Make 'Needs ID' Independent of BOTH 'Research Grade' AND 'Casual' Statuses

There is an existing feature request for the idea of separating whether something “Needs ID” from whether it’s “Casual” – see https://forum.inaturalist.org/t/make-captive-cultivated-not-automatically-no-id-needed/112. However, this feature request is for more than that.

I propose that the question of whether an observation “Needs ID” is separated from the question of whether the observation is “Research Grade” OR “Casual”. An observation should be “Needs ID” until:

  • the community agrees on an ID of a taxon with no infrataxa
  • OR the community agrees that the community taxon is as good as it can be

Once one of the above criteria are met, the observation would no longer be considered “Needs ID”. If this condition must be labeled, it could be called something along the lines of “ID Verified”, “Identified”, etc.

Whether an observation is separately “Casual”, “Research Grade”, or “Neither” would otherwise follow all of the same criteria as they do today.

Examples:

  1. “Species ABC” has two accepted subspecies. When the community agrees an observation is of “Species ABC” and the observation otherwise meets the criteria for “Research Grade”, the observation becomes “Research Grade” but remains “Needs ID” until (a) the community agrees the ID cannot be improved, or (b) the community agrees the observation is of “Subspecies XYZ”.

  2. When the community agrees an observation is of “Genus SquareCircleTriangle” and the observation is of a captive/cultivated organism, the observation becomes “Casual” but remains “Needs ID” until (a) the community agrees the ID cannot be improved, or (b) the community agrees the observation is of “Species XYZ” (or “Subspecies XYZ” in some cases).

This is an improvement over the current design in two ways:

  • Captive/cultivated organisms that can be more granularly identified can be located and identified easily by those who wish to spend time and energy on them.
  • Wild organisms at the rank of species that could be identified more granularly can be located and identified easily by those who wish to spend time and energy on them.

EDIT: One thing to think about with this is: in such a system, does the threshold for when an observation automatically flips to “Research Grade” stay at the rank of species? Or does the threshold move to a higher rank? I’d previously worded my examples to suggest an intrinsic answer to that question (move to a higher rank), but I’ve since edited them to decouple my proposal from that question. It’s not directly relevant and a separate conversation in its own right. This proposal could handle either way though.

EDIT 2: See https://forum.inaturalist.org/t/make-needs-id-independent-of-both-research-grade-and-casual-statuses/52795/23?u=regnierda for a comparison of how this would impact the logic used to decide what is “Casual”, “Needs ID” (the thing that is today called that), and “Research Grade” (spoiler alert: It doesn’t at all), as well as what the logic to determine the newly proposed indicator could/would be.

EDIT 3: See https://forum.inaturalist.org/t/make-needs-id-independent-of-both-research-grade-and-casual-statuses/52795/29?u=regnierda for a more visual mockup of the logic and what this proposal might look like in the user interface.

In your proposed system, it would be important for me to be able to distinguish observations that have one ID vs. those with multiple IDs.

I think this one might cause a riot lol

9 Likes

Agreed on the riot.

I definitely prefer the current system for this aspect where observations to go to RG (and not remain at Needs ID) when they reach species level, regardless of the presence of any infrataxa.

This aspect of the proposed system would lead to a lot more IDing work and a larger Needs ID pool. IDers would need to be familiar with infrataxa(whether they even existed for a taxon, and, if so, how to ID them) to take an observation out of Needs ID. Many IDers would not ID to that level either on principle or because they don’t have the expertise.

This aspect would also likely:

  • Increase the incidence of users without strong expertise picking/agreeing to an infrataxon just to get the observation out of Needs ID, decreasing the accuracy of IDs.
  • It would also likely increase the usage of the “No” check box for CID which is more work and can cause issues/confusion.
15 Likes

I went back and cleaned up my original post as it was written late at night and I kind of rushed through editing it for the sake of going to bed. Hopefully it now more clearly conveys the nuances (or doesn’t convey unintended nuances) related to this proposal.

I’m not sure what you mean by this. Could you elaborate? As proposed, I don’t imagine the ability to distinguish between observations with one ID vs. multiple would be impacted at all.

Or are you saying that there’s currently a lack of such an ability and in the proposed system you would need that to be an ability? If so, can you elaborate on why the need for that would increase?

Can you elaborate on this? Are people opposed to both themselves and others ID’ing things beyond species?

In the proposed system, observations would continue to go to Research Grade when they reach species level. In fact they could even potentially begin to go to Research Grade at even higher ranks than that (if so desired). Remember, this proposal separates “Needs ID” from both “Casual” and “Research Grade”.

