When you ask “do you wait for more opinions, or follow one that seems most confident”? the correct answer should be NEITHER if what you mean by “follow” is “agree.” Agreeing with an identification of a species puts it up to Research Grade, which implies a level of confidence in its accuracy. You shouldn’t agree with anyone’s ID unless you bring something to that agreement. You don’t have to do major research. Often that can just be looking at other examples of the species they suggest and thinking “ok that looks right to me.” And that’s fun because you learn something in the process. (Yes, I realize others say the criteria should be more stringent.)
10 Likes
Then there are the vagrants/accidentals. There are bird species from Asia that show up in western Alaska. More famously, there’s the Steller’s Sea Eagle that showed up in central Alaska in 2020, in the mountains over 200 miles/320 km from the ocean. It then moved on to the southern U.S. and then the northeast U.S. There’s a locally well-known case of a ruby-throated hummingbird photographed feeding on flowers in a small Alaska village 70 miles north of the Arctic Circle.
I’ve also witnessed sudden and lasting range expansions in Alaska. The large butterfly, Compton’s Tortoise Shell, had only been recorded in central Alaska once or twice before 2003 and then suddenly they were everywhere in their thousands, if not tens of thousands. They’ve been common regulars ever since.There was a similar sudden range expansion of Red-breasted Nuthatches in the 1990s, going from rarely seen to often seen to this day. But the cautionary “it doesn’t occur here” has, more than once, made me doubt my eyes or ears and dismiss my own observation.
Actually, joining iNat has taught me to pay much closer attention to just about everything.
5 Likes
This is common with reptiles where introduced things are widespread. My favorite two examples are a brown anole (Anolis sagrei) in Whitehorse, Yukon Territory, Canada that hatched out of someone’s potted plant purchased at a Lowe’s. Another example is a lizard found in a shipping container in Poland that someone asked me to identify based on the clue that the container probably originated on the west coast of South America. I was able to confirm the species occurred in Guayaquil, Ecuador, one of maybe three Pacific ports in South America.
7 Likes
Vagrant birds usually come with a story in the notes - which I would expect, as a courtesy to the identifiers they rely on.
And multiple supporting obs from enthusiastic birders.
For instance - this came up today.
Or the Eastern Cape chameleons which have established on the Cape Peninsula.
4 Likes
It’s actually easier on identifiers when things have been established since there’s a history to rely on. When it’s a one-off observation of an escaped pet it is much more challenging to ID. However, going back to the original topic of what to do when there is disagreement, more opinions and more voices are usually valuable as what I have noticed is that the community can more easily come to a correct ID than any one or two identifiers. Sure, it’s harder to overturn an erroneous ID, but the community is pretty effective at getting the answer correct.
3 Likes
When it comes to trusting someone’s expertise, I do think it is relevant to look at that identifiers record of ID’s for that particular species. If they have IDed multiple observations and those agree with other verified descriptions, I would be more ready to endorse their ID
2 Likes
But it’s also worth keeping in mind that everyone makes mistakes. I’ve recently been going through RG observations for some species and have found a not-insignificant number which were clearly incorrect - including some where one or more identifier(s) were among the top identifiers for the species. So I would emphasise that any agreement with a previous ID should be based on checking the ID, not blindly agreeing - however experienced the previous identifier.
9 Likes
Since we’ve been talking a bit about range (and hoping this isn’t too far off-topic)…
While I think about it, does anyone know how the geomodel works in terms of changes in range? I tried looking once at IDs which were supposedly outside the known range, but came across so many that clearly had a well established population in the area which the model didn’t know about that I gave up. Is it a fixed model of range based on what was known at some set point in time, or does it get updated periodically, or what?
3 Likes
See the (long and interesting) discussion in the comments. It is on my daily list to check for the Cape Peninsula - and I allow for - we can expect at least 10-20% of our Cape Flora species to only fit in a single point, and thus produce anomalous results
https://www.inaturalist.org/blog/99727-using-the-geomodel-to-highlight-unusual-observations
Looking for the Grampians I find wildfire!
https://www.theguardian.com/environment/2024/dec/30/as-the-grampians-fires-approached-20-kangaroo-joeys-took-shelter-in-a-living-room-experts-say-others-arent-so-lucky
PS Geomodel Anomalies for Grampians National Park 430 obs - where you will have to judge if you share my problem, or if yours are anomalies.
