Out Of Control

To a first approximation, science is about differences between groups: if death is a major motivator for human behavior, then we should expect people who have been reminded of their mortality to act differently than people who were not. Sometimes, a group is only conceptual: if the Higgs boson exists, then when slamming two protons together there should be more photons observed in the aftermath, compared to the amount we predicted. Sometimes, we generate groups after the fact: if we plot star colour and luminosity on a Hertzsprung-Russell diagram, stars naturally settle into four major groups. Sometimes, the “difference” we care about is that there is no difference at all: if a cosmetic is safe to use, then if we compare a group of people who use it to people who don’t, we should observe no difference in health. These divisions are so common that we often neglect to clearly delineate our groups: “does a daily dose of Aspirin prevent strokes?” implies that people who take Aspirin are less likely to get a stroke than people who do not.

At some point these groups must be clearly delineated, however; when they are not, a common problem in epidemiology, we lose our ability to find differences between them. Worse, fuzzy groups allow us to manufacture differences that don’t exist, say by classifying legitimate data as illegitimate “outliers” to get the results we want. This “differences between groups” metaphor is surprisingly powerful, to the point that’s a good solution to the demarcation problem. A core claim of astrology is that people differ based on the day they were born; if we divide people into those groups, yet fail to find differences, then astrology cannot be true. The Myers-Briggs personality test claims that we can divide people into specific groups, yet studies that use difference to reconstruct groups have failed to see those groups materialize.

By convention, if we are testing whether or not some change leads to a difference, we call the group we don’t change the “control” group. This group is often conceptual, thanks to frequentist statistical techniques, but that only works if the tools we use to find difference are perfectly calibrated; if they are not, the data might be biased and you’d never know. As a result, lacking a control group is considered a reason to suspect your results.

I apologize if all that was painfully obvious; I grasped these concepts way back in Junior High, well before I was legally allowed to drive. Still, I needed to type it out to convey the pain of what comes next.

[Read more…]

If At First You Don’t Succeed

Beginning in August 2017, the trio wrote 20 hoax papers, submitting them to peer-reviewed journals under a variety of pseudonyms, as well as the name of their friend Richard Baldwin, a professor emeritus at Florida’s Gulf Coast State College. Mr. Baldwin confirms he gave them permission use his name. Journals accepted seven hoax papers. Four have been published.

Does that sound familiar? It should.

The three academics call themselves “left-leaning liberals.” Yet they’re dismayed by what they describe as a “grievance studies” takeover of academia, especially its encroachment into the sciences. “I think that certain aspects of knowledge production in the United States have been corrupted,” Mr. [Peter] Boghossian says. Anyone who questions research on identity, privilege and oppression risks accusations of bigotry.

Yep, after attempting to discredit all of gender studies by publishing a fake paper in a pay-to-publish journal, and being dismayed that no-one thought gender studies had been discredited, Boghossian and crew decided to repeat the experiment, only bigger. There is a unique spin on it this time, however.

While fat activism has disrupted many dominant discourses that causally contribute to negative judgments about fat bodies, it has not yet penetrated the realm of competitive bodybuilding. The author introduces fat bodybuilding as a means of challenging the prevailing assumptions of maximally fat-exclusionary (sports) cultures while raising fundamental ontological questions about what it means to “build a body.” Specifically, he advocates for imagining a new classification within bodybuilding, termed fat bodybuilding, as a fat-inclusive politicized performance and a new culture to be embedded within bodybuilding.

Baldwin, Richard. “Who are they to judge? Overcoming anthropometry through fat bodybuilding.” Fat Studies (2018): 1-13.

That’s one of their hoaxes. But if you read it carefully, you can see a legitimate point.

