Indeed. Though I do find Chrome is very good at preventing infection with these messages. I'm in a bit of a fix as my antivirus crashed both my computers with a recent update, had to roll back a restore on both of them, now I am scared to update the AV in case it happens again. LOL. Suppose I either get a different AV or bite the bullet with the upgrade.
Same happened to me. Next thing I knew I was logged out of my Account. Logged back in just fine. Any cause for concern about my safety on mTurk?
I think the malware was being served by one of the image hosts, so outside mTurk control and probably the requester, too. As ever, be cautious on line
indeed. not doing his HITs ever again. The second time I was given the message I decided it was no coincidence. thanks for the heads up everyone.
Behind the ProductRnR I read again all these messages of this thread, all of them full of complaints, and it just pop up in my mind that it's possible behind ProductRnR to be someone of Amazon's staff How else could be explained the protection that this requester benefits? That's the ProductRnR principle: Let them squawk - We go on stoling their work !
I don't like RNR either. I did a few of the flag videos, they were ALL RIGHT! and I got half of them rejected, WTF! Then I did the ones flag an image, and I got most right, but there were some that were hard to decide, I think I chose right, yet I still got them rejected. Generally I don't work with them anymore. I've never gotten banned from them, just a few rejections here and there.
This thread is good for a laugh. Simple solution to a simple problem: Don't do these HITs. If no one does them, they will eventually go away. Reporting them to the FBI -- L M F A O
I rely wonders, even after the availability of turkopticon, turkers r doing the hits from productrnr. They have only themselves to blame.
I would say so because I just did 1000+ of the relevant pictures V2 and They are slowly but surely approving. I have 250+ from my first batch all approve, and my 400+ day is just now approving since it's 3 day auto-approve. Never had problems with him yet.
I guess they have a newer scoring system. They don't reject hits anymore. Instead they grade the person's work. If it falls to an unacceptable level, they are banned for life. So watch those Quality scores. If they get near the limits, then you are doing something wrong. Really they should have more detailed instructions. There are too many borderline situations displayed in the images to do perfect work. And yes, I've done a few thousand of their hits and plan to do multiple thousands more in the future.
Quality score issue After that old time when this requester gave mass rejection and no communication, about one year ago I received an email with apologies from him and things have been starting to go better. I've done over 30k hits in time(Identify adult images) with no rejection because they have introduced this system with Quality scores. So far so good, but recently my Quality score 1.1 reached to 13. Even if all of my other quality scores have excellent value regarding the red limits, I can't work anymore on them due to this damn score 1.1. I've messaged the requester for many times by now asking what is that about, but to no avail. It seems he has returned to the old bad habit - zero communication. Does anybody know what this score is about? Are there any workers in the same situation (on the last reviews on turcopticon I see similar complaints) Hoping they have a look around here and give us a reply.
I usually keep track of the 1.2 Quality score - which I believe is to make sure you getting the check correctly. My 1.1 is 6, don't know if it changes much for me. Could be a time thing. The faster you are, the less quality your work is? Just a guess, but I can not think of a different Quality score he can get from you. I lost one of his quals, but I think he either reset the scores or changed the hit. I have not seen a RNR that I am not qualified to do (as of now!)
This is ProductRnR who have been submitting all the ProductRnR HITs. First of all, thanks for working on ProductRnR hits. We have been getting emails asking questions about our hits. We would like to provide an overview here. 1. Guidelines for our hits - Guidelines provide important information on how to label each hit correctly. In order to have high quality labels, a worker is required to read through guidelines thoroughly and pay close attention to each label. Please pay very close attention to the highlighted words in the guideline. They are about the differences between different categories. We have tried our best to make the guideline clear and simple, and we welcome any feedback and/or clarification about the guideline. 2. Quality control - Labeling quality is critical to us. We look at quality from different perspectives and calculate quality scores based on them. In order to maintain high quality score, a worker is required to follow the guideline. - From our past experience, workers using automatic scripts to do pre-selection tend to degrade their label quality. - When a worker does not meet our bar, then the worker will not be qualified to work on our tasks any more. But, the worker can still continue to work on tasks from other requesters in Mturk. - We are working on a service to provide workers with ongoing feedback with examples of labels they have done incorrectly. So please stay tuned. Hope this clarifies.