PPS Geomodel is updated each time the CV updates - so about once a month.
2 Likes
Exactly!
I mean, I do find it useful to check someone’s background if they provide an ID for something I know is a bit tricky or is not so common and distinctive that the average user is likely to be familiar with it.
It can help to get a sense of how reliable/well-founded the ID is likely to be and decide what I want to do next (such actions might include: consulting resources, asking the user about their ID, tagging other users for additional confirmation if something seems questionable, doing nothing and simply waiting to see what other members of the community suggest, etc.).
But someone’s expertise should not be the sole basis for agreeing with their ID. The top expert in the world may still make mistakes: because they overlooked something, because they are tired or in a hurry or distracted, because they chose the wrong item from the dropdown menu, etc. It happens to every single one of us.
I have said this before, but I think it bears repeating: while agreeing with someone’s ID may be meant to indicate appreciation and respect, there is a sense in which it often actually feels like the opposite. If I put time and effort into an ID (researching the taxon, acquiring the skills to interpret literature that is often not designed for field identification, etc.), it doesn’t feel respectful of my work if someone spends 2 seconds looking at a picture, decides that it looks good and I must know what I am talking about and clicks agree without doing even minimal research of their own – it is like copying someone’s answer to a math problem without working through and understanding the math oneself. (Again: not about gatekeeping or some idea that if I had to suffer then others must also; I would be delighted if my efforts spare someone wrong turns or make the process easier for others in the future – but taking credit for something without understanding it is not the same thing as following a map someone else has provided.)
6 Likes
Thanks for those links. I’m still not completely clear how the model works, even after reading that page - but it doesn’t seem particularly useful in my context. Of the plant obs listed as anomalies, all that I’m familiar enough with to judge are correct (there are a handful that I really don’t know enough to assess). In fact, a significant proportion of them seem to be my observations…
Yes, a huge burned area (around 75-80K ha) and the fire still burning, though mostly contained, I believe. The best I can say is that at least there hasn’t been much damage to property. From a nature-lover’s point of view, it’s pretty horrible, though the regeneration can be amazing to watch (if we get enough rain, unlike the last year or so). I live in a nearby town but have been well out of danger, thankfully.
2 Likes
I definitely agree that no one should agree with another user based on their perceptions of that other user’s expertise. Of course, even experts make honest mistakes, but they also make “accidental” mistakes.
For instance, I’m the top IDer of anoles (I think?) though I really only ID those in the US. I think I’ve IDed >125k in my day. The IDs are usually pretty easy in the US (for me), though I do sometimes find that I’ve missed a character or have someone disagree on an old ID that I made before I became more experienced on iNat - an “honest” mistake.
What I also find, at about the same frequency that I am corrected on an “honest” mistake, is some times where I either misclicked/typed (eg, hit agree when I didn’t mean to) or maybe agreed to a CID when I thought I was agreeing to a leading ID. In these mistakes, it’s obvious to me that I wouldn’t have made the ID based on any past thought process - it’s just a data entry error. When someone does thousands and thousands of IDs, a few of these “accidental” mistakes are going to get in no matter what! So, “if you see something, say something” - even the best IDers might misclick and otherwise not realize it!
10 Likes
Misclicks? Oh yes! The observations I ID’d as Canada Gooseberry and Olympic Grasshopper both had feathers. Oops!
(Goose and Gull respectively, in case you’re wondering.)
So anyway, @gljcrsmith1 – use a little caution when agreeing. (Advice I should heed a little more when IDing.)
4 Likes
Re accidental mis-IDs - absolutely yes. I once identified a Eupeodes hoverfly as “Woodpeckers and Allies”. I have no explanation.
