Join 3,375 readers in helping fund MetaFilter (Hide)


Study Predicts Political Beliefs With 83 Percent Accuracy
February 16, 2013 5:51 PM   Subscribe

New research, published yesterday in the journal PLOS ONE, suggests what mom and dad think isn’t the endgame when it comes to shaping a person’s political identity. Ideological differences between partisans may reflect distinct neural processes, and they can predict who’s right and who’s left of center with 82.9 percent accuracy.
posted by nevercalm (65 comments total) 14 users marked this as a favorite

 
I'm still discounting bloody conservatives.

Being born THAT way is no excuse.
posted by taff at 5:55 PM on February 16, 2013 [1 favorite]


The article closes on that point.
posted by Marisa Stole the Precious Thing at 6:04 PM on February 16, 2013


Yeah, I know. I'm just reinforcing my huffiness.
posted by taff at 6:10 PM on February 16, 2013 [1 favorite]


So if conservatism is not a choice can it cured? Probably not a "pray-it-away" type of therapy though. Perhaps somebody will make a drug, one can only hope.
posted by Long Way To Go at 6:18 PM on February 16, 2013 [1 favorite]


Gene therapy?
posted by taff at 6:26 PM on February 16, 2013


1. The researchers apparently do not know or understand that there exist political orientations besides liberal and conservative, which is specious, moronic and fatally biases the results by excluding potentially meaningful categories
2. The researchers seem to conflate liberals with Democrats and conservatives with Republicans, which is a nauseating marriage of intellectual incuriosity and a banal legitimation of the inevitability and naturalness of our current political hegemony
3. One of the study designers is identified as "Dr. Darren Schreiber, a researcher in neuropolitics at the University of Exeter" despite the fact that neuropolitics sounds totally made-up
4. The study has an incredibly small N, and you know what they say about studies with incredibly small Ns (they are useless)
5. The researchers apparently did not include people who were not registered with either major party in their study
6. This study is so full of unjustified assumptions, arbitrary exclusions, old-fashioned shitty research design and outright nonsense that it's impossible for any serious individual to take it seriously
posted by clockzero at 6:26 PM on February 16, 2013 [43 favorites]


But clockzero, without truly moronic neuroresearch like this, what would cable news and 20/20 and Dateline fill their hours with?
posted by incessant at 6:40 PM on February 16, 2013 [7 favorites]


7. But it's pretty obvious conservatives and/or Republicans are idiotic fraidy-cats, so the study's probably right anyway.
posted by uosuaq at 6:41 PM on February 16, 2013


The study itself is equivocal about whether one is born with a political ideology pre-stamped on the brain:
One intriguing remaining puzzle regards the direction of causality. One might infer that the differing brain structures identified by Kanai et al. suggest genetic foundations for the differences in ideology. However, recent work has shown that changes in cognitive function can lead to changes in brain structure. For instance, applicants who worked to learn the map of London in order to pass a knowledge test required of potential cab drivers demonstrated significant growth in their hippocampus, a brain region related to memory formation.

Although genetic variation has been shown to contribute to variation in political ideology and strength of partisanship, the portion of the variance in political affiliation explained by activity in the amygdala and insula is significantly larger, suggesting that acting as a partisan in a partisan environment may alter the brain, above and beyond the effect of the heredity. The interplay of genetic and environmental effects may also be driving the observed correlations between the size of brain regions and political affiliation. Further untangling the roles of party, ideology, genes, and neurocognition will be essential for advancing our understanding of political attitudes and behavior. The ability to accurately predict party identification using only neural activity during a risk-taking task suggests that investigating basic neuropsychological differences between partisans may provide us with more powerful insights than the previously-available traditional tools of psychology, sociology, and political science.
The way I read it, the model is predictive about your current political ideology, but there is nothing here that would lend much credence to the idea that one could predict future political ideology from the brain scan of a child.

In any event, the study does not undermine the earlier observation that parental political affiliation is predictive of one's political affiliation. Rather, the new study gives an explanation -- differences in brain structure -- that is itself potentially explained on either a nature or nurture theory. If the brain structure is determined by genes, then the ultimate explanation is nature. If the brain structure is determined by environment, then the ultimate explanation is nurture. And, of course, the answer could be some of both.
posted by Jonathan Livengood at 6:44 PM on February 16, 2013 [3 favorites]


This study is so full of unjustified assumptions, arbitrary exclusions, old-fashioned shitty research design and outright nonsense that it's impossible for any serious individual to take it seriously

Considering I am neither "Left" or "Right" oriented, I found this to be a real knee-slapper. People don't ever change political beliefs or have beliefs that are not defined by modern-day conveniences? This "study" has so many flaws and silly assumptions, that it's almost not worth mentioning, but the idea that people never change or evolve amuses.

Does the study assume that people just fall into a rut because they refuse to think or empathize and then just not exercise all the parts of the brain equally so that some of it works and the rest atrophies; so what we would call a "Left-wing" brain is a sign of a certain cerebral laziness while a "Right-wing" brain is a sign of another kind of the same basic mental slacker-ism and politics is just a self-inflicted brain dysfunction?
posted by Alexandra Kitty at 6:57 PM on February 16, 2013 [3 favorites]


The study has an incredibly small N, and you know what they say about studies with incredibly small Ns (they are useless)

I would be practically the last person to try to defend this sort of study, and if it were not late on saturday, I would probably actually try to take a critical look at it. But their N is _massive_ for an FMRI study, 82 subjects, to the point where what I'd be worried about is that their effect size must be tiny.
posted by advil at 7:08 PM on February 16, 2013 [8 favorites]


Re: clockzero. They split up the whole of the political spectrum into a classification task to get to two outputs because it makes the statistics easier, not because they have a particular third-party bias. They have a fairly solid foundation in previous research as to tasks which will predict for conservatism, which is in their lit review. There's enough subjects to get significance. The statistics checks out, and as Jonathan Livengood mentions, doesn't really lay a great big claim towards causality. It creates a predictive model that's better than the state of the art, that's all it does.

One thing I would've liked in the paper which I couldn't find in it is the whole bit about the dimensionality reduction of political preference. That is, figuring out which of the many many factors that act as input to political preferences ends up being the most important. There's established statistics which exists for telling us which variables are most important in a predictive model like this, and I didn't read them doing any of it, even though they talked about that whole dimensionality problem. I only gave the paper a cursory reading, though: it may be in there.
posted by curuinor at 7:09 PM on February 16, 2013 [2 favorites]


This might sound snarky, but it is not intended to be: Can someone explain why it should be surprising that distinct thoughts arise from distinct neural processes? (Or is the big deal the fact that they have made progress in identifying the actual processes themselves?)
posted by compartment at 7:15 PM on February 16, 2013 [1 favorite]


but the idea that people never change or evolve amuses.

