like to transform into my dinner. I’ve already managed to prepare and combine the ingredients, so all I have left to do is cook them. How am I to solve this problem of cooking my food? Well, I need a good source of heat. Right now, my best plan is to get in my car and drive around for a bit, as I have noticed that, after I have been driving for some time, the engine in my car gets quite hot. I figure I can use the heat generated by driving to cook my food. It would come as no surprise to anyone if you have a couple of objections with my suggestion, mostly focused on the point that cars were never designed to solve the problems posed by cooking. Sure, they do generate heat, but that’s really more of a byproduct of their intended function. Further, the heat they do produce isn’t particularly well-controlled or evenly-distributed. Depending on how I position my ingredients or the temperature they require, I might end up with a partially-burnt, partially-raw dinner that is likely also full of oil, gravel, and other debris that has been kicked up into the engine. Not only is the car engine not very efficient at cooking, then, it’s also not very sanitary. You’d probably recommend that I try using a stove or oven instead.

“I’m not convinced. Get me another pound of bacon; I’m going to try again”

Admittedly, this example is egregious in its silliness, but it does make its point well: while I noted that my car produces heat, I misunderstood the function of the device more generally and tried to use it to solve a problem inappropriately as a result. The same logic also holds in cases where you’re dealing with evolved cognitive mechanisms. I examined such an issue recently, noting that punishment doesn’t seem to do a good job as a mechanism for inspiring trust, at least not relative to its alternatives. Today I wanted to take another run at the underlying issue of matching proximate problem to adaptive function, this time examining a different context: directing aid to the great number of people around the world who need altruism to stave off death and non-lethal, but still quite severe, suffering (issues like alleviating malnutrition and infectious diseases). If you want to inspire people to increase the amount of altruism directed towards these needy populations, you will need to appeal to some component parts of our psychology, so what parts should those be?

The first step in solving this problem is to think about what cognitive systems might increase the amount of altruism directed towards others, and then examine the adaptive function of each to determine whether they will solve the problem particularly efficiently. Paul Bloom attempted a similar analysis (about three years ago, but I’m just reading it now), arguing that empathetic cognitive systems seem like a poor fit for the global altruism problem. Specifically, Bloom makes the case that empathy seems more suited to dealing with single-target instances of altruism, rather than large-scale projects. Empathy, he writes, requires an identifiable victim, as people are giving (at least proximately) because they identify with the particular target and feel their pain. This becomes a problem, however, when you are talking about a population of 100 or 1000 people, since we simply can’t identify with that many targets at the same time. Our empathetic systems weren’t designed to work that way and, as such, augmenting their outputs somehow is unlikely to lead to a productive solution to the resource problems plaguing certain populations. Rather than cause us to give more effectively to those in need, these systems might instead lead us to over-invest further in a single target. Though Bloom isn’t explicit on this point, I feel he would likely agree that this has something to do with empathetic systems not having evolved because they solved the problems of others per se, but rather because they did things like help the empathetic person build relationships with specific targets, or signal their qualities as an associate to those observing the altruistic behavior.

Nothing about that analysis strikes me as distinctly wrong. However, provided I have understood his meaning properly, Bloom goes on to suggest that the matter of helping others involves the engagement of our moral systems instead (as he explains in this video, he believes empathy “fundamentally…makes the world worse,” in the moral sense of the term, and he also writes that there’s more to morality – in this case, helping others – than empathy). The real problem with this idea is that our moral systems are not altruistic systems, even if they do contain altruistic components (in much the same way that my car is not a cooking mechanism even if it does generate heat). This can be summed up in a number of ways, but simplest is in a study by Kurzban, DeScioli, & Fein (2012) in which participants were presented with the footbridge dilemma (“Would you push one person in front of a train – killing them – to save five people from getting killed by it in turn?”). If one was interested in being an effective altruist in the sense of delivering the greatest number of benefits to others, pushing is definitely the way to go under the simple logic that five lives saved is better than one life spared (assuming all lives have equal value). Our moral systems typically oppose this conclusion, however, suggesting that saving the lives of the five is impermissible if it means we need to kill the one. What is noteworthy about the Kurzban et al (2012) paper is that you can increase people’s willingness to push the one if the people in the dilemma (both being pushed and saved) are kin.

