That accuracy rate seems to be a bit of a stretch. Most people, even the experts, won’t give themselves an accuracy rating that high for something so broad. There’s so many variables in play when it comes to identifications, including range, habitat, what specific features are important, the specific anatomy, intra and interspecific variation, species complexes, how well known to science it is, etc. to say that you can have a near perfect accuracy rating without having any sort of mistakes, which are often a common occurrence in a field involving what is basically a 4-billion-year-old runaway self-improving chemical reaction that often breaks any sort of rules imposed on it.
11 Likes
I work mainly with Noctuid moths, and unless I really know the species and the collector, a typical confirmation generally takes at least one or two minutes. Sometimes a confirmation can take an hour or more. I like to be thorough.
@mmmiller, my identification rule of thumb is to pass over research grade identifications, unless I see something that looks wrong. I rarely add a third confirmation. Sometimes, if it takes me a long time, someone will beat me to the confirmation, and I add a third just because I’ve done all that work!
6 Likes
lol… I can identify with taking a long time with an identification. I once initiated an ID suggestion for an observation with the intent on disagreeing with the first ID made. When I finally hit ‘submit’ on the ID, I found that someone had made the same suggestions as mine … and hour earlier. That told me how long I’d been checking resources, writing up a cogent post, etc.
I do think that being the 3 or 4th + agreement could easily serve a function in certain situations. Say you’re helping someone determine between difficult species (flycatchers, for example). They might - as a personal exercise, go to a bunch of observations and see if they can come to the same identification on those photos. Or you’re teaching young people how to identify common backyard birds and you give them the same exercise. I might spend some time refining my own butterfly identification skills this winter and do something similar although I will likely got to genus level ids rather than RG ones. More bang for the buck, so to speak. I learn something. An observation gets forwarded.
But in those kinds of situations people aren’t making thousands of IDs and I can easily imagine some personal benefit they may be getting even if I don’t know what that might be.
4 Likes
How can you tell you have an accuracy rate greater than 99.5%? Wouldn’t you need a higher authority to be verifying them all in order to know your accuracy rate?
6 Likes
I was accidentally instrumental in getting some observations a lot of confirming identifications. I assigned my students to identify a certain number of iNaturalist posts. I viewed it good practice for them plus a way to get more posts to Research Grade. It only partly worked. Some students who lacked confidence in their ability to identify or were just lazy learned they could follow other, more confident students and agree with their identifications. Some observations ended up with 4 or 5 identifications. Fortunately, the people being followed around were mostly good and/or IDing easy plants, so the chains of identification were correct. (Yes, I checked every one.)
If I teach (or try to teach) a plant ID class on-line again, I’ll give the same assignment, I think, but in a more controlled way. Each students will get certain species to identify or a certain time period to identify.
8 Likes
You are awesome.
11 Likes
A bird flashes across the road. “Robin!” the birder exclaims, or “Red-bellied Woodpecker!” We can ID some species as fast as we can see them, just as we can understand written words as fast as we look at them.
I can totally believe that some people can very quickly identify some things observed in some places. Glance at the picture, glance at the map, done! In fact, I get frustrated because I can identify some things faster than the computer can register the identification. (I see that “failed to save identification” window a lot.) I have to pause and give the computer time to catch up. So 5 second ID’s? Entirely possible – for some things in some places.
Do I make mistakes? Yes. That usually isn’t a result of speed per se. It’s a result of not knowing that species X is very similar to species Y, or not knowing that the range of species Y is entirely east of the Pacific coast states. When I’m corrected I pay attention and stop identifying certain things or learn how to do it right, and often go back and change my identifications. What’s my error rate? (How would I know??) I suspect around 5% over all, zero or nearly so for some species but then there are others. (I had a bad day recently; too many corrections to make.)
Reminds me of years ago my colleague and I drove with a botanist who knew we mapped the distribution of certain species on roadsides during long field trips. “You can’t identify Poa secunda (a small bluegrass) while traveling at 60 mph!” she’d say as we recorded it, and we’d explain about color and habitat and why we knew what it was. She’d sigh and agree, but soon she’d say, “You can’t identify Juncus balticus (a rush) at 60 mph!” and we’d explain that the particular shade of dark green in a swale in eastern Oregon always indicated that species.
Of course, a lot of times I go slower. I look up species in our floras, check ranges in various sources, or just stop to debate whether I see enough in that picture to tell it’s really what I suspect it is. The thing is, there are literally millions of photos of plants, birds, and other things I can probably identify that are “Needs ID,” and I’d like to change a significant number of them to “Research Grade.” Not enough people identify plants! So sometimes I’ll select an easy species and ID all 300 or whatever of them, celebrating the rare moments when I can type in a new name instead of just clicking on “agree” again again again.
11 Likes
Might I suggest the following…
https://forum.inaturalist.org/t/prairie-wildflowers-at-full-speed/9754
2 Likes
I believe it. My mother used to be able to spot dewberries below the level of the road as we sped along on road trips, just from the occassional tiny glimpse of red from an unripe one as we zoomed past.
5 Likes
Speaking only for myself, I will add another confirmatory ID to the Castillejas posted on iNat because I’m a moderator for that genus. It is, as you say, not really necessary, but it also doesn’t hurt. In addition, I’ve seen occasional posts where 3-4 people have confirmed an incorrect identification. So there’s that as well, though I know that was not your question.
6 Likes
I often add a third ID if I see one of the two confirming IDs is the original observer. Also some observations with multiple IDs are done in good faith by those who are looking at multiple thumbnails on a page and have no easy way to see how many IDs have already been put on the observation.
4 Likes
WOW. I have never seen such a massive spam of mis-identifications. I took a break from IDing, maybe too long, haha.
2 Likes
Thank you, thank you, thank you for adding identifications, @markegger ! The only Castilleja identifications I believe are the ones you’ve identified. Please keep up the confirming (and correcting) identifications.
2 Likes
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.