Are Associations Attitudes?

If there’s one phrase that people discussing the results of experiments have heard more than any other, a good candidate might be “correlation does not equal causation”. Correlations can often get mistaken for (at least implying) causation, especially if the results are congenial to a preferred conclusion or interpretation. This is a relatively uncontroversial matter which has been discussed to death, so there’s little need to continue on with it. There is, however, a related reasoning error people also tend to make with regard to correlation; one that is less discussed than the former. This mistake is to assume that a lack of correlation (or a very low one) means no causation. Here are two reasons one might find no correlation, despite underlying relationships: in the first case, no correlation could result from something as simple as there being no linear relationship between two variables. As correlations only measure linear relationships, distributions that resemble bell curves would tend to yield correlations equal to zero.

For the second case, consider the following example: event A causes event B, but only in the absence of variable C. If variable C randomly varies (it’s present half the time and absent the other half), [EDIT: H/T Jeff Goldberg] you might end up with no correlation, or at least a very reduced one, despite direct causation. This example becomes immediately more understandable if you relabel “A” as heterosexual intercourse, “B” as pregnancy, and “C” as contraceptives (ovulation works too, provided you also replace “absence” with presence). That said, even if contraceptives aren’t in the picture, the correlation between sexual intercourse and pregnancy is still pretty low.

And just in case you find that correlation reaching significance, there’s always this.

So why all this talk about correlation and causation? Two reasons: first, this is my website and I find the matter pretty neat. More importantly, though, I’d like to discuss the IAT (implicit association test) today; specifically, I’d like to address the matter of how well the racial IAT correlates (or rather, fails to correlate) with other measures of racial prejudice, and how we ought to interpret that result. While I have touched on this test very briefly before, it was in the context of discussing modularity; not dissecting the test itself. Since the IAT has recently crossed my academic path again on more than one occasion, I feel it’s time for a more complete engagement with it. I’ll start by discussing what the IAT is, what many people seem to think it measures, and finally what I feel it actually assesses.

The IAT was introduced by Greenwald et al in 1998. As per its namesake, the test was ostensibly designed to do something it would appear to do fairly well: measure the relative strengths of initial, automatic cognitive associations between two concepts. If you’d like to see how this test works firsthand, feel free to follow the link above, but, just in case you don’t feel like going through the hassle, here’s the basic design (using the race-version of the test): subjects are asked to respond as quickly as possible to a number of stimuli. In the first phase, subjects will view pictures of black and white faces flashed on the screen and asked to press one key if the face is black and another if it’s white. In the second phase, subjects will do the same task, but this time they’ll press one key if the word that flashes on the screen is positive and another if it’s negative. Finally, these two tasks are combined, with subjects asked to press one key if the face is white or the word is positive, and another key if the face is black or the word is negative (these conditions then flip). Different reaction times in this test are taken to be measures of implicit cognitive associations. So, if you’re faster to categorize black faces with positive words, you’re said to have a more positive association towards black people.

Having demonstrated that many people seem to show a stronger association between white faces and positive concepts, the natural question arises about how to interpret these results. Unfortunately, many psychological researchers and laypeople alike have taken a unwarranted conceptual leap: they assume that these differential association strengths imply implicit racist attitudes. This assumption happens to meet with an unfortunate snag, however, which is that these implicit associations tend to have very weak to no correlations with explicit measures of racial prejudice (even if the measures themselves, like the Modern Racism Scale, are of questionable validity to begin with). Indeed, as reviewed by Arkes & Tetlock (2004), whereas the vast majority of undergraduates tested manifest exceedingly low levels of “modern racism”, almost all of them display a stronger association between white faces and positivity. Faced with this lack of correlation, many people have gone on to make a second assumption to account for this lack, that assumption being that the implicit measure is able to tap some “truer” prejudiced attitude that the explicit measures are not as able to tease out. I can’t help but wonder, though, what those same people would have had to say if positive correlations had turned up…

“Correlations or no, there’s literally no data that could possibly prove us wrong”

Arkes & Tetlock (2004) put forth three convincing reasons to not make that conceptual jump from implicit associations to implicit attitudes. Since I don’t have the space to cover all their objections, I’ll focus on the key points of them. The first is one that I feel ought to be fairly obvious: quicker associations between whites and positive concepts are capable of being generated by merely being aware of racial stereotypes, irrespective of whether one endorses them on any level, conscious or not. Indeed, even African American subjects were found to manifest pro-white biases in these tests. One could take those results as indicative of black subjects being implicit racist against their own ethnic group, though it would seem to make more sense to interpret those results in terms of the black subjects being aware of the stereotypes they did not endorse. The latter interpretation also goes a long way towards understanding the small and inconsistent correlations between the explicit and implicit measures; the IAT is measuring a different concept (knowledge of stereotypes) than the explicit measures (endorsement of stereotypes).