Family always has your back in that way…

The reason for this increase in pushing when dealing with kin, rather than strangers, seems to have something to do with our altruistic systems that evolved for delivering benefits to close genetic relatives; what we call kin-selected mechanisms (mammary glands being a prime example). This pattern of results from the footbridge dilemma suggests there is a distinction between our altruistic systems (that benefit others) and our moral ones; they function to do different things and, as it seems, our moral systems are not much better suited to dealing with the global altruism problem than empathetic ones. Indeed, one of the main features of our moral systems is nonconsequentialism: the idea that the moral value of an act depends on more than just the net consequences to others. If one is seeking to be an effective altruist, then, using the moral system to guide behavior seems to be a poor way to solve that problem because our moral system frequently focuses on behavior per se at the expense of its consequences. 

That’s not the only reason to be wary of the power of morality to solve effective altruism problems either. As I have argued elsewhere, our moral systems function to manage associations with others, most typically by strategically manipulating our side-taking behavior in conflicts (Marczyk, 2015). Provided this description of morality’s adaptive function is close to accurate, the metaphorical goal of the moral system is to generate and maintain partial social relationships. These partial relationships, by their very nature, oppose the goals of effective altruism, which are decidedly impartial in scope. The reasoning of effective altruism might, for instance, suggest that it would be better for parents to spend their money not on their child’s college tuition, but rather on relieving dehydration in a population across the world. Such a conclusion would conflict not only with the outputs of our kin-selected altruistic systems, but can also conflict with other aspects of our moral systems. As some of my own, forthcoming research finds, people do not appear to perceive much of a moral obligation for strangers to direct altruism towards other strangers, but they do perceive something of an obligation for friends and family to help each other (specifically when threatened by outside harm). Our moral obligations towards existing associates make us worse effective altruists (and, in Bloom’s sense of the word, morally worse people in turn).

While Bloom does mention that no one wants to live in that kind of strictly utilitarian world – one in which the welfare of strangers is treated equally to the welfare of friends and kin – he does seem to be advocating we attempt something close to it when he writes:

Our best hope for the future is not to get people to think of all humanity as family—that’s impossible. It lies, instead, in an appreciation of the fact that, even if we don’t empathize with distant strangers, their lives have the same value as the lives of those we love.

Appreciation of the fact that the lives of others have value is decidedly not the same thing as behaving as if they have the same value as the ones we love. Like most everyone else in the world, I want my friends and family to value my welfare above the welfare of others; substantially so, in fact. There are obvious adaptive benefits to such relationships, such as knowing that I will be taken care of in times of need. By contrast, if others showed no particular care for my welfare, but rather just sought to relieve as much suffering as they could wherever it existed in the world, there would be no benefit to my retaining them as associates; they would provide with me assistance or they wouldn’t, regardless of the energy I spent (or didn’t) maintaining social relationship with them. Asking the moral system to be a general-purpose altruism device is unlikely to be much more successful than asking my car to be an efficient oven, that people to treat others the world over as if they were kin, or that you empathize with 1000 people. It represents an incomplete view as to the functions of our moral psychology. While morality might be impartial with respect to behavior, it is unlikely to be impartial with regard to the social value of others (which is why, also in my forthcoming research, I find that stealing to defend against an outside agent of harm is rated as more morally acceptable than doing so to buy recreational drugs).  

