This topic has >200 comments to a very well defined feature request (or rather bug report) and there are very clear solutions proposed (such as the quoted above from @bouteloua ) that can be implemented by a simple change of wording.
Identification to species level with clear lack of evidence is from my short iNat experience with insect observations a major source of questionable IDs. I often find myself leaving genus or higher-level identification for such observations. This action is currently reported in a misleading way. I hope this can be fixed as suggested above.
I would like a feature change here too. There are many times that lab work is needed to get to species, so it should move to RG at genus. But often times, the original poster goes AWOL or doesnāt respond for some reason; and so the only way to not put it into the system incorrectly, but get it out of the ID pile, is to click the 'nope.
BUT
To me, there is disconnect between the wording of the āPotential Disagreementā popup box, and then the note it puts on your own id.
The popup box wording:
What then it says in the new ID note, even though the popup clearly states it may just be canāt be determined by evidence provided:
IMPO It would be nice to either:
Just change that ā(Username) disagrees this is (species)ā to ā(Username) disagrees this is (species) or evidence provided is not enough to be conclusiveā >>> this would fall in line with the popup question then, where it allows either disagreement OR not enough evidence for that ānopeā selection
Change the popup box to allow three options:
āI donāt know if it is this (species) or not, but def this (higher level you selected)ā (already exists)
āIt is not this (species)ā (already exists)
āIt may be this (species), but these photos** alone are not enough to conclusively ID it.ā
aka splitting up the two variations that are currently combined into one, to allow then the correct note to show under our disagreeing ID.
Iām guessing option 1 is easier from coding end, but it is less descriptive, so option 2 would, in my opinion, be nicer for clarity purposes.
In the end, if the comment it leaves matches the text in the popup that would still be better in my opinion.
I of course write notes (everyone I know in the things I help with do!) but the reality is a lot of this stuff are users briefly here, or not here anymore, or people just not caring/checking so relying on the original observer to fix their ID in my experience rarely works.
**ETA: see Star3ās reply - āphotosā should be switched to āevidenceā
I like this, but Iād recommend one minor verbiage change here:
Iād change āthese photos aloneā to āthis edidence aloneā, to also cover observations using:
audio recordings,
gifs,
spectograms (though technically those are usually uploaded as photos)
etc
Yes I like that change!! I often ask people for notes, or if they remember something about it, to get something to species. Or sometimes knowing the location & time of year myself, is enough to confirm something (such as a blurry bird photo from a spot I know that does indeed have a lot of sandhill cranes)
i know some dislike added complexity but i would like to have these three different options too. Iāve noticed people again explicitly disagreeing with IDs because they think it canāt be confirmed with the photo, but sometimes when they are clearly wrong and itās actually pretty easy to get an ID from the photo. Personally I donāt think people should disagree unless they think it is specifically wrong, in most cases. At least not on an observation of an active user without discussion first.
A newer thread addressing the same topic (https://forum.inaturalist.org/t/doesnt-inat-commit-a-basic-logical-error-when-an-identifier-suggests-a-higher-order-taxon/44144) has just been shut down by a moderator, sadly, because this issue persists and causes a lot of unnecessary confusion and frustration. I was writing my suggestion for a change of iNatās process (which currently violates basic logic) when the discussion was closed.
I believe my suggestion is essentially the same as @fogartyfās. Iām pasting it below.
Please note that there is no point in repeating the identifierās taxon choice on the buttons for the different options explaining possible reasons.
Intention of the identifier:
āIām fairly certain itās the higher taxon Iām suggesting but I personally donāt know what it could be within this taxon. Iām not disputing the original choice.ā This is current option 1. Text on button:
āI donāt know what it could be within the taxon Iām suggesting. Iām not disputing the original choice.ā Consequence:
The taxon is displayed as a normal ID suggestion. The displayed/registered ID for this observation is not changed.
Intention of the identifier:
āIām fairly certain it is not the original choice but it is within the higher taxon Iām suggestingā. This expresses clear disagreement with the original choice. It is what current option 2 ends up implementing but the current wording on the button for option 2 means something completely different. Text on button:
āI disagree with the original choice.ā Consequence:
The taxon is displayed as an ID suggestion with the usual statement about disagreement. The displayed/registered ID for this observation is downgraded to the taxon suggested by this identifier.
Intention of the identifier:
āIām fairly certain that an ID at the proposed taxon level is not possible based on the provided evidence but it is within the higher taxon Iām suggesting.ā This serves the case where the identifier is aware that it may well be the original choice but more evidence is needed to be sure. It includes allowance for the possibility that the ID may be improved by additional or better photos, or photos of the habitat, or a DNA sequence, etc. This is what the current button text for option 2 sounds like but not what iNat implements afterwards. Text on button:
āI disagree that an ID at the originally chosen taxon level is possible based on the provided evidence.ā Consequence:
The taxon is displayed as an ID suggestion with a new statement like, ā[identifier] disagrees that an ID at the [taxon] level is possible based on the provided evidenceā. The rest is a bit tricky. I believe it would be fairest if this option counts as a āNoā vote in the evaluation of the quality grade of the community taxon, but the displayed/registered ID for this observation remains unchanged.
