Why Evolution is True is a blog written by Jerry Coyne, centered on evolution and biology but also dealing with diverse topics like politics, culture, and cats.
Reader Pliny the in Between is on a roll lately at the Evolving Perspectives site; here’s the artist’s satirical take on the Templeton World Charity Foundation’s big initiative to study the metaphysics of the Trinity. The cartoon’s title is “Much physics was sacrificed in the making of this panel.” (Click to enlarge.)
In the last three days I’ve visited two renowned Southern restaurants.
The first is the Busy Bee Cafe, featuring southern soul food, particularly its famous fried chicken. Founded in 1947, the cafe was frequented by many civil rights leaders, including Martin Luther King, Jr., back in the days when Atlanta was a locus for anti-segregation activism. The cafe remains the same since then, and has been visited by many other famous and famished customers, whose pictures line the walls.
You can get an idea of food from the to-go menu on the site, though it lacks certain items (e.g., red velvet cake). Here’s the modest exterior:
The interior. It’s a happy and friendly place. Who wouldn’t be happy when they’re about to tuck into a plate of fried chicken?
My dinner: fried chicken, fried green tomatoes, a fantastic carrot souffle (I have no idea how they make it), accompanied by corn muffins and washed down with “the table wine of the South,” sweet iced tea. It was a fantastic meal. The carrots and tomatoes were succulent; the chicken juicy and fried to perfection. The Busy Bee serves 2000 pounds—a full ton—of fried chicken each week.
I was too full to eat dessert, so ordered the banana pudding with vanilla wafers (a classic southern dish) to go. The “small” size weighed over a pound, I’d guess!
Here’s a long excerpt from several television shows about the cafe, including the first by chef Emeril Lagasse:
Yesterday I lunched (not the right word for Southern food!) at the Barbecue Kitchen in College Park, about two blocks from my hotel. Note the chimney for venting the wood smoke. Don’t ever go to a BBQ joint that doesn’t have a chimney, because they’re not smoking over wood.
Lunch: chopped pork with BBQ sauce, creamed corn, sweet potato souffle (note the marshmallow sauce), a biscuit, sweet tea, and peach cobbler for dessert. All of this was only $7.65!
More photos of eats in the offing. . .
As I recall, for a long time the University of Texas has fought for the right to have gun-free campuses, but that has now failed. According to the Houston Chronicle, Governor Greg Abbott signed a bill permitting students, faculty, and staff to practice “concealed carry” (carrying handguns that are hidden) on all state campuses. (According to the excerpt below, this also holds for private universities). It will take effect on August 1 of next year.
Here’s the bill, and the relevant section is below:
| (b) A license holder may carry a concealed handgun on or | ||
| about the license holder’s person while the license holder is on the | ||
| campus of an institution of higher education or private or | ||
| independent institution of higher education in this state. | ||
| (c) Except as provided by Subsection (d), (d-1), or (e), an | ||
| institution of higher education or private or independent | ||
| institution of higher education in this state may not adopt any | ||
| rule, regulation, or other provision prohibiting license holders | ||
| from carrying handguns on the campus of the institution. | ||
Ah, the gunfights we can look forward to at Texas schools!
The really sick thing is the list of items the students are NOT allowed to have:
Student handbooks at Texas public schools show an interesting juxtaposition between the items that will stay banned from dorm rooms while guns will be legal. The majority of the prohibited objects are considered fire hazards.
Most kitchen appliances like crock pots, toasters or even blenders are expressly disallowed at some Texas colleges. The government will trust students with guns next year, but they won’t be able to handle candles or even incense (an icon of liberal college students?).
Texas A&M even goes as far as to single out toy Nerf guns for a ban, and the University of Houston outlaws water guns. The rule makes sense of course. Seeing someone running around on campus firing a toy a gun could definitely cause some issues.
Other banned items (there’s a gallery at the site): incense, homemade furniture, toasters, George Foreman grills, halogen lights, black lights, power tools, wireless routers, pets, and (at Texas A&M) slingshots. Many of these are seen as fire hazards, while guns are only life hazards. Slingshots not ok, concealed guns ok—seriously?
Meanwhile, Slate reports a humorous pushback by students at the University of Texas at Austin.
When the 2016 fall semester begins, University of Texas alumna Jessica Jin wants to see legions of dildos parading across the quad. Jin’s response, via Facebook: “You’re carrying a gun to class? Yeah well I’m carrying a HUGE DILDO.” In an invitation to a “Campus (DILDO) Carry” protest, Jin encourages UT Austin students to strap “gigantic swinging” fake penises to their backpacks when campus carry takes effect—an act that could earn the carrier a $500 fine for a misdemeanor display of obscene material.
. . . More than 4,000 people have responded in the affirmative to the Facebook event, and Jin is confident that she can deliver the necessary goods. “I know that quality dildos, especially super large ones, can be a little pricey,” she writes. “If we can gather enough willing participants, I will personally take it upon myself to put in the time and legwork to find a dildo supplier sponsorship.”
What kind of nation is this where you can get fined for carrying a dildo (their sale was illegal in Texas until 2008), but it’s okay to carry a hidden Glock?
h/t: Robin
Beating deaths occur regularly, and for many reasons, but here’s one that can be imputed only to an extremist form of faith. As reported by both Reuters and CBS News/AP, two parents, Bruce and Deborah Leonard), have been charged with first-degree manslaughter in the beating death of their son Lucas (19), and assault on his brother Christopher (17) at the Word of Life Church in Chadwicks, New York. Their crimes? Being “sinful”: the beatings were administered to try to get the teenagers to admit their sin. Apparently the family drove the boys to a hospital and claimed that they had suffered gunshot wounds, but that’s just dumb, because there was no indication of such wounds, which of course don’t look like beating at all.
Four other church members, including the victims’ sister, were also arrested and jailed. All defendants are shown below.
The beatings were horrible:
Brutal beatings that left one teenager dead and his brother seriously injured Monday at a New York church were part of what members considered a “counseling session,” according to police.
New Hartford Police Chief Michael Inserra said Wednesday that both Lucas and Christopher Leonard were subjected to hours of physical punishment at the Word of Life Church “in hopes that each would confess to prior sins and ask for forgiveness.”
. . . O’Neill said Lucas Leonard died Monday after he was beaten the night before at the Word of Life Church in New Hartford, which is 80 miles northwest of Albany.
He was said to have “died a violent death over several hours.”
Police did not reveal if any weapons were used, but did say “feet and fists were involved.”
The six church members were arraigned Tuesday and sent to Oneida County Jail. At the arraignment, it was revealed that both teens suffered injuries to their abdomens, genitals, backs and thighs. Bail for the Leonards was set at $100,000 each and for the four other defendants at $50,000 each. All pleaded not guilty.
The defendants:
There are no atheists, at least no sane ones, who claim that all the ills of the world would be cured if we only got rid of religion. That’s nonsense: to paraphrase Steve Weinberg, in this world good people (many of them believers) do good things, and bad people do bad things. But there are some cases where religious beliefs motivate people to do bad things that they wouldn’t have done otherwise. That is undeniable. This is such a case, and it cost a young man his life. The defendants are the American equivalent of those ISIS members who torture apostates and Yazidi for their own “sins”: not accepting the proper faith.
h/t: David
Reader Stuart Coyle from Oz sent a variety of photos, including one of a domesticated mammal. I’m glad that Australian readers chip in so much. But I’ll request more photos from readers as the tank is getting low.
My wife, two cats and I recently moved to Samford Valley, a town about half an hour out of Brisbane. I have taken a few photographs of the local wildlife that you and your
blogwebstite [JAC: my alteration] audience may find interesting. I’m not much of a photographer but I think these are mostly passable.The photos include a python over 2m long, that our cats stupidly decided to go after. I was bitten by one of the cats because I tried to get him away from a snake that could swallow him whole.
The work week is drawing to an end, and posting will be light for the next few days as I’ll be attending the Atheist Alliance of American conference in Atlanta. In the meantime, perhaps my co-writers will fill in. If not, talk among yourselves. Meanwhile in Dobrzyn, the long shadow of a morning cat presages the weak slant of winter light.
Hili: Do you see the long shadow of a morning cat?
A: I do. This is the longest shadow in your career.
Hili: Czy widzisz długi cień porannego kota?
Ja: Widzę, to najdłuższy cień w twojej karierze.
At the beginning of September I wrote about a paper in Science produced by a large group called “The Open Science Collaboration.” That paper reported the repeatability of 100 papers whose results were published in three prestigious psychology journals. My brief summary of the conclusions is below, though my original post gave a lot more data:
Now this doesn’t mean that the original studies were wrong, for of course the replications could have produced the wrong answer. But given that the replication studies generally used larger sample sizes than did the original work, it suggests that there are endemic problems with the way science is adjudicated and published. My suspicion is that the main cause is a bias toward publishing positive rather than negative results, combined with “p-hacking”: looking for those statistical analyses that give you probability values lower than the cutoff needed to reject the null hypothesis, or a tendency to collect data only up to the point where your “p” values become significant, and then stopping and writing a paper.
