Add functionality that actively encourages cleanup of IDs (e.g., maverick, flagged IDs)

[Note: the following has been edited in response to comments below]

When a user logs onto inaturalist, and if the user has identifications (or annotations?) that appear likely to be incorrect and not previously addressed by the user, have a notification circle with a number in it indicating the number of items at issue. When the user hovers over the circle, it will display a message something to the effect of “you have possible misidentifications we would like you to review.” When the user clicks on the circle, it will take them to an Identify window where they can examine these identifications. Possible misidentifications could be indicated by maverick status, sub-maverick status (two other IDs following and conflicting with the user’s ID), observations somehow flagged as very unlikely correctly IDed (due to wrong range for example). New functionality to flag observations is also suggested (but would need to to be highly restricted to certain items to prevent abuse). While this functionality would be primarily used to fix a user’s incorrect IDs, it would also encourage those with what they believe to be correct maverick or sub-maverick IDs to try to convince the other identifiers to change their IDs or solicit help from other identifiers to override the other IDs.

Dealing with large numbers of incorrect IDs is a huge problem for identifiers and I think the above would be helpful in reducing that burden.

maverick doesn’t necessarily mean wrong. i think i would be annoyed by a glaring banner if i had intentional maverick IDs. i think the kind of flags you’re talking about – since they would basically be tasks or reminders assigned by others – could be easily abused, though tasks created by me for me might be useful for some.

i would not mind having easy access to several pre-defined variants of the Explore or Identify screens – or some other new screen – that included a variant to find my maverick IDs though.

13 Likes

I’d also point out that once an ID has maverick status, that means that the community has already ID’d the observation to something else, so the maverick ID isn’t actually doing any harm. It may still be useful for the user to learn from, but glaring banners or similar are more likely to provoke irritation than learning.

Finally, iNat’s taxon page shows ‘similar species’ to let you know what other species can be confused with that one. It’s a very useful resource for those learning to ID, and I believe that it uses maverick observations as an input, though I may be wrong on that point. So for that purpose, cleaning up mavericks would actually be detrimental.

7 Likes

@rowan_m - A maverick ID can be maverick but still limit the rank, no?
Leaving them there to support similar species seems a bit counterproductive if so. This could anyhow be trained on historical maverick IDs … leaving them in place doesn’t seem essential one would imagine.

I like the idea of being able to find your own maverick IDs to review through the Identify portal as @pisum suggests. Could also imagine the new notifications update might help grapple with this issue.

3 Likes

We have a way to view mavericks: https://www.inaturalist.org/identifications?user_id=mertensia&category=maverick , but replace “mertensia” with your iNaturalist username. They’re not inherently wrong. Mine exist there because I believe those plants are identified incorrectly.

18 Likes

On that note, such a warning is likely to have a chilling effect on experts looking to right some Research Grade wrongs, especially if there is a taxon with a large number of those due to bad AI advice.

8 Likes

I follow my notifications. If there is disagreement, I will … agree, or withdraw, or leave my maverick because you haven’t convinced me yet.

11 Likes

Wow thanks for that link…had some cleaning up to do apparently :)

7 Likes

I would LOVE this functionality. There are so many observations that get stuck at a higher taxonomic level that could be addressed by this. Being able to Flag bad identifications would solve this, but could also be made reversible to avoid the system being abused. For example, if I Flag the ID of a user but they disagree with that, they could check a box that votes against that Flag… similar to the little boxes that lurk below asking whether the data is correct or the ID can be improved.

3 Likes

isn’t there already a way to ‘flag’ an ID you feel is wrong? Add your own ID to whatever taxonomic level you feel comfortable, and if you are confident in disagreement, explicitly disagree with the other ID (if relevant). I get that this gets a little fudgy when you aren’t easily able to note to which taxonomic level you disagree… but I feel like adding another way to disagree with IDs would just be more confusing/complicated. Also there are other options - marking ‘additional ID needed’ or tagging in others you think know that species…

8 Likes

I have to disagree with you on this one: I think this would deeply put off users, especially those who are new and might construe it as a personal attack or those who are in most need of it and will probably never see it because they don’t use the platform anymore. The Maverick system works the way it is supposed it and as was said already means an observation is research grade. It would speed up Research Grade observations to a small extent but I feel that it’s more likely to keep people from using iNaturalist.

The cost/benefit ratio of having a few extra data points at a faster rate isn’t worth losing a contributor who may provide hundreds or thousands of valuable posts in the future by putting them off. We’ve been working at this for a couple weeks and have recruited new people to assist and they have been great at overriding incorrect ID’s if they feel they are wrong. I think this is the proper way to go as it does not overtly call out a user in a way that might feel insulting.

I’ve encouraged people NOT to change an ID in a few cases if I can’t convince them if they don’t agree, especially if they are new and more likely to be put off. I think building a good relationship between new users and identifiers is key for the success of iNaturalist and try take extra time if they are new to explain an ID if they have it incorrect.

My two (or ten) cents on this one.

