No comment

At the risk of beating the issue to death, I offer yet another post on the question, “why don’t scientists comment on scientific articles?” Previous reflections stood within the larger context of scientific impact and article-level metrics, and I’ve also attempted some superficial analysis of commenting behavior at PLoS, BMJ, and BMC. More recently (and this is why the topic is on my mind again), a room full of bright minds at the PLoS Forum (including Cameron Neylon and Jon Eisen) scratched their heads over it and came up with pretty much the same conclusion as everyone else who’s ever thought about the problem — the costs simply outweigh the benefits.

The costs, in principle, are minimal. You might need to register for an account at the journal website and be logged on, but then all that’s needed is little more than what most of us already do multiple times a day with our email — type into a box and click “submit”. (In practice, there may be nonsensical, hidden costs that make you wonder what the folks at those journals were smoking.) So the perception that the cost-benefit equation doesn’t work speaks more to the lack of benefit than anything else.

Photo by jamesclay on flickr


Read more of this post

A brief analysis of commenting at BMC, PLoS, and BMJ

As announced on FriendFeed and Twitter, a writing collaboration between me and the inimitable Cameron Neylon has just been published at PLoS Biology, “Article-level metrics and the evolution of scientific impact”! (Loosely based on a blog post from several months ago.)

One of the many issues Cameron and I touched on was the problem of commenting. Most people probably aren’t aware of the problem; after all, commenting is alive and well on the internet in most places you look! But click over to PLoS or BioMed Central (BMC) and the comment sections are the digital equivalent of rolling tumbleweed.

As we mention briefly in the article, comments have great potential for improving science. For one thing, they’re a form of peer review, but without the month-long wait and seemingly arbitrary review criteria. Readers, authors, and other evaluators can also get a sense of what people think about the article. The ideal is certainly tantalizing — vigorous, rigorous debates over the finer scientific points as well as the overarching conclusions with participation both from experts in the field as well as informed laypeople, always with intelligence and civility!!!1!11!!one!! But let’s not kid ourselves — the worst-case scenario is all too easy to imagine and would probably look something like the discussions over at YouTube.

And this would be positively urbane. (From PhD comics)


Read more of this post

Scientific discourse as an epic FAIL

A post on FriendFeed pointed me to this blog post in Adventures in Ethics and Science discussing a particularly infuriating example of just how broken the current system of scientific publishing can be. The epic tale is presented by Prof. Rick Trebino in a PDF document (above) outlining “How to Publish a Scientific Comment in 123 Easy Steps”. This version includes his second addendum in which he gives many excellent (and some painfully obvious) suggestions for how to improve the system.

Here’s a preview:

1. Read a paper in the most prestigious journal in your field that “proves” that your entire life’s work is wrong.

2. Realize that the paper is completely wrong, its conclusions based entirely on several misconceptions.  It also claims that an approach you showed to be fundamentally impossible is preferable to one that you pioneered in its place and that actually works.  And among other errors, it also includes a serious miscalculation—a number wrong by a factor of about 1000—a fact that’s obvious from a glance at the paper’s main figure.

3. Decide to write a Comment to correct these mistakes—the option conveniently provided by scientific journals precisely for such situations.

6. Prepare further by writing to the authors of the incorrect paper, politely asking for important details they neglected to provide in their paper.

7. Receive no response.

15. Write a Comment, politely explaining the authors’ misconceptions and correcting their miscalculation, including illustrative figures, important equations, and simple explanations of perhaps how they got it wrong, so others won’t make the same mistake in the future.

16. Submit your Comment.

17. Wait two weeks.

18. Receive a response from the journal, stating that your Comment is 2.39 pages long. Unfortunately, Comments can be no more than 1.00 pages long, so your Comment cannot be considered until it is shortened to less than 1.00 pages long.

20. Remove all unnecessary quantities such as figures, equations, and explanations.  Also remove mention of some of the authors’ numerous errors, for which there is now no room in your Comment; the archival literature would simply have to be content with a few uncorrected falsehoods.  Note that your Comment is now 0.90 pages.

21. Resubmit your Comment.

22. Wait two weeks.

23. Receive a response from the journal, stating that your Comment is 1.07 pages long. Unfortunately, Comments can be no more than 1.00 pages long, so your Comment cannot be considered until it is shortened to less than 1.00 pages long.

And so the saga begins. Really, the whole thing makes my blood boil.

