Problems with wrong suggestions

Welcome to the forum!

2 Likes

Please only do this if you are certain it can’t be identified from the photo, and that you are certain a conversation amongst the community would support that conclusion. Being an expert is not enough to be certain, as there might be characters in a localised population that might be evident to a field worker that had worked closely with that population. It stands to reason that if you don’t know the minds of others in the community, you can’t say with certainty that they don’t know of such characters, so start from a position of assuming they can determine, and then work back from there (ie they appear to be a new inexperienced user that had utilised a CV suggestion, then maybe explicit disagree but tag them and ask why they think species etc)

6 Likes

Actually, I prefer the opposite. I have made the experience that users, which for whatever reason think they are sure it hast to be a certain species, will contact me and I will retract my evaluation if their argument sounds somewhat valid. But for example look at Araniella curcubinita, which cannot be determined in the field due to similar sister species, but look at how many research grate observations there are on iNat of this species. Many people are all too ready to ID to species and them not being experts I understand that there sometimes is not much knowledge of possibilities of getting apparently very prominent species like Araniella curcubinita mixed up.

7 Likes

That’s not the usage I was talking about for the green option. In some cases users want to add an ID to an observation; they know it’s definitely that genus/family/etc, but unsure of the species, so the green option can be a way of still contributing (as one example of its use).

5 Likes

@tiwane

Maybe I’m in the minority on this then, or perhaps just on the wrong platform…

2 Likes

I’d been wondering for a while what the purpose of these prompts are!

I think these prompts are pretty unclear with the current wording.

7 Likes

I support that point, and I am sure I have clicked on the wrong option in the past because of that. Thank you for bringing it up!

2 Likes

Consider adding that species to this post: https://forum.inaturalist.org/t/computer-vision-clean-up-wiki/7281
I do think using the orange button is warranted in situations like these because otherwise they’re impossible to fix.

Here is a feature request for clarifying the wording (along with a lengthy discussion about how it’s meant to be used, with some clarification here): https://forum.inaturalist.org/t/change-wording-used-by-the-system-when-downgrading-an-observation-to-an-higher-level-taxa/3862

4 Likes

What it comes down to is that it’s not totally clear what the green and orange buttons mean, despite the many discussions which led to the current wording, and the discussion @upupa-epops linked to.

4 Likes

What I don’t get is that everyone agrees the wording is confusing, it’s a simple change to reword it, and yet a good year down the track and we are still having the same arguments. This aggravates my OCD far too much… I quit…

I don’t think it is just an issue of wording (but admittedly the problem with question vs. answer does not help). As far as I understood you don’t feel anyone has the right to downgrade an observation belonging to the observer no matter how certain he|she is? I feel, as soon as you upload the observation to a (scientific) database used by many, it is not anymore just “your” observation. It has a higher purpose now then just your species count (hobby observation would be an alternative for that I guess). It therefore should be as reliable as possible, meaning that corrections should also be made to the best of knowledge. Pushing it back from research level if you think it is justified puts it back in the pool for others to evaluate, even if your idea turns out to be wrong. Leaving a quite likely wrong ID at RG does not give you that option but instead might lead more nature enthusiasts on the wrong track.
So yeah, while phrasing is an issue, there is definately also an underlying disagreement on how to deal with these kind of situations.

2 Likes

I get either of the angles here:

  • A: “I disagree with the ID that this is species x. It simply isn’t x.”
  • B: “I disagree that an ID of species x is possible. It simply can’t be said to be x.”

But one of the issues is that we already have another tool for B in the Data Quality Assessment annotation “Based on the evidence, can the Community Taxon still be confirmed or improved?”–and yet when the community taxon is already incorrectly granular, that tool is no use.

(This is probably already expressed in the feature request thread, so if so my apologies.)

I’m at the point where I would support totally removing the buttons. Just allow people to add identifications period. “Fixing” identifications is done the same way as they get established, by the consensus of the entries.

You know that a species level ID is not possible, enter a genus (or whatever level) ID, write a comment to get the attention of the species level identifiers, recruit others to add a genus ID etc.

Forcing users to say this it not x to fix something (ie the orange button) when what is really true is it may not be x is not an appropriate model.

Right now trying to fix potentially wrong RG records means users either effectively get no votes in the algorithm (the green button) or a super veto vote (the orange button), they should get 1 vote.

3 Likes

Maybe instead of just quitting, I delete as well…that’s another argument that could do with a rehash…

Yeah, there as a case a while back where someone kept overruling IDs on a bird bone repeatedly insisting it was impossible for anyone to identify to species and tanking everyone’s IDs.

Of course, it turned out that it absolutely was identifiable to species, and it eventually was, but that person’s actions didn’t act as a positive contribution.

2 Likes

yikes! thats a long thread.
Is there a summary somewhere?
I scanned it but…

I think if it remains in place it needs to use very simple wording.
Any of the options with long clauses become quickly muddy to me.
I imagine even more so for non native English speakers.

The more complex clauses or examples could be revealed through some sort of “more info” interaction.

Wouldn’t this whole thing be resolved though if iNaturalist just enabled expert users with override powers like iSpot ? I guess thats all discussed in another long thread or two :)

i think this thread has sort of gone off the rails, and a lot of the discussion seems to be built on a false premise – that an identifier came over to this observation and made a higher-level id with a disagreement.

i did a little test over at: https://www.inaturalist.org/observations/44493954. as far as i can tell, if a user makes a higher-level id with disagreement, the disagreement sticks, even if the previous species-level id is deleted and added anew. and there is simply no evidence that in the observation noted in the original post that there was a higher-level id made with a disagreement.

so i think that, along with a mistaken disagreement over the adult annotation, suggests to me that any problems noted in the original post were probably due to lack of familiarity with how to use the system more than anything.

i suggest we close this thread, and let any discussions about the green and yellow buttons continue on existing threads noted above.

7 Likes

At the risk of getting another thread off the rails or even locked, that topic has an even longer and more colourful debate history around the site. In short, the site have made it clear they will not be implementing an iSpot style system. Any system the site did do (and it does not seem a priority) would solely be based on the amount of interactions on the site.

3 Likes

As @upupa-epops said, this can happen when the observer did not actually enter an identification, but instead there is a placeholder, and then the observation is labelled “Unknown” initially. It is very common for people to add general identifications to “Unknown” observations. Ideally, they would copy the placeholder into a comment, but the effect would still be that the name has changed from the name in the placeholder. It’s not clear to me that this is the situation you are talking about, but if it is, the solution is to make sure that you have an actual identification of your own on the observation, rather than simply a placeholder. If there is a placeholder, you need to enter an actual identification.

2 Likes

Thanks all! Looks like this question has been answered.