6 ms·

This probably varies by field, but the "large project" thing can be gamed too. So, for example, in biomedicine you often have lots of people on a paper who mig

by o09rdk 7y ago

This probably varies by field, but the "large project" thing can be gamed too. So, for example, in biomedicine you often have lots of people on a paper who might only read a draft, make some trivial suggestions, and then be added as an author. As a result, there's this pressure for large groups to form, where everyone is added and everyone can cite each other. This doesn't mean the projects are bad, but it does lead to individuals with large citation counts primarily because they find ways to add themselves to everything, regardless of their level of effort. People should get credit where it's due, and large projects involve lots of people. But what defines a "project" has become very vague. I've become extraordinarily disillusioned with academics. Science gets done but the rewards seem to filter preferentially to those who are able to game the system, and the system exists out of a need to make one's self look as productive as possible, in areas where contributions are generally necessarily tiny or nonexistent, even among very competent people, because the problems are hard and because so many people see the same things at the same time.

cbanek 7y ago
I don't disagree with what your saying, but I think LSST's builder concept is actually quite amazing and the opposite side of that coin. For people building the telescope (think hardware, software, logistics, everything before the science can be done), many of whom are not academics, and don't typically get authorship or write papers, it's great to get credit for working on the project in a formal, public way. You don't even have to edit or provide some kind of task directly related to the paper either, which I agree can get somewhat clique-ish.
- abricot 7y ago
  Wouldn't it make more sense to adopt the way Hollywood does crediting by then crediting everyone but with a note on how they contributed?
- Bartweiss 7y ago
  The builder concept is actually really appealing. Academia can tend towards the same problem as consulting, where tasks get sharply split between "credit-producing" and "not worth doing". Answering questions about your past papers; looking over someone else's proposed methodology; or cleaning up an internal tool into one you can share are all great tasks for advancing the field, but none of them bolster a CV, earn grants, or help you get tenure. If you want credit for them, you usually have to commit lots more time to the task, like running a formal discussion, becoming an author, or polishing the tool into an OSS contribution. All too often, the result is siloed projects and work abandoned as soon as it's published. (How many papers offering some novel twist on priming or ego depletion could have been turned into replication-and-extension if past authors had been involved?) Especially in astronomy, with large projects and lots of non-PhD team members, this makes so much sense. (I believe something similar may happen at LIGO - if not formally then at least in practice?) If work is going to be judged by authorship, it's only fair to recognize that at a certain point the groundwork and floating aid people give is comparably valuable to the act of writing up some chunk of the text.
  - godelski 7y ago
    This definitely happens on LIGO. You have hundreds of authors. My optics professor in undergrad was never a first author but he sure is an author on a lot of papers.
mschuster91 7y ago
> Science gets done but the rewards seem to filter preferentially to those who are able to game the system, and the system exists out of a need to make one's self look as productive as possible This is largely due to the current model of science funding, and not just in the US. Here in Germany, many involved in public (i.e. at an university, not inhouse r&d at a company) science only get 1-year-limited chain contracts with no real security and low pay, as many grants and funds are also available only in the same time frame which means it's a constant hustle for funding, especially at third parties. IMO there is only one way to solve this problem: politics have to irrevocably allocate fixed chunks of money for public scientifical investment for long terms (think 10 or 20 years) to give the scientists and universities actual security in their planning and staffing. This would also solve the problem that e.g. NASA has with each President reversing course. No wonder that the last Moon visit was decades ago when the priorities get completely turned over every 4-8 years.
- cr0sh 7y ago
  > This would also solve the problem that e.g. NASA has with each President reversing course. No wonder that the last Moon visit was decades ago when the priorities get completely turned over every 4-8 years. This is also one of the arguments against promoters of so-called "term limits" for congressmen and others; imagine this kind of churn in priorities occurring with major and minor public works projects! We don't have to wonder too much - we can already see it at the state level with governorships changing; one recent large change of this kind is with California's high-speed rail system. For all of it's "boondoggle-ry" and problems, I don't think the way it's been "axed" lately will be of help to completing it. In fact, it might just be a self-fulfilling prophecy for its opponents. That's only one example; I'm sure others in other states could be easily found as well if one were to look. Ultimately, that kind of thing would only get worse with term limits on representatives to Congress, because federal funding for such large scale projects is needed - and that would end up likely in flux, and ultimately scuttle projects that depend on steady funding to be completed. One could argue that individual state projects should only be funded by the state itself, but that notion of state self-sufficiency went out with the end of the Civil War. I also tend to wonder if - under a term-limited system - such a thing as the interstate highway system could have ever been built. It doesn't seem likely.
Breza 7y ago
Academics have basically reinvented the webring.
ajuc 7y ago
The problem is with the binary attribution. Either you're an author of everything in the paper or you're not an author at all. Software world solved this issue with version control systems like git. And if scientists write papers in latex or other text-based formats it's trivial to use version control system for that too. Then when you quote a fragment you do "git blame" on it, and you see who created and edited this fragment, so you can quote only the relevant people instead of authors of the whole book. This would make it much harder to abuse quotation rankings. Additional benefits - when a paper is found to contain manipulated data or other errors - it's trivial to check who did it, so only that person's career is done.
- godelski 7y ago
  Except that's not how science is done. Some people are bad writers and just don't touch the paper at all. Sometimes you have a grad student do all the work for a paper and someone else writes it, or the majority (common in the first year or two, or with undergrads). Just because they had no or little commits should they not be the main author? They did the work after all. Another example, my advisor doesn't like git. While writing papers I and my collaborators use git but send an email copy to my advisor. Clearly he's going to be on the paper because he's my advisor but you'll see zero commits from him. I think it's just too easy to think that technology solves this in a trivial way. It's complicated. You have people from different eras working on things. And this is in a CS program, mind you. In different fields it gets much worse very quick. Side note: go look at papers from top tier universities. You'll notice that they frequently cite colleagues at their University. Is this because they are gaming the system? Is it because they are doing the most related research (which is VERY common for a single University to work close)? Or is it a combination. In all likelihood it's a combination because citations matter. The h index is used in your performance because this is meant to be how impactful your paper is, but the system can definitely be manipulated (and likely isn't happening for malicious reasons nor necessarily unethical reasons)
  - antpol 7y ago
    Not to mention the politics of adding prominent names from your University/Institute to the list of authors to improve the chances of paper acceptance to competitive journals and conferences