I actually find this idea to be dangerous in its incorrectness, too. If you are convinced that people are pretty much hardwired against reason or discussion (which goes against my person experience, as well), it devolves into power moves to gain political control. I'm not saying that society never requires some of that, but if that's all it becomes, then there's a pretty scary future ahead of us.
posted by SpacemanStix at 7:19 PM on February 16, 2013


"Study shows Democrats use left side of brain and Republicans use the right"
posted by panaceanot at 7:32 PM on February 16, 2013


But SpacemanStix, they don't say that. What they do say is that they can have better predictive power than the state of the art with respect to being able to just have some random person come into the lab, do the fMRI, and be able to predict their party affiliation based upon that. The state of the art before big brain machines was 68%, because you could just guess based upon parents' political affiliation, the state of the art now is 83%.

Just because it's in the brain doesn't mean that it's "hardwired", since human brains are neuroplastic. The research has basically nothing to say about how political preferences change over time.
posted by curuinor at 7:33 PM on February 16, 2013 [4 favorites]


Well, I'll be very interested to take this test. I've been both a Democrat and a Republican.
posted by sidi hamet at 7:55 PM on February 16, 2013


previously: Your brain on pseudoscience: the rise of popular neurobollocks

82 subjects / 83% accuracy seems convincing to me. If the the results were reproduced independently, I might believe them.

This might sound snarky, but it is not intended to be: Can someone explain why it should be surprising that distinct thoughts arise from distinct neural processes? (Or is the big deal the fact that they have made progress in identifying the actual processes themselves?)


I doubt anyone is surprised that "thoughts arise" from neural processes. Human behavior and personality seems so diverse and complicated, and the brain such a black box, to me, that I'm a bit surprised they are able to find the same specific processes in 82 different people capable of predicting something somewhat arbitrary like their political preference. There is a difference between 'identifying a brain process' and locating brain activity. Assuming a real correlation between fMRI patterns and political preference has been found, it doesn't explain very much of what the brain is actually doing. Knowing the other behaviors that these same regions correlate with perhaps starts to paint a better picture.
posted by Golden Eternity at 7:56 PM on February 16, 2013


neuropolitics?!

An echo, not a choice?
posted by carping demon at 7:57 PM on February 16, 2013 [1 favorite]


1. The researchers apparently do not know or understand that there exist political orientations besides liberal and conservative, which is specious, moronic and fatally biases the results by excluding potentially meaningful categories

That doesn't make any sense, sorry.

Study: Because of huge pile of existing lit over here, and this other pile of existing lit, we think liberals and conservatives will show different neural structures. We looked at their brains and they do.

Omitting libertarians, Christian communitarians, MOVE, or other fringe groups shouldn't bias a study of liberals and conservatives. If you wanted to, you could use similar theory to predict neural differences between conservatives and libertarians, and see if you were right. They weren't interested in doing so.

2. The researchers seem to conflate liberals with Democrats and conservatives with Republicans, which is a nauseating marriage of intellectual incuriosity and a banal legitimation of the inevitability and naturalness of our current political hegemony

This would have been a reasonable criticism in 1960, when there were significant numbers of conservative Democrats. As of 2006, though, only *looks* six percent of Democrats were conservative and only two percent of Republicans were liberal.

3. One of the study designers is identified as "Dr. Darren Schreiber, a researcher in neuropolitics at the University of Exeter" despite the fact that neuropolitics sounds totally made-up

That's perhaps the world's silliest criticism. We know that he's a researcher in neuropolitics because he's an author of this study, which is clearly neuropolitics.

4. The study has an incredibly small N, and you know what they say about studies with incredibly small Ns (they are useless)

Advil has already noted that the N is not small for fmri, and I'll take his/her word on it. In any case, studies with small N's aren't useless, they just have less power, so you need bigger effects for them to stand out against the relatively higher statistical noise.

5. The researchers apparently did not include people who were not registered with either major party in their study

Which is smart if you want to examine your predictions about neural structures in liberals and conservatives. It would certainly be dumb if you were trying to examine predictions about the neural structures of moderates or people too unsophisticated to have any coherent ideology, but then they weren't trying to do that.
posted by ROU_Xenophobe at 8:00 PM on February 16, 2013 [18 favorites]


People don't ever change political beliefs or have beliefs that are not defined by modern-day conveniences?

People obviously change political beliefs sometimes, and it would be interesting to see whether the neural structures of people who had moved were more like their original or current ideology, and what factors influenced this balance.

But that's a different study than this one.

A few people even have beliefs that are not defined by modern day conveniences. This study didn't examine them.
posted by ROU_Xenophobe at 8:04 PM on February 16, 2013


but the idea that people never change or evolve amuses.

This confused me greatly until I reread it and realized that the last word was not "anuses."
posted by leftcoastbob at 8:08 PM on February 16, 2013 [2 favorites]


In addition to what clockzero listed, there's also a correlation/causation problem. I can easily believe that living in a constant state of offendedness and fear brought on by regularly watching Fox news would, over time, cause changes in the brain.

I'm not joking. Why would anyone be surprised that the emotional climate one was raised in would affect activity in the amygdala, etc? It's not hardwiring, any more than soldiers are hardwired to develop brain activity consistent with PTSD.
posted by justsomebodythatyouusedtoknow at 8:54 PM on February 16, 2013 [2 favorites]


That doesn't make any sense, sorry.

Study: Because of huge pile of existing lit over here, and this other pile of existing lit, we think liberals and conservatives will show different neural structures. We looked at their brains and they do.


But they didn't look at the differences between liberals and conservatives, they looked at the differences between registered Democrats and registered Republicans. Those party affiliations have a more complex relationship than the study design acknowledges with any conception of liberality and conservatism besides tautological redundancy.

2. The researchers seem to conflate liberals with Democrats and conservatives with Republicans, which is a nauseating marriage of intellectual incuriosity and a banal legitimation of the inevitability and naturalness of our current political hegemony

This would have been a reasonable criticism in 1960, when there were significant numbers of conservative Democrats. As of 2006, though, only *looks* six percent of Democrats were conservative and only two percent of Republicans were liberal.


I don't know where you got those figures, but in any case I think you're again treating "liberal" and "conservative" as synonymous with being a registered Dem or Republican. If you conflate them, why use a general adjective in the first place?

3. One of the study designers is identified as "Dr. Darren Schreiber, a researcher in neuropolitics at the University of Exeter" despite the fact that neuropolitics sounds totally made-up

That's perhaps the world's silliest criticism. We know that he's a researcher in neuropolitics because he's an author of this study, which is clearly neuropolitics.


My point is that neuropolitics seems like an insubstantial discipline because its prominent members do crappy work which hides behind a veneer of sophistication which only partially masks the fact that it's utterly specious.

4. The study has an incredibly small N, and you know what they say about studies with incredibly small Ns (they are useless)

Advil has already noted that the N is not small for fmri, and I'll take his/her word on it. In any case, studies with small N's aren't useless, they just have less power, so you need bigger effects for them to stand out against the relatively higher statistical noise.


