For example, I download all the content of Academia@SE, later analysis it in a data mining paper, and submit the paper in the end. Is it OK to do so? Do I have to ask the permission from the administrator of the website? And does he or she have rights to forbid my academic use? Thank you.
-
Maybe you should site it; better safe than sorry– LemonCommented Jun 5, 2013 at 3:35
-
@sidht Yes I will surely cite it.– ZiyuanCommented Jun 5, 2013 at 3:57
-
To clarify: your question was originally titled "Is it legal/ethical to use data grabbed from a certain website in a paper?", and @Piotr edited it to "a Stack Exchange site". Are you interested in the situation for SE sites specifically? Or do you want to know about general sites, with SE just being an example?– Nate EldredgeCommented Jun 5, 2013 at 12:24
-
@ziyuang Excuse me for an overzealous edit. However, title of the question didn't match the content. So I second Nate with question (as SE may be not a typical example).– Piotr MigdalCommented Jun 5, 2013 at 14:02
Add a comment
|
3 Answers
Your University may have an Institutional Review Board (IRB) that reviews how you conduct experiments. This board may be known by various names (Ethics Committee, Experiment Review Board, Human Subjects Research, etc.) but they are generally the ones that you would go to to consult about whether what you are doing is within the scope of ethical behavior and good treatment of human subjects data.
As StackOverflow and associated StackExchange repository data is available under Creative Commons Attribution Share-Alike (as @piotr_migdal linked above) and is publicly available, your IRB will probably tell you, "It's fine" and not require review. However, it depends on the IRB and the institution and the nature of the data.
There are entire research disciplines built on scraping web sites, software repositories, and social media, so don't feel bad for doing it.
-
+1 for IRB approval. It's fairly easy to make the "seatbelt study" argument and say that something posted to a site like StackExchange is intended to be publicly viewable. But it never hurts to have IRB approval, and your university may indeed require it (mine does for all research - I have an IRB approval form for a purely simulation study around here somewhere).– FomiteCommented Jun 12, 2013 at 23:07
Disclaimer: I am not a lawyer. If you are seriously concerned about this issue, you should consult one; your institution probably has intellectual property lawyers on staff.
There is a general principle that "you can't copyright facts". Wherever you get your data set, you probably can legally publish any analysis of that data, without requiring anyone's permission. However, you may not be able to legally reproduce the data itself.
Of course, by standard academic ethics, you must properly cite and attribute the source of the data. And if you can't guarantee that the data will remain accessible, it could affect the reproducibility of your results and hence the quality of your paper.
answered Jun 5, 2013 at 4:10
-
SE license all user content under CC SA, so here there is no problem with copyright. Commented Jun 5, 2013 at 10:00
-
@PiotrMigdal: Before your edit to the question's title, I read it as asking about data from Internet sources in general, not just SE. Commented Jun 5, 2013 at 12:25
-
For me it was not clear as well, just I matched title to the content. But an answer by @Shion interprets it as SE-specific. And anyway, I wrote a comment, it is not a downvote. Commented Jun 5, 2013 at 14:04
I believe you can do it with StackOverflow data, as long as you cite/attribute it properly. This article affirms it. However, I do not know whether this can be extended to the rest of StackExchange. A question to the mods or to the support team might help you clarify.
-
Thank you. So for other internet sites, do you mean I need to ask beforehand anyway, because there are no such licenses usually?– ZiyuanCommented Jun 5, 2013 at 4:09
-
This depends on the privacy policy of the particular website. There are ways to get around it. You can email the mods of the websites in question and if you are a credible researcher, often they will give you permission to use their data, often in a de-identified dataset format.– ShionCommented Jun 5, 2013 at 4:49