by o09rdk 7y ago

cbanek 7y ago

I don't disagree with what your saying, but I think LSST's builder concept is actually quite amazing and the opposite side of that coin. For people building the telescope (think hardware, software, logistics, everything before the science can be done), many of whom are not academics, and don't typically get authorship or write papers, it's great to get credit for working on the project in a formal, public way. You don't even have to edit or provide some kind of task directly related to the paper either, which I agree can get somewhat clique-ish.

abricot 7y ago
Wouldn't it make more sense to adopt the way Hollywood does crediting by then crediting everyone but with a note on how they contributed?
Bartweiss 7y ago
The builder concept is actually really appealing. Academia can tend towards the same problem as consulting, where tasks get sharply split between "credit-producing" and "not worth doing". Answering questions about your past papers; looking over someone else's proposed methodology; or cleaning up an internal tool into one you can share are all great tasks for advancing the field, but none of them bolster a CV, earn grants, or help you get tenure. If you want credit for them, you usually have to commit lots more time to the task, like running a formal discussion, becoming an author, or polishing the tool into an OSS contribution. All too often, the result is siloed projects and work abandoned as soon as it's published. (How many papers offering some novel twist on priming or ego depletion could have been turned into replication-and-extension if past authors had been involved?) Especially in astronomy, with large projects and lots of non-PhD team members, this makes so much sense. (I believe something similar may happen at LIGO - if not formally then at least in practice?) If work is going to be judged by authorship, it's only fair to recognize that at a certain point the groundwork and floating aid people give is comparably valuable to the act of writing up some chunk of the text.
- godelski 7y ago
  This definitely happens on LIGO. You have hundreds of authors. My optics professor in undergrad was never a first author but he sure is an author on a lot of papers.