I don't think advil is necessarily wrong (I don't really know), but I'm not sure how relevant the standards, uh, of fmri are to a study of this type. Different disciplines have different standards for what a good-sized N is because they measure different kinds of phenomena, and to the extent that this work is sociological the N is very small. Also, more importantly, I think you may be mistaken about the utility of small Ns in general: smaller-N studies don't have less "power," they're more likely to capture anomalous statistical patterns which are then regarded as equivalent in representative terms to larger-N studies, which is bad. There isn't more statistical noise in those cases, but less, because there's fewer data points. I'm not sure your grasp of statistics is serving you well here.

5. The researchers apparently did not include people who were not registered with either major party in their study

Which is smart if you want to examine your predictions about neural structures in liberals and conservatives. It would certainly be dumb if you were trying to examine predictions about the neural structures of moderates or people too unsophisticated to have any coherent ideology, but then they weren't trying to do that.


They didn't examine neural structure, they examined patterns of brain excitation. Kind of a big difference.

I don't think you have any idea what you're talking about, dude. Sorry.
posted by clockzero at 9:27 PM on February 16, 2013 [3 favorites]


Wow. The snarky hate for conservatives at the start of the thread literally just gave me a stomachache.
posted by Conspire at 9:44 PM on February 16, 2013 [1 favorite]


justsomebodythatyouusedtoknow: Don't read the press release. Never read the press release. I have no idea what the fuck the press release people are doing. They don't say a goddamn thing about being hard-wired. Nobody ever said anything about being hard-wired. In fact, they interview the first author and he specifically mentioned that brains aren't hard-wired that way.

clockzero: They have a citation for the very big correlation between political party and political views. They also acknowledge the problem in the paper. They would acknowledge that this is a valid criticism of their paper, but they will tell you: there's a big damn correlation. A biiiig damn correlation, and as they note, it's a more significant one in California where they did the study.

I feel that their statistical reasoning for saying that their result is not noise would not necessarily hinge upon the N. Although their study is certainly above average for an fMRI study, the effect holds up under cross-validation, and it specifically doesn't overfit on cv. So you would have to make the criticisms against a study that holds up under cross-validation: is the model holding up perfectly well against something that isn't what the experimenter is talking about? You might repeat your point about the conflation of conservatives and liberals vs. democrats and republicans, but that's not the ultimate point of the study. Again, it's a simplification of the classification task. It's not a very easy statement to make, that there's a viable third party in the United States, or that there's an incredible surfeit of conservative Democrats and liberal Republicans in the US. If you wish to make that statement, make it with evidence, just as they made their statement with evidence that says that most Democrats are more liberal and most Republicans are more conservative.

I will probably agree on you that the "neuropolitics" stuff is bullshit. But it seems to me that the regressions are valid and the modelling also valid, and that is all they need in order to claim that their model is now the state of the art.
posted by curuinor at 9:51 PM on February 16, 2013 [3 favorites]


to the extent that this work is sociological the N is very small.

The study is not "sociological" in any relevant respect, as far as I can see, and I encourage you to look at the large literature that exists on the necessary sample size for fmri (it is a fairly important issue because scanner time is very expensive -- this study probably ran somewhere between 24000 and 80000 for just scanner time alone, back of the envelope).

Also, I'm afraid I'm with rou on the effect of size. Smaller studies most definitely have less power by the standard interpretation of that term -- likelihood of making a type II error (and "power" doesn't deserve scare quotes.)
posted by advil at 10:15 PM on February 16, 2013 [3 favorites]


I don't think advil is necessarily wrong (I don't really know), but I'm not sure how relevant the standards, uh, of fmri are to a study of this type. Different disciplines have different standards for what a good-sized N is because they measure different kinds of phenomena, and to the extent that this work is sociological the N is very small. Also, more importantly, I think you may be mistaken about the utility of small Ns in general: smaller-N studies don't have less "power," they're more likely to capture anomalous statistical patterns which are then regarded as equivalent in representative terms to larger-N studies, which is bad. There isn't more statistical noise in those cases, but less, because there's fewer data points. I'm not sure your grasp of statistics is serving you well here.

Small N-studies most definitely do have less power. From a sociology perspective, you might be thinking about how shitty it is when lit reviews or meta-analyses deal incorrectly with low Ns or something. Though this study is about a phenomenon that sociology has an interest in, for the variables it's actually measuring and the methods it uses to measure them, the N is higher than the norm.

I've worked in a language lab where an N of 82 would have been considered just about right. I've also worked in a vision lab where an N of 82 would have been considered ridiculous overkill- if your effect is worth writing about, you shouldn't need that many to hit significance. It is not atypical for fMRI studies to have an N in the twenties, which according to this study is about right. An N of 82 isn't setting off any alarm bells on my end.

The statistical tests used by the researchers in the paper discussed already take into account N*. They corrected for multiple comparisons. They partialled out other sources of variance (age, gender, and income) before letting their variable of interest take a swing at it. From my perspective it looks mostly OK statistically.

The only sin that I see is that they had a set of planned comparisons that didn't look significant, so they went back and expanded the regions of interest to be compared. They corrected for the increased number of voxels in their comparisons (which is good), but I don't know whether they just kept on increasing incrementally until they found a sweet spot that worked (which would not be good).

*Maybe you aren't worried about a test resulting in a Type I error; maybe you're worried about a potentially nonrepresentative sample compromising external validity. If you're concerned about that, maybe you should rail against basically every fMRI study ever rather than writing this one in specific off.
posted by Jpfed at 10:18 PM on February 16, 2013 [6 favorites]


That's perhaps the world's silliest criticism. We know that he's a researcher in neuropolitics because he's an author of this study, which is clearly neuropolitics.

If I may speak for clockzero, I think the point is that when we hear a scientist works in a certain field, and that a certain study is in a certain field, we are supposed to believe the study is trustable Science™ because obviously someone in a field knows what they are doing. But in this case, the field is made up by the scientist to lend credibility to the study that creates the field. It's a chicken-and-egg problem. The right thing to do is to do the study that only afterwards, when it has endured serious criticism, creates the field, not claim expertise in a field before it really exists to ensure the study catches on.

Other than that, I think this survey didn't do nearly enough to define and isolate the political belief variable. I grant that might be impossible, but that's no excuse for a useless study. In many fields of study there are studies that cannot be performed yet. It's okay to wait.
posted by michaelh at 12:12 AM on February 17, 2013 [1 favorite]


I often think it's comical
How Nature always does contrive
That every boy and every gal
That's born into the world alive
Is either a little Liberal,
Or else a little Conservative!
Fal lal la!
posted by Aiwen at 1:01 AM on February 17, 2013


"I don't think advil is necessarily wrong (I don't really know), but I'm not sure how relevant the standards, uh, of fmri are to a study of this type. Different disciplines have different standards for what a good-sized N is because they measure different kinds of phenomena, and to the extent that this work is sociological the N is very small. Also, more importantly, I think you may be mistaken about the utility of small Ns in general: smaller-N studies don't have less "power," they're more likely to capture anomalous statistical patterns which are then regarded as equivalent in representative terms to larger-N studies, which is bad. There isn't more statistical noise in those cases, but less, because there's fewer data points. I'm not sure your grasp of statistics is serving you well here."
clockzero, you know I love you, but this is mathematically illiterate bullshit and we can do better.

