Recruiting more identifiers

I found INaturalist in February and joined straight away. It has taken me a couple of months to start contributing identifications, but I am now doing so, concentrating on my home country (Wales) and sticking to taxa I know.
I am eager to contribute, but I am very much aware of my limitations. I am glad to have read this thread as I am now more certain that my efforts are wanted and welcomed.


I fully agree! I’ve had some good help on wild bromeliads and cactuses which are the sorts of things that get cultivated and moved around the world a lot. For other groups, I agree that accessing regional experts is essential. I’m doing what I can to spread knowledge of iNat amongst scientists in the Brazilian university where I work. When they realise that it can be a useful tool to support their research (I’ve been sending them this paper) they are interested!


Carrie - I had an email exchange last year with someone in that category: thousands of high quality, diverse observations, and single digit IDs. Despite their impressive general knowledge, this individual felt that there were far more qualified people to handle the IDs, and they just weren’t comfortable jumping into the ID path. It’s too bad as it is a loss for the community, but I thought that I’d share the feedback.


One thing I try to do is for every ID that is confirmed by another user, I head over to their page and do the same for a few of theirs. Might be a good practice for all of us and would put a little dent in the lack of IDers posted by the OP.


It’s much more rewarding clicking ‘Agree’ for easy observations rather than spending 20minutes keying out a plant species by reading descriptions.

Reward for Effort; I’m not sure if I would agree to it, but it’s a suggestion that has proven to work.
Some suggestions of reward may be:
• Titles for profile
• Increased file size of uploads from 2048px to 3840px
• Graphic Art for Profile - Golden iNat logo or something?
• Point system – The more uncommon a species has been identified the more points. Whereas a common bird may receive 1 point an extremely rare plant iNat 1st may receive 100points!
• 2 second Audio on profile?


The only problem with a point system/gamifying things is that people may then ignore observations of common things, thinking that it’s not worth their time IDing to get the 1 point. Say it takes 5 seconds to ID a common bird (look at pic, think a little, click agree/type in name). For 100 common birds that’s over 8 minutes to get 100 points. Conversely, it might only take 5-6 mins to key out/ID a rare insect/plant. So they’ve invested less of their time to get the same reward. This could also create a bit of a bystander effect where every person thinks ‘I won’t ID this common thing, not worth my time for the 1 point, someone else will do it anyway’, so then noone does it because they all assume someone else will.

On a personal level I love gamification and am constantly checking ID leaderboards to see where I am, but I think it does have its issues too.


Also, the boundaries/definitions of what is and isn’t a ‘rare’ species can be very blurry/arbitrary. Is a Sulphur-crested Cockatoo more or less rare than an Australian King Parrot?

While I like the idea of a point system, this may cause bias for identifiers to disproportionately identify organisms as rare organisms when in reality they are a common organism that just looks similar. I know this is true for myself because I love when my own observations are ID’d as something rare so I am always wanting to hit the agree button straight away. However sometimes the person who made the ID for me clearly was either just guessing or missing the more common species it actually is. To avoid this I’ll always check the profile and number of IDs made for whoever IDs my observations.


A post was split to a new topic: Group accounts?

When I first started, I knew very little. The more that was identified for me, more I determined to be able to “pay it forward”, so I have actively pursued literature, sources and connections that help me learn. Because it takes two IDs to make RG (typically) I came up with the idea that I would be contributing “my share” if I could hit 2x IDs to obs ratio. However, I am aware of the imbalance of identifiers to observers, so I recently shifted my goal to 3x…having said that, I find it easy to reach that, so I now use it as a measure of how much I “need to get out more”!