“You have just as much value to me as anyone else; even people who aren’t alive yet”

To top this discussion off, it is also worth mentioning those pesky, unintended consequences that sometimes accompany even the best of intentions. By relieving deaths from dehydration, malaria, and starvation today, you might be ensuring greater harm in future generations in the form of increasing the rate of climate change, species extinction, and habitat destruction brought about by sustaining larger global human populations. Assuming for the moment that was true, would that mean that feeding starving people and keeping them alive today would be morally wrong? Both options – withholding altruism when it could be provided and ensuring harm for future generations – might get the moral stamp of disapproval, depending on the reference group (from the perspective of future generations dealing with global warming, it’s bad to feed; from the perspective of the starving people, it’s bad to not feed). This is why the slight majority of participants in Kurzban et al (2012) reported that pushing and not pushing can both be morally unacceptable courses of action.  If we are relying on our moral sense to guide our behavior in this instance, then, we would unlikely be very successful in our altruistic endeavors.

References: Kurzban, R., DeScioli, P., & Fein, D. (2012). Hamilton vs. Kant: Pitting adaptations for altruism against adaptation for moral judgment. Evolution & Human Behavior, 33, 323-333.

Marczyk, J. (2015). Moral alliance strategies theory. Evolutionary Psychological Science, 1, 77-90.

Examining Some Limited Data On Open Relationships

Thanks to Facebook, the topic of non-monogamous relationships has been crossing my screen with some regularity lately. One of the first instances involved the topic of cuckoldry: cases in which a man’s committed female partner will have sex with, and become pregnant by, another another man, often while the man in the relationship is fully aware of the situation; perhaps he’s even watching. The article discussing the matter came from Playboy which, at one point, suggested that cuckoldry porn is the second most common type of porn sought out in online searches; a statement that struck me as rather strange. While I was debating discussing that point – specifically because it doesn’t seem to be true (not only does cuckold porn, or related terms, not hold the number 2 slot in PornHub’s data searches, it doesn’t even crack the top 10 or 20 searches in any area of the world) – I decided it wasn’t worth a full-length feature, in no small part because I have no way of figuring out how such data was collected barring purchasing a book 

“To put our findings in context, please light $30 on fire”

The topic for today is not cuckoldry per se, but it is somewhat adjacent to the matter: open relationships and polyamory. Though the specifics of these relationships vary from couple to couple, the general arrangements being considered are relationships that are consensually non-monogamous, permitting one or more of the members to engage in sexual relationships with individuals outside of the usual dyad pair, at least in some contexts. Such relationships are indeed curious, as a quick framing of the issue in a nonhuman example would show. Imagine, for instance, that a researcher in the field observed a pair-bonded dyad of penguins. Every now and again, the resident male would allow – perhaps even encourage – his partner to go out and mate with another male. While such an arrangement might have its benefits for the female – such as securing paternity from a male of higher status than her mate – it would seem to be a behavior that is quite costly from the male’s perspective. The example can just as easily be flipped with regard to sex: a female that permitted her partner to go off and mate with/invest in the offspring of another female would seem to be suffering a cost, relative to a female that retained such benefits for herself. Within this nonhuman example, I suspect no one would be proposing that the penguins benefit from such an arrangement by removing pressure from themselves to spend time with their partners, or by allowing the other to do things they don’t want to do, like go out dancing. While humans are not penguins, discussing the behavior in the context of others other animals can remove some of less-useful explanations for it that are floated by people (in this case, people might quickly understand that couples can spend time apart and doing different things without needing to have sex with other partners).

The very real costs of such non-monogamous behavior can be seen in the form of psychological mechanisms governing sexual jealousy in men and women. If such behavior did not reliably carry costs for the other partner, mechanisms for sexual jealousy would not be expected to exist (and, in fact, they may well not exist for other species where associations between parents ends following copulation). The expectation of monogamy seems to be the key factor separating pair-bonds from other social associations – such as friendship and kinship – and when that expectation is broken in the form of infidelity, it often leads to the dissolution of the bond. Given that theoretical foundation, what are we to make of open relationships? Why do they exist? How stable are they, compared to monogamous relationships? Is it a lifestyle that just anyone might adopt successfully? At the outset, it’s worth noting that there doesn’t seem to exist a wealth of good empirical data on the matter, making it hard to answer such questions definitively. There are, however, two papers that discuss the topic I wanted to examine today to start making some progress on those fronts. 