13 Likes

This is more of an issue for the vast backlog of observations that have gotten taxonomically stuck due to inactive users. I completely agree about the importance of communicating why an ID is suspect, but, if there’s no one to communicate that to, then what’s the solution? Currently, it’s finding more users by tagging them and hoping they respond, but this isn’t viable for taxa that don’t see much engagement on here.

Hypothetical example: an observation from 2015 that has 4 IDs, all objectively erroneous, from users who are no longer active. Now say an expert in this group wishes to correct this, but their Maverick ID doesn’t even un-Research Grade this, let alone move it to the appropriate taxon. There aren’t any other regular users who work with this hypothetical taxon, so nobody else to tag (of course, it would take 12 new IDs to actually establish a new commmunity ID… yikes).

Consider how much time is being spent attempting to clean up an observation like that. As someone who has done so countless times in a similar group that I work on (Indo-Pacific corals), a better solution needs to exist. This problem will only get worse over time.

9 Likes

@Dan_Johnson and I just spent 2 weeks cleaning up a couple thousand cicada ID’s which is what prompted this and other related posts. I’ve identified nearly 2000 to species during this time and Dan easily 4 times that many. Like I said, it’s my two cents after running into the problem several hundred times during this period.

6 Likes

At one point I had over 40 maverick IDs to my name. I even wrote a journal post about it (https://www.inaturalist.org/journal/cabintom/28051-tom-the-maverick). The fact is, I’ve probably made hundreds of IDs that were initially maverick, but they’ve mostly all been sorted out. I’m now down to only 5, with just a couple of instances where I changed or withdrew the maverick ID.

It would be terribly annoying to have something “actively encouraging” me to clean-up an ID when in fact the ID I’ve given is correct.

8 Likes

Thanks everyone for the comments on this suggestion. Sounds like it’s far from a winning suggestion at this point, but perhaps with some refinement, we can get some agreement on some sort of functionality in this direction. Just as a background, the idea for this request was motivated by doing some 8000 or so cicada IDs over the past three weeks or so. I saw huge numbers of really bad IDs, such as “Genus Cicada,” a non-American genus applied to American species. Many of these IDs were made by people who continue to visit iNaturalist at least occasionally. Many of them I assume were probably unaware that they had outstanding mavericks out there that they would be happy to correct if they were aware of them.

Concerning the various comments asserting that if an ID achieves “maverick” status, it’s no longer a problem, because the community ID overrides it: It actually is still a big problem for advanced identifier, because an extremely useful tool is the ident_taxon_id command line option. This option allows listing all observations identified as a taxon. This allows finding for example an observation that was identified as a species by an expert in disagreement with another identification and thus the community ID is not what the expert IDed it as. When used, it will list everything having one of those IDs, including all the incorrect mavericks. If those mavericks were cleaned out, it would reduce the number of observations the identifier has to weed through.

Great point about mavericks being sometimes correct mavericks. I think some sort of notification system for both the good and bad would be useful. For the bad one’s, it would be obviously useful for those that made them to consider withdrawing them. For the good, it might be handy to have some sort of reminder to help motivate the identifier to round up help and perhaps contact the conflicting IDers to help get the maverick established as the correct ID.

Rather that a “Glaring colored bannered,” how about a more gentle, more subtle little circle on the banner which when hovered over has the message “you have observations with potential issues” (or something like that). When clicked on, it could provide them with a listing of observations to look at, perhaps in Identify.

I also imagine that included in a list of IDs to be addressed would be those that that have two conflicting IDs (or maybe even one), especially if the user has never checked that observation after those other IDs were made.

Perhaps there would also need to a button were the user can acknowledge seeing the items but chooses to do nothing for whatever reason. So next time they go to their “issues” list, they wouldn’t see them again.

As for flags, I would expect this would be restricted to glaring errors, such as geographically not occurring anywhere near the observation.

I may revise the writing of my initial suggestion following comments.

5 Likes

As someone who has a few maverick IDs that I feel are correct, I also disagree in principle, and I think the way that @mertensia suggests to look for mavericks is pretty useful.

5 Likes

Did you see I addressed the subject of good maverick IDs in my response?

1 Like

I think a button somewhere on the homepage for quickly viewing mavericks, as well as an explanation for what they are would be useful. The cicada identification sounds extremely frustrating!

3 Likes

Or due to users not looking at all the pics.
Like this one. Less than half the photos (only 3 of 7) feature the butterfly that this observation was RG for until I came & dissented. I will go down with my Maverick ship on that one*.

*Unless the photos are changed. :slightly_smiling_face:

4 Likes

I feel like I don’t see all the facets of this discussion. Still, maybe this is a useful idea: The profile page for every user could show some “statistics” on their observations, e.g. the number of Mavericks. Hovering over the description (Mavericks) could bring up an explanation of what it means and how the user can help make iNat better. Clicking the number itself would bring up a list of these observations by that user.

On the subject of “false community IDs” of inactive users, one could define a procedure by which the observation would be made casual. The user would be informed and a time frame set. I’m also thinking of observations where it’s unclear what the subject of the photo(s) is.

4 Likes