Allow curators to view an account's DQA votes, and allow for flagging and hiding of DQA votes and Annotations

[graysquirrel]

Platform(s) All

URLs N/A

Description of need:
This is a two-in-one request because the need and functionality for both is very similar, but if it is better to split them, I can do that.

At the basic level, the issue is that users can make significant changes to other people’s observations, which are beyond the power of curators to alter.

There has been a steady increase in instances of users abusing the DQA and annotation systems, and there is currently very little that can be done to address the problem aside from pestering staff to manually delete them.

Although both annotations and DQA votes can be outvoted, this involves recruiting the time and attention of a large number of other users, and thus a single bad actor can monopolize a disproportionate amount of effort that could be better spent elsewhere.

Given that we can now hide bad-faith IDs instead of having to recruit others to outvote them, it seems logical that a similar system could and should exist for these as well.

In the case of DQA votes, there does not seem to be a way for curators to see the votes a particular user has placed, so it’s not possible to even monitor and countervote.

Some recent examples of DQA abuses:

• An account who uploaded clearly captive animals, marked them as “wild”, and, when countervoted, created 8+ alt accounts exclusively to add more DQA votes in their favor. I was able to suspend all the accounts, but the votes stayed until removed by staff.
• Several accounts who - for reasons I cannot discern - have decided to spend their time marking zoo animals as “wild”. Since this is across multiple taxa and regions, it’s difficult to find the particular observations that have been interfered with.
• Retaliatory DQA votes - users who retaliate against others for IDs, flags, or curatorial actions they dislike, by going to the other user’s observations, choosing a few at random, and downvoting their DQA. Some of my own observations were recently targeted in this way.

As for annotations, it seems like some teachers are now requiring their students to add a certain number of annotations for assignments.

As you might guess, this results in observations like this:

Although these are easier to find, and can be downvoted, it is frustrating when blatant abuse of the system prevents the correct annotations from being added on otherwise useful observations.

Feature request details:
Basically, I would like the ability for curators to view a user’s DQA activity in a similar manner to how we can view a list of a user’s comments. I do think this part should be restricted to curators only.

I would also like anyone to be able to click on a DQA vote or an annotation and submit a flag for abuse of the system.

Finally, I would like curators to be able to take action and hide or remove flagged DQA votes or annotations.

(And while we’re at it, the ability to see timestamps for when a vote was cast could be very helpful as well!)

[insectobserver123]

I was going to request this if you didn’t, I think I was investigating some of the same DQA abuse incidents you were.

The lack of a way to even see all DQA votes of a user makes it very difficult to determine if a pattern of DQA misuse exists, or what the pattern is. This is one of the only suspendable offenses where curators are unable to effectively track a user’s pattern of behavior

It also makes it impossible to find and outvote all the DQA abuse even after suspending someone for DQA misuse

[michaelpirrello]

Oh, I’ve been running across this quite a bit as I remove inaccurately applied EoP: Track annotations without understanding the “why” behind users selecting all the annotations. Has anyone posted about this to the Educators section of the Forum?

[Jewlsinnature]

I think the timestamp is a very important part of this request.

• helps to get an idea of the voting timing in relation to the ID’s
• helps to see the order of the voting between different voters
• helps to see if the vote was casted prior additional information about the observation was provided (like clarifications regarding wild/captive)

I like this feature request but definitely think only curators and staff should have access to this info. I actually thought that this was something curators already had access to!

[cthawley]

I agree with parts of this request, including the usefulness of curators viewing an account’s DQA votes and their timestamps. I also agree that DQA/Annotation abuse are really annoying.

However, I do not support the ability to flag and hide DQA votes and Annotations myself for several reasons.

Hiding is the functional equivalent of deleting for both DQA votes and Annotations. It would be very hard to make this process transparent to most users given the current layout of the site/app (ie, it would not be obvious to users what had happened if a vote/annotation was hidden).

Votes and annotations can also be quite subjective, and users can disagree on them - there often would not be as much evidence to say that one vote/annotation was vandalism/made in bad faith vs. just incorrect as compared to an ID or joke photo. Applying guidelines to DQA votes and annotations consistently would be challenging.

Also this would just be a major expansion of curatorial powers. It would increase the number of flags, increase the number of actions needed, and increase potential for abuse.

Additionally, the DQA and Annotations really don’t have the potential for much of the “badness” that spurred the creation of hiding functionality for IDs, comments, and pictures. As I understand it, the main reason this functionality was created was to deal with serious insults, suspendable offenses (racism, sexism, anti-LGBTQ), etc. While any intentionally incorrect data entry could be grounds for suspension, fake DQA and Annotations just don’t have the repercussions in terms of these serious violations of guidelines that IDs, comments, and pictures do. Adding fake votes can certainly be a form of harassment (and I’ve been targeted with intentionally incorrect DQA votes myself), but I don’t think this causes the same level of issues as some of the more nasty hidden comments/IDs/pics out there. I think the need for this function with DQAs/annotations is lower.