I started in my IDing by trying to tackle easy ones, but it was always difficult to beat the experts to them. It occurred to me that if the experts focus on obs that are over a week old, us novices can tackle the easy stuff and they just pick up on the more difficult ones. By “having a go” blind, and putting our guesses in on difficult ones as comments, often the experts would give us tips on how to know for sure etc. I am doing that at the moment with NZ moths, and learning heaps, thanks to Neville and co…

I guess what I am saying, is that for many taxa we do have enough expertise, but the use of that expertise might not necessarily be efficient…


I’ve been trying to do some more identifying the past couple of days since I haven’t done a whole lot in a while. I’ve found what’s easiest for me as a non-expert is to just work on one species at a time. One that has unique markings and is easily identifiable when they are visible, but isn’t being worked on much for some reason.

I start off by setting my filters to only show that one species, and then I can go through them quickly by clicking agree with the correct ones and skipping over the rest. Then I go up a level to genus and pick out all of those matching observations and can copy and paste the same explanation and links indicating what markings I’m looking at to id it to species. Then I can go and look at all the incorrectly identified ones and have a different response to paste as an explanation for them. If I get through all of those I might pull up ones that are already research grade that I haven’t reviewed to do a quick scan, because sometimes there are mistakes there too. I just hope my copying and pasting responses doesn’t get flagged as spamming ever.

This lets me get through more ids quickly to clear them from “needs id”, and I’m less likely to make a mistake than if I was jumping around from species to species. Unfortunately, it can also put me at the top of the leaderboard for a genus and make me look like an expert, which I’m definitely not! I’m just going after some of the the low hanging fruit to try and contribute.


here are some stats, just to help inform the conversation (note that some of the numbers might not tie exactly because i pulled them one by one):

life type verifiable r grade % (of v) needs ID % (of v) all diff (all-v)
bird 2958507 2817798 95.2 140711 4.8 3706099 747592
amphib 312756 257090 82.2 56666 18.1 346178 33422
rept 581161 526571 90.6 54590 9.4 637797 56636
mammal 511469 433626 84.8 77843 15.2 682877 171408
fish 245483 195266 79.5 50217 20.5 290325 44842
mollusk 323144 182487 56.5 140627 43.5 339365 16221
arachnid 455131 150780 33.1 304351 66.9 472686 17555
insect 4244179 2340634 55.1 1903552 44.9 4494011 249832
plant 6886471 3933625 57.1 2952853 42.9 7985769 1099298
fungi 974731 298815 30.7 675917 69.3 1015882 41151
protozoan 22306 7116 31.9 15190 68.1 23092 786
unknown 357304 1095 0.3 356207 99.7 994334 637030
subtotal 17872642 11144903 62.4 6728724 37.6 20988415 3115773
all 18286822 11338887 62.0 6948481 38.0 21431192 3144370
diff (all-subttl) 414180 193984 46.8 219757 53.1 442777 28597

(by the way, who knew that if you copied and pasted from Excel into the forum that the forum would format a table for you?)

i wasn’t sure how to pull in verifiable that had just one ID… but that might be a useful stat

another way of looking at the data:

life type verifiable % (of subttl) rg % (of subttl) needs ID % (of subttl)
bird 2958507 16.6 2817798 25.3 140711 2.1
amphib 312756 1.7 257090 2.3 56666 0.8
rept 581161 3.3 526571 4.7 54590 0.8
mammal 511469 2.9 433626 3.9 77843 1.2
fish 245483 1.4 195266 1.8 50217 0.7
mollusk 323144 1.8 182487 1.6 140627 2.1
arachnid 455131 2.5 150780 1.4 304351 4.5
insect 4244179 23.7 2340634 21.0 1903552 28.3
plant 6886471 38.5 3933625 35.3 2952853 43.9
fungi 974731 5.5 298815 2.7 675917 10.0
protozoan 22306 0.1 7116 0.1 15190 0.2
unknown 357304 2.0 1095 0.0 356207 5.3
subtotal 17872642 100.0 11144903 100.0 6728724 100.0
all 18286822 11338887 6948481
diff all-subttl 414180 193984 219757

a breakdown of Needs ID plants:

