Well that was an interesting experience

Small Hyde enters the world.

All well and healthy. More later.

How Americans distrust science

This week's Times Week in Review section sports a lead article by an ex-physicist, James Glanz, whose thesis is that the failure of the Large Hadron Collider in Geneva is reminiscent of the giant abandoned pyramids of the Mayans: a symbol of our hubris in trying to understand too much.

Glanz doesn't use the word "hubris," with all the Greek mythological implications it entails, but substitutes the "your business model was bad" term "overreach." Nonetheless, his point is clear: we were trying to probe too far into the limits of the cosmos, and so naturally we were doomed for failure.

Well excuuuuuuse us for trying.

Glanz doesn't seem to have a problem with the experimental question itself: whether the Higgs boson (a) exists, and (b) underlies the concept of mass. (I'm no physicist, but I think finding out whether a particle is responsible for one of the most salient characteristics of our physical world is yeah, pretty important.) Instead his beef is that the electrical and magnetic problems with the LHC are a sign of our unwonted temerity in probing such questions.

I call serious bullshit. As the saying goes: If I knew what I were doing, it wouldn't be called an experiment. Yes, it's a vicious disappointment that there's no data out of the LHC yet, leaving hundreds of physics grad students with theory-only theses.

But if an experiment beset with unexpected problems is a sign that we are overreaching, then I guess I'm pretty much all washed up as a scientist, because most of my experiments are overrun with unexpected problems. I call myself a scientist because these problems don't make me don't throw up my hands and head for the bar (ok, sometimes) but instead cause me to find solutions, to troubleshoot, to analyze whether the problems are inherent to the experimental question or simply to the method in use.

Glanz's article reflects (embarrassingly, for an ex-scientist) the basic fear with which many Americans view science. If it doesn't work out right the first time, it's because we wanted too much! We tried to learn Something We Shouldn't Have. We angered the gods.

Just last night Dr Hyde and I were talking about our collective failure to teach the right lessons about science in school--that science is not so much a collection of facts as a form of rigorous critical thinking. I think Glanz's article is an extension of that problem: he equates a failure to obtain facts on schedule with a failure of science.

If more Americans understood that this idea was perniciously wrong, we might not be in the era of creationism - global warming denalism - disrespect for reality that we are. Thanks for helping, "Dr" Glanz.

Texting while driving

Normally when I am doing experiments, they only occupy about three-fourths of my brain. The remaining brain is answering email, thinking about other experiments, thinking about dinner, humming a little tune, whatever.

So you would think that I could perfectly well carry on a conversation and answer a few questions about what I'm doing while I do experiments. But no. With people watching, I forget which way caps twist to tighten. I forget to press important buttons. I drop things, like my experimental prep.

Today two people were watching me prep and (try to) carry out experiments, and it was borderline disastrous. (Some of this had to do with bad luck for the day, but then again why is it always on the days that my prep is a total failure that people are watching? They never seem to come by when I am my usual genius whiz self.)

And it got me to thinking, why is it so damn hard to do experiments when people watch? I have four explanations:

1) Observers make you self-conscious, leading to obvious problems with routine tasks of coordination that proceed most effectively without conscious control.

2) The cortical load imposed by answering even very easy questions about your experiments is far higher than one realizes.

3) A more hypothetical possibility: that there is a non-trivial burden placed on your brain simply by being in a social situation--i.e. with observers--so that even if those observers are silent, a surprising portion of your brain is being used to monitor them.

4) Though I routinely can spare a quarter of my brain for email etc., there are brief periods (as short as 5-10 s) during which I use all of it to focus on the experiment at hand, and any social burden is problematic.

Of course these are not mutually exclusive, though I would love to know whether anyone has ever studied the possible existence of effect #3.

The whole situation reminds me of the recent brouhaha over texting while driving (ed: who the hell was surprised that texting is bad for driving?!) and the attendant reminders that the "hands-free" policy espoused by many states is essentially useless, because it turns out that the mental load of carrying your end of the conversation, rather than the physical act of holding the phone, is the problem.

I wish the Senate would pass a law preventing me from having to train up other people on my experiments when I am very pressured for time by the impending bomb in my belly, and also all the rest of the time too because let's face it, that shit's annoying. If I get a petition going, you people will be the first to know.

Well there goes my Friday night

Yes, yes, scientists do amazing and interesting research that pushes the limits of our knowledge for the betterment of humanity.

Also, stuff on zipper-trapped penises and flatufonia. Thanks (I think), anonymous graduate students running NCBI ROFL.

Apparently this is Whine Week

When I am typing a document in Microsoft Word, it turns out that I never want to insert the phrase SPECIAL DELIVERY.

SPECIFIC AIM, however...

Two points

1) If you don't like a book, write as caustic a book review as you choose.

2) I think it's a pity when ScienceBlogs, which has won prominence in the NY Times' Science section, and represents a selection of scientists who have chosen to do a blog-form of outreach to the public, turns into a childish, cryptic den of finger-pointing, particularly on the very topic of....outreach to the public.

If I were the scientifically engaged public (which is how I'd have described myself circa 1993, before I started doing lab work), and clicked over to SciBlogs after reading my usual Tuesday Science Times, I would sure as hell click right back to more pleasant, not to mention more comprehensible, waters. And I say that as someone whose blood pressure is roughly doubled by reading Tierney's sad excuse for a "science" column.

Post about science. Post about the culture of science. Post about your life as a scientist and how it functions, or doesn't. Post about your life to give non-science readers a window into how three-dimensional scientists are.

But for chrissakes, don't post about your backroom spats with other science bloggers (whether or not they're still on SB!) and waste one of the better opportunities we have of engaging laypeople.

Drs. I and Z, I think you're capable of much more interesting work. I fully agree that you should call out misogyny when you see it. But the full-scale SB war, with blog posts being lobbed like hellfire missiles, coated in links and comments-on-comments, is boring and ugly, and worst of all, utterly uninteresting to the public. It's not like we have so very many chances to bring them on board.

Things I learned in grad school 4: Thinking vs doing

One of the most difficult aspects of science for me to deal with when I entered graduate school was empiricism. I came to a love of science from the intellectual, not the practical, end. I liked the questions, ideas, and logic. I was under the impression that, because scientific phenomena underlie our whole existence, any sort of practical problem could be solved by application of reason and logic.

I probably spent the first three years of grad school unlearning this idea.

For a long time, the idea that logic would triumph held me captive to massive frustration. It turns out to be, appropriately enough, right in principle but wrong in application.

I spent years kicking tables out of anger at problems like electrical noise in my setup. Now, there are some straightforward rules governing electricity. You learn a number of them, probably all the relevant ones, in college physics if not earlier. You can apply them all in a logical, orderly manner. You can still have inexplicably bad electrical noise.

You can learn some special rules of thumb for the particular configuration you have. Still, electrical noise.

What drove me up a tree was that invariably I'd get advice from a senior student along the lines of, "Well, just try a bunch of things like this and see if it helps." Why? There is no earthly reason that I should have to try random things. I should be able to solve this problem using a multimeter and my brain.

As I say, I wasted a lot of time on frustration before learning that in fact, some problems are best solved empirically, by trying a bunch of stuff and seeing what works. You can reason your way back to an explanation (maybe!), but some problems are more or less intractable without just trying things out, even things that you think shouldn't matter.

I know plenty of good scientists who entered grad school from the opposite pole, that of being do-ers who love nothing more than to plunge in and try stuff, and think about the results later. Those people usually do really well at carrying out experiments, and their challenge is often to make sure those experiments are thoughtfully designed.

If you are headed for or starting graduate school, I'd encourage you to contemplate which of these--thinking or doing--you consider your strength. I certainly knew heading in to grad school that I loved scientific concepts, but struggled with benchwork. If you know which half you find easier, recognize how much you need practice with the other half, and find someone to help you with it. Perhaps you have a fellow labmate who has the skills you lack, or a friend in a neighbor lab. This person's approach to problems may seem totally foreign to you, but that's exactly why you need them.

I was thinking of this today as a colleague and I were troubleshooting a completely bizarre problem with some fancy equipment we both use in his lab. We spent at least half an hour fighting two seriously disconcerting glitches, one of which seemed hardware-related and perhaps unfixable (and which would cast doubt on several published papers that use this hardware), and the other of which seemed software-related but equally unfixable.

Five years ago, I would have given up, or thrown a fit, or called up the company to yell at them, probably uselessly. Instead I suggested we try changing a software parameter that had nothing to do with anything we were having trouble with. Why? Not because I thought it was relevant, but because I have learned respect for empirical outcomes.

Presto. Both glitches cleared up. (Not with the first irrelevant parameter we changed, mind you! More like the 4th.) The other postdoc and I gaped at each other. We tested some more. Solved.

We tried to come up with any possible reason this parameter might have caused these problems. Nothing. But you know what? I don't really care (though I might still go back with some BNCs and an oscilloscope to see if I can figure it out). It's clear that the software has a bizarre bug that causes this inexplicable behavior. And now we know how to steer clear.

This is not to claim that I never get frustrated with inexplicable problems any longer. Ha ha, no. But that graduate school taught me to take seriously the attitude, "Just try it and see what happens."

More simple requests

Pubmed, you have some good qualities. Also, some not so good ones. To wit:

--Why must the cursor automatically jump to the search field when a page loads? Like all right-thinking, i.e. keyboard-shortcut-using, people, I am a fan of using the space bar to page down. But on each new results page, the space bar just....adds space, unless I click somewhere else on the page first.

--When I type the first few letters of a recently-performed search into the search field, you bring up a helpful drop-down of my recent searches. I key down to the desired one and press Return to put it in the search box.
(a) You immediately re-run that search. Sometimes I wanted to add something else to that search term.
(b) Not only do you immediately re-run that search, for reasons beyond me you sometimes fail to run the full search term, but instead use only the fragment that I had typed in. For example, when I type in "Ph" and then key down to select "Physioprof" from the drop-down, sometimes you produce Physioprof's papers. Sometimes you produce every single goddamn paper with "ph" in it. All 8 million of them. [This is just an example, people. In reality, "physioprof" produces no search results in PubMed. Don't waste your time.]

--Select some unique author ID method and implement it. I don't care which one. But if I have to try searching for a paper by "Xu X" ever again, I will maim you.

I'm lookin' at you, VWR

Welcome to ConglomerateScienceSupplyCo!

Please indicate your country to continue.

But you know my IP address.

Please select from the following menu:

[List of all countries starting with A, because we don't discriminate] or
[List of all countries starting with United States and United Kingdom, because let's face it]

The first version annoys me because I have to scroll all the way to the bottom. The second version annoys me because it reminds me that I am a spoiled American. Also, why the hell do I have to do this at all? You know my damn IP address. You could tell me what university I'm coming from, if you so chose.

Thanks for your selection! Would you like us to remember this selection for future visits?

Go right ahead, although you've never succeeded in remembering it before and I am doubtful that you even know how to place a cookie on my computer, given that you don't seem to have figured out that you know my goddamn IP address.

Referencing dross

Everyone's had this problem: you need to cite a paper in your field. The paper is totally ghastly. You know it's crap, and probably (?) your readers do too.

But you can't not cite it, because it's one of the few precedents relevant to your work--perhaps the only precedent for a certain facet of your work. Perhaps it's well-cited within your field and your failure to cite it would be seen as a glaring omission; perhaps your advisor or a reviewer has told you to cite it. Whatever the cause, you have to reference this bit of rubbish.

You can
1) Cite it. "Our data provide one of the first demonstrations of bleg-de-blug (Loser et al.)"
2) Bury it. "Previous research has not focused on the area of bleg-de-blug (Loser et al.)"
3) Trash it. "With the exception of a dubious report (Loser et al.), our data provide the first demonstration of bleg-de-blug."
4) Trash it cryptically. "Bleg-de-blug is without precedent in this field (Loser et al.)"

Sigh.

And thus does everyone's h-index creep up.

---------------

For the other junkies out there, my odds run thusly:

2:1 Looming indictment/impeachment for ethical problems related to Troopergate.
3:1 Levi's inked a book deal.
6:1 Being blackmailed by Levi or someone else, and needs to raise cash.
10:1 Sex tapes.
18:1 Another pregnancy, hers, Willow's, or Todd's mistress's.
100:1 Actually thinks this is her best route to the presidency.

Reporting the story

Presumably most of you have read Gina Kolata's take, in the Times, on why cancer research hasn't cured cancer yet. It's the super-conservative grant review process that rewards only incremental research and variations on well-established themes, it turns out.

Well, yes, that's something that all of us, cancer researchers or not, complain about fairly regularly: the requirement for in-depth preliminary data that signals "The first aim is all but complete" before funding shows up.

But I think she buried the real story. Anyone else catch the graph?


Let's take a couple of years here for illustrative purposes.

In 1996, we spent about $3 bn on NCI total, of which $0.6 bn went to "Other expenditures--non research." That's 20%.

In 2005, we spent about $5 bn on NCI total, of which $1.5 bn went to "Other expenditures--non research." That's 30%. Aka, a 50% increase in proportion of non-research expenditures.

Now, there are plenty of other remarks to be made about the article, including "Biology is really complex and simplistic formulations like 'The War on Cancer' are downright silly;" and "You see how funding flatlined and then actually sank? That might have something to do with reviewers becoming conservative." Neither of which was ever addressed in Kolata's article.

But looking at that graph, all I can ask is: where is this non-research money going? Is this due to an increase in overhead/indirect costs? And if so, why? Did it really get that much more expensive to administer and run science buildings starting around the year 2000?

For 2005 alone, if these non-research costs could have been kept at their original 20%, there would be an extra half-billion dollars freed up for cancer research funding. At, say, $250,000 per project/year, that's an extra 2000 projects that didn't happen. In 2005 alone.

This article felt like one of those talks where the speaker doesn't address some really fundamental issue up front, and I can't even process the rest of the talk because I'm so focused on this problem raised by the graph on the third slide.

Leave the epic similes to Virgil, sir

A PI to a classroom of students and postdocs, regarding a scientific issue of signal propagation:

“Here’s an analogy. It’s like when you’re on the same floor as the secretaries, and there’s only one thermostat, and the secretaries, who are always young women with very little body fat and they’re wearing kind of skimpy clothing, are always turning the thermostat way up, and it affects you even though you’re at the other end of the hall, but your feedback doesn’t change the thermostat.”

Cruel to be kind

Good old Blog Fodder the Tech, who astounded me with his many iniquities before shaping up as a reasonable lab member, has announced his plans to attend graduate school.

BF had previously mentioned grad school as a possible future, but had acknowledged that he was just as interested in business school (poll: which requires a better grasp of basic arithmetic?). Now, however, he is actually studying for the general and biology GREs (and taking a Princeton Review course). Last week he came into lab on Saturday to do a couple of small lab tasks and grind away at a practice exam.

The problem is, of course, that BF lacks the fundamental curiosity, let alone the determination, to make it through grad school. He asks questions....about the tasks he is given. Perhaps once a month, he asks a question about the research presented in lab meeting, but these questions are not typically experimental in nature--they are more likely to run along the lines of, "Does this relate to phenomenon X that I read about in Science/the Times/the Economist?" Sometimes it's an interesting question, but it's pretty much never what I would expect from a first-year grad student--i.e., a question about experimental details, or whether one could follow up with Experiment Z, or even just what the heck the significance of an observation is.