It would indeed increase the quantity of observations in the “Needs ID” pool, but why would anyone need to change their behavior? If one doesn’t want to learn infrataxa, then they don’t have to. They could mark observations of species with infraspecies as reviewed and move on, same as today (just at a different rank). As long as there’s community agreement, it still goes to GBIF.

I’d argue that it’s the current system that incentivizes people to pick/agree to taxa just to get the shiny green “Research Grade” badge, or to misuse the Community ID DQA question. But both systems, today’s and this proposed one, both have to deal with the same problem either way. I don’t see how this would make that worse if what today is one dimension is separated into two.

2 Likes

I am sure there are some cases where below-species ID matters, but I suspect that this varies a lot by taxon and what question(s) people are interested in. In most cases, it seems like the curious observer’s question of “what is this?” is adequately answered by a species ID. This will also be sufficient for most researchers’ questions. (And if it isn’t, I suspect the researcher will probably want to make and/or verify the below-species IDs themselves.) If someone is invested in below-species IDs, either in general or for a research question, they can use the current Identify set-up to find observations for their species of interest and add subspecies or other IDs to them:

The proposed change seems like a rather large change for an improvement to a small number of cases

6 Likes

What happens if infrataxa are added to a taxon that previously had none? Do all IDs need to be reindexed? Do those observations get marked as unreviewed?

This is definitely true. Most invertebrate groups I see have infra-specific taxa added very sporadically, and those that exist often have a dubious taxonomic standing. I prefer the current system, which is centered around the species ID (just as taxonomy is based on the species concept).

I think changing such a fundamental part of how iNat works would be very confusing and not very helpful.

6 Likes

I’m not a fan of the system being proposed here, but I do feel like Casual observations with a species-level community ID and those without one do need to be separated in some way.

7 Likes

Re obs with Id to species level, but infrataxa are not determined

Could there be a separate label for these, something like RG (finer ID to subspeciecies still needed) or two labels:

  • RG
  • needs ID to subspecies/variety
2 Likes

After some thought I realized that the concept of when (at what rank) an observation with community agreement automatically flips to “Research Grade” is a separate idea, so I’ve once more updated my original post to clarify that it’s not within the immediate scope of my proposal (but may, or may not, be something to consider along with it).

@egordon88 @guerrichache FYI, in case this changes whether this still has your vote.

@bouteloua @cthawley @bugbaer @thomaseverest @raymie I’m curious if this changes anything for you?

This only works if you know the species that have subspecies already. This proposal would allow you to find any observations that could be further refined to infraspecies without querying for a particular species.

I’m curious if anyone actually has (or is able to obtain) some numbers on how many observations are RG but are of a species that has infraspecies. I wonder whether it’s really a small number of cases.

Yeah, for this to work, anytime an infraspecies is added or removed, I think you’d have to reindex some species-or-below-level observations to assess whether “Needs ID” needs to be reapplied or removed.

Regarding reviewed status (and “community ID” DQA votes, which you didn’t mention but which would be impacted similarly) we already have that concern today, so that could be solved (or remain unsolved) however the community sees fit.

I’m curious what you don’t like about it. One of the biggest things this does is to accomplish just the very thing you just said needs to be done.

That’s what I’m proposing. Today we have just one set of mutually exclusive labels. In the proposed system, we’d have two sets of labels indicating (1) whether the observations ID can be further refined, and (2) whether the observation is (or has potential to become) eligible for export to GBIF.

I say labels to refer to them in an easy to understand way, but if people think that’s too many labels, “Needs ID” status could displayed in some manner other than where it lives today (as a badge on the observation).

1 Like

At this point, I’m somewhat confused since the original wording of the request has been changed so many (7) times, which I don’t think is ideal for Feature Requests since it makes the conversation quite difficult to follow and difficult for voters as well. As such, I don’t have a more detailed response to offer.

4 Likes

Totally a fair point, and my apologies in that respect. There’s not like a draft document that I can attach (I don’t think?) so it was the best way I could think of to clarify the proposed idea.

To complicate things further, I interact with the forum a lot from my phone, which sometimes makes editing well more difficult for me. :man_facepalming:t3:

To summarize, the material difference in substance between the original wording and the current wording is as mentioned in my last reply, that the notion of when something automatically flips to RG has been decoupled from the proposal. It was never really my core concern and only snuck in as I was writing examples scenarios, scenarios which were from the beginning intended to illustrate the possibilities of, but not necessarily define, the proposal.

2 Likes

I think I’ve understood the core proposal here, which is simply that Needs ID should depend only on 1) whether there are more specific taxa into which an ID can be refined, and 2) what the community has voted on the “Can Still Be Improved” annotation.

