3 ms·
More important: > We have also received consistent feedback encouraging us to promptly discard responses to study email. We agree, and we will delete all respo
by kstrauser 5y ago
More important:
> We have also received consistent feedback encouraging us to promptly discard responses to study email. We agree, and we will delete all response data on December 31, 2021.
I wrote one of the blogs posts that got linked here on HN, and I have some strong feelings about that. None of them are joy, though. I think it’s good and appropriate that the study is deleting all the data; since it was collected by misleading methods, I don’t think it was valid. I’m not happy that a study covering an important subject, and led by researchers who had good motivations, went so far off the rails in the first place that it had to be axed.
Edit: I wrote more about my response to this whole situation at https://honeypot.net/post/dealing-with-princetons-flawed-privacy-research/ https://honeypot.net/post/dealing-with-princetons-flawed-pri... .
- gpm 5y agoIgnoring the ethical concerns, all the data they collected was completely worthless, because many of their subjects were contacting eachother and responding to it with the knowledge that it was a mass email sent with a variety of presumably fraudulent names.
- charcircuit 5y agoEarly data may be less biased than later data. The data should just be released anyways since the damage has already been done.
- seoulmetro 5y agoIt definitely was valid. Probably the most valid data you're going to get if you want to test for this thing. Your emotions are getting in the way of your logic. I personally don't see what the problem is. People should be allowed to send whatever emails they like, it's up to you to reply to them. If they had sent emails out asking how everyone's day was going, would you get upset? What about how many employees they had?
- ad404b8a372f2b9 5y agoResearchers are held to a higher ethical standard than random people because experimenting on people is morally dubious unless you follow strict guidelines. I feel like there is a lot of hyperbole wrt the harm that was done by this study but on the other hand I think it's clear researchers shouldn't have free reign to manipulate people as they see fit just because it's through email.
- seoulmetro 5y agoI'm not talking about moral standard when I talk about the first sentence. I'm talking about data. And this data is flawless. They shouldn't be allowed to manipulate people, but that doesn't discredit the results. The same way it would be very bad for us to put normal, average people in control of jetliner cockpits as a scientific test to see if they can fly them properly with guidance. But if that managed to happen, the data would still be valuable. I wouldn't call sending one-off emails to people to be manipulation.
- jffry 5y agoGiven how widespread word of this thing got, how could researchers possibly distinguish responses to their email that were from people who were not aware it was research, versus responses from people who had become aware what was happening? The same goes for people who didn't respond. Did they not respond because they heard about this being research, or did they not respond for other reasons? This data is the opposite of flawless, it is poisoned, and any attempt to draw conclusions from the responses they got would be junk.
- seoulmetro 5y ago>Given how widespread word of this thing got, how could researchers possibly distinguish responses to their email that were from people who were not aware it was research, versus responses from people who had become aware what was happening? By limiting results by time to anything received before everyone found out? It's pretty easy to set a time period for that. It has the benefit that people who find out will not want to participate or will complain. >This data is the opposite of flawless, it is poisoned, and any attempt to draw conclusions from the responses they got would be junk. Nope. Even just limiting it to 48hours would provide great data.