Monday, November 05, 2012

Where I blog now...

If you are one of the occasional people who randomly encounter this blog, and are wondering what the hell happened to my blog, come drop by my current home at the science pub The Finch and Pea.

Saturday, March 22, 2008

Adaptive Complexity is Moving - Update your links

I'm completely moving this blog over to Scientific Blogging. If you follow this blog via RSS, the new link is:

Adaptive Complexity.

If you read this through the DNA Network, you'll see the new feed shortly.

My Scientifc Blogging column is here.

If you're waiting for the long-delayed next installment on the science in Pynchon's Against the Day, it will go up there (soon, I hope).

Why am I moving? Scientific Blogging is a great and growing community of science bloggers, as opposed to an isolated blog on blogger.

On top of that, I get to write about science for a broader audience. The NSF's Science and Engineering Indicators consistently show that people's acceptance of well-established, mainstream scientific findings - that humans are the product of evolution, that the continents move, that the universe began with the Big Bang - all correlate strongly with education. This is a major reason why scientists should be reaching out to the broader public to get people informed and excited about science.

So come on over and visit my new blogging home, and check out some of the other great writers there. And if you're so inclined, come on over and start your own science blog.

Friday, March 07, 2008

Statisticians are verifiably insane

R is a great free statistical software package. I'm trying to use it frequently enough to get comfortable with the basic commands and data objects, so that I don't have to dig through the help manual every time I want to do some small thing. I'm trying to break my Excel habit - if you want real statistics, you don't use Excel, right?

I'm writing my paper, and need to make a simple bar graph. The basic kind, with error bars, which you see in every single issue of every science journal published today. In about 1 minute I can make one, using Excel:

I'm thinking, 'wow, that was easy in Excel, but I should really learn to do it in R.' Of course in R, 35 minutes later, I still haven't figured out how to get the chart to look the way I want. So I go to the trusty R Help mailing list, confident that someone else has shared my frustration. And there I learn that statisticians hate these charts. Do we need any more evidence that statisticians are insane?

(I understand their point, but in most cases - come on! Bar charts are great!)

Thursday, March 06, 2008

Bad Science Journalism: The Myth of the Oppressed Underdog

There is a particular narrative about science that science journalists love to write about, and Americans love to hear. I call it the 'oppressed underdog' narrative, and it would be great except for the fact that it's usually wrong.

The narrative goes like this:

1. The famous, brilliant scientist So-and-so hypothesized that X was true.

2. X, forever after, became dogma among scientists, simply by virtue of the brilliance and fame of Dr. So-and-so.

3. This dogmatic assent continues unchallenged until an intrepid, underdog scientist comes forward with a dramatic new theory, completely overturning X, in spite of sustained, hostile opposition by the dogmatic scientific establishment.

We love stories like this; in our culture we love the underdog, who sticks to his or her guns, in spite of heavy opposition. In this narrative, we have heroes, villains, and a famous, brilliant scientist proven wrong.

I'm sure you could pick out instances in science history where this story is true, but more often it is not. You wouldn't know this from the pages of our major news media though; in fact you'd probably get the impression that the underdog narrative is the way science works. And many journalists may think that too; after all, most of them read (or misread) Thomas Kuhn when they were in college, and Kuhn brought this kind of narrative to a new high. The impression this narrative leaves is that science only progresses by the efforts of brave individuals who are willing to wither the wrath of the scientific establishment.

Why is this narrative about science wrong? Let me illustrate with the nearest example I have at hand right now: a piece out of the 2007 Best American Science and Nature Writing: "The Effeminate Sheep", by Jonah Lehrer, a piece that was originally published in Seed magazine. The piece is about Joan Roughgarden, an accomplished Stanford Biologist, and the transgendered author of the book, Evolution's Rainbow, about the sexual diversity that we find in nature. As we learn in her book, gay sex is quite popular in the animal world - bonobos, fish, giraffes, whales, and big-horn sheep make humans look incredibly prude.

This raises a question: being gay has obvious evolutionary fitness consequences - without modern medicine, you have to have heterosexual sex to have offspring. So is homosexuality in nature just a freak occurrence, a case of bad genes; or is it something that is in some way adaptive and therefore under selective pressure? Does this mean that there are problems with our current understanding of sexual selection in evolutionary theory?

(Important aside: let's make it clear right now that my intention is not to knock Dr. Roughgarden or her research, or her book - I'm talking here about how science is presented to the public. In fact I feel a little school pride in Roughgarden's accomplishments - she and I both spent some of our educational careers at the University of Rochester, though obviously not at the same time.)

Getting back to our underdog narrative, take a look at how Lehrer sets up the story. After giving a very brief introduction to Darwin's ideas about sexual selection, using the classic example of peacock tails, he writes:

"Darwin's theory of sex has been biological dogma ever since he postulated why peacocks flirt. His gendered view of life has become a centerpiece of evolution, one of his great scientific legacies."