Conceptually, fat bodybuilding emerged from applying that lens to a prototype: a disruptive “fathletic” event, the “Fattylympics.” The Fattylympics was an act of cultural disruption undertaken as a nonprofit community event in East London in 2012 to satirize the Olympics and offer a different take on “sport, bodies, community, [and] protest” (…). The Fattylympics ultimately relies on (Judith) Butlerian parodic performance, which has been effectively utilized as a culturally disruptive tool, especially with regard to gender/queer activism (…). Here, as Monaghan, Colls, and Evans (2015) explained, “Fattylympics illustrated the possibility of claiming a public space for resisting the dominant anti-fat ethic of sport and physical activity, constructing an alternative value set for active bodies and critically understanding the relationship between fat and health” (117).

“Baldwin” (2018), pg. 3-4

The bit about Judith Butler is pure nonsense that should have been caught during peer review, but their overall proposal is rooted in legitimate body-positive activism. Look at pictures of female weight lifters, and you’ll find two basic body types. The first has a “conventional” body type with minimal fat, not too dissimilar from Michelle Rodriguez or Ronda Rousey.

Type-1 Weightlifters, via Google Image Search.

Type-2 Weightlifters, via Google Image Search.

But there’s a second type, with the stocky barrel-chest that’s more typical of “World’s Strongest Man” events. Women like this are incredibly rare in pop culture; the only example I can think of is Zarya, and she’s a fictional videogame character. The net result is that we’re discouraging or minimizing an entire class of women because they don’t look the way we expect them to. At the same time, it’s clear body fat is not much of a factor in weight-lifting performance. So if we wanted to break body stereotypes, “fat bodybuilding” is a great choice.

“We understood ourselves to be going in to study it as it is, to try to participate in it,” Ms. [Helen] Pluckrose says. “The name for this is ethnography. We’re looking at a particular culture.”

Each paper “combined an effort to better understand the field itself with an attempt to get absurdities and morally fashionable political ideas published as legitimate academic research,” Mr. Lindsay wrote in a project summary. Their elaborate submissions cited and quoted dozens of real papers and studies to bolster the hoax arguments. […]

The trio say they’ve proved that higher ed’s fixation on identity politics enables “absurd and horrific” scholarship. Their submissions were outlandish—but no more so, they insist, than others written in earnest and published by these journals.

The Dunning-Kruger effect is when you are so ignorant of what you’re ignorant of that you think you’re knowledgeable. But if you don’t know anything about gender studies, how can you tell a legitimate paper from a hoax? By doing extensive research to write a hoax paper, yet nonetheless accidentally creating a legitimate one, Boghossian, Lindsay, and Pluckrose have proven beyond a shadow of a doubt they know jack-shit about gender studies. You will not find a better example of the Dunning-Kruger effect than that trio!

Mr. Boghossian doesn’t have tenure and expects the university will fire or otherwise punish him. Ms. Pluckrose predicts she’ll have a hard time getting accepted to a doctoral program. Mr. Lindsay said he expects to become “an academic pariah,” barred from professorships or publications.

Yet Mr. Lindsay says the project is worth it: “For us, the risk of letting biased research continue to influence education, media, policy and culture is far greater than anything that will happen to us for having done this.”

Oh, I sincerely hope the trio are made academic pariahs. I also hope they achieve enough self-awareness to realize the true reason why.


[HJH 2018-10-03]: I had plans to revise to tack on an addendum. After all, the original paper was about bodybuilding, not weight-lifting, and there’s still the obvious retort “but their goal was to fool you into making a legitimate paper, so aren’t you admitting they succeeded?”

And then I read their methodology, and I realized I didn’t have to.

Specifically, over the course of a year we wrote twenty academic papers and submitted them to significant peer-reviewed academic journals in these fields with the hopes of getting them published. Every paper combined an effort to better understand the field itself with an attempt to get absurdities and morally fashionable political ideas published as legitimate academic research. Some papers took bigger risks in this regard than others. […]

We wrote academic papers targeting (mostly) highly ranked, peer-reviewed journals in fields we are concerned might be corrupted by scholarship biased by “grievance studies.” These papers were submitted to the best journals we could find, given constraints of the journals’ aims and scopes, and then we used the feedback we received about them from editors and peer reviewers to improve them and our future papers. […]

Each paper was submitted to higher-ranked journals first and then down a line of suitable alternatives until one of the following occurred: it was accepted; it was deemed too unlikely to succeed for reasons we came to understand to continue with it; or we ran out of time.