In general researchers try to use the smallest number of tests possible while still being able to say significant things about those tests in order to not waste money and time. Especially when using expensive tests like fMRI, researchers will often do a statistical power analysis, which is a real thing that does not need to be scare quoted, in order to see how large their sample size actually needs to be to detect an effect. They undoubtedly needed to do such a power analysis for the grant that got them the funding to do this, but what is important for us as readers to look at is the t-test they performed on the data they generated and its corresponding p-value. For the their two results they showed a t-test statistic of 3.153 for Democrats having a more active Left Posterior Insula with a corresponding p-value of 0.002 and a test statistic of 3.774 for Republicans having a more active Right Amygdala with a corresponding p-value of <0.001.

That P-value is a statistical measure of the significance of their data. It is defined roughly as the probability of obtaining a test statistic at least as extreme as the one that was actually observed, assuming that the null hypothesis (That your thinking is bullshit) is true. That the P-value they obtained is so dramatically low (where being below 0.05 is standard for most disciplines) indicates that what their data seems to indicate is in fact quite significant and not likely to be the result of random chance. With fMRI studies specifically one does need to be careful though as, at least at the scales fMRIs look at, the brain can be an awfully huge sample size to look at and it is thus really easy to generate small p-values that are in fact bullshit. Indeed, when you look at the whole brain and go fishing for significance you can get dead salmon to find love in pictures of human faces with p-values even lower than the ones these authors found [PDF]. The problem that the dead salmon helpfully exposed though really has nothing to do with the N used, but sample size in a much more abstract sense that breaks the usefulness of hypothesis testing as it is properly performed, explained helpfully here. However, the authors address this pretty solidly with the basic design of their study, looking only at these two regions of the brain specifically without doing too much fishing.

While it is worthwhile to be suspicious of these kinds of studies, as there are indeed a lot of bullshit fMRI papers out there, the discipline has cleaned itself up significantly over the last couple of years and there is a reason this paper was published in PLOS ONE. They make a very good case for what they are showing is real and a decent case for that being interesting. You can read the actual paper here:
Red Brain, Blue Brain: Evaluative Processes Differ in Democrats and Republicans
Liberals and conservatives exhibit different cognitive styles and converging lines of evidence suggest that biology influences differences in their political attitudes and beliefs. In particular, a recent study of young adults suggests that liberals and conservatives have significantly different brain structure, with liberals showing increased gray matter volume in the anterior cingulate cortex, and conservatives showing increased gray matter volume in the in the amygdala. Here, we explore differences in brain function in liberals and conservatives by matching publicly-available voter records to 82 subjects who performed a risk-taking task during functional imaging. Although the risk-taking behavior of Democrats (liberals) and Republicans (conservatives) did not differ, their brain activity did. Democrats showed significantly greater activity in the left insula, while Republicans showed significantly greater activity in the right amygdala. In fact, a two parameter model of partisanship based on amygdala and insula activations yields a better fitting model of partisanship than a well-established model based on parental socialization of party identification long thought to be one of the core findings of political science. These results suggest that liberals and conservatives engage different cognitive processes when they think about risk, and they support recent evidence that conservatives show greater sensitivity to threatening stimuli.
posted by Blasdelb at 1:56 AM on February 17, 2013 [8 favorites]


"In addition to what clockzero listed, there's also a correlation/causation problem. I can easily believe that living in a constant state of offendedness and fear brought on by regularly watching Fox news would, over time, cause changes in the brain."
So long as we're actually critically reading the actual fucking paper, this is the second paragraph of the discussion as well as the papers it cites,
"One intriguing remaining puzzle regards the direction of causality. One might infer that the differing brain structures identified by Kanai et al. suggest genetic foundations for the differences in ideology. However, recent work has shown that changes in cognitive function can lead to changes in brain structure [50], [51]. For instance, applicants who worked to learn the map of London in order to pass a knowledge test required of potential cab drivers demonstrated significant growth in their hippocampus, a brain region related to memory formation [52]."
Neuroplasticity: Changes in grey matter induced by training
Does the structure of an adult human brain alter in response to environmental demands1, 2? Here we use whole-brain magnetic-resonance imaging to visualize learning-induced plasticity in the brains of volunteers who have learned to juggle. We find that these individuals show a transient and selective structural change in brain areas that are associated with the processing and storage of complex visual motion. This discovery of a stimulus-dependent alteration in the brain's macroscopic structure contradicts the traditionally held view that cortical plasticity is associated with functional rather than anatomical changes.

Training induces changes in white-matter architecture
Although experience-dependent structural changes have been found in adult gray matter, there is little evidence for such changes in white matter. Using diffusion imaging, we detected a localized increase in fractional anisotropy, a measure of microstructure, in white matter underlying the intraparietal sulcus following training of a complex visuo-motor skill. This provides, to the best of our knowledge, the first evidence for training-related changes in white-matter structure in the healthy human adult brain.

Acquiring “the Knowledge” of London's Layout Drives Structural Brain Changes
The last decade has seen a burgeoning of reports associating brain structure with specific skills and traits (e.g., [1,2,3,4,5,6,7,8]). Although these cross-sectional studies are informative, cause and effect are impossible to establish without longitudinal investigation of the same individuals before and after an intervention. Several longitudinal studies have been conducted (e.g., [9,10,11,12,13,14,15,16,17,18]); some involved children or young adults, potentially conflating brain development with learning, most were restricted to the motor domain, and all concerned relatively short timescales (weeks or months). Here, by contrast, we utilized a unique opportunity to study average-IQ adults operating in the real world as they learned, over four years, the complex layout of London's streets while training to become licensed taxi drivers. In those who qualified, acquisition of an internal spatial representation of London was associated with a selective increase in gray matter (GM) volume in their posterior hippocampi and concomitant changes to their memory profile. No structural brain changes were observed in trainees who failed to qualify or control participants. We conclude that specific, enduring, structural brain changes in adult humans can be induced by biologically relevant behaviors engaging higher cognitive functions such as spatial memory, with significance for the “nature versus nurture” debate.
posted by Blasdelb at 2:16 AM on February 17, 2013 [1 favorite]


clockzero: Also, more importantly, I think you may be mistaken about the utility of small Ns in general: smaller-N studies don't have less "power," they're more likely to capture anomalous statistical patterns which are then regarded as equivalent in representative terms to larger-N studies, which is bad. There isn't more statistical noise in those cases, but less, because there's fewer data points. I'm not sure your grasp of statistics is serving you well here.