The first study (Rubin & Adams, 1986) examined martial stability between monogamous and open relationships over a five-year period from 1978-1983 (though precisely how open these relationships were is unknown). Their total sample was unfortunately small, beginning with 41 demographically-matched couples per group and ending with 34 sexually-open couples and 39 monogamous ones (the authors refer to this as an “embarrassingly small” number). As for why the attrition rate obtained, two of the non-monogamous couples couldn’t be located and five of the couples had suffered a death, compared with one missing and one death in the monogamous group. Why so many deaths appeared to be concentrated in the open group is not mentioned, but as the average age of the sample at follow up was about 46 and the ages of the participants ranged from 20-80, is possible that age-related factors were responsible.

Concerning the stability of these relationships over those five years, the monogamous group reported a separation rate of 18%, while 32% of those in the open relationships reported no longer being together with their primary partner. Though this difference was not statistically significant, those in open relationships were nominally almost twice as likely to have broken up with their primary partner. Again, the sample size here is small, so interpreting those numbers is not a straightforward task. That said, Rubin & Adams (1986) also mention that both monogamous and open couples report similar levels of jealously and happiness in those relationships, regardless of whether they broke up or stayed together. 

However, there’s the matter of representativeness….

It’s difficult to determine how many couples we ought to have expected to have broken up during that time period, however. This study was conducted during the early 80s, and that time period apparently marked a high-point in US divorce frequency. That might put the separation figures in some different context, though it’s not easy to say what that context is: perhaps the monogamous/open couples were unusually likely to have stayed together/broken up, relative to the population they were drawn from. On top of being small, then, the sample might also fail to represent the general population. The authors insinuate as much, noting that they were using an opportunity sample for their research. Worth noting, for instance, is that about 90% of their subjects held a college degree, which is exceedingly high even by today’s standards (about 35% of contemporary US citizens do); a full half of them even had MAs, and 20% had PhDs (11% and 2% today). As such, getting a sense for the demographics of the broader polyamorous community – and how well they match the general population – might provide some hints (but not strong conclusions) as to whether such a lifestyle would work well for just anyone. 

Thankfully, a larger data set containing some demographics from polyamorous individuals does exist. Approximately 1,100 polyamorous people from English-speaking countries were recruited by Mitchell et al (2014) via hundreds of online sources. For inclusion, the participants needed to be at least 19 years old, currently involved in two or more relationships, and have partners that did not participate in the survey (so as to make the results independent of each other). Again, roughly 70% of their sample held an undergraduate degree or higher, suggesting that the more sexually-open lifestyle appear to disproportionately attract the well-educated (that, or their recruitment procedure was only capturing individuals very selectively). However, another piece of the demographic information from that study sticks out: reported sexual orientations. The males in Mitchell et al (2014) reported a heterosexual orientation about 60% of the time, whereas the females reported a heterosexual orientation a mere 20% of the time. The numbers for other orientations (male/female) were similarly striking: bisexual or pansexual (28%/68%), homosexual (3%/4%), or other (7%/9%).

There are two very remarkable things about that finding: first, the demographics from the polyamorous group are divergent – wildly so – from the general population. In terms of heterosexuality, general populations tend to report such an orientation about 97-99% of the time. To find, then, that heterosexual orientations dropped to about 60% in men and 20% in women represents a rather enormous gulf. Now it is possible that those reporting their orientation in the polyamorous sample were not being entirely truthful – perhaps by exaggerating – but I have no good reason to assume that is the case, nor would I be able to accurately estimate by how much those reports might be driven by social desirability concerns, assuming they are at all. That point aside, however, the second remarkable thing about this finding is that Mitchell et al (2014) don’t seem to even notice how strange it is, failing to make mention of that difference at all. Perhaps that’s a factor of it not really being the main thrust of their analysis, but I certainly find that piece of information worthy of deeper consideration. If your sample has a much greater degree of education and incidence of non-heterosexuality than is usual, that fact shouldn’t be overlooked.