2 Likes

I’m not really sure what you changed, but I still disagree with

This is rather complicated because how infrataxa (which I think is a botany term) are used varies drastically from group to group. The entire system (in iNat and taxonomy) is based on the species level, so I think it should stay like that.

2 Likes

I didn’t realize infrataxa was a botany term! Wikipedia tells me that the equivalent zoological term is trinomial name or trinomen. Thanks for teaching me something new today!

So just to be clear, you’re saying that, if for example an observation with two votes of Oreohelix strigosa were to have both a “Research Grade” label (and go to GBIF) and a “Needs ID” label (encouraging people to look at it further to determine if it can be ID’ed to, say Oreohelix strigosa depressa or Oreohelix strigosa cooperi or any other subspecies), you would dislike that? (I don’t know this species or its subspecies personally, I just picked them because it looks like that’s your most observed subspecies. Feel free to substitute any other applicable taxa in that example.)

Is there something intrinsic to that idea which you dislike? Or is it primarily because it differs from the current design (and thus is disruptive)?

For many taxa (at least many vertebrate taxa, which is what I am most familiar with) there are several subspecies that cannot be differentiated by morphology. Already, there is a group of users that often misapply subspecies-level identifications. Typically they base these IDs off of something, such as a published range, but they don’t understand or don’t account for varying levels of uncertainty.

For example, some users might give a subspecies designation to a Seaside Sparrow observation based on this range map:


But these range boundaries are approximate, and genetic work is ongoing. Already, I believe genetic testing has shown that the range of maritima extends further south into NC than presented on this map.

I think the current system works. Subspecies can be and are often applied where appropriate. But I think the proposed system would exacerbate subspecies identifications being applied where they are not appropriate.

7 Likes

The current system “works” because we’ve adjusted to what it can and can’t do well. Use cases it doesn’t “work” for are:

  • a knowledgeable and conscientious ID’er wanting to locate RG grade observations (just in general, not a particular species they’ve specifically written a search query for in that moment) that could be further ID’ed to subspecies/infraspecies or marked as “as good as it can be”
  • an ID’er wanting to locate casual grade observations (e.g. cultivated plants) that can be ID’d more granularly

These things can be accomplished in the current system, sure, but in a much more cumbersome fashion than finding other “could be ID’d further” scenarios, and I don’t feel like there’s really an intrinsic need for it to be like that.

Regarding people ID’ing incorrectly or beyond their expertise, as you mention, that is something today’s system has to deal with already. I can understand where you’re coming from in saying that it might exacerbate that issue, but I believe that can be dealt with effectively. Perhaps the proposed system’s “Needs ID” status could be conveyed using wording that essentially conveys “I can be ID’ed further, but ID me responsibly”. Or if such wording cannot be crafted, perhaps the “Needs ID” status goes from being a prominent label to something less prominent (but still surface-able via the Identify and Explore search pages) – perhaps it gets moved to the community taxon section, or gets removed from the observation page entirely. There are lots of ways this could be addressed in the proposed system, just as there are lots of ways it could stand to be addressed in the current system.

Using search queries to find observations for your two provided scenarios, in my opinion, is less cumbersome than dealing with erroneously applied subspecies IDs.

Regarding your second bullet specifically, I agree with @raymie:

2 Likes

Hm… So am I correct in understanding that you are:

  • for enabling people to more easily distinguish between Casual observations with a species-level community ID and those without one;
  • against enabling people to more easily distinguish between Casual observations with a lowest-possible-taxa community ID/DQA vote and those without one; and
  • against enabling people to more easily distinguish between Research Grade observations with a lowest-possible-taxa community ID/DQA vote and those without one?

I’m guessing those who said “I like that things are based on the species level” feel similarly? (That question isn’t directed at you, per se, but at the audience that is people who have already replied to and might again read this thread.)

Given the information shared by you and that others have mentioned, I can understand your concerns, though I can’t say I agree with the conclusion it seems you’ve reached from them. But that’s okay. I don’t have to reach the same conclusions as you, just as you don’t have to reach the same conclusions as me. You sharing your thoughts has certainly helped me to understand those concerns better.

I still think this proposal could be implemented in such a way that these very valid concerns are addressed effectively, but I’ll have to think on that a bit more and come back.

1 Like

Yes. Subspecies in mollusks are quite dubious. They are almost always eventually found to be either synonyms or distinct species. So ideally they aren’t really used, and changing the whole system to be centered around them doesn’t seem worth the trouble (although other groups do it differently).

This is also a major reason.

4 Likes

Yes, basically. Though, I don’t see that latter two as that difficult under the current system. So instead of wording it as “against making it easier…”, I’d word it as “against making it too easy to add erroneous lower IDs.”

1 Like