In order to appreciate the next criticism of this conceptual leap, there’s an important point worth bearing in mind concerning this IAT: the test doesn’t measure where two concepts are associated in any sense whatsoever; it merely measures relative strengths of these associations (for example, “bread” might be more strongly associated with “butter” than it is with “banana”, though it might be more associated with both than with “wall”). This importance of this point is that the results of the IAT do not test whether there is a negative association towards any one group; just whether one group is rated more positively than another. While whites might have a stronger association with positive concepts than blacks, it does not follow that blacks have a negative association overall, nor that whites have a particularly positive one either. Both groups could be held in high or low regard overall, with one being slightly favored. In much the same way, I might enjoy eating both pizza and turkey sandwiches, but I would tend to enjoy eating pizza more. Since the IAT does not track whether these response time differentials are due to hostility, these results do not automatically seem to apply well to most definitions of prejudice.

Finally, the authors make the (perhaps politically incorrect) point that noticing behavioral differences between groups – racial or otherwise – and altering behavior accordingly is not, de facto, evidence of an irrational racial biases; it could well represent the proper use of Bayesian inference, passing correspondence benchmarks for rational behavior. If one group, A, happens to perform behavior X more than group B, it would be peculiar to ignore this information if you’re trying to predict the behavior of an individual from one of those groups. In fact, when people fail to do as much in other situations, people tend to call that failure a bias or an error. However, given that race is touchy political subject, people tend to condemn others for using what Arkes & Tetlock (2004) call “forbidden base rates”. Indeed, the authors report that previous research found subjects were willing to condemn an insurance company for using base rate data for the likelihood of property damage in certain neighborhoods when that base rate also happened to correlate with the racial makeup of that neighborhood (but not when those racial correlates were absent).

A result which fits nicely with other theory I’ve written about, so subscribe now and don’t miss any more exciting updates!

To end this on a lighter, (possibly) less politically charged note, a final point worth considering is that this test measures the automaticity of activation; not necessarily the pattern of activation which will eventually obtain. While my immediate reaction towards a brownie within the first 200 milliseconds might be “eat that”, that doesn’t mean that I will eventually end up eating said brownie, nor would it make me implicitly opposed toward the idea of dieting. It would seem that, in spite of these implicit associations, society as a whole has been getting less overtly racist. The need for researchers to dig this deep to try and study racism could be taken as heartening, given that we, “now attempt to gauge prejudice not by what people do, or by what people say, but rather by millisecs of response facilitation of inhibition in implicit association paradigms” (p.275). While I’m sure there are still many people who will make a lot about these reaction time differentials for reasons that aren’t entirely free from their personal politics, it’s nice to know just how much successful progress our culture seems to have made towards eliminating racism.

References: Arkes, H.R., & Tetlock, P.E. (2004). Attributions of implicit prejudice, or “Would Jesse Jackson ‘fail’ the implicit association test?” Psychological Inquiry , 15, 257-278

Greenwald, A.G., McGhee, D.E., & Schwartz, J.L.K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74, 1464-1480

I Know I Am, But What Are You? Competitive Use Of Victimhood

It’s no secret; I’m a paragon of mankind. Beyond simply being a wildly-talented genius, I’m also in such peak physical form that it’s common for people to mistake me for a walking Statue of David with longer hair. As the now-famous Old Spice commercial says, “Sadly, you are not me”, but wouldn’t it be nice for you if you could convince other people that you were? There’s no need to answer that; of course it would be, but the chances of you successfully pulling such a feat off are slim to none.

The more general point here is that, in the social world, you can benefit yourself by strategically manipulating what and how other people think about you and those around you. Further, this manipulation is going to be easier to pull off the less objectively observable the object of that manipulation is. For instance, if I could convince you that my future prospects are good – that I would be a powerful social ally – you might be more inclined to invest in maintaining a relationship with me and giving me assistance in the hopes that I will repay you in kind at some later date. However, I would have harder time trying to convince you I have blue eyes when you can easily verify that they are, in fact, brown.

