The recent posts about assisting organizations to help hindering cheaters in competitions / tests (e.g. PRIMES and a Polish Math Olympiad) made me think:

Is it our responsibility to police for these questions?

Next thing we know, every organization that has some introductory math tests (that are sophisticated and unique in their opinion, I'm pretty sure) requests we police math.SE - that's their responsibility to take care of in my opinion.

Now, math.SE isn't the first StackExchange to face issues of this kind, mother StackOverflow has similar questions such as

Heck, there's even an ethics tag at meta.SO. Honestly, I only skimmed the answers there for now, but I think there are some good points that also apply here, especially:

If you give pedagogical answers to pedagogical questions, the matter of ethics does not come up. -- dmckee

$\rightarrow$ Questions from an exam/contest/etc. are usually very -like, so they are supposed to be answered with hints only anyway. The internet has become part of our daily lives - deal with it.

at no time did SEI actually alert an institution to academic dishonesty or assist one in identifying a user -- Tim Post♦

$\rightarrow$ the latter would actually violate the SE privacy policy

  • $\begingroup$ There was a good answer that I hope gets undeleted (whether or not I agree with its proposals). $\endgroup$ – zyx Oct 7 '13 at 22:15
  • $\begingroup$ @zyx Indeed, Asaf made some important points about the beyond-SE math community. Lacking 10k I cannot tell whether anything was added after my reply explaining why it was deleted though... $\endgroup$ – Tobias Kienzler Oct 8 '13 at 8:02

My thoughts:

  1. It is not our responsibility; the moderators do not actively hunt for competition questions, and we don't expect users to do so either.
  2. But it never hurts to play nice: if it is brought to the attention of the moderators that a question is from a competition, we may do something about it. As a user you are free to follow the request by PRIMES or any other organisation, or you can choose not to pay attention to it.

Let me add something about "privacy policy": assisting an educational institution in identifying cheating may or may not violate the privacy policy. I quote from the privacy policy:

Other than on Careers, we won’t share your personal information with non-agent third parties unless we are required to do so by law, or if we believe in good faith that disclosure is reasonably necessary to protect our property, rights or those of third parties or the public at large.

While it is true that SE (and the moderators) will not volunteer information about individual users to (say) universities, it is unclear (and not for me to decide) whether an event where SE assist an educational institution in such a way can arise, as best to my knowledge no university has sued StackExchange for users' IP addresses yet.