Their most common major was in gettin’ down

In general, from this limited peek into the less-monogamous relationships and individuals in the world, the soundest conclusion one might be able to draw is that those who engage in such relationships are likely different than those who do not in some important regards; we can see that in the form of educational attainment and sexual orientation in the present data set, and it’s likely that other, unaccounted for differences exist as well. What those differences might or might not be, I can’t rightly say at the moment. Nevertheless, this non-representativeness could well explain why the polyamorists and monogamists have such difficulty seeing eye-to-eye on the issue of exclusivity. However, sexual topics tend to receive quite a bit of moralization in all directions, and this can impede good scientific progress in understanding the issue. If, for instance, one is seeking to make polyamory appear to be more normative, important psychological differences between groups might be overlooked (or not asked about/reported in the first place) in the interests of building acceptance; if one views them as something to be discouraged, one’s interpretation of the results will likely follow suit as well.

References: Mitchell, M., Bartholomew, K., & Cobb, R. (2014). Need fulfillment in polyamorous relationships. Journal of Sex Research, 21, 329-339.

Rubin, A. & Adams, J. (1986). Outcomes of sexually open marriages. The Journal of Sex Research, 22, 311-319.

Punishment Might Signal Trustworthiness, But Maybe…

As one well-known saying attributed to Maslow goes, “when all you have is hammer, everything looks like a nail.” If you can only do one thing, you will often apply that thing as a solution to a problem it doesn’t fit particularly well. For example, while a hammer might make for a poor cooking utensil in many cases, if you are tasked with cooking a meal and given only a hammer, you might try to make the best of a bad situation, using the hammer as an inefficient, makeshift knife, spoon, and spatula. That you might meet with some degree of success in doing so does not tell you that hammers function as cooking implements. Relatedly, if I then gave you a hammer and a knife, and tasked with you the same cooking jobs, I would likely observe that hammer use drops precipitously while knife use increases quite a bit. It is also worth bearing in mind that if the only task you have to do is cooking, the only conclusion I’m realistically capable of drawing concerns whether a tool is designed for cooking. That is, if I give you a hammer and a knife and tell you to cook something, I won’t be able to draw the inference that hammers are designed for dealing with nails because nails just aren’t present in the task.

Unless one eats nails for breakfast, that is

While all that probably sounds pretty obvious in the cooking context, a very similar set up appears to have been used recently to study whether third-party punishment (the punishment of actors by people not directly affected by their behavior; hereafter TPP) functions to signal the trustworthiness of the punisher. In their study, Jordan et al (2016) has participants playing a two-stage economic game. The first stage was a TPP game. In this game, there are three players: player A is the helper, and is given 30 cents, player B is the recipient, and given nothing, and player C is the punisher, given 20 cents. The helper can choose to either give the recipient 15 cents or nothing. If the helper decides to give nothing, the punisher then has the option to pay 5 cents to reduce the helper’s pay by 15 cents, or not do so. In this first stage, the first participant would either play one round as a helper or a punisher, or play two rounds: one in the role of the helper and another in the role of the punisher.

The second stage of this game involved a second participant. This participant observed the behavior of the people playing the first game, and then played a trust game with the first participant. In this trust game, the second participant is given 30 cents and decides how much, if any, to send to the first participant. Any amount sent is tripled, and then the first participant decides how much of that amount, if any, to send back. The working hypothesis of Jordan et al (2016) is that TPP will be used a signal of trustworthiness, but only when it is the only possible signal; when participants have an option to send better signals of trustworthiness – such as when they are in the roll of the helper, rather than the punisher – punishment will lose its value as a signal for trust. By contrast, helping should always serve as a good signal of trustworthiness, regardless of whether punishment is an option.