They had twenty papers going at once, yet by their own admission they made 48 “new submissions.” It’s not clear if “new submissions” includes the original submission, so let’s be charitable and say it does. That means that, on average, each paper went through one and a half rounds of peer review. Peer review is probabilistic: reviewers can vary substantially in terms of how much effort and scrutiny they put in, so if you keep submitting a paper over and over you might get lucky and get lazy reviewers. When you’re submitting twenty papers, you make that much more likely for one of them. When you’re editing your papers according to reviewer feedback to make them better fakes, you raise the odds of that even higher. On top of that, after those edits they’d take the paper to another journal with less prestige, and presumably lower standards for peer review.

It’s like watching evolution in action. The authors kick out what they think are nonsensical ideas; since they know jack-shit about the field they’re trying to discredit, some of those turn out to be legitimate by accident, or nearly so. These do well in peer review, though from the looks of it even their best work needed a second round; it took five months to get their first acceptance, yet the median review time is about three months. Either way, the best of the bunch get edited, accepted, and then published. The failures die out or get edited until they join these “successes.”

In reality, the methodology is heavily rigged to generate “success.”

Speaking of which, let’s look at what counts as a success. Here are the articles they got published:

Wilson, Helen. “Human reactions to rape culture and queer performativity at urban dog parks in Portland, Oregon.” Gender, Place & Culture (2018): 1-20.

Smith, M. “Going in Through the Back Door: Challenging Straight Male Homohysteria, Transhysteria, and Transphobia Through Receptive Penetrative Sex Toy Use.” Sexuality & Culture (2018): 1-19.

Richard Baldwin, “Who are they to judge? Overcoming anthropometry through fat bodybuilding”, Fat Studies, DOI: 10.1080/21604851.2018.1453622, published online on 10 April 2018.

Baldwin, Richard. “An Ethnography of Breastaurant Masculinity: Themes of Objectification, Sexual Conquest, Male Control, and Masculine Toughness in a Sexually Objectifying Restaurant.” Sex Roles (2018): 1-16.

Of those four, two were retracted within days of the news coming out. That’s a damn quick turnaround! Say what you will of the peer review process, but quickly scrubbing nonsense from the scientific record isn’t what you’d expect if the field of gender studies was lax about rigor.

Er, sorry, I mean “grievance studies,” the term Boghossian et al. use. What does that term mean, anyway? Emphasis mine:

The specific problem we targeted has various names in various quarters and is difficult to pin down. Careful academics would refer to it as “critical constructivism” and/or “blank slatism” and its scholars as “radical constructivists.” (In this sense, it is the descendants of postmodernist and poststructuralist thought from the mid 20th century.) Pundits have termed it “academic leftism” or “cultural studies” and identify it with the term “political correctness.”

We prefer to call it “grievance studies” because many of these fields refer to themselves as “[something] studies” and because they operate primarily by focusing upon and inflaming the grievances of certain identity groups.

Uh, “critical constrictivism” and “blank slatism” have nothing in common with each other, and the latter doesn’t exist except as a straw. “Academic leftism” is bad, according to three self-proclaimed “left-leaning liberals?” “Political correctness” has no academic meaning at all. “Grievance studies” has as much coherence as ghosts!

Even if we steel-person the argument and go with “grievance studies” as “focusing upon and inflaming the grievances of certain identity groups,” how does promoting increased acceptance of overweight people fit under that banner? How does making men less homo- and trans-phobic via anal sex toys “focus” and “inflame grievances” in certain groups? How about observing a unique pattern of sexism in “breastaurants?” None of their published papers qualify as “grievance studies” papers, for the most charitable definition of “grievance studies,” so they cannot draw any conclusions about the rigor of that field. Even if their methodology was absolutely perfect, these three still cannot prove what they claim to.

Shit, I’ve seen ghost hunters with a more coherent world view. Is this what organized skepticism has been reduced to?!


[HJH 2018-10-04]: Looks like someone else came to the same conclusion as I did, only on a different paper:

I read the article that Hypatia accepted, “When the Joke Is on You: a Feminist Perspective on how Positionality Influences Satire.” In my opinion, if the citations are legitimate and the descriptions of others’ views are accurate (something which I am not in a position to determine at this time), the editors of Hypatia have nothing to be particularly ashamed of. Most of the twenty-page paper is a reasonable synthesis of others’ ideas about oppression and humor. It may not be groundbreaking (as one of the reviewers points out), but it is not ridiculous. It seems to me that only on the last page of the paper are there certain statements that could be interpreted as outrageous, but they are so vague that a much more charitable alternative interpretation would be reasonable. In short, assuming accurate representations of others’ views and legitimate citations, one’s opinion of Hypatia should not be affected by its publication of this paper.

Now I know some of you won’t believe me. So please, read the paper for yourself. It’s right here (look for the document titled “HOH2 Typeset”). You can also read the referee reports and editors comments here (look for the document titled “HOH2 ReviewerComments”). Let me know what you think.

As that last paragraph implies, Boghossian and friends have released their manuscripts to the public. Now you don’t have to take my word for it.

Little Lies and Big Truths

Brett Kavanaugh lied. Yet, as I just pointed out, Republicans are still fighting hard to put him on the Supreme Court, ignoring any damage to the (admittedly quite cracked) political neutrality of the court.

The most obvious explanation is that Kavanaugh is one of their own. Jeff Flake declared “I’m a conservative. He’s a conservative;” Kavanaugh shored up his Republican support by spinning conspiracy theories about a vast Democratic coalition trying to take him down, conspiracies we see echoed by Donald Trump, Mitch McConnell, and Lindsay Graham; and the arbitrary deadline of one week was to prevent a partisan “fishing exposition.”

Partisanship doesn’t explain everything, though. Take Donald Trump: he wasn’t much of a Republican, has been at odds with his own party and allies repeatedly, yet is still enjoying broad support from Republicans of all stripes. There’s got to be something more at work here.

A special-access lie is a deliberately false statement based on facts about which the speaker is thought to have special access. A good example of such a lie is Bill Clinton’s notorious false claim that he “did not have sexual relations with that woman” (i.e., Monica Lewinsky). […] A common-knowledge lie is quite different. This is a false assertion about facts to which the speaker has no special access. … For instance, Trump often pointed to information that was supposedly in the public domain to support his claims, even if it was easily demonstrable that such supporting evidence did not exist (e.g., his claim that his election victory was “the biggest electoral college win since Ronald Reagan,” or his claims regarding the size of the crowd at his inauguration). As such, the ideal-typical case of this type of lie is one in which the speaker not only knows the statement is false, but she knows her listeners also know that she knows the statement is false; it is thus common knowledge that the statement is false.

Hahl, Oliver, Minjae Kim, and Ezra W. Zuckerman Sivan. “The Authentic Appeal of the Lying Demagogue: Proclaiming the Deeper Truth about Political Illegitimacy.” American Sociological Review 83.1 (2018): 7-9.

A tweet by one of the study authors suggested what that “more” could be. “Common knowledge” lies are false statements that are either known to be lies or could easily be verified to be a lie. Why do these types of lies exist? They signal something to the listener.

In particular, whereas the speaker of a special-access lie is implicitly upholding the norm of truth-telling, the common-knowledge liar is implicitly attacking this norm. Following Frankfurt (2005), such a liar is a type of “bullshit artist”: he is publicly challenging truth as a prescriptive norm. … Insofar as a speaker seems capable of distinguishing between truth and falsehood and yet utters a statement everyone knows is false, the speaker is flouting the norm of truth-telling and inviting his listeners to endorse such violations. Indeed, listeners are complicit in the norm violation as long as they do not challenge him—and especially if they applaud him.

Hahl (2018): 9.