  • $\begingroup$ There was a disclosure to third parties of information derived from private user data, in at least one case. The third party listed two or more usernames and inquired about the cardinality of the set of persons using those accounts. The response was not "privacy policy prevents us from answering that". It would be useful to know whether this type of disclosure can happen in the future. $\endgroup$ – zyx Oct 7 '13 at 16:52
  • $\begingroup$ Thanks for you thoughts on this - I'm not sure how far the privacy policy covers this, but I severely doubt a mere "we suspect this user participates in our event and tries to brake the rules" is sufficient (as compared to breaking actual laws), though I wouldn't be surprised if at least a mod or SE dev would forward mails to keep said user's privacy respected while offering a way to potentially solve such an issue. Concerning playing nice, could you specify the "do something about it"? Would that be e.g. locking such a post until after the contest? $\endgroup$ – Tobias Kienzler Oct 7 '13 at 19:31
  • 2
    $\begingroup$ @TobiasKienzler: like you said, we cannot help the organisers identify the potential cheaters; though they often identify themselves when they submit an answer cribbed from MSE directly. In general the Moderators do not (and cannot, per the Mod agreement) make decisions on release of personal information. Any such requests would be processed by a rep of SE Inc. So the extent to which we can "do something about it" is exactly like you said: we can lock a question, we can add a moderator notice, we can delete the existing answers. Not much beyond that. $\endgroup$ – Willie Wong Oct 8 '13 at 7:43
  • 2
    $\begingroup$ I believe that the case @zyx refers to is what transpired here. The way that the privacy policy was written, the confirmation given by Qiaochu was "technically allowed". But since then (and possibly because of that), SE Inc has clarified to the moderators not to do that in the future. So I will say that the particular type of disclosure that zyx had in mind in his first comment should not happen again in the future. $\endgroup$ – Willie Wong Oct 8 '13 at 9:08
  • $\begingroup$ In general, the current policy is that if there's any potential for private information to be used, we clear any putative statement with the Community Team before we make them. Generally this will be done inside the Math-Mods chat room. $\endgroup$ – Willie Wong Oct 8 '13 at 9:10
  • $\begingroup$ Thanks. (@WillieWong ) The comment was aimed at the abstract behavior and the future, and hoping to avoid particular names and old events, but yes, that's the situation I had in mind. It was clear yet fuzzy at the time that disclosure could be technically consistent with the then-policy (which talked about IP addresses IIRC), but the straightforward interpretation covers anything computed from private data. The current policy does not fully clarify how broadly the words "private information" are understood, and whether "rights of third parties" refers exclusively to legal rights. $\endgroup$ – zyx Oct 8 '13 at 15:08
  • $\begingroup$ Chat link does not work (probably needs mod/dev privilege). Did you mean to convey that SE Inc's current interpretation includes things derived from the private data in addition to the data itself? (I'd include anything some random other modestly privileged user could not determine if asked, cold, to obtain the information in a limited time using only the human interface to SE. What can be done using Prism-like monitoring and archiving of the public data, or sophisticated unraveling of the undisplayed data exchanged while browsing, is essentially private information.) @WillieWong $\endgroup$ – zyx Oct 8 '13 at 15:20
  • 1
    $\begingroup$ @zyx: SE's instructions to moderators are quite clear, "Moderators should generally not provide any information related to or derived from emails, real names, IP addresses to any third parties." Stuff that a random individual could see if he was at the right place at the right time is not private, no matter how you read the rules. In regards to how SE employees and their lawyers see the situation, well, I cannot speak for them and am not privy to that information. $\endgroup$ – Willie Wong Oct 8 '13 at 15:45
  • $\begingroup$ Stuff that a random individual could see if he was at the right place at the right time is not private. This is not the place to discuss all forms of privacy, but that is a (technical) legal concept of privacy, while we are discussing whether the SE policy and informal user notions of privacy align. Under the italicized definition SE could supply a user's entire comment history, including all deleted ones and edit-revisions, and all deleted postings, just as things a passerby could have noticed who coincidentally Maxwell daemoned his passings-by to observe every one of these many events. $\endgroup$ – zyx Oct 8 '13 at 15:54
  • $\begingroup$ Thanks for quoting the chat. They do now include derived data in so many words, and that is an important specification. (To add to previous comment, your entire history of email hashes is something a suitably equipped random person walking by at opportune times could have observed, and that would be of significant interest to third parties wanting to make an ID. It seems to be covered by SE's policy but the "theoretically public at some point" standard is not realistic one when applied to a possibly large and evolving archive of user postings.) $\endgroup$ – zyx Oct 8 '13 at 16:00
  • 2
    $\begingroup$ @zyx: I will not debate with you on what should be private. While I often strongly disagree with your world view, that is besides the point. I do not write the privacy policy. SE does. And they are very clear about what is considered public (scroll down to the section "Information You Choose to Display Publicly on the Network"). If you don't like it, you should stop using the site. Harassing me about it is not going to amount to anything. $\endgroup$ – Willie Wong Oct 9 '13 at 7:25
  • $\begingroup$ @WillieWong, the last is one of an unfortunately long stream of aggressive comments from you to me in the meta, in situations where I neither debated nor proposed to debate, much less "harassed" you. A reading of my preceding words "this is not the place to..." should indicate that I was stating your very point, on not debating what privacy is or should be. I cannot fathom how you consider clarification of what is SE's current policy, and elaboration of your comments based on unpublished information, as off-topic, since you initiated the posting about it. $\endgroup$ – zyx Oct 9 '13 at 10:39
  • $\begingroup$ It is also a misreading of anything I wrote to imagine that it is a complaint about the current SE policy. I really think you ought to reconsider and delete your comment with the loaded and patronizing "if you dont like it, ....". If you do so, feel free to remove these two replies. I really, really don't care for the unearned personalized comment trolling that has been your pattern in the last several interactions, nor for the spectacle (not a new one) of having moderators be the ones instigating attacks. @WillieWong $\endgroup$ – zyx Oct 9 '13 at 10:44
  • 1
    $\begingroup$ Personally I didn't read any aggression into @zyx's comments, but I can merely imagine what actual misbehaviour mods may be exposed to as "thanks" to them spending their spare time moderating SE for free. There's the saying "to a cop, everyone's a suspect", and I suspect a lesser version applies to moderators as well - they have to develop the skill to read and extrapolate subtext in order to defuse unpleasant situations preferably before they even occur. That means they may err sometimes, but in total their efforts really make the internet a better place. $\endgroup$ – Tobias Kienzler Oct 12 '13 at 7:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .