I totally agree. I think changes could be made to the system to correct these problems, and are preferable to changes asking people to alter their behavior voluntarily. I wonder if changes could be made to the collection project parameters that would help without affecting other important collection projects, since collection projects are the vehicle for these events.
The pinnacle of bad student IDs: https://www.inaturalist.org/observations/5890862
(Anyone feel free to help, if I did the math right we still need 8 more IDs to override all the students.)
wow, thatâs ⌠i mean i donât think thereâs ever been a plant in the history of iNat with half that many IDs added. Wow. Also obviously a muskrat :)
wow, is right! Iâm thinking an even better place for non-Penang-related student/duress/contest user discussions might be here though: https://inaturalist.trydiscourse.com/t/duress-and-contest-users-discussion-thread/422/3 that way we can all find such examples (I added a muskrat âagree!â for this one, too) in a topic heading that will be a little more universally relevant and easier to search. After this particular challenge gets cleaned up these conversations will still be very useful and may get lost or forgotten here.
Hey everyone,
Just wanted to give you an update about the Penang project. @carrieseltzer and I chatted with the organizers via video for about an hour yesterday, California time. They were incredibly nice, gracious, and apologetic about what they describe as the âhavokâ this project created, the scale of which took them totally by surprise. Some quick takeaways:
-
the purpose of this project was to encourage the kids to get outside and take notice of the living things around them, and I do believe this was accomplished, although it was detrimental to those of us who were wading through the observations. I do think thatâs important to remember. They said the kids had a lot of fun and were seeing many things theyâd never noticed before.
-
they disseminated documents, including how-tos that told the kids to take photos of wild organisms, to not use computer vision suggestions and instead make high-level IDs, etc. but apparently this message was lost and/or misunderstood by some students. Also some of the teachers who were not directly involved apparently gave advice such as âgo photograph vegetables in markets.â So kind of like a game of telephone, the message became garbled. Weâll be reviewing those documents so we can provide feedback.
-
while there was a tangible prize involved, the organizers believe most of the students werenât aware of it and were instead motivated less by a tangible prize and more by wanting to just do well (we were assured no grades were involved) and due to competition between the two school districts in the area. Iâm paraphrasing but the organizers told us that in the area it is very important to do what teachers ask you to do and to excel.
-
moving forward with the actual City Nature Challenge, they wonât be promoting it in secondary schools but more to the general public and to college students, especially those studying biology. Theyâre also going to try and recruit biology students and other local experts to ID observations in the project as practice for CNC and to help clean up the data from this event. The upside to all of this is that we have a lot of data now and, if properly identified, can help iNat a) train its vision model on species more prevalent in Asia so itâs more accurate there and b) know which species are likely to be captive/cultivated in Penang.
My two my main conclusions after this conversation are:
-
There are things iNat can do with code, design, and messaging to mitigate similar events (I doubt we can prevent them totally), and weâre looking into them. If you have ideas, please post them in General or make a concrete feature request.
-
It is incredibly important to always keep the âassume people mean wellâ Community Guideline in mind. I know some frustrations were vented by iNat users, and I was frustrated as well, but if there were any insults or put-downs directed at the organizers or participants, they were unwarranted. I also apologized on the communityâs behalf if they received any untoward messages and the organizers were very gracious. The vast majority of folks involved in the event acted in good faith and the organizers were quite willing and open to discussing the issues with us.
Thanks Tony, that helps a lot and makes it feel like this event can be productive in the end.
Do you happen to know why they set the Quality Grade settings to Needs ID and Casual rather than Research Grade and Needs ID?
We didnât get into that but I imagine itâs due to unfamiliarity with the platform.
Thanks, Tony. Looking over the aftermath, Iâve seen a bunch of the students going back and going to top-level IDs, like âPlantâ, which is pretty constructive. (Some have switched off community ID, which is less constructive, but oh well.)
Having seen similar experience play out on Wikipedia, I know it can be very frustrating for both teachers and staffers caught in the middle: you have students on one hand, and community on the other, whom youâre sort of answerable for but have very little direct ability to control, becoming increasingly aggravated as cultures collide.
Do you know if the organizers saw or made use of the Teacherâs Guide before this deployed? I started to write down some of my thoughts from experience here and on Wikipedia and found that the guide hit the major points so well I didnât see the use in it! Of course, just because itâs there doesnât mean people will read it or use it, but I think itâs something Iâd try to push to organizers of all major events, whether or not theyâre strictly speaking âteachersâ.
Many thanks to Tony and Carrie, as well as all of the curators that devoted some time (and headaches) to managing the influx of observations.
Hear hear!
This seems like a really intelligent and practical idea for preventing things like the âPenang incidentâ from happening again, especially during the upcoming City Nature Challenge. For CNC this year, I will be in town but out making observations during the observation days of CNC, and will be trying to ID NYC observations during the identification days of CNC, so I wonât have time to spend struggling with attempting to clean up an avalanche of misleading stuff from parts of the world where the AI does not yet know what is what.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.