The evolution of scientific impact

Photo by cudmore on Flickr

In science, much significance is placed on peer-reviewed publication, and for good reason. Peer review, in principle, guarantees a minimum level of confidence in the validity of the research, allowing future work to build upon it. Typically, a paper (the current accepted unit of scientific knowledge) is vetted by independent colleagues who have the expertise to evaluate both the correctness of the methods and perhaps the importance of the work. If the paper passes the peer-review bar of a journal, it is published.

Measuring impact

For many years, publications in peer-reviewed journals have been the most important measurement of someone’s scientific worth. The more publications, the better. As journals proliferated, however, it became clear that not all journals were created equal. Some had higher standards of peer-review, some placed greater importance on perceived significance of the work. The “impact factor” was thus born out of a need to evaluate the quality of the journals themselves. Now it didn’t just matter how many publications you had, it also mattered where.

But, as many argue, the impact factor is flawed. Calculated as the average number of citations per “eligible” article over a specific time period, it is highly inaccurate given that the actual distribution of citations is heavily skewed (an editorial in Nature by Philip Campbell stated that only 25% of articles account for 89% of the citations).  Journals can also game the system by adopting selective editorial policies to publish articles that are more likely to be cited, such as review articles. At the end of the day, the impact factor is not a very good proxy for the impact of an individual article, and to focus on it may be doing science – and scientists – a disservice.

In fact, any journal-level metric will be inadequate at capturing the significance of individual papers. While few dispute the possibility that high journal impact factors may elevate some undeserving papers while low impact factors may unfairly punish perfectly valuable ones, many still feel that the impact factor – or more generally, the journal name itself – serves as a useful, general quality-control filter. Arguments for this view typically stem from two things: fear of “information overload”, and fear of risk. With so much literature out there, how will I know what is good to read? If this is how it’s been done, why should I risk my career or invest time in trying something new?

What is clear to me is this – science and society are much richer and more interconnected now than at any time in history. There are many more people contributing to science in many more ways now than ever before. Science is becoming more broad (we know about more things) and more deep (we know more about these things). At the same time, print publishing is fading, content is exploding, and technology makes it possible to present, share, and analyze information faster and more powerfully.

For these reasons, I believe (as many others do) that the traditional model of peer-reviewed journals should and will necessarily change significantly over the next decade or so.

Article-level metrics at PLoS

The Public Library of Science, or PLoS, is leading the charge on new models for scientific publishing. Now a leading Open Access publisher, PLoS oversees about 7 journals covering biology and medicine as well as PLoS ONE, on track to become the biggest single journal ever. Papers submitted to PLoS ONE cover all areas of science and medicine and are peer-reviewed only to ensure soundness of methodology and science, no matter how incremental. So while almost every other journal makes some editorial judgment on the perceived significance of papers submitted, PLoS ONE does not. Instead, PLoS ONE leaves it to the readership to determine which papers are significant through comments, downloads, and trackbacks from online discussions.

Now 2 1/2 years old, PLoS ONE boasts thousands of articles and a lot of press. But what do scientists think of it? Clearly, enough think highly of it to serve on its editorial board or as reviewers, and to publish in it. Concerns that PLoS ONE constituted “lite” peer review seem largely unfounded, or at least outdated. Indeed, there are even tales of papers getting rejected from Science or Nature because of perceived scientific merit, getting published in PLoS ONE, and then getting picked up by Science and Nature’s news sections.

Yet there is still feeling among some that publishing in PLoS ONE carries little or no respectability. This is due in part to a misconception of how the peer review process at PLoS ONE actually works, but also in part because many people prefer an easy label for a paper’s significance. Cell, Nature, Science, PLoS Computational Biology – to most people, these journals represent sound science and important advances. PLoS ONE? It may represent sound science but it’s up to the reader to decide whether any individual paper is important.

Why is there such resistance to this idea? One reason may be tied to time and effort to impact: while citations always have taken some time to build up, a journal often provides a baseline proxy for the significance of a paper. A publication in Nature on your CV is an automatic feather in your cap, and easy for you and for your potential evaluators to judge. Take away the journal, and there is no baseline. For some, this is viewed as a bad thing; for others, however, it’s an opportunity to change how publications – and people – are evaluated.

Whatever the zeitgeist in particular circles, PLoS is clearly forging ahead. PLoS ONE’s publication rates continue to grow, such that people will eventually have to pay attention to papers published there even if they pooh-pooh the inclusive – but still rigorous – peer review policy. Recently, PLoS announced article-level metrics, a program to “provide a growing set of measures and indicators of impact at the article level that will include citation metrics, usage statistics, blogosphere coverage, social bookmarks, community rating and expert assessment.” (This falls under the broader umbrella of ‘post-publication peer review’.) Just how this program will work is a subject of much discussion, and certain metrics may need a lot of fine-tuning to prevent gaming of the system, but the growing consensus, at least among those discussing it online, is that it’s a step in the right direction.

Essentially, PLoS believes that the paper itself should be the driving force for significance, not the vehicle it’s in.

The trouble with comments

A major part of post-publication peer review such as PLoS’s article-level metrics is user comments. In principle, a lively and intelligent comment thread can help raise the profile of the article and engage people – whether it be other scientists or not – in a conversation about the science. This would be wonderful, but it’s also wishful thinking; as anyone who’s read blogs or visited YouTube knows, comment threads devolve quickly unless there is moderation.

From rustylime.com

From rustylime.com

For community-based knowledge curation efforts (think Wikipedia), there is also a well-known 90-9-1 rule: 90% of people merely observe, 9% make minor or only editorial contributions, and 1% are responsible for the vast majority of original content. So if your audience is only 100 people, you’ll be lucky if even one of them contributes. Indeed, experiments with wiki-based knowledge efforts in science have been rocky at best, though things seem to getting better. The big question remains:

But will the bench scientists participate? “This business of trying to capture data from the community has been around ever since there have been biological databases,” says Ewan Birney of the European Bioinformatics Institute in Hinxton, UK. And the efforts always seem to fizzle out. Founders enthusiastically put up a lot of information on the site, but the ‘community’ — either too busy or too secretive to cooperate — never materializes. (From a news feature in Nature last September on “wikiomics”.)

Thus, for commenting on scientific articles, we have essentially two problems: encouraging scientists to comment, and ensuring that the comments have some value. An experiment on article commenting on Nature several years ago was deemed a failure due to lack of both participation and comment quality. Even now, while many see the fact that ~20% of PLoS articles have comments as a success, others see it as a inadequate. Those I’ve talked to who are skeptical of the high volume nature of PLoS ONE tend also to view their comments on papers to be a highly valuable resource, one not to be given away for free in public but disclosed in private to close colleagues or leveraged for professional advancement through being a reviewer.

Perhaps the debate simply reflects different generational mindsets. After all, people are now growing up in a world where the internet is ubiquitous, sharing is second-nature, and almost all information is free. Scientific publishing is starting to change, and so it is likely that current incentive systems will change, too. Yet while the gulf will eventually disappear, it is perhaps at its widest point now, with vast differences in social norms, making any online discourse potentially fraught with unnecessary drama. As Bora Zivkovic mentions in a recent interview,

It is not easy, for a cultural reason, because a lot of scientist are not very active online and also use the very formalised language they are using in their papers. People who have been much more active online, often scientists themselves, they are more chatting, more informal. If they don’t like something they are going to say it in one sentence, not with seventeen paragraphs and eight references. So those two kinds of people, those two communities are eyeing each other with suspicion, there’s a clash of cultures. The first group sees the second group as rude. The second group views the first group as dishonest. I think it will evolve into something in the middle, but it will take years to get there.

When people point to the relative lack of comments on scientific papers, it’s important to point out the fact that online commenting has not been around in science for very long. And just as it takes time for citations to start trickling in for papers, it takes time to evaluate a paper in the context of its field. PLoS ONE is less than three years old. Bora notes, “It will take a couple of years, depends on the area of science until you can see where the paper fits in. And only then people will be commenting, because they have something to say.”

Brush off your bullshit detector

The last argument I want to touch on is that of journals serving as filter for information. With millions of articles published every year, it can seem a daunting task keeping up with the literature in your field. What should you read? In a sense, a journal is a classifier, taking in article submissions and publishing what it thinks are good and important papers. As with any classifier, however, performance varies, and is highly dependent on the input. Still, people have come to depend on journals, especially ones with established reputations, to provide this service.

Now even journals have become too numerous for the average researcher to track (hence crude measures like the impact factor). So when PLoS ONE launched, some assumed that it would consist almost entirely of noise and useless science, if it could be considered science at all. I think it’s clear that that’s not the case; PLoS ONE papers are indeed rigorously peer-reviewed, many PLoS ONE papers have already had great impact, and people are publishing important science there. Well, they insist, even if there’s good stuff in there, how am I supposed to find what’s relevant to me out of the thousands of articles they publish every year? And how am I supposed to know whether the paper is important or not if the editors make no such judgment?