He’s going to have a hell of a time convincing his boyfriend he’s gay now.

As I’ve written about before, one of those fuzzy concepts open to manipulation is victimhood. Given that legitimate victimhood status can be a powerful resource in the social world, and victimhood requires there be one or more perpetrators, it should come as no surprise that people often find themselves in disagreement about almost every facet of it: from harm, to intent, to blame, and far beyond. Different social contexts – such as morally condemning others vs. being morally condemned yourself – pose people with different adaptive problems to solve, and we should expect that people will process information in different ways, contingent on those contexts. A recent paper by Sullivan et al. (2012) examined the matter over the course of five studies, asking about people’s intuitions concerning the extent of their own victimhood in three contexts: one in which there was no harm being done, one in which a group they belong to was accused of doing harm, and one in which another group was accused of doing harm.

In the first study, 49 male undergrads were presented with a news story (though it was actually a fake news story because psychologists are tricksters) that had one of three conclusions: (a) men and women had equal opportunities in modern society, (b) women were discriminated against in modern society, but it was due to their own choices and biology, or (c) women were discriminated against, and this discrimination was intentionally perpetrated by men. Following this, the men indicated on a 7-point scale whether they thought men or women suffered more relative discrimination in modern society (where 1 indicated men suffered less, 4 indicated they suffered equally, and 7 indicated men suffered more). When confronted with the story where women were not discriminated against, men averaged a 1.69 on the scale; a similar set of results was found when women were depicted as suffering from self-inflicted discrimination, averaging a 1.87. However, when men were depicted as being the perpetrators of this discrimination, the ratings of perceived male oppression rose to 2.61. When men, as a group, were accused of causing harm, they reacted by suggesting they were themselves a victim of more discrimination, as if to suggest that the discrimination women faced wasn’t so bad.

What’s curious about those results is that men didn’t rate the discrimination they faced, relative to women, as more equal when the news article suggested equality in that domain. Rather, they only adjusted their ratings up when their group was painted as the perpetrators of discrimination. The information they were being given didn’t seem to phase them much until it got personal, which is pretty neat.

A similar pattern of findings arose for women in a following experiment. One-hundred forty-two women read a fictional news story about how men were discriminated against when it came to hiring practices, and this discrimination either came from other men or women. Following that, the women filled out the same 7-point scale as before. When men were depicted as responsible for the discrimination against other men, women averaged a 5.16 on the scale, but when women were depicted as being the cause of that discrimination, that number rose to a 5.42. While this rise in ratings of victimhood was smaller than the rise seen with the men, it was still statistically significant. The difference in the scale of these results might be due to the subjects, (the males were undergrads whereas the women were recruited on Mturk) or perhaps the nature of the stories themselves, which were notably different across experiments.

Sexist male behavior is the cause of all the problems for women across society. On the other hand, 65% of women might not hire a man. Seems even-handed to me.

Two of the five experiments also examined whether one group discriminating against another in general was enough to trigger competitive victimhood, or whether one’s own group had to be the perpetrator of the discrimination to cause the behavior. Since they had similar results, I’ll focus on the one regarding race. In this experiment, 51 White students read a story about how Black students tended to be discriminated against when it came to university admissions, and this discrimination was perpetrated either predominately by other White people, or by Asian people. Following this, they filled out that same 7-point scale. When Blacks were being discriminated against by Asians, the White participants averaged a 2.0 on the scale, but when it was Whites discriminating against Black students, this average rose to 2.78. What these results demonstrate is that it’s not enough for some group to just be claiming victimhood status; in order to trigger competitive victimhood, your group needs to be named as the perpetrator.

These results fit neatly with previous research demonstrating that when it comes to assigning blame, people are less likely to assign blame to a victim, relative to a non-victim or hero. When people are being blamed for causing some harm, they tend to see themselves as greater victims, likely in order to better dissuade others from engaging in punishment. However, when people are not being blamed, there is no need to deflect punishment, and, accordingly, the bias to see oneself as a victim diminishes.

