The community aspect and individual aspect aren’t necessarily completely independent of one another in difficult taxa where IDing capacity is limited.
I do a lot of IDing for a taxon that often can’t be ID’d to species level from photos (bees). Most IDers in my region don’t make much use of the “ID cannot be improved” box. One reason for this is that even the specialists have certain genera they are more confident with than others, so just because one person can’t take the ID further, another person might.
However, one effect of this is that the “needs ID” pile is always getting bigger and it is difficult to sort out those observations that multiple people have looked at and really can’t be ID’d any more specifically because the photos don’t show enough detail.
So clicking “ID cannot be improved” – if you have the knowledge to assess whether a more specific ID is theoretically possible – can (at least in principle)* serve as a way to signal to other IDers that they don’t need to spend time on that observation and can look at observations that haven’t been reviewed instead. From an IDer standpoint, I see using the DQA button as not being about making as many observations RG as possible, but about making it possible for IDers to filter observations and find the ones that are in need of attention without having to look at every observation oneself. Reducing the “needs ID” pile can help the remaining observations get seen.
This doesn’t mean I think RG observations are necessarily correct (I find wrong RG observations regularly, even in taxa that are generally unproblematic to ID), but in most cases it is likely that at least one knowledgeable person has looked at the observation. If the DQA has been used to make an observation RG at a rank above species, it is even more likely that someone knowledgeable has looked at it, because observers and generalist IDers don’t generally use this button.
*Now, in practice, I have found that in my region there are a number of specialists who will ID everything in their particular area of interest, whether or not it is RG. This has never struck me as a particularly effective way to distribute IDer effort in a taxon where there is a chronic shortage of IDers – although given the rate of mis-IDs for bees, I can understand why some people may feel they need to do so. (But do we really need a fourth confirming ID on one of the handful of bee species that is usually ID’d correctly, when there are hundreds of observations that have been sitting for months or even years and clearly not been looked at by anyone?)
There’s a handfull of damselfly species where I go trough the RG obs aswell. Just because I like them a bit more than other ones. But I often see obs that are RG, and are just… wrong. Even RG obs of American species in Europe. Which is just insane. And every single time it’s just the ob using a suggested ID, and 1 person agreeing with that ID.
It just baffles me that that’s enough to be RESEARCH GRADE… like how low is the standard. All you need is 1 person with bad intentions, or isn’t very knowledgeable about the species, and they can mess up a whole lot.
At the very least make it so you need 2 agreeing people (excluding ob) when ob uses a suggested ID. Or maybe even for any observation. Why is the goal to have as many RG obs as possible. Isn’t it much better to have less obs at RG, but be sure they’re actually correct, rather than have a lot of RG obs, but have a good amount of wrong ones.
I also often see the suggestion system suggest species which don’t appear in Europe. I see atleast 5 a day. Which is rather ridiculous.
This has been discussed before – if 3 IDs were required, there are lots of taxa that would never be able to become RG because there simply aren’t enough skilled IDers.
I didn’t say most RG observations are wrong. Most are probably correct in most taxa (in certain taxa most IDs are probably wrong, though these taxa usually have a low percentage of RG observations). As a rough guidance for whether someone has reviewed the observation (and thus as a sorting mechanism for which observations should be prioritized when IDing), it is reasonably effective.
I see the problem as being not that the threshold for RG is too low, but that too many users uncritically accept whatever the CV suggests and uncritically click “agree” on IDs suggested by others. The user interface unfortunately encourages this behavior by making it very easy to accept an ID (and reward users with the coveted green “RG” label). I would want to see a solution that includes better onboarding and encourages different user behavior, not one that makes additional work for IDers who can’t keep up as it is.
What about only for certain taxa? Like the big ones. Or maybe even give that power to the curators. So they choose how many agreeing votes you need.
Strongly agree with you here.
I used to be one of these people too, well for my own obs. I enjoy mushrooms, and taking pictures of em etc. But I can’t ID them at all. Thankfully a person from Switzerland IDs them for me. And I always just used agree with him, so it would get to RG. (Seeing there’s not many people IDing them in Europe) But he then told me that you’re not supposed to do that. And I’m pretty sure it’s against TOS aswell.
But if it wasn’t for him, I wouldn’t have ever known.