As others have pointed out, this is completely wrong. Lower N studies indeed have lower power, all other things being equal (why the scare quotes?). Your N has nothing to do with being "more likely to capture anomalous statistical patterns which are then regarded as equivalent in representative terms to larger-N studies." Whether you have a nonrepresentative sample is a function of your sampling *method*, not your sample. If you sample correctly, you must get weird samples every now and then. That's why we have statistics. If your sampling *method* is giving you a bad small sample, then increasing your N won't help. You'll just get a bigger bad sample.

Blasbelb: indicates that what their data seems to indicate is in fact quite significant and not likely to be the result of random chance.

There is no principled method within the frequentist statistical paradigm to make such a claim. Scientists often think this is the way to interpret a p value, but it simply cannot be interpreted this way. The logical jump from "the probability of obtaining a test statistic more extreme than the one observed, given the null hypothesis" to "the results are not likely to be the result of random chance" is not allowed.

My academic work is specifically in this area, so it is a bit of a personal crusade on my part to keep people from saying things like this...
posted by Philosopher Dirtbike at 2:17 AM on February 17, 2013 [2 favorites]


clockzero: I don't think you have any idea what you're talking about, dude. Sorry.
Take that, you Professor of Political Science, you!--Internet commenter.
posted by Sonny Jim at 2:50 AM on February 17, 2013 [3 favorites]


Participants in the study played a simple chance-game exercise where their guesses could lead to them gaining or losing a small amount of money, up to 80 cents. Their was a slight difference in how people who were registered as Democrats or Republicans reacted emotionally to playing the game. Republicans seemed to have a stronger reaction to losing 80 or 40 cents. The researchers were able to see these differences in their reactions in fMRI scans.

This provides no evidence at all about whether or to what extent the differences in their reactions can be attributed to genetics, "partisan environment" or any other possible explanation. It just seems like people registered as Republicans tended to take the game a little more seriously than Democrats. (And, hey, the fMRI scanner was working!) It tells us nothing at all about why.

Using fancy equipment does not transform BS research like this into science. There are people using brain imaging technology to do interesting research into how brains organize themselves and process information and what can go wrong with them. This study does not fall into that category.

I'm also curious about their sample. It was pretty young, and heavily Democratic. There were 60 Democrats and 22 Republicans, with the average ages for Democrats and Republicans just over 22 and 28, This makes me think the participants in the study were college students. College students do tend to vote for Democrats, with humanities and social science majors leaning more Democratic, and business and economics majors more likely to be Republican. If this is true, the Republicans who participated in the study would be likely to be business majors, so the conclusion we might be able to draw would be that business majors take money games more seriously than people in other majors, and they're also more likely to be Republicans.
posted by nangar at 8:27 AM on February 17, 2013 [1 favorite]


adamdschneider's Law: for any study in any research area drawing any conclusions whatsoever, there exists someone on MetaFilter who will tell you it is bullshit.
posted by adamdschneider at 9:02 AM on February 17, 2013 [2 favorites]


I don't know about this "adamschneider's Law" ... sounds like bullshit to me.
posted by Jonathan Livengood at 9:35 AM on February 17, 2013 [1 favorite]


Genuine thanks to all the people who made erudite and illuminating comments about statistics, and on my haphazard interpretation of what ROU_xenophobe said/meant about Ns and statistical power and about Ns in general. I'll be the first to admit that I'm not an expert in statistics, and my remarks about that were hastily-written to boot. Also, ROU_xenophobe, I didn't mean to be a jerk, hope I was not. It just happens naturally sometimes. I think I misunderstood you. (And thank you for linking to the actual article, Blasdelb, I hadn't found it before, and it does address a number of the concerns I articulated).

So, the statistical methodology might be totally fine in this study. That wasn't really the focus of my critique or interest anyway. My personal opinion is that the model the researchers use is arbitrary in its conflation of liberal with registered Democrat and conservative with Republican. The insight of the research, it seems, is actually that there exist two distinct patterns of neural activity in the amygdala and insula related in some way to the neuropsychological process of risk assessment which correspond strongly with political party affiliation in the US. That is interesting. However, it seems strange that the researchers had only two variables for political orientation. That's what I think is specious: the design seems to gratuitously constrain the potential for discovery, and does so in a way that may uncritically legitimate and reproduce existing political and ideological biases such as the idea that there are two and only two meaningful political orientations.

Also, Blasdelb, I did not know that you love me but I'm glad to hear it.
posted by clockzero at 10:53 AM on February 17, 2013


My personal opinion is that the model the researchers use is arbitrary in its conflation of liberal with registered Democrat and conservative with Republican.

It's noisier than it has to be, but very safe in 2013.

However, it seems strange that the researchers had only two variables for political orientation.

Just guessing, but reading between the lines it looks like what happened is that some of them had done a study on risk-taking, and then at some later time the whole gang ran the participants through the voter file to get their party registration.

You were not a jerk
posted by ROU_Xenophobe at 1:11 PM on February 17, 2013 [1 favorite]


what mom and dad think isn’t the endgame

I'm sure that I'm not alone in saying: mom and dad's thoughts were at loggerheads. Which led to my liberation politics: a pox on ALL their houses.

Later extended to organized religion (the Sophia tromping hand-in-hand with 'leaders' (mild image chosen to be delicate) throughout and all over human history).

Man will never be free until the last king is strangled with the entrails of the last priest. - vive Diderot
posted by Twang at 2:16 PM on February 17, 2013


These studies are always bizarre to me as someone from the UK, where politics break down very differently to the US. 'Liberal' and 'conservative' are so obviously not natural kinds to me that I find it difficult not to read these studies as a reductio ad absurdem of gender-based brain difference studies. The fact that both can be predicted from observable neurological differences shows that 'male' and 'female' is just as natural a binary pairing as 'Republican' and 'Democrat', ie not at all what the fuck.
posted by Acheman at 2:36 PM on February 17, 2013


> The researchers seem to conflate liberals with Democrats and conservatives with Republicans, which is a nauseating marriage of intellectual incuriosity and a banal legitimation of the inevitability and naturalness of our current political hegemony

> This would have been a reasonable criticism in 1960, when there were significant numbers of conservative Democrats. As of 2006, though, only *looks* six percent of Democrats were conservative and only two percent of Republicans were liberal.


> It's noisier than it has to be, but very safe in 2013.

Assuming that Republicans are conservative is actually a problem here, because the word has more than one meaning. The authors were following up on research indicating that conservative people are more risk-adverse. This almost true by definition if you mean 'conservative' in the sense of being conservative about your money or being cautious about change. But Republicans and political conservatives are not necessarily conservative in this sense — they can sometimes advocate pretty radical policies, especially if they're Tea Party Republicans — and liberals can be cautious and conservative with their money.