There is one part of the paper that bothered me in a big way: the authors’ suggestions about which groups face more victimization objectively. As far as I can tell, there is no good way to measure victimhood objectively, and, as the results of this experiment show, subjective claims and assessments of victimhood themselves are likely to modified by outside factors. For example, consider two cases: (a) a woman suggests that her boyfriend is physically abusing her, vs. (b) a man suggesting that his girlfriend is physically abusing him. Strictly in terms of which claim is more likely to be believed – regardless of whether it’s true or not – I would put the man’s claim at a disadvantage. Further, if it is believed, there are likely different costs and benefits for men and women surrounding such a claim. Perhaps women would be more likely to receive support, where a man might just be painted as a wimp and lose status among both his male and female peers.

Whether that pattern itself actually holds is besides the point. The larger issue here is that this strategy of claiming victimhood may not work equally well for all people, and it’s important to consider that when assessing people’s judgments of their victimhood. The third-parties that are assessing these claims are not merely passive pawns waiting to be manipulated by others; they have their own adaptive problems to solve when it comes to assessment. To the extent that these problems entailed reproductive costs and benefits, selection would have fashioned psychological mechanisms to deal with them. A man might have more of a vested interest in concerning himself with an attractive woman’s claim to victimhood over a sexually unappealing man, as preferentially helping one of the two might tend to be more reproductively useful.

How often do you come across stories of knights rescuing strange “dudes in distress”, relative to strange damsels?

It should be noted that claiming victimhood is not the only way of deflecting punishment; shifting the blame back towards the victim would likely work as well. The results indicated that competitive victimhood was not triggered in those contexts, presumably because there was no need for it. That’s not to say that they two could not work together – i.e. you’re the cause of your own misfortune as well as the cause of mine – but rather to note that different strategies are available, and will likely be utilized differently by different groups, contingent on their relative costs and benefits. Further work is going to want to not only figure out what those other tactics are, but assess their effectiveness, as rated by third-parties.

I’d like to conclude by talking briefly about the quality of the “theory” put forth by the authors in this paper to explain their results: social identity theory. Here is how they define it in the introduction:

Individuals are motivated to maintain a positive moral evaluation of their social group…we argue that when confronted with accusations of in-group harm doing…individuals will defensively attempt to bolster the in-group’s moral status in order to diffuse the threat.

As Steven Pinker has noted, explanations like these are most certainly not theories; they are simply restatements of findings that need a theory to explain them. Unfortunately, non-evolutionary minded researchers will often resort to this kind of circularity as they lack any way of escaping it. To suggest that people have all these cognitive biases to just “feel good” about themselves or their group is nonsense (Kurzban, 2010). Feeling good, on its own, is not something that could possibly have been selected for in the first place, but even if it could have been, it would be curious why people wouldn’t simply just feel good about their social group, rather than going through cognitive gymnastics to try and justify it. I find the evolutionary framework to provide a much more satisfying answer to the question, as well as illuminating future directions for research. As far as I can tell, the “feel good” theory does not.

References: Kurzban, R. (2010). Why everyone (else) is a hypocrite. Princeton, NJ: Princeton University Press.

Sullivan, D., Landau, M.J., Branscombe, N.R., & Rothschild, Z.K. (2012). Competitive victimhood as a response to accusations of ingroup harm doing. Journal of Personality and Social Psychology, 102, 778-795.

What Are We To Make Of The Term “Race”?

In the language of biology, race has no hard definition. The most basic taxonomic classification that we as humans get without resorting to “eyeballin’ it” is species, the most frequently referred to definition being: a group of organisms capable of interbreeding and producing fertile offspring. All races of humans definitely fall into the same species category (if they didn’t, we’d hardly be calling ourselves humans). Additionally, in terms of our cognitive functioning, it’s unlikely that people were ever selected to encode the races of other people, given that they were not likely to travel far enough to ever really encounter someone of a different race (Kurzban, Tooby, & Cosmides, 2001), not to mention the matter of what selective advantages the coding of other races would bring being unanswered.

So surely that means race is simply an arbitrary social construct with no real underlying differences between groups, right? Well…

“Have fun with that, buddy. I’m going to sit this one out.”

The answer is both a “yes” and a “no”, but we’ll get to that in a minute. Let’s return to that definition of a species first. There is a hypothetical population of mice (A1), all from the same area of the world. Half of that population is randomly selected and moved to a new area (A2), so the two groups are reproductively isolated. It’s unlikely to two groups of mice will evolve in the same direction, as each group will have to deal with different selection pressures and drift. Let’s further say that each generation, you took a sample of mice from A1 and A2 and attempted to breed them, to see if they produced viable offspring. Turns out they do, leaving you quite unsurprised and able to publish your results in “who cares?” monthly . If you continued this experiment long enough, eventually you’d find that some percentage of the mice from the two groups would probably fail to successfully produce viable offspring.