The Science paper set off a flurry of self-scrutiny and self-recrimination as scientists begin to worry that, even if they’re not psychologists, the problem may cut across disciplines—and probably does. I’ve long thought that studies in ecology and evolutionary biology, particularly field work or experimental work that doesn’t involve DNA sequencing (sequencing data are easily replicated), may also be largely unrepeatable, both for the reasons given above and because field and lab results may be particularly sensitive to experimental or environmental conditions. Next month I am in fact going to a meeting of biologists and editors to address the issue of replication in my field.
In the meantime, lots of articles have come out highlighting the problem. Before people conclude that the replication problem is a big problem for all of science, I’d suggest that some fields, particularly molecular biology, may be largely immune, because a). they’re easily replicated and b). are very often the building blocks for future work, so researchers not only have the incentive to get their results right, but their studies will automatically be replicated as a first step in other people’s followup work. Deciphering the DNA code, the subject of Matthew Cobb’s new book, for instance, automatically involved other people calibrating their system using the codons worked out by earlier researchers. Science journalists should realize this before sounding a general alarm.
That alarm, however, was sounded by writer Regina Nuzzo in an article in Nature about replication called “How scientists fool themselves, and how they can stop.”
The summary is shown in her diagram of the problem and possible solutions:
Here’s my take on Nuzzo’s analysis, which is by and large pretty good, and on her solutions, which are somewhat problematic but still worth considering.
THE PROBLEMS
The first line, “cognitive fallacies,” is pretty self-explanatory. It’s simply doing experiments that are contaminated by confirmation bias: neglecting alternative hypotheses and data inimical to your favored hypothesis. That is the epistemic method of most religions.
The “Texas sharpshooter” problem involves, among other things, p-hacking, and also doing a gazillion different tests on a diversity of data, some of which will be significant by chance alone, and then seizing on those as your publishable results.
“Asymmetric attention” is self-explanatory. Nuzzo gives two examples:
A 2004 study observed the discussions of researchers from 3 leading molecular-biology laboratories as they worked through 165 different lab experiments. In 88% of cases in which results did not align with expectations, the scientists blamed the inconsistencies on how the experiments were conducted, rather than on their own theories. Consistent results, by contrast, were given little to no scrutiny.
In 2011, an analysis of over 250 psychology papers found9 that more than 1 in 10 of the p-values was incorrect — and that when the errors were big enough to change the statistical significance of the result, more than 90% of the mistakes were in favour of the researchers’ expectations, making a non-significant finding significant.
“Just-so storytelling”, in which you make up a story post facto to explain your results, seems to me less of a problem. If your “story” is simply something you say in the discussion to rationalize a result you didn’t expect, well, others readers (and presumably the reviewers of the paper) should catch that, and realize that it’s just a rationalization. Things become more serious if you pretend that that result was your initial hypothesis, and then confirm it with the data, which is an inversion of your scientific history and basically dishonest. But while I see a lot of the former tactic in evolution, it’s not a big problem, for we all know when somebody’s grasping at straws or rationalizing. The former problem I haven’t seen—but we wouldn’t see it anyway unless the “hypothesis” that is tested is not a priori obvious.
THE SOLUTIONS
Nuzzo offers four solutions, one of which is already in play and another that seems unrealistic. Two are feasible.
“Devil’s advocacy,” considering and testing alternative hypotheses, is part of all science, and should be ingrained in every researcher. “How might I have gone wrong?” is a question all good scientists ask themselves, and then we test to see if we’ve erred. Now some people don’t do that, but they’re often caught by reviewers of their papers or grants, who ferret out the hypotheses that are neglected. That’s why every paper should have at least two good reviewers familiar with the field, and grants should have at least four people who scrutinize proposed research. This issue doesn’t seem to be a big problem unless journals and funding agencies do a sloppy job of reviewing papers and proposals. At least in the US, the two major granting agencies (NSF and NIH) are very careful at vetting proposals using in part a “devil’s advocacy” approach.
“Team of rivals”, getting your scientific opponents to collaborate with you in hopes that opposing views will help bring out the truth, is in principle a good idea but will rarely work in practice, at least in my field. For one thing, there would be authorship fights: who gets the credit? Also, who wants to drop their research to work on somebody else’s problems? Anyway, Nuzzo gives one example of such a collaboration in psychology, which didn’t appear to work so well.