mschuster91 7y ago

> Science gets done but the rewards seem to filter preferentially to those who are able to game the system, and the system exists out of a need to make one's self look as productive as possible This is largely due to the current model of science funding, and not just in the US. Here in Germany, many involved in public (i.e. at an university, not inhouse r&d at a company) science only get 1-year-limited chain contracts with no real security and low pay, as many grants and funds are also available only in the same time frame which means it's a constant hustle for funding, especially at third parties. IMO there is only one way to solve this problem: politics have to irrevocably allocate fixed chunks of money for public scientifical investment for long terms (think 10 or 20 years) to give the scientists and universities actual security in their planning and staffing. This would also solve the problem that e.g. NASA has with each President reversing course. No wonder that the last Moon visit was decades ago when the priorities get completely turned over every 4-8 years.

cr0sh 7y ago
> This would also solve the problem that e.g. NASA has with each President reversing course. No wonder that the last Moon visit was decades ago when the priorities get completely turned over every 4-8 years. This is also one of the arguments against promoters of so-called "term limits" for congressmen and others; imagine this kind of churn in priorities occurring with major and minor public works projects! We don't have to wonder too much - we can already see it at the state level with governorships changing; one recent large change of this kind is with California's high-speed rail system. For all of it's "boondoggle-ry" and problems, I don't think the way it's been "axed" lately will be of help to completing it. In fact, it might just be a self-fulfilling prophecy for its opponents. That's only one example; I'm sure others in other states could be easily found as well if one were to look. Ultimately, that kind of thing would only get worse with term limits on representatives to Congress, because federal funding for such large scale projects is needed - and that would end up likely in flux, and ultimately scuttle projects that depend on steady funding to be completed. One could argue that individual state projects should only be funded by the state itself, but that notion of state self-sufficiency went out with the end of the Civil War. I also tend to wonder if - under a term-limited system - such a thing as the interstate highway system could have ever been built. It doesn't seem likely.

Breza 7y ago

Academics have basically reinvented the webring.

ajuc 7y ago

The problem is with the binary attribution. Either you're an author of everything in the paper or you're not an author at all. Software world solved this issue with version control systems like git. And if scientists write papers in latex or other text-based formats it's trivial to use version control system for that too. Then when you quote a fragment you do "git blame" on it, and you see who created and edited this fragment, so you can quote only the relevant people instead of authors of the whole book. This would make it much harder to abuse quotation rankings. Additional benefits - when a paper is found to contain manipulated data or other errors - it's trivial to check who did it, so only that person's career is done.

godelski 7y ago
Except that's not how science is done. Some people are bad writers and just don't touch the paper at all. Sometimes you have a grad student do all the work for a paper and someone else writes it, or the majority (common in the first year or two, or with undergrads). Just because they had no or little commits should they not be the main author? They did the work after all. Another example, my advisor doesn't like git. While writing papers I and my collaborators use git but send an email copy to my advisor. Clearly he's going to be on the paper because he's my advisor but you'll see zero commits from him. I think it's just too easy to think that technology solves this in a trivial way. It's complicated. You have people from different eras working on things. And this is in a CS program, mind you. In different fields it gets much worse very quick. Side note: go look at papers from top tier universities. You'll notice that they frequently cite colleagues at their University. Is this because they are gaming the system? Is it because they are doing the most related research (which is VERY common for a single University to work close)? Or is it a combination. In all likelihood it's a combination because citations matter. The h index is used in your performance because this is meant to be how impactful your paper is, but the system can definitely be manipulated (and likely isn't happening for malicious reasons nor necessarily unethical reasons)
- antpol 7y ago
  Not to mention the politics of adding prominent names from your University/Institute to the list of authors to improve the chances of paper acceptance to competitive journals and conferences