Under the assumption that conservatives would be more risk-adverse, based on previous research, the authors expected to find behavioral differences in how they played the game (Democrats would be more likely to take risky strategies) and differences in the level of activity of four different areas of the brain. They found no behavioral differences in Republicans and Democrats' approaches to playing the game, and no differences in activity in the areas they originally targeted. They only found significant differences in activity after they expanded their search to include larger areas of the brain. (They found some differences near two of the areas where they were originally expecting them.)

The "southern strategy" brought social conservatives into the Republican party, but I'm not sure how well that corresponds to cautious or conservative as personality trait. This may be part of the reason they didn't find the patterns they were originally expecting.
posted by nangar at 5:00 PM on February 17, 2013 [2 favorites]


Quote from one of the scientists commenting on the findings in the linked article: “If you went to Vegas, you won’t be able to tell who’s a Democrat or who’s a Republican, but the fact that being a Republican changes how your brain processes risk and gambling is really fascinating,” says lead researcher Darren Schreiber, a University of Exeter professor who’s currently teaching at Central European University in Budapest. “It suggests that politics alters our worldview and alters the way our brains process.”

This seems 100% backwards. Although the article goes on to suggest that there could be some impact of political choices on brain morphology and function similar to that seen in London taxicab drivers, it seems much more likely that people choose their politics based on their neural predispositions and that changing against this predisposition would be the exception born of strong external motivating factors (like mum and dad's often and aggressively pushed politics).
posted by Mental Wimp at 7:38 PM on February 17, 2013


people choose their politics based on their neural predispositions

So people living in red states are neurally predisposed to "conservatism," and people living in most urban areas are genetically predisposed to "liberalism?" Seems to me political preferences would be formed from social relationships and cultural environment mostly.
posted by Golden Eternity at 11:29 PM on February 17, 2013


It's great that in this enlightened year of 1890 1942 2013, scientists finally have the courage to explore the idea that Blacks Jews Conservatives are biologically more primitive than the rest of us, due to their excessively large lips noses amygdalas.
posted by wolfdreams01 at 10:14 AM on February 18, 2013


wolfdreams01, you haven't actually read the paper have you? Because if you could point to anything the researchers have actually said to indicate that that has any relationship to what they are exploring or how they feel that would be something actually worth posting in this thread.
posted by Blasdelb at 10:18 AM on February 18, 2013


"So people living in red states are neurally predisposed to "conservatism," and people living in most urban areas are genetically predisposed to "liberalism?" Seems to me political preferences would be formed from social relationships and cultural environment mostly."

This is not as strong a correlation as you seem to think it is.
posted by Blasdelb at 10:21 AM on February 18, 2013


wolfdreams01, you haven't actually read the paper have you? Because if you could point to anything the researchers have actually said to indicate that that has any relationship to what they are exploring or how they feel that would be something actually worth posting in this thread.

Oh, of course! Now I see my mistake. The scientists aren't actually coming out and saying that conservatives are more primitive, they're simply saying that they have more primitive characteristics - since amygdalas are the part of the brain almost universally considered to be associated with more primitive instinctual responses. I'm sorry; I should have recognized that distinction right away, especially since it's so familiar. It's sort of like on Metafilter, where you can't directly call somebody a "hate-filled idiotic liar" but you're totally allowed to say "Gosh, that comment sounds like something a hate-filled idiotic liar would say."

You are absolutely right - since nobody was ever directly accused of being more primitive or inferior and it was only strongly insinuated, I am totally convinced now that this is a completely impartial study. Eugenic hair-splitting for the win!
posted by wolfdreams01 at 10:41 AM on February 18, 2013


You.. you still haven't even looked the paper we are discussing have you?

The word 'primitive', much less the phrase 'more primitive characteristics', never appears in either the paper we are actually talking about, nor its supporting information, nor the news article about it linked in the FPP. What the hell are you talking about?
posted by Blasdelb at 11:07 AM on February 18, 2013


I'm sorry Blasdelb, I thought you were being sarcastic, but I realize now that you genuinely don't know what the amygdala does. My apologies - I've dated a lot of psychologists and the occasional neuroscience PhD, so over time I've just became so accustomed to certain concepts that I guess I automatically assumed that they had entered the vernacular. My apologies - I'll address your response seriously here.

This is a slight oversimplification, but the amygdala is the portion of the brain responsible for more primal "instinctual" responses - fear, aggression, hunger, etc. When the amygdala is activated, it makes us "downshift" and use only our lower brain. That means we sometimes cannot think of a logical solution to a problem. So when the scientists in this article say that conservatives use their amygdalas more often in response to problems, they are actually using scientific terminology to veil the degrading subtext of their thesis, which is that conservatives are more instinct-driven and animalistic than liberals.

Does what I was saying make more sense to you now?
posted by wolfdreams01 at 11:33 AM on February 18, 2013


You could think of the study as couching some slur against conservatives in the language of neurology. Or you could think of it as presenting conclusions backed by evidence. If you don't believe that the study accurately characterizes differential activation of the amygdala, then you can rest assured that the next PI to measure this will fail to replicate and the facts will win.

Or you could just recognize that the study could be right and the value judgment is on your end. The amygdala is super-important and super-useful! Just because it may be activated more in conservatives under some circumstances says nothing about whether conservatives are well-adapted to their environment.
posted by Jpfed at 11:47 AM on February 18, 2013


Also, if Blasdelb was actually ignorant of the functions of the amygdala, I will eat my hat.
posted by Jpfed at 11:47 AM on February 18, 2013 [4 favorites]


You could think of the study as couching some slur against conservatives in the language of neurology.

Thanks, that's exactly how I think of it.

Or you could think of it as presenting conclusions backed by evidence.

Given that similar "Incontrovertable Scientific Evidence!" claims were historically used to assert the atavistic character of both Blacks and Jews, I will pass on following that line of reasoning.

If you don't believe that the study accurately characterizes differential activation of the amygdala, then you can rest assured that the next PI to measure this will fail to replicate and the facts will win.

Great - problem solved! Other commentors here have already addressed the numerous flaws inherent to this study (the exceptionally small data set, the fact that one of the people running it made up a scientific field just so he could claim expertise, etc) so there's no need for me to rehash it in detail. I'll just sit back and wait for more studies to expose this as the claptrap that it is.
posted by wolfdreams01 at 12:01 PM on February 18, 2013


"Please note that the text below was written by the feeling-being ‘Peter’ and feeling-being ‘Vineeto’ while they lived in a pragmatic (methodological), still-in-control/same-way-of-being Virtual Freedom before becoming actually free."
*bbthbthbht*

The authors actually take the time to explain their rational for exploring these two regions explicitly, as it has non-intuitive relevance to the statistical validity of all of it, and it bears no relationship to your fevered paranoia or who you say you've dated. Here let me just go ahead and paste the most relevant portion into this thread so that you don't even need to click the damn link,
"A large body of research suggests that liberals and conservatives differ on important psychological characteristics [1]. For example, conservatives demonstrate stronger attitudinal reactions to situations of threat and conflict. In contrast, liberals tend to be seek out novelty and uncertainty [1]. Moreover, Democrats, who are well known to be more politically liberal, are more risk accepting than Republicans, who are more politically conservative [2]. While ideology appears to drive reactions to the environment, environmental cues also influence political attitudes. For instance, external threats prime more conservative attitudes among liberals, moderates, and conservatives [3].