So telling people who do the same, would be an effective way. But the thing is, I just don’t have the time to type or copy paste that every single time. I see hundreds of people doing it. It’s one of the most common occurances. And it’s up to iNat to solve it. Not us.
For some taxa, it would be thrilling just to be told that they can be identified that far from photographs. It’s mighty discouraging to be told, essentially, that you just wasted your time taking pictures from lots of different angles and that you would have had equal (lack of) success with just one picture.
Coming late to the party, but about the “good as it can be”
I used to be very hesitant of using it, but that did change. I will use it for species the current literature cannot tell appart without genital examination. We are a handful of spider IDers now, who help each other out (sometimes extensive tagging each other) and clean up this needs ID pile, also by means of marking an observation “as good as it can be” if applicable… So it gets of the plate of the other IDers. However, we usually only do it for at least genus level observations with two IDs that will then go to RG… from my experience, this will satisfy most observers instead of frustrating them.
I also tick the box often, for taxon where I am sure it can’t be identified further, for exemple when flowers are told apart and they aren’t present, ect. Sometimes I’d like if identifiers used it more often though, I’ve taken the habit of doing it myself on my obs. But I see why “general” identifiers would hesitate to use it, as I have to admit this I use it less often for taxa I don’t usually ID.
Hello! Could I interest any lovers of Eupatoriineae to visit Unknown / Eupatoriineae to make some IDs, particularly in those with no taxon at all? Thanks!
A state “identified but not confirmed” seems ok to me. Such an ID won’t be used as machine learning material, so no harm done on this side.
These observations are still valuable though, when considered in the context of other observations. If, for example, any possible mix-up woukd be with a species of another continent.
For me, the main problem isn’t exactly that I too many notification stack up when I’m gone for a few days as that there is NO handy way to deal with them a few at a time. I mean, there’s no way to dismiss the ones I’ve seen and don’t want to see again. Right now, if I have 600 notifications (not counting the posts from people I follow), I either have to set aside time to do all 600 at once, or I have to hit “more” again and again (without looking at the observations!) until I get far enough down the list I can hope to finish them in one sitting, then come back later for more. Of course, since the report of number of notifications never includes the post from people I follow, I never have any idea how many pages I have to just click “more” for.
(If I deal with a couple pages of notifications, mostly skimming past ones I don’t need to investigate, when I come back later they’re still all waiting and I have to go through all the ones I already saw before finding the ones I didn’t.)
I’d really like a quick way, from the main page itself, to dismiss each notification. Something very fast, a button to click on the dashboard itself, so I can go through the big pile up of notifications in small doses.
I know I could open each observation and click “unfollow” and hope they go away, but that misses the point. I like to see agreeing ID’s. Right now I can just skim past the observations that reach “Research Grade,” if I was confident about them, figuring that either I was right or at least I did no harm. If I have a question about an RG observation, I can open it and look at it, but mostly I don’t need to take the time. Using “unfollow” would greatly slow down my checking. Not just double or triple but hugely multiply the time for most RG observations. A “dismiss” button on the dashboard! Need a “dismiss” button!
I suspect we’re all pleased at iNaturalist hitting the milestone of 200 million verifiable observations - I certainly am. But it’s a little scary that it only took two years to go from the first 100 million to the second 100 million. So many IDs needed! And what if we grow another 100 million in less than two years? Eek!
So, I’ve revised my identification strategy and I’m just trying to make as many as I can as quickly as possible. That leaves little time for annotations (I had resolved to make more this year) or working on Unknowns (I used to do lots of that) or learning how to ID new species. Instead, I’m focusing on easy-to-ID species and expanding my usual area of effort. At least for now - I reserve the right to change my mind at any point!
Anyone else energized by hitting the 200 million mark?
This was my primary response, to be honest – not excitement, but a sense of foreboding.
We have to pay attention to both sides of the equation – the system falls apart if the increase in observations is not accompanied by a similar increase in IDing capacity.
One thing that I miss in these periodic announcements about iNat growth is some indication that there is a long-term strategy for ensuring that such an inbalance does not arise – that is, a strategy for actively encouraging IDers (particularly specialist IDers) and supporting them in their activities, whether these are technical changes to better cope with persistent challenges like notification management or solutions for larger questions like better onboarding for observers or resources and tools to assist users in acquiring IDing expertise or aid in collaboration and networking.