One other point is that DQAs are much more easily counterweighted than IDs due to the majority rules nature of the voting. So curators can often fix an issue with one vote (which would be faster than hiding/deleting). Annotations are their own problem (since they can’t be changed, just downvoted). I support changing that aspect of Annotation function (separate thing!), but it does make me worry that hiding Annotations would just lead to a bunch of flags asking curators to fix wrong annotations since it would be possible, not hiding/deleting just those made in bad faith.

Additionally, I think adding this functionality would just make both the website and apps very complex - the functionality would be quite difficult to find given the current setup of DQA and annotations.

[graysquirrel]

@cthawley You make some good points. I am remembering some recent situations where a couple of curators had strongly-held disagreements over what “wild” meant - so this could certainly add potential messiness to those sort of issues.

@cthawley

One other point is that DQAs are much more easily counterweighted than IDs due to the majority rules nature of the voting.

One of the problems I’ve been encountering is that people keep making new accounts, and it doesn’t matter if they get suspended almost immediately because the DQA vote stays after suspension.

Perhaps some kind of option where a user who is suspended specifically for abuse of these features has their votes “suspended” (hidden, basically) as well? With an option to unsuspend votes along with the account, in case of unwarranted suspensions.

@cthawley

the DQA and Annotations really don’t have the potential for much of the “badness” that spurred the creation of hiding functionality for IDs, comments, and pictures.

I agree it doesn’t have the potential to be as in-the-face and shocking, but it can still be a pretty big avenue of harassment. Especially for users who may not check their observation statuses often, and never realize all their stuff has effectively been hidden from the community.

[cthawley]

I do agree that bad faith DQA votes are a problem, but I think that this will be addressed at least somewhat by the feature request to make the votes viewable. If they are viewable, the votes are pretty easily/quickly counteracted by one (or two) curators by countervoting. In bulk cases, I would guess that curators might not even bother with the extra clicks to hide and add a justification if countervoting has essentially the same result (negating the DQA vote). I currently see lots of cases where curators don’t hide IDs that they could and just add a disagreeing ID instead presumably because it is much quicker. I think, if the request is considered, it could make sense to implement the DQA vote viewing feature first and see how well that allows the community to address DQA vote issues before considering hiding functionality.

On a side note, this overall issue also intersects with there not being notifications for DQA votes. If DQA votes ever do generate notifications, this will also surface bad faith DQA voting more quickly (though it will also make it more annoying for people who get a lot of notifications).

I agree that some feature to delete/hide in bulk bad faith DQA votes from suspended users seems reasonable on one level. But I also think this would be challenging to implement and a bit of a change in how iNat generally handles user content and probably require staff involvement. It seems like a big change for an issue that is comparatively rare.

[dysm]

I don’t feel great about being able to hide votes either. I think outvoting them is better and often works, so I’d prefer to keep it that way.

But I think notifications for votes, viewing a history of votes, seeing the time a vote was made and especially being able to flag them would all be extremely helpful, and probably would make it so the hiding function wouldn’t be needed as much as people feel like it might be. Because right now, I think any additional ability to deal with DQA abuse would be helpful, so maybe if we get the ball rolling and start little, bigger things like hiding or removing other people’s votes won’t seem so pressing.

[charlie]

@graysquirrel

• An account who uploaded clearly captive animals, marked them as “wild”, and, when countervoted, created 8+ alt accounts exclusively to add more DQA votes in their favor. I was able to suspend all the accounts, but the votes stayed until removed by staff.
• Several accounts who - for reasons I cannot discern - have decided to spend their time marking zoo animals as “wild”. Since this is across multiple taxa and regions, it’s difficult to find the particular observations that have been interfered with.
• Retaliatory DQA votes - users who retaliate against others for IDs, flags, or curatorial actions they dislike, by going to the other user’s observations, choosing a few at random, and downvoting their DQA. Some of my own observations were recently targeted in this way.

Arent all of these already a violation of the iNat rules? I am skeptical that trying to set up flagging for DQA is the answer for this, though maybe it makes sense. I definitely don’t think curators should be able to hide DQA they don’t agree with! Some of the bad actors here are curators. I would support that for a smaller subset of mods but not all of the curators many of whom are more taxonomists than mods. I do think being able to see DQA votes of a user and flag from the end of the user’s page if there’s problematic behavior… is a good idea.

I pretty much agree with all of @cthawley 's posts. And also @dysm ’ s post.

Sock puppeting is pretty consistent problem but it’s hard to ban all of the sock puppets and admin have been fairly lenient on sock puppeting and in at least one case letting someone return after doing a LOT of it.

