Readers prefer papers with data-sets attached to them, because you can check the data and see how the conclusions of the paper stand up. That preference may be part of what drives the higher citation rates.
The main reason for that is that there is an established reputation system for papers, namely journals, but there isn't one for content formats that aren't papers - data-sets, code, videos, images etc.
Part of what Academia.edu is doing is building a new reputation system in science based on audience metrics: how many people have read your papers, from which countries; how many followers you have etc. We are then attaching various kinds of scientific media to that reputation system in order to accelerate sharing.
The idea is that if you can build your reputation by sharing your data-sets, you're likely to consider doing it.
Both I suppose, I think it would be interesting to browse Wikipedia and have access to actual peer reviewed science (with all the accompanying data). I am aware the wiki articles are supposed to be sourced, but that can be hit and miss.
Graduate student myself: This a cry for business that will fall on deaf ears in academia. Nobody would ever publish half-baked code that 'does the job', for the fear of getting judged. And none of us have time to prettify the code.
Academia.edu started with a lot of promise but is frankly a useless placeholder of your name and Email ID now. ResearchGate has surpassed them long ahead in useful features and Mendeley is already very good. They are just making some money from the academic jobs ads and I frankly don't see a future for them if they are happy being almost the exact same product for the past 3 years I have used them.
You're right to focus on rewards. The rewards have to be there for academics to share their data-sets and code. The feedback loop needs to be closed, so that if you share your data, and that data has an impact, that impact is reflected back on you. You can then take that impact and show it to your tenure/grant committee, who can reward you for it. There are a bunch of startups, Academia.edu being one of them, that are very focused on closing that feedback loop.
Mendeley and ResearchGate are both great companies, but if you look at Quantcast/Compete/Alexa, you'll see that Academia.edu is noticeably larger than them in traffic. Academia.edu also has more users. I'm sorry to hear that Academia.edu doesn't satisfy your use-case though....
Why should people put their data here rather than Dryad (http://www.datadryad.org/), which assigns DOIs and doesn't require registration to download data sets, and is a non-profit organization, governed by a board including PLoS and Science and powered by completely open source software?
If your papers are already on Academia.edu, and that is where you share your research output, it's useful to be able to share your data-sets and code too.
Of course that doesn't prevent you from uploading data to other places on the web, such as the site you mention...
Simply providing the ability to upload data only solves a small part of the problem. We run the CRAWDAD wireless network data archive (http://crawdad.org/) and the hardest parts are: convincing people to share data; ensuring that the data can be shared (much of the time this is not possible due to consent, data protection, etc); sanitising the data; and finally (and most time-consumingly) creating appropriate metadata so that the data are meaningful to other researchers.
The Research Data Alliance (http://rd-alliance.org/) is trying to solve many of these problems.
Agreed. There are many aspects to changing the scientific ecosystem in such a way that sharing data-sets is the default, not the exception.
My sense is that the most critical aspect here is the incentives aspect. I think if sharing data-sets can enhance a scientist's reputation, and make it easier for them to demonstrate to grant committees that they have had an impact on the field, scientists will put in the work to curate and share data-sets.
13 comments
[ 4.6 ms ] story [ 39.5 ms ] threadReaders prefer papers with data-sets attached to them, because you can check the data and see how the conclusions of the paper stand up. That preference may be part of what drives the higher citation rates.
The main reason for that is that there is an established reputation system for papers, namely journals, but there isn't one for content formats that aren't papers - data-sets, code, videos, images etc.
Part of what Academia.edu is doing is building a new reputation system in science based on audience metrics: how many people have read your papers, from which countries; how many followers you have etc. We are then attaching various kinds of scientific media to that reputation system in order to accelerate sharing.
The idea is that if you can build your reputation by sharing your data-sets, you're likely to consider doing it.
Academia.edu started with a lot of promise but is frankly a useless placeholder of your name and Email ID now. ResearchGate has surpassed them long ahead in useful features and Mendeley is already very good. They are just making some money from the academic jobs ads and I frankly don't see a future for them if they are happy being almost the exact same product for the past 3 years I have used them.
Mendeley and ResearchGate are both great companies, but if you look at Quantcast/Compete/Alexa, you'll see that Academia.edu is noticeably larger than them in traffic. Academia.edu also has more users. I'm sorry to hear that Academia.edu doesn't satisfy your use-case though....
Of course that doesn't prevent you from uploading data to other places on the web, such as the site you mention...
The Research Data Alliance (http://rd-alliance.org/) is trying to solve many of these problems.
My sense is that the most critical aspect here is the incentives aspect. I think if sharing data-sets can enhance a scientist's reputation, and make it easier for them to demonstrate to grant committees that they have had an impact on the field, scientists will put in the work to curate and share data-sets.
I heard that http://bohr.launchrock.com/ is working on something similar.