This must be region or taxon specific, because I almost never see this. I only use the word āalmostā because I donāt trust my memory to be 100% correct. In actual fact, I donāt recall ever seeing it even once in all my years of identifying Butterflies for Ontario. Iāve actually seen the opposite a number of times, where experts have categorically indicated in their comment that it isnāt the species identified, explained why it isnāt, and then inexplicably clicked option 1 to leave the ID unchanged.
In my region/taxon, Iām one of the few people who click the second option to bump IDs up to higher taxon. I think itās perfectly correct and acceptable to bump the ID to a higher level when thereās uncertainty on the ID (ie. it may be correct, but it may be incorrect). Itās not unusual to have cases where it could be one of 3 species - all more or less equally plausible.
for those starting to comment on this thread again in 2023 in an attempt to improve the ancestor disagreement workflow, you really should read the blog post linked below, and use that as a starting point for additional discussion (rather than starting again from scratch).
I think part of the frustration here (apart from the wording) is that there appears to be a pattern in some cases of this method being applied for the purpose of āclearing the Needs ID queueā rather than indicating true disagreements.
The way it works: Someone makes an ID that is poorly supported by the pictures but may still be true. The āID queue cleanersā put a disagreement on it to bump it to genus, and then mark it āas good as it getsā to force it to RG at genus. When questioned about why the disagreement, often it is revealed that they donāt actually disagree that it could be that species (iNat counts it as a vote against it in the CID though) but the pictures donāt allow them to make a positive confirmation and they didnāt want it lingering in Needs ID forever.
The problem with this is that it discounts additional lines of evidence such as field observations that are not captured in the pictures, and assumes nobody else will be able to find identifying features in those pictures either. The observer ends up being frustrated about this. The method used may send some valuable records into RG at a higher level than what they could be. How many experts out there will think of looking through RG observations stuck at genus level to weigh in on those?
At the root of it appears to be the desire to āclean upā the Needs ID pool. Iām not sure how to address that better. If it was used solely for cleaning up wrong IDs, Iām convinced it wouldnāt breed so much discontent.
I canāt speak for anyone else, but this doesnāt apply to me. I have no problem with an indeterminate observation sitting in the Needs ID state forever. I punt it up to the genus level (or higher) precisely because it wonāt stay in the Needs ID state for very long. Someone with much more modest identification skills (or none at all) will likely come along and agree with the original ID and make it research grade when it doesnāt warrant a species level ID. I regularly see people agreeing with the original (species level) ID even after Iāve explained why it isnāt warranted. I see almost nobody coming along and clicking the āthis is as good as it getsā button - in my experience, most people donāt even know it exists.
I agree with you that there are quite a few people who seem overly concerned about too many observations remaining in the āNeeds IDā pool. When I ask them why, they tell me it will waste the time of the IDers. To which I reply āThatās me, and Iām perfectly capable of keeping track of which observations Iāve reviewed and which I havenāt, so please donāt do me any favours.ā
If there are additional lines of evidence, the observer is free to present them, at which point I will reconsider my (higher level) ID. In my experience, this is relatively rare. And even if additional evidence does exist, in my opinion, a photo observation should either stand or fall based on the photo. If the photo doesnāt support the ID, how is the observation any different from an observation that has no photo at all? I have nothing against āsight onlyā observations - I make a lot of them myself - but an observation with no photo is treated as ācasualā on iNat, regardless of what the observer claims they saw. Why should a blurry, or partially obscured, or otherwise ambiguous photo be treated any differently? And I regularly see people misidentifying perfectly clear photos (Iāve even seen experts do this). Thatās why having more than one set of eyes looking at the photos is so powerful. If people make mistakes identifying photos that they have in their possession, which they can zoom in on and ponder at their leisure, then I submit that we shouldnāt put too much faith in what they think they remember seeing when they briefly saw something in the field.
Does it? Iām pretty sure that anyone is free to add a more specific ID, and push it back down to species level, or present arguments as to why the higher level ID is unwarranted.
Do you have actual examples of this happening, or is this purely hypothetical? I ask because I have a very narrow view of what goes on in iNat. I only look at observations for a specific set of taxa in a specific region. But I look at almost all of them that fall in that scope (Needs ID AND Research Grade). Within that scope, Iām one of the few people who regularly bumps IDs up to higher levels. I donāt see any of the activity you describe. Iām not saying it isnāt happening elsewhere, but Iāve seen no evidence to support your concerns.