7 Likes
I would argue that, at least in the case of bugs (where I focus most of my time), there really isn’t any evidence beyond range to make many ID calls. The example I like to use is Luna Moth- is there really evidence in the 10s of thousands of iNat records of them to rule out them being Mexican Moon Moths? Not really, the two look virtually identical in photos. We just base the ID on range. We could make a list of literally thousands of cases like this- species that look the same in photos, but differ in ecology, DNA, and range. It would be sort of weird and gatekeeper-y to refuse to allow any of them to be ID’d to species, when the ranges aren’t known to overlap.
For as popular as it is to declare “identifying by range” is a problem, an actual alternate-universe iNat where IDs must be defensible without reliance on range would be sort of a nightmare. Sure, you can tell if something is an American Robin or not without needing range as a datapoint, but once you get to insects, there’s almost always some “lookalike” that everyone is knowingly or unknowingly discounting as a possibility based on range alone. All it will take is someone kicking a hundred thousand RG insects back to genus with comments suggesting some alternate species from another region that the photo technically can’t rule out, and “identifying by range” will become much more popular I suspect. lol
I guess there’s a difference in my mind between saying “this clear photo of a Brambling can’t be a Brambling because it’s out of range” (bad) and “this photo of a Nomophila from North America shouldn’t be labeled Rush Veneer, because the literally-identical-in-photos Lucerne Moth is known to be common here, while the Rush Veneer is only known from a different hemisphere” (perfectly fine).
10 Likes
It may also be wise to check if someone added an ID on their own, or with the help of the computer vision suggestions. When you go to suggest a species, a list appears with suggestions, right? If you choose from this list rather than type the name yourself, your ID will have this symbol that looks like a shield with stars, next to the time of posting (i.e. 2h, 1d, 3y).
In the case of conflicting IDs, I tend to trust someone who types the name by him/herself over someone who picks the name from the list of suggestions, since there seem to be many people who simply pick the first name that appears in the list of suggestions and trust the computer vision is always correct, whereas someone who actually types probably had the correct name in mind before the suggestions list was even opened, or the suggestions list might not even have the correct name at all.
and wait until I’m sure before adding an ID
I always try to ID my observations as far as I can go, and then leave it to more experienced people to go further. If someone else adds a more precise ID, I’m very happy, but I don’t agree with them just for the sake of getting my observation to RG. After all, like others have pointed out, you’re only supposed to add IDs that you are personally certain about. I just leave it and wait for others to agree and get the observation to RG.
1 Like
Why would you value ‘I can typing’ over someone who has found a more efficient way to do it? This isn’t a spelling bee.
8 Likes
I identify on my iPad mostly. It is way more onerous to pull up the keyboard and type in an ID than to use the CV suggestion to autofill my ID. I’ve gotten so used to doing it this way that I almost never ID by typing the ID any more.
6 Likes
Some people really do let the CV do their IDing. Therefore, having a bit of caution around the CV-provided names seems good to me. Of course, most of us who do a lot of identifying also use the CV names just for speed and/or spelling.
2 Likes
In normal situations I don’t value one over the other. If the CV suggestions actually suggest the correct species, why not just click it? I do that all the time as well. Even when there are conflicting IDs, I don’t necessarily value one over the other, but I do keep it in mind when trying to figure out why different people added different IDs.
Imagine this situation:
Person A uploads a photo with a very general ID, like “Butterflies”.
Person B knows the species and adds an ID by typing the name, since this species is very rare and the photo was taken from a weird angle, so the CV suggestions don’t recognize it and don’t suggest the correct name.
Person C comes along and picks the first suggestion from the CV suggestions as their ID. This could be for whatever reason: new to inat, trying to get their number of IDs up, believing CV suggestions are always correct, a misclick, not thinking straight or whatever.
Of course, the opposite could happen, too. The correct species could be the first in CV suggestions and someone else simply thinks they know it better and they type a (wrong) species name. I’ve seen both situations.
So, again: I don’t mean to say we shouldn’t value IDs done by choosing from the CV suggestions, but I’ve regularly experienced that it’s helpful to keep in mind when there are conflicting IDs.
2 Likes