There you have the classic start of the narrative: Darwin, our brilliant scientist, came up with a theory about evolutionary sexual selection, which has been dogma among biologists ever since.

But this story isn't true: Darwin's theories about selection took some time before they were widely accepted (in fact, Darwin's claim that all living species share a common ancestry was accepted before his ideas about selection). And even then, they weren't taken as dogma; researchers have been actively studying the subject for a long time. The theory of sexual selection has undergone heavy scrutiny and extensive modification, including an effort to put it within the mathematical framework of game theory - a development which didn't take place until 100 years after Darwin proposed sexual selection. Biological dogma ever since Darwin? Hardly! (Take a look at this book on the development and status of theories of sexual selection.)

But Lehrer doesn't bother to tell his readers any of this; it would spoil the underdog narrative. It's time to introduce the underdog scientist ready to overturn it all:

"Despite this new evidence [of gay sex among animals], sexual selection theory is still stuck in the nineteenth century. The Victorian peacock remains the standard-bearer. But as far as Roughgarden is concerned, that's bad science: 'The time has come to declare that sexual theory is indeed false and stop shoehorning one exception after another into a sexual selection framework ... To do otherwise suggests that sexual selection theory is unfalsifiable, not subject to refutation.'"

Roughgarden is the underdog against the scientific establishment - an establishment stuck in the nineteenth century, so willing to protect its pet theory that it will go so far as to make it unfalsifiable.

What is this revolutionary idea that the establishment is hostile towards? Roughgarden believes that homosexual behavior is an adaptive trait, one preserved by natural selection to play an important role in group cohesion, or as Lehrer puts it, "gayness is a necessary side effect of getting along." To support this, Roughgarden marshals examples of unorthodox sexual arrangements in many different species, and explains that these arrangements actually promote evolutionary fitness in complex animal societies. She has presented this evidence in a popular book, Evolution's Rainbow, and in a review article in the journal Science (subscription required).

Her ideas are unquestionably bold. She and her co-authors set themselves an ambitious task to replace the current theory of sexual selection:

"We think that the notion of females choosing the genetically best males is mistaken. Studies repeatedly show that females exert choice to increase number, not genetic quality, of offspring and not to express an arbitrary feminine aesthetic. Instead, we suggest that animals cooperate to rear the largest number of offspring possible, because offspring are investments held in common. We therefore propose replacing sexual selection theory with an approach to explaining reproductive social behavior that has its basis in cooperative game theory."

They go on to present their new mathematical formalism for their ideas. This is how science is supposed to work, incidentally. Roughgarden wrote a popular book, but didn't expect that to be a substitute for genuine scientific papers. She knows she has to convince her scientific peers, and to do that, she wrote a technical paper, spelling out the mathematical basis for her novel ideas. The next step is for the people who understand those mathematical details to check them out, work them over, and see how persuasive they are.

But that's not how the underdog narrative goes. According to Lehrer, by this point it should have been an open-and-shut case, if it weren't for the hostile scientific establishment:

"Despite Roughgarden's long list of peer-reviewed articles in prestigious journals, most evolutionary biologists remain skeptical of her conclusions... In the absence of something conclusive, most scientists stick with Darwin and Dawkins."

In the underdog narrative, it is wrong for the establishment to remain skeptical, which in reality is exactly the opposite of how science is supposed to work. It is not like a courtroom where innocence is the presumption; in science, a novel idea is unfounded until proven otherwise. And Roughgarden's publication record, impressive as it is, is not evidence that her hypothesis is correct. Nor is the fact that her review article was published in Science evidence that her idea is true. It does mean though that she's put something together serious enough to deserve a hearing. (One more nitpicking point - "scientists are sticking with Dawkins"? Dawkins may be the authority on the subject to someone like Lehrer who has probably read only popular books on evolution. Dawkins has done some professional work in this field, but he's not the reigning authority.)

Our underdog narrative is almost complete: we have a reigning scientific authority, Darwin, whose ideas are entrenched dogma among an establishment that is skeptical of our underdog scientist, whose ideas are so obviously true that they would have been accepted if it weren't for the closed-mindedness of the defenders of orthodoxy.

And closed-minded they are: Lehrer, instead of summarizing the real critiques that Roughgarden's paper generated (those with access to Science can read them here and here), suggests that biologists are unwilling to abandon their dogma of sexual selection and view homosexuality as anything but "sexual deviants" and "statistical outliers."