In my experience, thereās almost nobody cleaning up the āneeds IDā pool but me, and I donāt set things to research grade at higher than species level just to get them out of the pool. I just keep track of where I am in the pool. In my experience, the problem is more that observers want all their observations resolved down to species level and set to research grade - all nice and tidy. Theyāre the ones who want closure, if they are paying attention at all (some observers seem to submit observations and then never interact with them again, regardless of what IDs other people apply to them, or what questions we may ask).
Yes, I do, including on some of my own observations, but weāre not supposed to ācall outā specific ones by linking to them on the forum. I remember one where I had revisited the same patch of plants later in the year when they were in bloom and that was disregarded with the notion that there was no proof they were the same individuals. Same with observations of pupae and adults where I was told there was no proof that particular adult emerged from it. By now I have video footage of another one emerging and Iām sticking to my IDs. I suspect it depends on where and what you are identifying, or posting observations of. I guess we all have different experiences based on what corner of iNat weāre working on. To be clear, I have no problem at all with actual disagreements of wrong IDs.
Me neither - I just click āreviewedā and move on. I can see the point though that this only affects my personal view and it will still be there for the next person going through the pool.
Multiple photos of one individual scattered across many separate obs
Multiple species in one obs
A duplicate of ā¦ please delete
They need the observer to sort out the problem first, before we add IDs.
For those 3 cases, I ID as Life and Good as it can be - to take them out of Needs ID. Otherwise I battle with good intentions IDing the first picture and ignoring all the other species. Or ID the flower, but donāt realise the leaf obs belongs to that flower.
But the bulk where someone else can refine the ID, they must stay politely in the queue. Identifiers can choose to Mark as Reviewed, or Follow - as they prefer.
Ok, that sounds like someone taking my notion that āa photo observation should stand on its ownā to an unreasonable extreme. Is it one particular identifier, or are there multiple people following this strict regimen? Iāve had cases where people have followed up with additional evidence to support a species level ID and Iāve cheerfully accepted it and retracted my higher level ID (āthis is the adult that emerged from the pupaā, āthis is the host plant the larva was found onā, etc.). My attitude is that even if the evidence is incorrect/fabricated, itās worth accepting it just as an illustration to other observers that multiple lines of evidence can be followed to arrive at an ID (unless there is some good reason for doubting the ID that would be arrived at).
Rather than making major changes to how iNat works (and in the process, possibly robbing IDers of tools that are critical to them doing their āgoodā work), perhaps the ābadā behaviour of particular IDers should be addressed.
Thatās my biggest concern with these discussions. If people want to change the disagreement text that appears on various screens, I have no issue with that. If someone is arguing that there should be a 3rd disagreement option, Iām fine with that as well (though I donāt think it will do much to alleviate confusion). What concerns me are calls for the elimination of our ability to punt an ID up to a higher level when we donāt explicitly disagree with the ID, but disagree that we can be reasonably confident that the species level ID is correct. I need that ability to properly do what I do. Take that away, and Iāll be asking myself āwhat the heck am I even doing here anyways?ā. I donāt think thatās the train of thought you want to start in the minds of the experts who are doing ID work.
I donāt even click reviewed (typically), because in my mind, that would be equivalent of saying āIām done with this observationā, and if I havenāt put an ID on a Needs ID obs, then Iām probably not done with it. I donāt expect my own identification queue to be empty - ever. Thereās an ever growing backlog of observations that I couldnāt make a definite call on. Each day when I do identifications, I just scroll backward through however many pages of observations there are in my identification queue until I find where I left off in the previous session. I may waste a minute or two doing this, but thatās peanuts compared to the amount of time I spend actually looking at observations. (and yeah, I realize this is not a perfect system )
I guess if I were going to create a feature request out of this kind of discussion, it would be that iNat provide some way for us to flag observations TO OURSELVES. ie. features/tools to allow people who do IDs to better manage the observations they are working on. As far as I can tell, that doesnāt really seem something that iNat has put much effort into. I often have observations that require followup, and other than copying a URL into a text file, I canāt think of an easy way of flagging an observation for followup, or of quickly finding a particular spot in a list of observations (ie. my observation queue). If such features exist, Iād appreciate hearing about them.
Two things though, at least what appears to be consensus for my area:
If there are multiple pictures, I think they should be IDād to the common taxon, e.g. if all the pictures are flowering plants, they should be IDād as Angiospermae, not Life. Life adds to the unknown category, and there are those of us who go through casual unknowns every now and then. Adding Life as a disagreement puts all these into that pool and also disagrees with all future IDs then.