at any level at kingdom at class at family at species
2953589 109836 all 130606 all 184671 all 1391686
dicots 97425 asters 37142 common dandelion 6714
true mosses 13507 grasses 28870 ponderosa pine 4190
ferns 10084 mustards 10188 common blue violet 3714
monocots 6538 legumes 9858 American elm 3565
conifers 1000 roses 5539 coast live oak 3545
top 5 classes 128554 carrots 5046 red maple 3221
remaining classes 2052 sedges 3639 curly dock 3117
mallows 3587 eastern redcedar 3096
pines 3509 northern red oak 2977
mints 3319 black cherry 2836
top 10 families 110697 top 10 species 36975
remaining families 73974 remaining species 1354711

The local Society for Nature Conservation site of my country has a forum where users can ask for identification. The acceptance is really good.
I think that this could help with problematic identifications by having an additional channel to ask for help identifying an observation. This way hard to identify observations could get more attention.

The forum mentioned above recently got a new feature where identified observations can be directly submitted to the site, which is pretty nice.

Overall the site is far behind iNat though when it comes to usability and features.


I have to imagine that the sheer volume of ID requests that would ensue on iNaturalist would make a Forum approach pretty unworkable here.

But maybe if there was a way for an iNat observer to “flag” their own (and others’ ?) observations into a pool of “ID Help Requested” observations, allowing maybe 5 such flags per user to be active at any time (kind of like “votes” on this forum), and having that flag be a filter option in Identify and Explore – then potential identifiers could specifically filter for those observations (along with all the other usual filter options) that users considered to be their highest-priority in need of ID help.


You’re probably right that the volume could be way too much for iNat on an international level.

I really like your suggestion!
I found a similar request on the iNat Google Group: Identification Bounties


Yes, it is cooperation I value here. Written communication in the competitive sense may lead to misunderstandings and reduced participation.


What a great analysis! Lots of food for thought here.
Just curious, how did we get 1095 “unknown” observations to Research Grade?


they look like bacteria and viruses, like this:
look at the top species, in that category, and you’ll see what i mean:

the question for me is why there’s a difference between the subtotal and all. i’m not sure what’s going on there. i don’t think it has to do with the time it took me to pull those points in individually, since i doubt there were 414K verifiable observations added between the time i pulled started pulling with birds and when i finished with all. (maybe i had a typo somewhere?)

EDIT: the difference seems to be non-insect/-arachnid arthropods and other miscellaneous things like bryozoans and cnidarians…

1 Like

looking at the Needs ID plants at species level (which i assume means they just need a confirming ID to get to research grade), i think i see why there are a lot of things not at research grade here.

i bet common dandelion is just people not taking the time to be able to distinguish between Taraxacum officinale and the one or two other Taraxacums that might be the given area. we could probably fix that with a little bit of training. ditto with curly dock.

i bet a lot of the trees are going to be hard because some mix a lot (ex. oaks), and trees in general are hard to separate if you don’t get a lot of different photos/angles and other information (ex. is that an American elm leaf or a winged elm leaf?)

then there are the trickier things like violets. i saw that V. sororia var. missouriensis was upgraded to its own species recently in iNat. so unless you’re going to do a lot of detailed leaf measurements or rough guessing based on intuition and geography or just likeliness, what do you do? if someone already has it at V. sororia, can you ID it as Viola, and then just say it’s as good as it’s going to be?


there’s this red seeded dandelion that comes up in California but i haven’t really been able to verify it in Vermont, but the leaf characters seem iffy at best and the only real one is the seed color. So for most dandelion photos you can’t rule it out… but on the other hand it’s probably only a tiny proportion of the dandelion observations out this way… but I never really know what to do with them. @silversea_starsong was looking at that species, I think.

Violas are iffy, apparently they hybridize a lot, and it also sounds like they are being taxonomically vandalized by splitters. I wonder if they have sections or similar we can lump the look-alikes into. That has really helped to clean up some plant groups. But i am not a taxonomist.