Although some forum discussions have touched on this, I haven’t seen a focused discussion on the topic of when (or whether) to rely on observer notes when identifying observations. Generally, in my own identifying, I do my best to follow consistent protocols and adhere pretty strictly to the rule that if I can’t see species ID features, I can’t offer a species-level ID. Now and then, this will really offend observers who feel that my reluctance to add an ID is tantamount to calling them a liar, which isn’t my intent at all; it’s just that if I can’t see it, I can’t verify it.
How do you other identifiers navigate this? And what about species like lampyrids (fireflies/lightning bugs), whose identification relies on flash patterns that can’t be easily captured on camera or shared in still photos? How much stock do you place in observer notes, and how do you tactfully explain if “take my word for it” doesn’t cut it?
I don’t have a policy on whether to include observer notes in IDs. I do read the notes. I know that some observers (usually beginners) will copy or paraphrase published descriptions into their notes, presumably to give more information to the reader, but those don’t count as observational notes. So it depends on the wording, whether I “know” the observer, etc. I certainly take habitat notes into consideration, and measurements if given, and specifics of key characteristics that may not show in the photos. But I don’t feel I’m obligated to accept the notes, and I may say “I can’t make out the midrib from the photo, so I can’t make a definite ID.” If the person is that invested in an ID they can ask someone else or go back for more pictures.
I kind of take them as supporting evidence, and will include or discount them depending on things like the particular thing they’re talking about and how subjective or clear-cut it is, the observer’s experience level, whether what they note seems to support or contradict what the rest of the observation shows, etc. A note might take me from being 90% confident I know what something is to 100%, but I probably wouldn’t use it to go from 10% to 100%–and if it contradicts what I see, I wouldn’t defer to what the notes say (I had someone tell me quite confidently they saw the crest and heard the call of one species of bird, but as far as I could tell the photo was of a different bird, so I held my ground on it to their consternation).
Notes are not verifiable, but neither are things like location, or date, or if we ask someone clarification about the observation in some way.
If someone’s notes don’t feel compelling enough for me to narrow down an ID, I’ll usually frame it as “I don’t feel confident”; if it’s contradictory to the evidence I have I’ll try to redirect to the evidence presented in the photos (or so on) .
I was going to say this also, that location and date can provide information that helps us decide an ID and these are data that are not actually in the image itself. So we don’t always rely just on the photo evidence whether there are observer notes added or not.
That said, I might in some cases be swayed by an observer’s detailed notes about characteristics the photo didn’t capture but that situation is not all that common.
Being an eBird reviewer, I experience the full gamut of notes on observations of birds. They range from “seen well” to 5,000 characters describing the bird, complete with 10 full frame photos.
Many times the observer describes the bird, then attaches a picture of the bird that does not match the description they wrote. Usually it’s subtle details like bill shape or fine plumage details, but sometimes the colors completely don’t match, yet the observer is 100% confident that the bird they photographed is the same one they are describing.
It’s not just new or inexperienced birders that fall victim to this, even veteran birders with far more experience and knowledge than me have also shown me photos that don’t match their description. That makes me very hesitant to accept details provided by an observer in iNaturalist, unless they are very detailed and precise. Having gone through a few hundred thousand iNaturalist observations of birds, I will sometimes confirm an observation that’s borderline identifiable from the images based on observer notes, but only if it’s something that’s already known to be in that area.
Thanks for that link! That does address the same question, but I evidently didn’t search for the right words to make it pop up. I think it’s worth revisiting the topic (at the risk of redundancy), but now I’ll have two forum discussions I can direct people to if I need to support an “I can’t ID what I can’t see” comment.
I think the notes are really helpful, especially for traits that don’t show up in the picture, traits I wouldn’t expect to be there – calls, behavior, tiny features. If the person describes the lemma of a grass, I at least know that he knows that the lemmas matter so presumably he did actually look at them.
I have been going through a lot of audio observations. Most of them are bird songs/calls.
If I can’t hear the species ID features, I can’t offer a species-level ID.
It’s nice to see notes that people add. There are a number of them that say that the person saw the bird. But, if the song/call isn’t clear, I don’t give it a specific ID.