[wildskyflower]

Fully agree; heavy support to being able to view votes in some central portal and some way to flag them directly; if an account has done nothing but add bad DQA votes on other people content, there is not currently really any way to flag the users behavior at all, because you can only flag profile pages for ‘spam’. And right now if you find just 1 or 2 questionable votes there is no way to check if it is a pattern or not.

Some other reasons being able to hide votes for DQA is less important are:

1.) It is simple majority rather than 2/3+1, so it takes far less votes to outvote a bad DQA
2.) Part of the problem with outvoting bad-faith IDs is that especially in expert taxa it can be impossible to find enough people to outvote more than 1 or 2 bad-faith IDs to get an observation to RG unless someone blindly agrees. However you rarely need to be any kind of expert to decide if there is ‘evidence of an organism’ or not.

[chuuuuung]

@graysquirrel

Perhaps some kind of option where a user who is suspended specifically for abuse of these features has their votes “suspended” (hidden, basically) as well? With an option to unsuspend votes along with the account, in case of unwarranted suspensions.

A feature like that certainly would have been useful for when you were dealing with that user with 8+ sockpuppets. Tying the DQA vote/annotation hiding to a more drastic curator action (i.e. a suspension for bad faith DQA or annotations or sockpuppeting) should address the reservations raised by others in the thread.

And as you mentioned in your original post, a mechanism for flagging DQA/annotation abuse also seems prudent.

[Italopithecus]

Fortunately in my country this bad behaviour is not frequent and almost only limited to few cases of duress users who possibly had to fulfil a task for the university (maybe a digital herbarium).

I support your request.

I only add that, when possible, it would be better to reach the teacher to explain that forcing students to use iNat is one of the worst way to make them know its existance. It is pretty obvious that some of them will end up being in a hurry to make the required number of observations with everything they will find in a small area in the shortest time. As regards, I am afraid that for some teachers iNat could be a way to save time instead of having to correct the specimens of a herbarium.

[tiwane]

I think I’m for this, but it would take some design work to add flagging options and date/time of vote.

[charlie]

can we please get some differentiation between taxonomic curators and ‘referee’ mods set up before we expand the mod privileges of curators? It doesn’t make sense to just give taxonomists control over behavior moderation and it’s something that really needs to be addressed before mod power is increased.

[insectobserver123]

Don’t staff asses ones judgment in moderation issues prior to making them a curator?

[silversea_starsong]

Oh, suspending accounts doesn’t remove their DQA votes? That seems like an oversight that should be addressed regardless.

As for DQA votes, I have continuous gripes with folks who mark some of my wild observations “captive”, and while I would enjoy more power to contest these, I feel in the end it’s probably more likely to be equally abused by curators. If a curator can “undo” or remove DQA votes, then you have the opposite scenario where folks who want their observation to be able to circumvent any checks and balances, which is also malicious.

My take is that the site iNaturalist vote (for wild/captive) should be disabled if someone else has voted manually, because it means a single person can flip your observation to casual. This makes it the only DQA check that can be overturned by a single person in this way.

Otherwise, I would like to suggest that at the least, curators cannot overturn DQA votes on their own observations, as some manner of avoiding abuse from that side of the page.

[charlie]

@insectobserver123

Don’t staff asses ones judgment in moderation issues prior to making them a curator?

maybe they do now, but they didn’t used to, and there are tons of people who do taxonomy who probably shouldn’t be doing moderation and definitely sometimes misuse their power. Thre’s no inherent overlap between being a taxonomist and being a good moderator so there’s no reason to have them combined. It’s like saying all airplane pilots must also be plumbers.

@silversea_starsong

My take is that the site iNaturalist vote (for wild/captive) should be disabled if someone else has voted manually, because it means a single person can flip your observation to casual. This makes it the only DQA check that can be overturned by a single person in this way.

Otherwise, I would like to suggest that at the least, curators cannot overturn DQA votes on their own observations, as some manner of avoiding abuse from that side of the page.

Agree

[insectobserver123]

@charlie

there are tons of people who do taxonomy who probably shouldn’t be doing moderation and definitely sometimes misuse their power.

In my experience the taxonomy people try to stay out of moderation

@charlie

Thre’s no inherent overlap between being a taxonomist and being a good moderator

I disagree with this, taxonomy on iNat often requires discussion and compromise on taxon flags, which, like moderation, requires good interpersonal skills. Taxonomy flags also sometimes devolve into disputes requiring moderation

Editing common names is part of taxonomy curation, but also part of moderation, since trolls can add bogus names

[charlie]

well i won’t comment further beyond this, but your experiences sound different from mine. I’m glad the taxonomic curators have displayed good interpersonal skills towards you.

[insectobserver123]

I’m not saying all taxonomic curators have good interpersonal skills, but that you need good interpersonal skills to be a good taxon curator, and that the taxon curators who don’t have good interpersonal skills usually stay out of moderation