I hear you.
Sometimes when it is particularly obvious that they are not knowledgeable about the taxon in question and are just agreeing with the most recent suggestion, I have encouraged them to not agree to an ID unless they know how to verify it, but so far very few users have changed their behavior in response.
It has definitely made me inclined to provide more conservative IDs and put my “pretty sure but not completely confident” IDs in comments instead.
And that’s also why we need a “Thanks” or “
” or “
” button under each ID suggested, so that observers can express their gratitude towards identifiers, without seeing any tentative suggestion immediately propelled to “Research Grade” for no reason.
When identifiers feel the urge to bridle their contribution as a safeguard against blind --if well-meaning-- agreement… something is broken.
You two are those frustrating observers. There are a lot of us who do not blindly agree with suggestions, but if someone’s ID leads us to the right research path, it sure would be nice to be able to add that second ID and get RG rather than a first ID (verifying the comment) and hope that a third party comes along.
I’ve just been working on the Zingiberales. So far, it supports my hypothesis that one of the main problems with iNat IDs is just that people lack confidence. Almost every (Family) Zingiberaceae I came across was left at (Order) Zingiberales. This was not true of other families – If the actual suggested ID was Zingiberales (rather than a family-level disagreement), I don’t recall having to move from, say, Zingiberales to Costaceae or Zingiberales to Marantaceae. What other family in Zingiberales has the growth habit of Zingiberaceae? People lack confidence: if it doesn’t look like Musaceae, Heliconiaceae, Strelitziaceae, Costaceae, Marantaceae, or Cannaceae, then by process of elimination, they play it safe and leave it at order.
Sounds like we’re identifying to the limit of our knowledge. When I leave ID at a higher than species level, I wish I could go lower! But I can’t.
Not a matter of confidence in all cases. Often depends on what the photo shows or doesn’t show.
In the case I described, it shows a Zingiberales with a vegetative growth form that eliminates Musaceae, Heliconiaceae, Strelitziaceae, Costaceae, Marantaceae, and Cannaceae.
Yeah, fine, but that isn’t every case.
Or the people adding IDs for the order don’t happen to be familiar enough with it to do more than recognize the general group – it may be obvious to you that all other families in the order can be eliminated, but users who add broad level IDs to observations are generally not taxon specialists.
In some cases I suspect there may be uncertainty related to the taxonomic rank. When you have an order and family and genus that are all based on the same underlying name — and may have been given a common name like “gingers and allies” – it is understandable that not everyone might realize the huge difference in what they are selecting if they select the order rather than the family. Or they may assume that the name of the order means that everything in the order is clearly ginger-like.
Again, the taxonomic distinctions may be obvious to you, but it does not mean they are obvious to everyone adding IDs. Choosing the order may not mean lack of confidence at all, but lack of knowledge. This is very different.
I would estimate that, in my experience, at least 2/3 to 3/4 of observers agree with IDs that I add in a manner that suggests they are blindly agreeing (i.e., their initial ID did not suggest much thought or knowledge, the agreement comes quickly, without much time to research the taxon, they do not ask questions about the ID, they change it if someone suggests something different, etc.). These users agree regardless of any comment I might leave about the tentativeness of my ID.
By contrast, the portion of users I encounter who either thoughtfully agree with IDs or leave their ID at a higher level/merely withdraw wrong IDs as needed is very small. Because I mostly ID within a specific region, I learn to recognize these users and I am more willing to add a somewhat tentative ID for them because I know that I won’t be inadvertantly creating a situation that will be hard to fix later (for example, multiple wrong agreeing IDs from users who are no longer active or responsive by the time the mistake is discovered, for taxa where there are few expert IDers to begin with).
For what it’s worth, this is mostly in the context of identifying bees, which are extremely prone to misidentification in the first place because the CV is egregiously wrong the vast majority of the time. So bee people already spend a huge amount of time fixing wrong IDs. I do not have any formal training in bee identification, or taxonomy, and what I know I have picked up through my own efforts (field guides, scientific literature, tips from other users, etc.). I make most IDs at genus level or slightly above. This is not unusual for bees: the standard identification criteria are really not designed for photo-based ID, particularly photos that may lack detail or be out of focus. I am trying to learn to distinguish a few more species, but it is a slow and tentative process and I still make more really stupid mistakes than I would like.
I’m sorry you find IDers like me “frustrating”, but you don’t get to judge my decisions about what IDs I am comfortable making. Adding an ID is an action that has consequences – I am responsible for that ID and what it does. Leaving a comment conveys the same information, but has different consequences. Nobody is obligated to provide an ID, ever. It is every user’s prerogative to assess the situation (own level of certainty, amount and quality of evidence provided, difficulty of ID, likelihood of a wrong ID being able to be corrected, etc.) and decide whether to add an ID and if so, at what level.
It isn’t cowardice. It isn’t about IDers obstinantly preventing observations from becoming research grade because they don’t trust observers. Knowledge isn’t an absolute either/or dichotomy: there is a wide spectrum of levels of certainty and this grey area isn’t something that can be represented precisely in a system that only allows for adding an ID or not adding an ID.
I will also leave a comment, not an ID - when it is new to me.
And tag in a taxon specialist.
Then I am willing to support their ID (because that is what I thought it was)
Or if I am still wavering, I will tag in a second to support the first.
It depends how willing I am to be supported to RG. Only single digit obs on iNat - then I don’t want to be on THAT leaderboard.
I know this is an issue, as I used to do so when I was new to iNat and started using it regularly in 2019. I assumed I was being agreeable and thankful. When I started reading the Forum regularly, I learned this was an error and I was embarrassed about it.
Indeed, there were so many complaints about this kind of usage of the Agree button that staff removed it. That created such an uproar by prominent IDers that it was reinstated pretty quickly.