In my opinion, a clear photo of traits or a clear audio recording is needed.
I don’t think this will dissuade people from entering observations.
And, I don’t accept an audio observation based on the person making the note about using the Merlin app. That app is a good tool, but it is very unreliable.
To me, to be fair, we need to choose either accepting the notes or not accepting them for every observation. Then identifiers can just say it is iNat policy for those who are offended. I would choose not accepting them. This comes from being in the deep end of the unknown pool of observations. If we agree to accept notes then that black blob in the sky will have to be identified, or the empty field the observer downloaded and said did have a Monarch in it, but it left before she took the photo. I think we should go solely by evidence we can see or hear ourselves.
While some iNat functionality agrees with you – observations without media are forever Casual, for instance – we also don’t generally make a point of knocking those back to “Life.”
Use the DQA checkbox for “evidence of organism” and call it good.
I don’t believe observations without any media became RG. Ones like the Monarch could if people are trying to be fair and believe what is said by the observer, or want to avoid conflict. I did mark it as no evidence. The user marked it as having evidence. A DQA war ensued. If there is a stated policy from iNaturalist, then the problem may improve.
Thanks for your comment, David. Even the smartest among us still make mistakes. We’re not bots . . . yet!
We’ve all had the experience of misplacing something, or making some mistake, and we say “What idiot did this?!” And the idiot was . . . the past version of you!
As has been said many times, it’s a great feature of iNaturalist that we can correct each other’s mistakes, compared to many other biodiversity databases where, from what I understand, mistakes can be much harder to correct.
I usually do not take the observers notes into account if I cannot verify the ID by other means as well at least to a certain degree of certainty. @ekmes comment summed it up really nicely.
I may have a strong hunch it might be a certain species due to what is visible in the photo or timing or location etc… some believeable information given by the observer (e.g. habitat) might be able to give this little bit of extra information to allow a species ID.
But what does it tell me if the observer for example states the ID was verified by genital examination if no picture of that is provided as well? Genital examination can be tricky and I don’t know how skilled the “examinator” is.. I would not agree based on such a statement.
Btw. I do not at all agree with the sentiment that those written extra information is on the same level of reliability as date or location. I too have run into observations that obviously had the wrong location or date, yes, but generally I would think those data will be correct enough (e.g. in the worst case in many cases it is not very relevant for the ID whether the location is maybe off by some kilometers). Date and even location are often automatically retrieved from the photo meta data anyways..
However, written information like size estimates are very often and very obviously extremely wrong in the arthropod world at least..
It’s amazing the number of observations that show a tiny blurred ‘something’ above a field. I do the same as you - mark it as Reviewed and move on.
Very occasionally when uploading my own observations I include something that I know I saw, even though the photo is pretty much unidentifiable. I do this rarely, for my own records, and I’d never expect anyone to verify my ID, as I know it’s not actually ‘evidence’.
I mostly agree with a lot of the things said here - notes can be really useful and if they help to identify an organism then certainly I will take them into account. But the photos/audio are the primary evidence, and if the notes contradict them then I will generally make some sort of comment when I add an ID (based on the photos/audio!).
It also definitely depends on what note is actually given though. If someone just says “identified by call” or “identified by genitalia dissection” then there is no extra evidence there that I can agree with. They are telling me how they got to their ID, not giving me anything to use to identify it. On the other hand, if someone says “call with three short descending notes” or “male genitalia with afa strongly indented on the right hand side” then this is actually some evidence that I can evaluate and can use to inform my own identification.
As an observer who does not identify/validate the observations of others (unless asked), I am perfectly fine with an identifier who declines to make an ID based upon field notes — based on their assessment of the specific note, or as a category of observational data they choose to ignore. Add a comment or just move along to the next observation.
I would be unimpressed with an identifier who disagreed with an ID simply because witness observations are “impure” — and was not explicit about the perceived flaws in the observation. If the rationale for disagreement is made explicit by comment… well, fine.
Rejection of all first-person witness observations by a non-witness “expert” — in order to follow the principle that some will be flawed — kind of sticks in my craw, but I understand the concern…