In the general case, the speaker is arguing that everyone lies, but no-one wants to admit it. By breaking that taboo, they flag themselves as speaking truth to power, even if they themselves are quite powerful. For instance, Kremlin propaganda doesn’t argue Russia is free of corruption, instead it argues every country is corrupt. Admitting to this truth gains your trust and allows them room to be corrupt, plus denies any way to actually fix corruption.

A minority—or even a majority under some conditions (…)—may privately disagree with publicly-endorsed norms, but a group’s established leadership (however formal or informal) tends to determine group membership, at least in part, based on compliance with such norms. Accordingly, individuals who seek social acceptance generally have an incentive to hide their deviance through public compliance and even to enforce a norm they do not privately endorse (…). […] Put differently, voters have two ways to determine a candidate’s authenticity. One
approach is to determine authenticity on the basis of the candidate’s sincerity or prosociality: inauthentic candidates are those who tell lies or who violate publicly-endorsed norms. A second approach for determining authenticity is based on the implicit claim of the lying demagogue – that is, publicly-endorsed norms are imposed rather than freely chosen. The lying demagogue thus claims to be an authentic champion of those who are subject to social control by the established political leadership.

Hahl (2018): 10-11

People may say they never got drunk in high school or college, but Kavanaugh is indirectly calling them liars. By lying about Dr. Blasey Ford’s testimony and Leland Ingham Keyser’s statements, he’s dog-whistling that every guy has forced themselves on women but few would admit to it. By saying he earned his seat at Yale through hard work when he didn’t, Kavanaugh is quietly saying he’s on the side of people with power and privilege.

What’s the larger truth in Kavanaugh’s case? I’m speculating now but I’d say there are three levels to it.

At the most basic level, it’s simply that it’s unacceptable to hold someone accountable for high school hijinks 35 years later, esp without evidence. And so when he claims there were no hijinks when everyone knows there were, he’s inviting his fellow partisans to help protect…

… him from being held to an unfair standard. They know he’s lying but they collude in the lie for a higher purpose.

Second, the larger truth may be the partisan battle, as evoked by his opening statement. Under this logic, the GOP are invited to collude in his lies bc he will be a reliable champion of the cause. The lies are in service of the larger truth that Democratic power is illegitimate.

Finally, and as suggested by our experiments, he may also be appealing to his fellow traditionalists’ anxiety about threats to their culture. What kind of real American doesn’t like beer, amirite? And what kind of loser doesn’t have too many beers once in awhile? The larger…

… truth then is that those high school hijinks were *good* and it’s wrong for these jerks to now cast aspersions on them. Of course these three logics are complementary. One, two, or three of them could be working for any one person.

No wonder Republicans have rallied to Kavanaugh’s side and, via their conspiracies, added falsehoods of their own. It also changes our rhetorical tactics.

Larger implication: Exposing lies is insufficient to reach across this kind of partisan divide. We have to look harder for the deeper implicit claims being made & why they resonate with those who seem unable to see the lies. They *can* see the lies but their *focus* is elsewhere.

Don’t Do This, Skeptics

Science is not kind to minorities. Discrimination can make them difficult to identify and count, which combined with the minority’s relative rarity makes it nearly impossible to gather accurate statistics; convenience samples are the norm. Their rarity mean few people are researching them, so the odds of minority overcoming their discrimination and surviving academia to become a researcher are very small. Conversely, the few number of researchers means one bad apple can cause quite a bit of damage, and there’s a good chance researchers buy into the myths about this minority and thus legitimize discrimination.  A lot of care needs to be taken when doing science writing on the topic.

If you want to learn how to do it properly, read Dr. Harriet Hall’s recent article on gender dysphoria in children and do the opposite of what she does. [Read more…]

Frequentists Don’t Get A Free Ticket

I’ve been digging Crash Course’s series on statistics. They managed to pull off two good episodes about Bayesian statistics, plus one about the downsides of p-values, but overall Adriene Hill has stuck close to the frequentist interpretation of statistics. It was inevitable that one of their episodes would get my goat, enough to want to break it down.

And indeed, this episode on t-tests is worth a break.