In “blind data analysis”, you shift your real data around or even add made-up data, and then do all the analysis on several “blind” data sets—which aren’t really blind, as most researchers know their data and the procedure also involves removing outliers, which you often know as well. Then, when you’re satisfied that you did the analysis as you wanted, lift the blind. (This is sort of like John Rawls’s “veil of ignorance,” where you make up moral rules for society without knowing which position you’ll eventually occupy in that society.) This method will work in some situations but not others, for, as I said, you’re often familiar with your real data. At any rate, it did work in one study:
[Astrophysicist Saul] Perlmutter used this method for his team’s work on the Supernova Cosmology Project in the mid-2000s. He knew that the potential for the researchers to fool themselves was huge. They were using new techniques to replicate estimates of two crucial quantities in cosmology — the relative abundances of matter and of dark energy — which together reveal whether the Universe will expand forever or eventually collapse into a Big Crunch. So their data were shifted by an amount known only to the computer, leaving them with no idea what their findings implied until everyone agreed on the analyses and the blind could be safely lifted. After the big reveal, not only were the researchers pleased to confirm earlier findings of an expanding Universe, Perlmutter says, but they could be more confident in their conclusions. “It’s a lot more work in some sense, but I think it leaves you feeling much safer as you do your analysis,” he says. He calls blind data analysis “intellectual hygiene, like washing your hands”.
The final method, transparency, has promise but also problems. It comes in two forms:
[Form one] Another solution that has been gaining traction is open science. Under this philosophy, researchers share their methods, data, computer code and results in central repositories, such as the Center for Open Science’s Open Science Framework, where they can choose to make various parts of the project subject to outside scrutiny. Normally, explains Nosek, “I have enormous flexibility in how I analyse my data and what I choose to report. This creates a conflict of interest. The only way to avoid this is for me to tie my hands in advance. Precommitment to my analysis and reporting plan mitigates the influence of these cognitive biases.”
I’m fully in favor of this: all data used in a paper, and methods of analysis, as well as experimental analysis, should be available to researchers INSTANTLY after a paper is published. This has long been the custom in the Drosophila community: it’s unthinkable not to share data, or even laboriously constructed genetic stocks, with colleagues and rivals. This is being used already by many scientists, and is mandated by some journals. Other journals, however, either don’t require such data storage or impose a year’s moratorium on it so you can milk your data for more papers before others get their hands on it. I don’t favor the moratorium; it’s just too bad if other people use your published data to their own ends.
Form one could, however, be construed as sharing your aims, methods, and analyses BEFORE you get your data, and that’s just not on. It makes you vulnerable to intellectual theft, and no researcher wants to do that. This is why grant proposals are strictly confidential, and why no grant reviewer is allowed to lift ideas from grants they’ve reviewed. It’s also why reviewing papers is confidential.
[Form two] An even more radical extension of this idea is the introduction of registered reports: publications in which scientists present their research plans for peer review before they even do the experiment. If the plan is approved, the researchers get an ‘in-principle’ guarantee of publication, no matter how strong or weak the results turn out to be. This should reduce the unconscious temptation to warp the data analysis, says Pashler. At the same time, he adds, it should keep peer reviewers from discounting a study’s results or complaining after results are known. “People are evaluating methods without knowing whether they’re going to find the results congenial or not,” he says. “It should create a much higher level of honesty among referees.” More than 20 journals are offering or plan to offer some format of registered reports.
This method, in which the methods and aims are confidential, seems to be gaining popularity. It’s like giving a grant proposal to a journal before you do the experiment, and if they approve of the analysis and methods, they’ll accept the paper no matter how the results turn out. That’s fine, but there are some caveats. First, during a study new experiments often crop up that you haven’t planned, and often those are the most exciting ones. (This has often happened to me.) How do you deal with those? I don’t see how.
Second, this deals only with the analysis of data; it doesn’t deal with the importance of the results. But no journal will publish any paper that’s soundly executed regardless of the results: every scientist knows that there are some highly visible “top-tier” journals where publication can make your career (e.g., Science and Nature), and journals like that aren’t going to publish on just any subject, as they specialize in “important” results. (Journals like PLOS One, which consider only methods and not importance, are venues for all sorts of work, important and not-so-important). For regular journals this means that there has to be pre-vetting of not just the methods of research and how data are analyzed, but whether or not the problem is interesting. This doesn’t seem to be considered in any such proposals for “transparency,” but I haven’t read them all.
Scientists will be chewing over this problem in detail over the next year or so, and it is indeed a problem, though more so for some areas than for others. The solutions, however, vary in quality and depend on the field. The “transparency” method seems to be gaining popularity, but to me it seems the least practical of all solutions.
h/t: Ben Goren
I think I’m suffering from Congenital Comic Blindness, as I seem unable to fully grasp the humor in religion-related cartoons (or even understand them), while readers don’t appear to have that problem.
Several readers sent me today’s Non Sequitur strip by Wiley Miller, and I’m not quite sure why it’s funny. I submit it for the readers’ consideration, knowing that most of you will get it and like it, but maybe someone could explain why it’s funny.
As far as I can see, it’s funny because the surfer, as an agnostic, doesn’t trust that the waters will remain parted for Moses et al., and so is surfing on them. But I don’t understand why that’s so riotously funny. See for yourself, and feel free to comicsplain.