Indeed, this is precisely what they found. When the first participant was only able to punish, the second participant tended to trust punishers more, sending them 16% more in the trust game than non-punishers; in turn, the punishers also tended to be slightly more trustworthy, sending back 8% more than non-punishers. So, the punishers were slightly, though not substantially, more trustworthy than the non-punishers when punishing was all they could do. However, when participants were in the helper role (and not the punisher role), those who transferred money to the recipient were in turn trusted more – being sent an average of 39% more in the trust game than non-helpers – and were, in fact, more trustworthy – returning an average of 25% more than non-helpers. Finally, when the first participant was in the role of both the punisher and the helper, punishment was less common (30% of participants in both roles punished, whereas 41% of participants who were only punishers did) and, controlling for helping, punishers were only trusted with 4% more in the second stage and actually returned 0.3% less.

The final task was less about trust and more about upper-body strength

To sum up, then, when people only had the option to punish others, punishment behavior was used by observers as a cue to trustworthiness. However, when helping was possible as well, punishment ceased to predict trustworthiness. From this set of findings, the authors make the rather strange conclusion that “clear support” was found for their model of punishment as signaling trustworthiness. My enthusiasm for that interpretation is a bit more tepid. To understand why, we can return to my initial example: you have given people a tool (a hammer/punishment) and a task (cooking/a trust game). When they use this tool in the task, you see some results, but they aren’t terribly efficient (16% more trusted and 8% more returned). Then, you give them a second tool (a knife/helping) to solve the same task. Now the results are much better (39% more trusted, 25% more returned). In fact, when they have both tools, they don’t seem to use the first one to accomplish the task as much (punishment falls 11%) and, when they do, they don’t end up with better outcomes (4% more trusted, 0.3% less returned). From that data alone, I would say that the evidence does not support the inference that punishment is a mechanism for signaling trustworthiness. People might try using it in a pinch, but its value seems greatly diminished compared to other behaviors.  

Further, the only tasks people were doing involved playing a dictator and trust game. If punishment serves some other purpose beyond signaling trustworthiness, you wouldn’t be able to observe it there because people aren’t in the right contexts for it to be observed. To make that point clear, we could consider other examples. First, let’s consider murder. If I condemn murder morally and, as a third party, punish someone for engaging in murder, does this tell you that I am more trustworthy than someone else who doesn’t punish it themselves? Probably not; almost everyone condemns murder, at least in the abstract, but the costs of engaging in punishment aren’t the same for all people. Someone who is just as trustworthy might not be willing or able to suffer the associated costs. What about something a bit more controversial: let’s say that, as a third party, I punish people for obtaining or providing abortions. Does hearing about my punishment make me seem like a more trustworthy person? That probably depends on what side of the abortion issue you fall on.

To put this in more precise detail, here’s what I think is going on: the second participant – the one sending money in the trust game, so let’s call him the sender – primarily wants to get as much money back as possible in this context. Accordingly, they are looking for cues that the first participant – the one they’re trusting, or the recipient – is an altruist. One good cue for altruism is, well, altruism. If the sender sees that the recipient has behaved altruistically by giving someone else money, this is a pretty good cue for future altruism. Punishment, however, is not the same thing as altruism. From the point of the view of the person benefiting from the punishment, TPP is indeed altruistic; from the point of view of the target of that TPP, the punishment is spiteful. While punishment can contain this altruistic component, it is more about trading off the welfare of others, rather than providing benefits to people per se. While that altruistic component of punishment can be used as a cue for trustworthiness in a pinch when no other information is available, that does not suggest to me sending such a signal is its only, or even its primary function.