These ideological differences between political partisans have been attributed to logical, psychological, and social constraints [4] and past scholarship has focused primarily on institutional political processes or individual policy preferences, rather than biological differences in evaluative processes. But recent work has revealed physiological correlates of the differential responses to risk and conflict by liberals and conservatives. Consistent with the previously identified attitudinal divergence, conservatives have more intense physical reactions to threatening stimuli than liberals [5]. Conversely, liberals had stronger physiological responses to situations of cognitive conflict than conservatives [6].

Risk taking, the tendency to select an action where there is an uncertain potential for a relatively large beneficial outcome but also the possibility of an adverse outcome [7]–[9] requires balancing conflicting drives to obtain reward and avoid possible losses [10]–[12]. Risk taking is also closely related to and influenced by subjective perception and apprehension of threat [13], [14]. Considering differential physiological responses to threat and conflict by liberals and conservatives, examining neural processes during risk-taking decision-making may be an important avenue for understanding the link between mental processes and political preferences.

The discovery by Kanai and colleagues [15] that four brain regions implicated in risk and uncertainty (the right amygdala, left insula, right entorhinal cortex, and anterior cingulate (ACC)) differed in liberals and conservatives provided further evidence that political ideology might be connected to differences in cognitive processes. In the context of risk-taking decision-making, the amygdala is thought to be important for the processing of affective attributes involved in decision making [16]–[18]. The insular cortex is involved in the representation of internal bodily cues crucial for subjective feeling states and in signaling potential changes in interoceptive state to possible decision-related outcomes [10], [11], [19], [20]. Further, intolerance of uncertainty is related to posterior insula functioning [11]. The ACC is involved in conflict and error monitoring and in action selection [21], [22]. Thus, the regions implicated in risk and conflict, cognitive processes during which liberals and conservatives have been shown to differ in physiological response, are the similar regions shown by Kanai et al. to differ structurally in liberals and conservatives. If patterns of brain activity in these regions during the evaluation of risks could dependably differentiate liberals and conservatives, then we would have further evidence of the link between mental processes and political preferences.
"
These authors are not operating in the same woo-based space as your two links and really have no business being judged using them.
posted by Blasdelb at 12:07 PM on February 18, 2013


This is not as strong a correlation as you seem to think it is.

I still see a lot of polarization in the "purple" map, just delineated by smaller regions than state boundaries. There could also be very conservative and liberal cultures relatively co-located (born again Christian fellowship, hippie commune, nerdy science community). And it seems to me whichever culture one happens to grow up in would have a huge impact on their world outlook.

I guess I could also see how relatively predetermined personality traits might lead people to separate into like-minded communities.
posted by Golden Eternity at 12:09 PM on February 18, 2013 [1 favorite]


"Great - problem solved! Other commentors here have already addressed the numerous flaws inherent to this study (the exceptionally small data set, the fact that one of the people running it made up a scientific field just so he could claim expertise, etc) ..."
If you had actually bothered to even read this thread you are commenting in you would have discovered, from people actually familiar with these techniques, that the authors not only have an exceptionally large number of participants for the type of study they are doing, but that the typical problem with these kinds of studies if anything has more to do with the data sets they provide being much to large to handle honestly with ordinary statistical tools than too small for, well, anything. Also, people making up scientific fields is pretty much how they all start.
"... so there's no need for me to rehash it in detail. I'll just sit back and wait for more studies to expose this as the claptrap that it is."
Oh please do, I am awfully curious to read this.
posted by Blasdelb at 12:19 PM on February 18, 2013


Also, people making up scientific fields is pretty much how they all start.

Yes, like Phrenology. That's exactly how they all start. I'm glad you see that.

Oh wait, by "how they all start," were you talking about new scientific disciplines? I thought you were referring to doctrines of scientific racism.
posted by wolfdreams01 at 12:47 PM on February 18, 2013


Yes, like Phrenology, and every other purportedly scientific discipline you could hope to think of. You can tell the difference between things like phrenology and things like molecular genetics, which started when this Nobel Laureate made it up so he could be an expert in it, by looking at the methods used to see if they actually show what they are purported to show, seeing if the results obtained are verified by results from other disciplines, seeing if the results shown are consistent across multiple labs, and a little bit of critical thinking. Indeed these authors make a solid case for both the validity and meaningfulness of their results, they cite research in other disciplines that supports different aspects of their model, and they cite previous research done in other fMRI labs that confirm different aspects of what they are doing.

If you actually read that wikipedia article, you will not find any entries related to you, your funny 'conservative' ideas about women, or your sad penchant for liberal baiting. I would still love to hear your rehash of what you feel is wrong with this study though, the more detail the better.
posted by Blasdelb at 1:52 PM on February 18, 2013


I've dated a lot of psychologists and the occasional neuroscience PhD,...

I'm stunned. I've never heard this justification for expertise before.
posted by Mental Wimp at 2:12 PM on February 18, 2013 [4 favorites]


Oh, of course! Now I see my mistake. The scientists aren't actually coming out and saying that conservatives are more primitive, they're simply saying that they have more primitive characteristics - since amygdalas are the part of the brain almost universally considered to be associated with more primitive instinctual responses.

I think some fee-fees have been hurted.
posted by Mental Wimp at 2:13 PM on February 18, 2013


If you actually read that wikipedia article, you will not find any entries related to you, your funny 'conservative' ideas about women, or your sad penchant for liberal baiting.

Blasdelb, I haven't mentioned anything negative about you in this whole thread, so I would like to know why you are making this argument so personal. Do you normally include ad hominem attacks in your debate tactics, or is it only when you feel intellectually outmatched? I ask because I'm not quite sure whether I should feel insulted or flattered. Either way, I flagged your comment, and I would appreciate if you stopped doing that.
posted by wolfdreams01 at 2:13 PM on February 18, 2013


Not sure if he's attacking you, personally, or your skimming-before-commenting. All he's doing is pointing out the facts of the experiment; you're the one who seems to be taking this personally.
posted by Marisa Stole the Precious Thing at 2:30 PM on February 18, 2013


How about y'all just let that side conversation drop entirely, yes. wolfdreams01, this has become largely a derail about What Wolfdreams Thinks About This and you need to cut this sort of thing out and pretty much step away from this thread now.
posted by cortex at 2:36 PM on February 18, 2013


1. The researchers apparently do not know or understand that there exist political orientations besides liberal and conservative, which is specious, moronic and fatally biases the results by excluding potentially meaningful categories

The authors of the study already anticipated this:

A critical unresolved problem common to studies of the formation of ideology on both individual and institutional levels is the process through which a high dimensional space of distinct values, preferences, or issues is reduced to a low dimensional ideological space [3]. It is even less clear why voters and their representatives in government should organize political attitudes into apparently constrained bundles that are relatively consistent over time [47].

