If someone refines their ID and the new taxon is a descendent of the old one, the old taxon should be marked “refined” instead of "withdrawn"

As an example, this observation: https://www.inaturalist.org/observations/48422376

JayAvery IDed something as Vascular Plants. He later IDed it as Ivy-Leaved Bellflower, which is a vascular plant. iNat automatically withdraws his old ID even though it’s accurate, which unhelpfully implies that he took it back or changed his mind, which is counter-intuitive. He didn’t withdraw his ID, he just narrowed it down.

Accurate coarse IDs should not be automatically withdrawn when a more specific ID is added by the same user, unless the newer specific ID doesn’t fall within the previous coarser ID.

Edit:

I don’t think a person should be able to have more than one active ID on an observation at a time.

If the more coarse ID has been replaced by a more specific ID (that fits within the more coarse ID) by the same user, I think the more coarse one should still be crossed out, but to reduce confusion I think it makes more sense for it to say something like “refined” instead of “withdrawn”.

  • If the identifier changes their ID to something incompatible with their previous ID (like changing “fungi” to “sparrowhawk”), “fungi” should be crossed out and marked withdrawn.
  • If the identifier changes their ID to something that fits within their previous coarse ID (like changing “fungi” to “gilled mushrooms”), “fungi” should be crossed out and marked refined.
  • If the identifier changes their ID to something more coarse (like changing “gilled mushrooms” to “fungi”), “gilled mushrooms” should be crossed out and marked withdrawn.

The criteria should be as @JayAvery described it: If the new taxon is a descendent of the old one, the old taxon should be marked “refined”.

I think this makes for a more coherent timeline of IDs on an observation. You’d be able to read that someone managed to improve their ID, and scroll down to find out how.

Edit again: I understand the people who are saying that technically an older coarse ID being replaced by a new more refined one is a withdrawal of the former ID, and I’m not disputing that on a technical, computer-brain level it is being withdrawn. I’m proposing this idea to make the observation’s ID timeline more human-readable. Regardless of how much sense it makes to a computer, if a human scanning a page sees a crossed-out identification they are likely to infer that it has been judged incorrect by someone, and that is not the case when someone has refined their ID.

@tiwane said: “I’ll just say that judging from emails to help@inat, the current system and UI does confuse a small number of people (or at least a small number of people who then write in about it).”

For everyone who writes in about something, there are many more who feel the same way but don’t write in.

Since this would help some people, and wouldn’t cause any problems, this seems like an improvement with no downsides that makes the iNat experience more coherent to more people.

See previous discussion: https://forum.inaturalist.org/t/correct-but-less-specific-identifications-should-not-be-crossed-out-upon-subsequent-improvements/12378/10

@cas4moss are you requesting that each user should be able to have more than 1 active ID on an observation if the subsequent ID was just a refinement? (e.g. active IDs for both “vascular plants” and “ivy-leaved bellflower”) Or just requesting some sort of visual change than the current crossed out/“withdrawn” message?

4 Likes

I don’t view this “withdrawing” as saying that the previous more general identification was wrong, just that the identifier has now provided a more specific one. I think it’s just one of those iNaturalist features to get used to, not a problem I’d like to see changed.

8 Likes

I think having >1 active ID/user isn’t really feasible and would cause issues in other areas.

I also do think that, in a way, the user has withdrawn their initial ID. If you accept that each user only gets one “vote”/ID for each observation, and think of an ID as a user’s best understanding of the organism’s identity, then they have withdrawn/replaced their previous understanding with a new, more accurate one.

Having some kind of visual or text designation that distinguishes a withdrawal from a replace would probably be fine and not too confusing.

But I also agree that while this system may be a little confusing when you see it the first time, it’s not too difficult to get used to and does have a decent logic to it. For me, it would be one of those things where, if there’s a slight clarification that is easy to implement, I’d be for it, but otherwise would think it isn’t a priority given limited iNat staff time.

6 Likes

Personally I enjoy watching an Observation be refined from something general to an RG species ID. The process can be interesting.

But I do see the difference between withdrawing an ID and replacing an ID. A change to the format, such as a message that the ID was changed (not withdrawn) by the user would be helpful, especially when there is a long discussion and the changes are not visible on the same screen. Anyone interested in the evolution (if you will) of the ID can then look for the user’s new ID.

The crossed out format would still be useful, as the replaced ID is no longer part of the Community ID process.

3 Likes

I think the latter? I agree with what @karen5lund said:

But I do see the difference between withdrawing an ID and replacing an ID. A change to the format, such as a message that the ID was changed (not withdrawn) by the user would be helpful, especially when there is a long discussion and the changes are not visible on the same screen. Anyone interested in the evolution (if you will) of the ID can then look for the user’s new ID.