That's not exactly what Roughgarden's critics are saying. The responses to her Science review included two major criticisms: that Roughgarden did not correctly characterize sexual selection as it is currently understood, and that some of her assumptions in her game theory model were wrong. Most writers did think that she had offered something interesting, although not something which completely negates the theory of sexual selection; several writers suggested that current theories could be modified to incorporate Roughgarden's ideas. Lehrer does quote one scientist who basically says just that. He quotes PZ Myers, a University of Minnesota biologist:

"I think much of what Roughgarden says is very interesting. But I think she discounts many of the modifications that have been made to sexual selection since Darwin originally proposed it. So in that sense, her Darwin is a straw man. You don't have to dismiss the modern version of sexual selection in order to explain sexual selection of homosexuality."

Our narrative would not be complete without a final look at our persevering underdog:

"Roughgarden remains defiant," Lehrer writes. And we learn the real source of the establishment's skepticism. "I think many scientists discount me because of who I am," the transgendered Roughgarden says. "The theory is becoming Ptolemaic. It clearly has the trajectory of a hypothesis in trouble."

We are left with the impression of scientists hanging on to a sinking theoretical ship, unable to move forward in their understanding because they have something personally against the underdog of the narrative.

It's amazing science makes any advances at all, with such closed-minded people in control of the field! But that's not really how things work. The real story is an example of science operating the way it is supposed operate: a researcher comes up with new and very interesting observations that seem to challenge our current understanding of an important problem. She works to put those observations under some sort of theoretical framework, and presents the results in a paper to her scientific peers. Her fellow scientists think the work is interesting, but remain unconvinced because the evidence or theoretical development is not yet sufficient to support the hypothesis. What should happen next is that our researcher should go out and collect more evidence, correct any mistakes in the analysis or make a persuasive reply to her critics, and try again. A major new idea, one which overturns an existing, well-supported theory, does not get established in one paper. There has to be follow up and debate, and if the idea holds up to scrutiny it will be accepted.

Beware the underdog narrative in science journalism. This narrative severely misrepresents how science really works. It's designed to elicit our sympathy for a not-yet-established theory, maybe one that is socially attractive, and to arouse our indignation against the staid community of eggheaded scientists. This underdog narrative plays on our emotions, it makes for a good read, and helps us feel good about ourselves when we stand up for our convictions. What gets lost is the scientific method, the idea that novel proposals need to be thoroughly vetted and tested, no matter how intuitively attractive they are. That vetting process is done by a dynamic community of smart, educated, competitive people, who care passionately about science. It's a community where everyone wants to come up with the next big theory that overturns long-held beliefs. But that's hard to do, especially in fields where all the low-hanging fruit has been picked over by really talented people for decades or centuries. If a new theory is being presented in the media as the centerpiece of an underdog narrative, you can bet the farm that this theory is not yet substantiated by the evidence.

I said I wasn't writing this to knock Dr. Roughgarden, but I'm going to renege on that promise just a little: based on how she's quoted in the piece, she does seem to be feeding the narrative. She's not giving her colleagues enough credit for giving her ideas a hearing. What non-heterosexual behavior among vertebrates implies about evolution is a fascinating question, one plenty of biologists would be happy to know more about. But it's a question that's only going to be settled by evidence.

This is not a real scientific conference

I'm all for open debate over the genuine science behind global warming - researchers who have research results that don't jive with the IPCC's consensus report should be free to present their stuff at real scientific conferences and not be ostracized just because their results are different.

But, as Real Climate points out, the Heartland Institute International Conference on Climate Change didn't even try to disguise the fact that it was a PR event, and not a scientific conference. The conference organizers invited speakers with this letter (PDF file), which clearly states that:

"The purpose of the conference is to generate international media attention to the fact that many scientists believe forecasts of rapid warming and catastrophic events are not supported by sound science, and that expensive campaigns to reduce greenhouse gas emissions are not necessary or cost-effective. " (emphasis mine)

There are big conferences which are basically promotional events (not necessarily a bad thing!) like the Bio-IT World Conference, but these are not to be confused with real research conferences like the Keystone Symposia or the Cold Spring Harbor Meetings, where the latest, and often unpublished research is presented primarily to other scientists, and not to the media or interested laypeople. And real scientific conferences aren't usually sponsored by organizations that have decided what the answer to a scientific question has to be beforehand (the Heartland Institute is specifically devoted to "discover, develop, and promote free-market solutions to social and economic problems" - I'm all for finding free-market solutions when they're feasible, but we ought to [gasp!] consider cases where the free market solution isn't the most efficient or beneficial).

John Tierny is having trouble understanding this distinction as well, implying that Heartland's advocacy simply counteracts government climate change-advocacy perpetuated by bureaucrats doling out funding at the NSF and NASA (never mind those pesky study sections!).

If you've never been to a real science conference, I suppose it can be hard to understand that, for all of the human failings that exist in science like anywhere else, there are strong institutional cultural mechanisms in place that make science work - like the fact that speakers at a meeting are chosen by a scientific organizing committee, and not by financial sponsors. Advocacy can play an important role in public policy, but research conferences aren't about advocacy.

(Real Climate has more.)