You can object to the authors using a 1-dimensional left-right, liberal vs. conservative model of ideology, but there is already a huge scholarly literature (some of which is cited in the paper) that predicts how liberal and conservative people will differ in their neurological processes. There is no similar body of scholar literature that would predict how people would differ in their neurological processes based on some other model of political ideology, such as the 2-dimensional Nolan Chart. Similarly, there are parsimonious testable hypotheses about how liberals vs. conservatives will differ in brain activity, but there are no similar parsimonious hypotheses about liberals and conservatives will differ from anarchists, libertarians, populists, authoritarians, fascists etc.

2. The researchers seem to conflate liberals with Democrats and conservatives with Republicans, which is a nauseating marriage of intellectual incuriosity and a banal legitimation of the inevitability and naturalness of our current political hegemony

The use of Democratic vs. Republican partisan registration as an indicator for liberalism vs. conservatism is not due to "a nauseating marriage of intellectual incuriosity," but the only indicator that was available to the researchers under the constraints of the study. The original research was a politically neutral fMRI study about risk taking. The researchers only traced the participants back to their partisan registration until after they had undergone the fMRI. The authors also address this concern:

To test a conjecture that ideological differences between partisans reflect distinctive neural processes, we matched publicly available party registration records with the names of participants (35 males, 47 females) who had previously taken part in an experiment designed to examine risk-taking behavior during functional brain imaging. Ideally, we would have also directly inquired about the individuals’ ideological self-identification and attitudes about a set of political issues. However, we were not able to re-contact the participants.

In addition, ideological self-identification and self-reported attitudes about political issues might actually be a less valid indicator of political ideology than actual political behavior, as reflected in voter registration and participation in partisan primaries. For one thing, a person's ideological self-identification might not match his or her self-reported political attitudes. For example, back in 1967, Lloyd Free and Hadley Cantril published a book that argued that many Americans are ideological conservatives and operational liberals who prefer to label themselves "conservative," even though they prefer the substance of liberal policies more. In addition, self-reported information about ideological views or political attitudes may be falsified by study participants based on what they think the researchers want to hear.

I also want to address the argument that the study is "a banal legitimation of the inevitability and naturalness of our current political hegemony." Yes, you could make the argument that the Democratic Party in the United States is really a center-right party, and the only true "liberals" in American society are people who are politically to the left of the Democratic Party. The authors of the current study could certainly test the robustness of their research by recruiting members of the Green and Peace & Freedom Parties (i.e., the only two "qualified parties" on the official California voter registration form to the left of the Democrats) to participate in a new fMRI study, but this could also be problematic, because (1) some Greens refer to themselves as "neither left nor right but Green" and (2) some members of the Green and Peace & Freedom Parties might prefer to refer to themselves as leftists or radicals instead of liberals. On the other hand, if the hypothesis of the current study is robust and Greens and Peace & Freedom members can be viewed as "more liberal" than the Democrats, then members of the Green and Peace & Freedom Parties should have even more active in the insula region of the brain than the Democrats in this study.

3. One of the study designers is identified as "Dr. Darren Schreiber, a researcher in neuropolitics at the University of Exeter" despite the fact that neuropolitics sounds totally made-up

Dr. Schreiber is only referred to as "a researcher in neuropolitics" in second-hand journalistic accounts of the study, not in the study itself or on his Exeter page or his faculty page. You can't blame a researcher for what journalists label him as, if he didn't ask to be labeled that way himself.

As for the charge that "neuropolitics sounds totally made-up," there is a Wikipedia entry that disagrees with you, although the article is currently rated as only C-class. There's a Timothy Leary book from 1977 called Neuropolitics, but I've been able to find references to "neuropolitics" in peer-reviewed scholarship in Google Scholar that go back as far as this 1982 journal article. In other words, "neuropolitics" is a label that Dr. Schreiber neither invented nor did he ever specifically apply to himself.

4. The study has an incredibly small N, and you know what they say about studies with incredibly small Ns (they are useless)

The sample size was sufficiently large to show an improvement in predicting political partisanship based on neurological data alone than the most common method (i.e., predicting based on the partisanship of one's parents). I'd hardly call that useless.

Just to be sure, I used an online app to calculate the 95% confidence intervals for both methods. The parental method was done with 347 respondents from a preexisting political science survey data set. I get a 95% confidence of 64.6% to 74.3%. The neurological method using the amygdala and insula was done with 82 test subjects. Using the same online app, I get a confidence interval of 74.75% to 91.05%. Since the confidence intervals do not overlap, you can state with 95% confidence that there is a statistically significant difference in the effectiveness of the neurological method of classification vs. the parental method (with neurological method being more accurate).

5. The researchers apparently did not include people who were not registered with either major party in their study

If the researchers had included people who marked "Decline to state" on the California voter registration form, they would have no way of telling whether these people lean liberal/Democratic or conservative/Republican. Aside from the fact that many political independents are really "disguised partisans", adding political independents to the study would have simply added statistical noise to the data which would have made it impossible for the researchers to prove or disprove their initial hypothesis at all.

In addition, if you had included members of third parties, such as Libertarians and Americans Elect, where would you have put them on the political spectrum? Or could you have put them on a 1-dimensional political spectrum at all? And even if you could come up with a perfect multidimensional model that captures every American's possible political views, there's no scholarly research that would suggest that this multidimensional model can be used to predict how people placed differently in the model would look different in an fMRI study.

6. This study is so full of unjustified assumptions, arbitrary exclusions, old-fashioned shitty research design and outright nonsense that it's impossible for any serious individual to take it seriously.

In the future, I'd actually engage a lot more with the actual research paper (as opposed to second-hand blog posts or university press releases) before I accused somebody of "unjustified assumptions, arbitrary exclusions, old-fashioned shitty research design and outright nonsense."
posted by jonp72 at 3:16 PM on February 18, 2013 [3 favorites]


adamdschneider's Law: for any study in any research area drawing any conclusions whatsoever, there exists someone on MetaFilter who will tell you it is bullshit.

jonp72's corollary to adamdschneider's Law: for any study in any research area drawing any conclusions whatsoever, there exists someone on MetaFilter who will tell you it is bullshit without reading the frickin' article.
posted by jonp72 at 3:19 PM on February 18, 2013 [4 favorites]


« Older Elizabeth Warren, in her new role as a member of t...  |  Having previously written an a... Newer »


This thread has been archived and is closed to new comments