The crossed out format would still be useful, as the replaced ID is no longer part of the Community ID process.

Or maybe even change “withdrawn” to “improved” if the ID that is replaced is still technically correct?

4 Likes

I’d be in favour of a visual change, have the “ID withdrawn” text changed to “ID refined” (or similar) if the new taxon is a descendent of the old one. I agree that allowing more than one ‘active’ ID per user per observation would be confusing to implement though.

13 Likes

“Refined”! Perfect. :D I will edit that into my original post. Thanks!

2 Likes

What is confusing in a different way is - that when I do in fact withdraw my ID, it shows up in the notifications as refined. But I did not refine to whatever, I withdrew my ID.

1 Like

Leaving aside Diana’s experience, I don’t see the problem.

If an observation can have only one active ID per user, then you have to withdraw your old coarser ID to replace it with a new, more precise one. “Replace” means taking something away and putting something different in its place.

If I change “Lepidoptera” to “Vanessa cardui,” it doesn’t say I disagree that the butterfly in the photo is a member of Lepidoptera. That would be annoying. It merely says that I withdrew my first ID because I chose to use my one vote for a better candidate.

BTW I think we all agree that having more than one ID by the same user of the same observation would be chaotic and serve no purpose.

4 Likes

I’ll just say that judging from emails to help@inat, the current system and UI does confuse a small number of people (or at least a small number of people who then write in about it). And I think “withdrawn” makes sense in that I withdrew an ID as my active ID for the observation. That being said, I’d be OK with changing the wording but I wouldn’t consider it a priority.

@cas4moss what do you suggest for a situation where the identifier changes their ID to something more coarse, say from a species to its parent genus?

2 Likes

Or if someone withdraws an ID on one day, and much later puts on a finer ID than they had before. Should the “withdrawn” retroactively change to “refined” (or whatever the next text is)?

2 Likes

I think both cases should be “Withdrawn”. First (tiwane’s) because the person withdraw the species level identification and the next because the first ID WAS withdrawn. The new was then a new ID with no current ID from that person.

2 Likes

I think in that situation I would probably call that “withdrawn”, because it doesn’t meet the criterion @JayAvery specified: “if the new taxon is a descendent of the old one.”

2 Likes

Yes, that sounds right to me. :) Text automatically updating when it meets particular criteria seems possible?

1 Like

I was wondering if this kind of improvement was ever considered. Glad that I found your feature request!

I’d really love to see “withdrawn” replaced with “refined” where a user has started with a coarse but non-conflicting taxon and subsequently chosen a more narrow taxon (e.g. kingdom to family; genus to species within the genus; species to subspecies; etc…)

Such a display change could also allow a user’s initial ID to keep the “Improving” label" when the ID is subsequently refined. Removing that “Improving” acknowledgement seems to dismiss the initial contribution of the user if he or she updates on account of agreement with a refined taxon. The current treatment that just shows “ID Withdrawn” robs the user of that small acknowledgement. I would wonder if that might be keeping some people from refining their IDs, as people generally have a psychological need for acknowledgement, and some more so than others.

Another potential benefit that doesn’t seem to have been mentioned is that it might help to encourage users to guess less (when they are uncertain of the species ID) and encourage users to select broader initial taxa of which they are certain (which assists in allowing more knowledgable identifiers to see it). This way there’s no “penalty” nor negative label. Because let’s face it: for some, “ID Withdrawn” is seen as a negative label.

Recently there was a thread about “rampant guessing of IDs”. Although it is not something that I’m terribly worried about, one user noted how when a user is too specific and wrong, it often creates additional work for others:

My experience agrees with that. I see that happening occasionally (most often with correcting plants mis-IDed as either Oregon Grape or Cascade Oregon Grape).

Another piece of data to take into account is that the vast majority of identification is largely done by a very small number of frequent (often knowledgable) identifiers. According to iNat’s own data, the “top 1000 IDers contributed 70% of all IDs”. iNat should consider whether there’s any incentivizing behaviour that creates less or more work for frequent identifiers.

And since iNaturalist does, in many ways “gamify” citizen science, and since some people are “keeping score” (if only mentally), this would be an improvement that would recognize how the current features might be interacting with those aspects of participation. iNat is essentially using a sticker system, however there isn’t much acknowledgement that they’re doing so. Better to realize, if not acknowledge the type of system, and then try to eliminate any perverse incentives which the present implementation creates.

It’s a small encouragement to users to ID in a circumspect manner simply by not punishing it; that kind of inaction seems a small cost.

1 Like