[Read more…]

Ridiculously Complex

Things have gotten quiet over here, due to SIGGRAPH. Picture a giant box of computer graphics nerds, crossed with a shit-tonne of cash, and you get the basic idea. And the papers! A lot of it is complicated and math-heavy or detailing speculative hardware, sprinkled with the slightly strange. Some of it, though, is fairly accessible.

This panel on colour, in particular, was a treat. I’ve been fascinated by colour and visual perception for years, and was even lucky enough to do two lectures on the subject. It’s a ridiculously complicated subject! For instance, purple isn’t a real colour.

The visible spectrum of light. Copyright Spigget, CC-BY-SA-3.0.

Ok ok, it’s definitely “real” in the sense that you can have the sensation of it, but there is no single wavelength of light associated with it. To make the colour, you have to combine both red-ish and blue-ish light. That might seem strange; isn’t there a purple-ish section at the back of the rainbow labeled “violet?” Since all the colours of the rainbow are “real” in the single-wavelength sense, a red-blue single wavelength must be real too.

It turns out that’s all a trick of the eye. We detect colour through one of three cone-shaped photoreceptors, dubbed “long,” “medium,” and “short.” These vary in what sort of light they’re sensitive to, and overlap a surprising amount.

Figure 2, from Bowmaker & Dartnall 1980. Cone response curves have been colourized to approximately their peak colour response.

Your brain determines the colour by weighing the relative response of the cone cells. Light with a wavelength of 650 nanometres tickles the long cone far more than the medium one, and more still than the short cone, and we’ve labeled that colour “red.” With 440nm light, it’s now the short cone that blasts a signal while the medium and long cones are more reserved, so we slap “blue” on that.

Notice that when we get to 400nm light, our long cones start becoming more active, even as the short ones are less so and the medium ones aren’t doing much? Proportionately, the share of “red” is gaining on the “blue,” and our brain interprets that as a mixture of the two colours. Hence, “violet” has that red-blue sensation even though there’s no light arriving from the red end of the spectrum.

To make things even more confusing, your eye doesn’t fire those cone signals directly back to the brain. Instead, ganglions merge the “long” and “medium” signals together, firing faster if there’s more “long” than “medium” and vice-versa. That combined signal is itself combined with the “short” signal, firing faster if there’s more “long”/”medium” than “short.” Finally, all the cone and rod cells are merged, firing more if they’re brighter than nominal. Hence where there’s no such thing as a reddish-green nor a yellow-ish blue, because both would be interpreted as an absence of colour.

I could (and have!) go on for an hour or two, and yet barely scratch the surface of how we try to standardize what goes on in our heads. Thus why it was cool to see some experts in the field give their own introduction to colour representation at SIGGRAPH. I recommend tuning in.

 

A Little Racist Butterfly

Researchers have noted that, for decades, prison sentences have been just ever-so-slightly more harsh for black people than white people.

As a whole, these findings undermine the so-called ‘‘no discrimination thesis’’ which contends that once adequate controls for other factors, especially legal factors (i.e., criminal history and severity of current offense), are controlled unwarranted racial disparity disappears. In contrast to the no discrimination thesis, the current research found that independent of other measured factors, on average African-Americans were sentenced more harshly than whites. The observed differences between whites and African Americans generally were small, suggesting that discrimination in the sentencing stage is not the primary cause of the overrepresentation of African-Americans in U.S. correctional facilities.

Mitchell, Ojmarrh. “A meta-analysis of race and sentencing research: Explaining the inconsistencies.” Journal of Quantitative Criminology 21.4 (2005): 439-466.

Not as widely noted: incarceration sorta behaves like a contagious disease. [Read more…]

Gaining Credibility

You might have wondered why I didn’t pair my frequentist analysis in this post with a Bayesian one. Two reasons: length, and quite honestly I needed some time to chew over hypotheses. The default frequentist ones are entirely inadequate, for starters:

  • null: The data follows a Gaussian distribution with a mean of zero.
  • alternative: The data follows a Gaussian distribution with a non-zero mean.

In chart form, their relative likelihoods look like this. [Read more…]