In multiple threads it has been mentioned that duplicates should still be IDād truthfully, ideally with a note pointing out it is a duplicate and a link to the other observation. I agree iNat needs a better way of handling these, e.g. adding a flag for duplicate maybe, but putting them at Life and as good as it gets puts them into the casual unknowns for those of us who do sort that pool. At the very least, there needs to be a note that it is a duplicate because if these show up on their own without any comments itās impossible to tell and looks like a mistake.
I would agree with this for the Needs ID pool. I have on occasion gone back through stuff I had skipped after learning a new ID skill to pick out those that I wasnāt able to ID before but can now. I was curious and checked my numbers and it turns out I have 10x more āreviewedā observations than identifications. That sounds like a lot, but most of my āreviewedā but not IDād observations are casuals and the vast majority is missing pictures (>70% of what I marked reviewed), are humans, or not an organism, and a good amount of opted-outs. I guess those who avoid going through casuals wonāt run into this as much.
Duplicates is more difficult for me - as Tony R asks South African identifiers not to ID duplicates (which should be deleted). What to do??
But on balance, more identifiers tripping over multiples (and duplicates) which I feel shouldnāt be IDed until the observer resolves the problem. Difficult if they have gone, dormant - while new identifiers come on board.
If it is Life and Casual - it is a workaround to say this obs is broken. If I follow iNat guidelines to Angiospermae - there is no workaround (I can only make use of Good as it can be for Life) - it remains in the Needs ID queue forever and for all subsequent identifiers. And oh so tempting to, go ahead, and ID that first photo.
I only come back to Needs ID, when notifications tell me someone else has moved the ID, or left a comment.
Yeah, Iām not sure thereās a good solution for duplicates under those circumstances. I sometimes point them out to the observer, but Iām doing that less and less since it doesnāt get results very often. Iāve also used the curator_notes field to leave a note (to myself) along the lines of āduplicate of [link]ā.
Because our database gets observations from multiple sources, labelling observations in iNat as duplicates only addresses part of the problem. I wrote some software that effectively āfindsā candidate duplicates in the aggregated data, which I then review. I search for several different duplication scenarios (single observer entering multiple observations on one platform, single observer entering multiple observations on different platforms, collaborating observers reporting the same thing, etc.). Then my software āhidesā the verified duplicates within our database, and cross-references them to their respective āprimaryā observations. As all this happens outside of iNat, it doesnāt really solve the problem within iNat - it only allows me to plow ahead, knowing that at least the duplicates in iNat arenāt impacting our database.
I guess everyone is going to come up with different workarounds to make iNat work for them. I donāt expect iNat to be perfect. It would never occur to me to ask for new features. My thinking is always along the lines of āwhy should they change iNat just for my benefit?ā. Iām happy enough that iNat generally does what itās supposed to do, which isnāt the case with some other (bug riddled) citizen science platforms.
This. This is the most frequent kind of ādisagreement,ā when the disagreer comments at all. But again, this comes down to the wording: āis the evidence provided enough to confirm this isā¦?ā does encompass this situation. Yes, it could be that species, but the evidence is not enough to confirm it. But as the OP pointed out, āI do not disagree with the species ID, which may well be correct, I simply want to specifiy that there is not enough evidence to reach that ID level.ā We seem to be going around in circles, when the real issue, as I understand it, is the mismatch between the wording of the options we are given and the wording of what comes up ater we choose.
āI donāt think there is enough evidenceā =/= āI disagree that it is that species.ā
You assume that nobody takes detailed notes in the field. I can take detailed notes a lot more easily than I can get a good picture.
I work with butterflies, which often donāt hang around long enough for a photo, let alone for someone to take detailed notes. Iāve been doing this for a long time, and I canāt think of an instance where somebody said they sat down and took detailed notes of a butterfly in hand (except for old historical specimen records, and thatās not what weāre talking about here). Typically, the stories run along the lines of āI saw it from about 10 feet away, in flightā. Like I said, I frequently see people overlooking details in photos they have in front of them - where the details are plainly visible. Why should I believe that they were any better at seeing those details when they took their notes?
But regardless of any of that, the fact remains that it doesnāt matter how good someoneās field notes are, on iNat, an observation without a photo will never be research grade (AFAIK). So why should a bad photo (that doesnāt support a species level ID) + the same field notes magically become an observation that can be IDād to species at Research Grade? To my mind, if that second scenario is going to reach Research Grade, it should only be at the lowest level that the photo supports.
Like I said, I have nothing against āsightā observations per se. But the fact remains that they canāt be double checked.
Not quite true - there are plenty of sound observations without any visual media that are RG. Some people post a hand-drawn sketch with or without field notes and I think thereās even a project for these. I donāt encounter them myself very often when doing IDs but seem to remember looking at the project at some point and noting quite a few research grade observations in it.