Sure, they can clean the floors, but that’s not really why I hired them

In the real world, people’s behaviors are not ever limited to just the punishment of perpetrators. If there are almost always better ways to signal one’s trustworthiness, then TPP’s role in that regard is likely quite low. For what it’s worth, I happen to think that the roll of TPP has more to do with using transient states of need to manage associations (friendships) with others, as such an explanation works well outside the narrow boundaries of the present paper when things other than unfairness are being punished and people are seeking to do more than make as much money as possible. Finding a good friend is not the same thing as finding a good altruist, and friendships do not usually resemble trust games. However, when all you are observing is unfairness and cooperation, TPP might end up looking a little bit like a mechanism for building trust. Sometimes. If you sort of squint a bit.

References: Jordan, K., Hoffman, M., Bloom, P. & Rand. D. (2016). Third-party punishment as a costly signal of trustworthiness. Nature, 530, 473-476.

Smart People Are Good At Being Dumb In Politics

While I do my best to keep politics out of my life – usually by selectively blocking people who engage in too much proselytizing via link spamming on social media – I will never truly be rid of it. I do my best to cull my exposure to politics, not because I am lazy and looking to stay uninformed about the issues, but rather because I don’t particularly trust most of the sources of information I receive to leave me better informed than when I began. Putting this idea in a simple phrase, people are biased. In these socially-contentious domains, we tend to look for evidence that supports our favored conclusions first, and only stop to evaluate it later, if we do at all. If I can’t trust the conclusions of such pieces to be accurate, I would rather not waste my time with them at all, as I’m not looking to impress a particular partisan group with my agreeable beliefs. Naturally, since I find myself disinterested in politics – perhaps even going so far as to say I’m biased against such matters – this should mean I am more likely to approve of research that concludes people engaged with political issues aren’t quite good at reaching empirically-correct conclusions. Speaking of which… 

“Holy coincidences, Batman; let’s hit them with some knowledge!”

A recent paper by Kahan et al (2013) examined how people’s political beliefs affected their ability to reach empirically-sound conclusions in the face of relevant evidence. Specifically, the authors were testing two competing theories for explaining why people tended to get certain issues wrong. The first of these is referred to as the Science Comprehension Thesis (SCT), which proposes that people tend to get different answers to questions like, “Is global warming affected by human behavior?” or “Are GMOs safe to eat?” simply because they lack sufficient education on such topics or possess poor reasoning skills. Put in more blunt terms, we might (and frequently do) say that people get the answers to such questions wrong because they’re stupid or ignorant. The competing theory the authors propose is called the Identity-Protective Cognition Thesis (ICT) which suggests that these debates are driven more by people’s desire to not be ostracized by their in-group, effectively shutting off their ability to reach accurate conclusions. Again, putting this in more blunt terms, we might (and I did) say that people get the answers to such questions wrong because they’re biased. They have a conclusion they want to support first, and evidence is only useful inasmuch as it helps them do that.

Before getting to the matter of politics, though, let’s first consider skin cream. Sometimes people develop unpleasant rashes on their skin and, when that happens, people will create a variety of creams and lotions designed to help heal the rash and remove its associated discomfort. However, we want to know if these treatments actually work; after all, some rashes will go away on their own, and some rashes might even get worse following the treatment. So we do what any good scientist does: we conduct an experiment. Some people will use the cream while others will not, and we track who gets better and who gets worse. Imagine, then, that you are faced with the following results from your research: of the people who did use the skin cream, 223 of them got better, while 75 got worse; of the people who did not use the cream, 107 got better, while 21 got worse. From this, can we conclude that the skin cream works?