The staff used to say they would design better onboarding training for new users, but that has not come to be. I have not seen that project even mentioned lately; but I do think it would result in a better user experience, better observations, and make life a little easier for IDers.
I find it sort of frustrating when IDers suggest what an organism is in comments but do not commit to making a suggestion.
But, I do understand why some do that, especially after years of reading discussion here in the Forum.
I don’t have an education in biology, so ‘pointing out a research path’ leaves me clueless about what to do with the information. I just leave it, hoping someone else who understands the comment comes along later.
This is where I have seen this. John Ascher will often ID my bees at the same taxonomic level I have (I only observe in my little garden so three years in, I can recognize “my” bees) but sometimes will add a note with an even better taxonomic level. I just assume this is his way of making a note to himself or to Jorge Merida, who is the native bee specialist for this area of MX.
I am not privy to his workflow - maybe he is coming back to review characteristics, maybe he just wants that note there for some study he knows will pick up the Observation, maybe something else - but these notes never frustrate me. On the contrary I find them intriguing and often go to look up whatever he has noted.
Some of my native bee Observations never make Research Grade because they stay at Genus, but I don’t think that matters, so long as it is accurate. Honestly I am just exceedingly grateful if one of the bee specialists reviews them. (I’m also grateful for biology students, bee aficionados, and all others who help.)
Each and every observer and identifier is different and also evolves with time. Hmmm… there should be a taxonomy and phylogeny for that. :D
The “Assume good intentions” rule is golden. There’s usually at least a wee bit of reasoning and concern behind every strange behaviour… even if not immediately obvious, even if unnerving at first. :)
–
Apart from a ‘like/thanks/love/thumbs up’ button to appease the Grateful Crowd (I’d definitely use it! So much unexpressed love for y’all!), a better onboarding process might also help alleviate the “blind-agreers” issue.
For example : for the first 2 or 3 times that some newcomer on iNat clicks the “Agree” button, display a message reminding them that “Keep in mind this is not a thank you button: you should agree on an identification only if you feel confident or have good reasons to blah blah blah”. Make that message impossible to dismiss without reading it in full the first time, add a ‘do not show again’ checkbox the second time. Simple, instructive, interactive stuff. Not some crucial info amidst a wall of text buried deep in some Help page.
I suppose we all want to see our records properly IDed and at Research Grade as quickly as possible. But patience is often needed. I’ve had iNat plant records that sat at “Needs ID” or with a wrong ID for years before someone came along and helped me with a good ID and I thanked them for that. If I wanted that good ID faster than that, well, that’s too bad for me and I should’ve done my homework and maybe tagged someone to help. That’s my burden and not the IDer’s.
Having worked in a research collection at a university for many years, I know things don’t always happen fast enough for some folks. Some physical specimens sit untouched for years or decades until an expert can look at them. Mistakes happen and IDs get revised long after the wrong one was assigned. Some never get properly IDed. Just the nature of the game. In the case of a virtual collection like iNat, the internet can make us more impatient because we expect results now, not years from now. But the curation process, whether physical or virtual, is more or less the same.
I’m sure you’re right! However, I wouldn’t know, so I’d ID it to Monocots.
Yes, I imagine there is often some tension between the expectations of laypeople and the working methods of scientists, as well. The desire for a label to put on what one saw vs. the need for the data to meet a certain standard of reliability.
This is one reason why some experts will comment “it is probably x” but not add it as an ID – one can often assess the various possibilities (at this location/in this environment/with this overall appearance etc.) and suggest a likelihood, but this is not the same as being able to verify it based on standard identification criteria or other specific evidence.
I understand why the agree button is so tempting for many users – I just wish I were more successful at explaining to observers what clicking it means and why they should be cautious about using it. I would definitely support better onboarding processes for users, which I’m sure would help with this to some degree, but I think there is also a need to educate users in a broader sense about the idea that their observations are part of a larger scientific dataset and the ways in which various iNat structures are meant to support those scientific needs.
I suspect that much of this is not necessarily evident at first, particularly for app users. I also suspect that the ID process is also rather mysterious to many people.
It’s easy to assume that making an ID is a straightforward, almost mechanical process of looking at an organism, applying a set of rules about the distinguishing characteristics, and arriving at an ID, with no detours or uncertainty. Certainly a lot of communication about science encourages us to think about it as neat and tidy affair that happens in labs far away from everyday life – we see the results in the form of discoveries, diagnoses, and applications, but the path to arrive at those results is typically not narrated (the mistakes, the unexpected data, the failed experiments, the revised hypotheses).
In reality, nature is messy, and science is an attempt to find order and develop models that most accurately describe extremely complex phenomena. I think many of the frictions and concerns that arise around IDs are connected with expectations that everything “should” be unambigously identifiable. We would all probably benefit from learning to be more comfortable with uncertainty and recognizing that it is normal and not a failure on the part of either the observer or identifier.
Absolutely true, this is why I personally would love to see a graded ID button, with the options ranging from 100% certain, through most probably to probably and perhaps. Yes, I know this has been discussed before and it would be complicated both to implement and use, but just thought I’d mention it in passing 
.
Alternatively, more than a “percentage” (c’m’on, nothing in life is “100% sure”), three buttons :
Why not block the observers from agreeing with IDers, and by doing that, blocking their possibility to take their own observations to RG-level? Or at least block them until they meet a pre-determined criteria, like being registered users for a minimum amount of time, or they have contributed with a minumum number of observations and/or IDs?
I can only speak for myself, but as a new user, and with limited knowledge and experience, I wouldn’t mind if it wasn’t possible for me to agree with IDs on my own observations.
iNat itself could use some trust levels and rewards, as we have on the Forum.
Not badges and bling, but you can’t access that function … until you have done the tutorial and ticked the boxes.