In the span of a single generation then, two groups that used to be the same species would then not (all) be the same species anymore. That wouldn’t happen because of any one sudden change, but would occur because of genetic differences that had been accumulating over time. This suggests that while there is less agreement over what counts as a race, relative to a species, the concept itself need not be discarded despite its fuzziness; it may actually refer to something worth considering, as evolution doesn’t share our penchant for neat and tidy categorization. The example also demonstrates that the term species is not without ambiguity itself, despite it’s clear definition. Consider that all the mice in A1 could successfully reproduce with all the other A1 mice (A1 = A1), all the A2 mice with other A2 mice (A2 = A2), but only a certain percentage of A1s could reproduce with A2s (some A1 = A2 and some other A1 =/= A2). This means that some of the A1 mice could be considered an identical and/or separate species from the A2 mice, depending on your frame of reference. (Another way of putting this would be that the difference between statistically significant and not statistically significant itself is not significant.)

Try to organize these by color, then tell me the exact moment one color transitions to another.

Now, obviously, it hasn’t even come close to that point when it comes to race in people. All humans are still very much the same species, and the degree of genetic diversity between individuals is rather small compared to chimps (Cosmides, Tooby, and Kurzban, 2003). The more general point is that just because that is true, and just because the definition of race generally amounts to an “I know it when I see it”, it doesn’t mean there are no genetic differences between races worth considering (contingent, of course, upon how one defines race, in all its fuzziness).

It is also important to keep in mind that percentage of genetic difference per se does not determine the effect those differences will have. For instance, Cosmides, Tooby, and Kurzban (2003) note:

Within population genetic variance was found to be [approximately] 10 times greater than between-race genetic variance (i.e. two neighbors of the same ‘race’ differ many times more, genetically speaking, than a mathematically average member of one ‘race’ differs from an average member of another).

On the same token, the variance in height between the average man and the average women (a few inches) is less than the variance in height within genders (a few feet). I don’t find such statements terribly useful. Sure, the statements may be matters of fact and they may tell us we share a lot more than we don’t, but they in no way speak to the differences that do exist. Men are taller than women overall, and that needs an explanation. Put another way, humans share more – much more – of their DNA with chimpanzees, relative to the amount they do not share. However, the amount they don’t share does not cease to be relevant because of that fact.

*This product has been tested on animals we only share 93% of our DNA with

The real question is in what domains do different groups tend to differ from each other and what are the extent of those differences? Are those differences in terms of mean values or variances of a trait? Are they confined to non-psychological factors, like skin and hair color? I will admit near complete ignorance of what an answers to those questions would look like, nor do I feel they’d be particularly easy to obtain in many cases. Some examples could include issues of lactose intolerance among certain populations, sickle cell anemia in others, and the odd fact that while rates of identical twinning tend to be constant across races, the rates of dizyogtic twinning can range from as low as 1/330 in Asian populations to 1/63 among African populations (Segal, 2000). However, the point of this post was not to answer those questions; rather, the point was to demonstrate that such questions need not be immediately shunned because of the definitional issues (of which there are, to restate, plenty of) and political implications that come with the term race.

So while race may be a term that gets an arbitrary or subjective definition across different contexts and people, and while individuals differ more than races do, that does not imply that such a term is useless in all situations. People may disagree on precisely what colors should be considered blue, red, or purple, but that doesn’t mean we should stop thinking about different colors altogether in favor of one single color. It should go without saying that just because differences might/do exist between groups of people in whatever form they do, that’s no justification to treat any person as a representative member of their group rather than an individual, but it’s probably something that should be said more often anyway. So there it is.

References: Cosmides, L., Tooby, J., & Kurzban, R. (2003). Perceptions of race. Trends in Cognitive Sciences, 7, 173-179.

Kurzban, R., Tooby, J., & Cosmides, L. (2001). Can race be erased? Coalitional computation and social categorization. Proceedings of the National Academy of Science, 98, 15387-15392

Segal, N.L. (2000). Entwined Lives: Twins, and What They Tell Us About Human Behavior. Plume.