(And no, this is not because he fears me; he doesn't ask these questions of anyone.)

When he told me recently that he was studying up for the GREs in preparation for grad school applications this fall, I asked him more about his research interests. In the past he has mentioned labs that do fairly similar work to our lab, but this time he mentioned some labs that are further afield. It's fine for your subject affinity to shift over time, but I think in this instance he's probably just hoping that the work of Other Labs is more interesting to him than the work of This Lab.

In any case, I tried, gently, to dissuade him from Ph.D. programs. I pointed out that if he's still thinking about going into business in some capacity (the business end of a pharma company, for example), then an M.S. would likely give him the training he needs for much less hassle.

"I just figure that if I'm going to grad school, I might as well go the whole hog," he replied. "Besides, I'd be happy to spend six years in [City With His Dream Grad Program]."

Blink. Blink.

Despite BF's many failings as a scientist, I have developed a certain fondness for him, the way one does with household items. So I should probably try again to steer him in the direction of the M.S., or even an entire life re-evaluation. (Note to readers, BF is loaded, so the cost of doing an M.S. should not be a hurdle.)

On the other hand, it's not as though he's coming to me asking for advice or help, and often it is easy to irritate people by profferring unasked-for advice, or so I am told. Since I've already suggested the M.S. to no effect, I could just wash my hands of the situation.

Still, though, I dread the moment that BF asks Advisor for a recommendation. The problem is that BF likely won't ask until it's actually grad school application time, i.e. November, by which time it will be harder for him to change his plans in response to Advisor's anticipated reaction ("You are applying to what?")

So I feel as though the Right Thing To Do would be to sit BF down and try again to persuade him to think this all through. It might be a waste of my breath, because he has a certain unshakeable confidence, but I think it's worth a try, on the theory that a little effort now might save some misguided university the cost of his tuition and stipend for 2-3 squandered years. Have any readers ever successfully talked someone out of applying to Ph.D. programs? Particularly someone who is a bit stubborn and ignorant, and perhaps also very bad at math?

Why do you write a training grant?

I've been funded off an NIH training grant (T32) at two different points previously, and I'm applying to be put on another training grant in the future.

My participation level in these training grants has been, at maximum, to write a paragraph of application explaining why I should be on one, and another page or so each year I'm on the grant to update NIH on my progress. Ridiculously easy.

But the person administering the training grant seems to do an awful lot of work. They write the grant, pick trainees, and generally deal with the flow of funds. My question is, what's in it for them? It seems like a giant hassle for very little professional benefit--perhaps you can stick one of your own trainees on the grant, but beyond that it just looks like scut work. Yet the two training grants I've been on have been administered by high-ranking PIs (and sniffing around T32s at other institutions on CRISP, this doesn't appear to be uncommon).

Do you just get departmental brownie points for writing/running a T32? Is it the sort of service activity that gets you out of jail for some other service activity, like being on a search committee? Is it just out of the greatness of your heart? (And if so, how to explain the presence of some T32 administrators who are well known not to have hearts at all?)

I'm grateful to have this funding, but can't imagine that "postdoc gratitude" inspires a lot of would-be grant-writers. So I have to assume there is something I don't understand about the apparently selfless action of these folks. Internet, please explain.

Your personal eigenvalue

This just cracked me up:

http://www.pnas.org/content/106/17/6883.full

PNAS publishes a letter on the "Eigenfactor" version of citation ranking, some attempt to come up with yet another version of impact factor. The fawning conclusion? That you can "publish in PNAS with the full knowledge that you are contributing to one of the most influential drivers of scientific progress."

Also, apparently, in JBC, which way outranks Cell.




Who amongst you doesn't want to know what those tantalizing unlabeled dots near Cell and Phys Rev B are? J Nobody Reads This?