A little bit of division tells us that, among those who used the cream, about 3 people got better for each 1 who got worse; among those not using the cream, roughly 5 people got better for each 1 who got worse. Comparing the two ratios, we can conclude that the skin cream is not effective; if anything, it’s having precisely the opposite result. If you haven’t guessed by now, this is precisely the problem that Kahan et al (2013) posed to 1,111 US adults (though they also flipped the numbers between the conditions so that sometimes the treatment was effective). As it turns out, this problem is by no means easy for a lot of people to solve: only about half the sample was able to reach the correct conclusion. As one might expect, though, the participant’s numeracy – their ability to use quantitative skills – did predict their ability to get the right answer: the highly-numerate participants got the answer right about 75% of the time; those in the low-to-moderate end of numeracy ability got it right only about 50% of the time.

“I need it for a rash. That’s my story and I’m sticking to it”

Kahan et al (2013) then switched up the story. Instead of participants reading about a skin cream, they instead read about gun legislation that banned citizens from carrying handguns concealed in public; instead of looking at whether a rash went away, they examined whether crime in the cities that enacted such bans went up or down, relative to those cities that did not. Beyond the change in variables, all the numbers remained exactly the same. Participants were asked whether the gun ban was effective at reducing crime.  Again, people were not particularly good at solving this problem either – as we would expect – but an interesting result emerged: the most numerate subjects were now only solving the problem correctly 57% of the time, as compared with 75% in the skin-cream group. The change of topic seemed to make people’s ability to reason about these numbers quite a bit worse.

Breaking the data down by political affiliations made it clear what was going on. The more numerate subjects were, again, more likely to get the answer to the question correct, but only when it accorded with their political views. The most numerate liberal democrats, for instance, got the answer right when the data showed that concealed carry bans resulted in decreased crime; when crime increased, however, they were not appreciably better at reaching that conclusion relative to the less-numerate democrats. This pattern was reversed in the case of conservative republicans: when the concealed carry bans resulted in increased crime, the more numerate ones got the question right more often; when the ban resulted in decreased crime, performance plummeted.

More interestingly still, the gap in performance was greatest for the more-numerate subjects. The average difference in getting the right answer among the highly-numerate individuals was about 45% between cases in which the conclusion of the experiment did or did not support their view, while it was only 20% in the case of the less-numerate ones. Worth noting is that these differences did not appear when people were thinking about the non-partisan skin-cream issue. In essence, smart people were either not using their numeracy skills regularly  in cases where it meant drawing unpalatable political conclusions, or they were using them and subsequently discarding the “bad” results. This is an empirical validation of my complaints about people ignoring base rates when discussing Islamic terrorism. Highly-intelligent people will often get the answers to these questions wrong because of their partisan biases, not because of their lack of education. They ought to know better – indeed, they do know better – but that knowledge isn’t doing them much good when it comes to being right in cases where that means alienating members of their social group.

That future generations will appreciate your accuracy is only a cold comfort

At the risk of repeating this point, numeracy seemed to increase political polarization, not make it better. These abilities are being used more to metaphorically high-five in-group members than to be accurate. Kahan et al (2013) try to explain this effect in two ways, one of which I think is more plausible than the other. On the implausible front, the authors suggest that using these numeracy abilities is a taxing, high-effort activity that people try to avoid whenever possible. As such, people with this numeracy ability only engage in effortful reasoning when their initial beliefs were threatened by some portion of the data. I find this idea strange because I don’t think that – metabolically – these kinds of tasks are particularly costly or effortful. On the more plausible front, Kahan et al (2013) suggest that these conclusions have a certain kind of rationality behind them: if drawing an unpalatable conclusion would alienate important social relations that one depends on for their own well-being, then an immediate cost/benefit analysis can favor being wrong. If you are wrong about whether GMOs are harmful, the immediate effects on you are likely quite small (unless you’re starving); on the other hand, if your opinion about them puts off your friends, the immediate social effects are quite large.

In other words, I think people sometimes interpret data in incorrect ways to suit their social goals, but I don’t think they avoid interpreting it properly because doing so is difficult.

References: Kahan, D., Peters, E., Dawson, E., & Slovic, P. (2013). Motivated numeracy and enlightened self-government. Yale Law School, Public Law Working Paper No. 307.