Here, I would like to point out the many tools available for filtering and ranking information on the web. At the most basic level, Google PageRank might be considered a way to predict what is significant and relevant to your search terms. But there are better ways. Subscribing to RSS feeds (e.g. through GoogleReader) makes scanning lots of article titles quick and easy. Social bookmarking and collaborative filtering can suggest articles of interest based on what people like you have read. And you can directly tap into the reading lists of colleagues by following them on various social sharing services like Facebook, FriendFeed, Twitter, and paper management software like Mendeley. I myself use a loose network of friends and scientific colleagues on FriendFeed and Twitter to find interesting content from journals, news sites, and blog posts. The bonus is that you also interact with these people, networking at its most convenient.

The point is that there is a lot of information out there, you have to deal with it, and there are more and more tools to help you deal with it. It’s no longer sufficient to depend on only one filter, and an antiquated one at that. It may also be time to take PLoS’s lead and start evaluating papers on their own. Yes, it takes a little more work, but I think learning how to evaluate papers critically is a valuable skill that isn’t being taught enough. In a post about the Wyeth ghost-writing scandal, Thomas Levenson writes:

… the way human beings tell each other important things contains within it real vulnerabilities.  But any response that says don’t communicate in that way doesn’t make sense; the issue is not how to stop humans from organizing their knowledge into stories; it is how to build institutional and personal bullshit detectors that sniff out the crap amongst the good stuff.

From nitot on Flickr

From nitot on Flickr

Although Levenson was writing about the debate surrounding science communication and the media, I think there’s a perfect analogy to new ways of publishing. Any response that says don’t publish in that way doesn’t make sense; the issue is not how to stop people from publishing, it is how to build personal bullshit detectors – i.e. filters. People should always view what they read with a healthy dose of skepticism, and if we stop relying on journals, or impact factors, or worse to do all of our vetting for us, we’ll keep that skill nicely honed. At the same time, we are not in this alone; leveraging a network of intelligent agents – your peers – will go a long way.

So continue leading the way, PLoS. Even if not all of the experiments work, we will certainly learn from them, and keep the practice and dissemination of science evolving for the times.

Can’t attend ISMB 2009? The next best thing.

One of the biggest scientific conferences each year is Intelligent Systems for Molecular Biology (ISMB), put on by the International Society for Computational Biology (ISCB). I had the pleasure of attending the conference in Toronto last year, meeting many familiar names in person and collaborating with a number of them to microblog the sessions. That latter activity was so successful that it caught the eyes of the conference organizers, and we were able to publish a paper in PLoS Computational Biology summarizing the conference.

Even better, the ISCB is embracing microblogging from the outset this year at its ISMB meeting in Stockholm, which is starting this weekend and will run until July 2. They will be auto-generating threads for each talk in the FriendFeed room for live coverage and open commentary and are advertising that fact prominently on the website for those interested in blogging the event. Their actions are in stark contrast to those of Cold Spring Harbor, who recently updated their policies to require bloggers and twitterers to register with CSH beforehand and get advance permission from each presenter they plan on covering.

Now that blogging, microblogging, and even twittering is becoming more commonplace, it behooves conference organizers to have an official policy. Even one that is restrictive is better than no policy, which can result in an awkward backlash when people on both sides are caught unawares. Clearly there is no one-size-fits-all approach, but for conferences that do not deal with sensitive material, an open and even actively encouraging stance such as the ISCB’s is certainly liberating for those of us who are drawn to these kinds of activities.

So if you can’t attend ISMB this year for whatever reason, you (and I) are in luck. They’re freely providing the next best thing – live microblogging and a searchable archive of posts (through FriendFeed). Even if you’ll be physically attending, your experience will be arguably better if you follow the FriendFeed room. Because there’s only one of you, but there are also many others like you.

So check it out, whether you’re there or not, and if you’re there, contribute a post or two! If you’re not there, you can still participate by commenting and asking questions. That’s the beauty of it – the benefits go both ways!

The home stretch, and next steps

Photo by zizzy on Flickr

Photo by zizzy on Flickr

I’ve been neglecting my blog lately, through no fault of its own. There’s been lots going on, and lots not going on as a result, so I just wanted to post a quick update.

I’m a little over a month away from handing in my dissertation, which means many hours that might have gone towards writing blog posts instead went towards writing the darn thing. Learning LaTeX wasn’t nearly as bad as I thought it would be, especially with a handy Stanford thesis template and some web resources. I just handed a copy of the draft to my committee for comments and hoped for a brief respite to take care of some random bits (like all those appendices, code documentation, etc…).

Of course, my advisor says, “Now that you have nothing to do, why don’t you write a paper?” So now I’m writing a paper… which is probably a good thing, since there does seem to be enough material for a paper, and I might as well write it while I’m in so-called “writing mode”. I have the feeling by the time this is over, though, even Twitter will seem like too much writing!

Most of the rest of the time is spent on activities related to Ultimate, which started up about a month ago. There’s been two tournaments, some practices, various pickup games which are pseudo-mandatory for me, and various workouts to get in shape for the season.

Even with those two things, there’s probably ample time to squeeze in a post or two. But the mental energy isn’t quite there. And maybe it’s also because I still spend a couple hours cooking and get 8-9 hours of sleep most nights. Some things just can’t be sacrificed. Like home made strawberry shortcake with freshly whipped cream. :)

So that’s mostly what I’ve been up to the last couple months. The next month will probably be much the same. And then?

Well, I think it’s official enough now that I can announce it: I’ve accepted an offer to join 23andMe as a scientific curator in late August! I’m very excited about working with them and hopefully will be able to contribute across multiple facets of the company.

The end – and a new beginning – is in sight!

How to get your hands on some stimulus

Photo by sgw on Flickr

Photo by sgw on Flickr

If you weren’t stimulated enough by my recent post about Stimulomics, I’ve got some more goodies to get you going. My advisor took the first half hour of our last group meeting to give us a giddy “civics” lesson. The reason for his exuberance soon became clear: thanks to the stimulus package, the NIH needs to spend a LOT of money, and spend it FAST.

How much money, you ask? Less than 2% of the entire stimulus but that still translates to a hefty sum. With the stimulus upwards of $800B, NIH is getting a little over $10B that it must spend by the end of two years – with the aim of creating jobs alongside accelerating research. (The NSF is getting some, too, but was ignored for the purposes of this discussion, probably because of the interests of the audience.) Apparently, there are those who doubt that the NIH is capable of spending that much money, and so the grant fruit, which usually clings tightly to the funding tree, is now weighed down on its branches, ripe for the picking.

A fair amount of the NIH stimulus is going towards renovating the NIH campus in Bethesda and updating its infrastructure, but the majority is going towards 4 major classes of grants (meaning there’s something – and maybe multiple somethings – for everyone). Update (thanks Andrew for the heads up): Except for the first type, all grants have funding allotments and due dates specific to their funding agency (of which there are dozens), so check this list for all the details. The amounts below – except for the Challenge grants – are quoted for the NLM.

  1. Challenge grants – $500K x 2 yrs, total award $200M
    Due 4/27/09
    I mentioned these in the Stimulomics post, but to recap briefly, these are peer-reviewed grants for new projects, of which the NIH wants to fund at least 200. There are something like 800 acceptable topics (PDF), so I wasn’t kidding when I said there was something for everyone. The research portion of these grants isn’t supposed to exceed 12 pages, so you’re almost crazy NOT to write one.
  2. Competitive revisions – ~$500K x 2yrs, total award $7M
    Due 4/21/09
    These are grants that significantly expand the scope or research protocol of currently funded projects. Itching to apply your method in a new area? Go one step further with your research aims? Want to fund more students or add another group as a collaborator to explore a related direction? This grant is for you. These are peer-reviewed, and I’m not sure what the page limit is, but it can’t be more than for Challenge grants.

  3. Administrative supplements – $100K x 2 yrs, total award $15M
    Due 5/10/10 5/15/10 (rolling basis)
    These are smaller grants to accelerate existing aims of currently funded research. So if you need to buy hardware, or to hire new people (or contractors like Stimulomics), this grant is yours for the taking. About half of the total award is going towards career and training, so NLM training grants can apply for slots that were previously withheld and terminal phase K99s can apply for 1 year extensions or $50K supplements. The key here is that there is no peer review; instead, the application is reviewed by the program officer for your existing grant. And they really want you to apply for one, wink wink. Best of all, the grant is limited to 2-5 pages.

  4. Summer research supplements – total award $1M
    Due 5/10/10 5/01/10 (rolling basis)
    These are grants that supplement existing funded projects for the purposes of providing summer research opportunities for students (high school and undergraduate) and science educators. Interestingly, the topic of priority is… *drumroll*… informatics! And of course, we all love science education. :)

To give one last plug for Stimulomics, the thing that’s attractive about hiring contractors is that it’s a deliberately temporary relationship, which is perfect for these two year grants. Who wants to hire people full time only to have to lay them off after two years? Awk.ward.

At any rate, you have to admit all these options are tempting. It almost – almost! – makes me want to stay in academia. So all you folks who are, go get yourself some stimulus!