Two expert IDs disagreeing with non-expert, what to do?

So we have many observations where we have two experts agreeing on an ID which disagrees with that of a non-expert. If there are no other obvious experts to give a third confirming ID, the observation is going to be stuck essentially without a reasonable community ID. Is there a standard way of trying to resolve this issue?

5 Likes

One thing that would go at least some of the way toward solving this would be changing the default search to Active, rather than Consensus ID (see my Feature Request). Right now, observations in this situation don’t just need a third expert–they need a third expert who knows to use “ident_taxon_id” or searches through upper-level taxa to find these observations in the first place.

5 Likes

a comment addressed to the maverick explaining why you think your ID is right, along with some resources to look at, often goes a long way. if they say they are unsure of how to modify their ids, then you can explain how to make a higher level id or withdraw their id altogether.

if that kind of approach doesn’t work, then you could always recruit a third identifier to help tip the community id over to yours.

10 Likes

While leaving ID comments to explain the ID works in some cases for active members and I’ve been trying to be better about this, many of the ID’s in question are on old, inactive accounts (2+ years) and the observations are never visited by the observer so there isn’t a chance for them to “agree” or “withdraw”. In that case a 3rd expert would be good and who knows that function. Unfortunately the experts that exist either aren’t very active or don’t use iNaturalist. We do recruit 3rd parties a lot. But we are talking 100’s of ID’s. And I haven’t even touched half the stuff @dan_johnson has reviewed yet.

4 Likes

i don’t think a third person needs to be an expert. anyone who has an interest in cicadas could help you tip the scales. for example, in this recent thread (https://forum.inaturalist.org/t/anyone-know-of-books-that-could-assist-in-iding-north-american-cicadas/14661), @jhousephotos expressed an interest in cicadas and has 1500 identifications for others. that kind of person seems like a great candidate to help in helping to review and identify cicadas, even if not an expert.

10 Likes

When I need 100% control over the dataset I export pieces of it and manipulate it outside of the iNat website. In addition to the above, there are some other proposed requests related to this:

9 Likes

Great point and I suppose I worded that poorly. If they’d like to help out we’d love the assistance!

6 Likes

You can probably train a few identifiers who already work on insects how to make the important distinctions and then sic them on the problem cases.

6 Likes

As one of the 2 “experts” on here who regularly identifies Indo-Pacific corals, I relate to this problem all too well. The most frustrating thing is when I tag a 3rd party to add an ID and, assuming I get any kind of response, it’s often something to the effect of “sorry, I don’t know”. I’ve just given up on a huge backlog of observations like this. Unless the observation is of some particular importance, it’s simply not worth the tedious effort of tracking down identifiers.

IMO, this is a major failing of iNaturalist as a repository of biodiversity data and no doubt a major reason for why there are so few taxonomists contributing here.

2 Likes

That’s an easy to understand frustration and one that I often have, along with the lack of reciprocal identifications, the fact that large numbers of observations never receive any identification at all, etc.

However, it’s important to keep in mind that the goal of iNat is mainly community engagement, to encourage people to get out and be a part of nature, even if just in a somewhat passive observational role.

The data, as useful as it is, is secondary to that main goal.

I find that keeping that in mind helps to alleviate some of the frustrations I sometimes feel on iNat.

Then there are the days when you wind up with a bunch of observations ID’d or get an ID for one you had no idea what was and it’s all good again.

10 Likes

Yes. iNaturalist is full of treasure, but we have to dig it out. It’s not finished and ready to use.

7 Likes

I like pisum’s idea. I appreciate the fact that you guys have cleaned up the IDs on hundreds of cicadas. A non-expert shouldn’t just take your opinion and rubber stamp it, but if you were to point out to her the salient features of a particular species, she could visit those problem IDs for that species and check it for herself. If you could provide a query or a list that would turn them up, and a quick lesson, that would appeal to a lot of us who’d like to know more about cicadas but aren’t ready to delve into a lot of literature or keys.

5 Likes

Here’s a relevant thread for broad requests like that: https://forum.inaturalist.org/t/help-me-identify-non-experts-welcome/2915

6 Likes

I’ve found that it’s pretty easy to develop a fan base for a particular group and teach them what characters to look for–in fact, it’s the teaching that often results in the formation of a fan base. I’ve been on both the teaching and the learning side, and from both perspectives I can’t overstate the value of sharing one’s expertise with others. If the desired goal is to get most every observation correctly identified to research grade for a group that your passionate about, then providing resources to make it easier for serious amateurs to identify things is really the only way to reach that goal. The number of professional biologists will never be able to keep up with the growing public interest in documenting biodiversity–sharing ones knowledge is the only viable option, IMO.

15 Likes

As mentioned, the problem is usually caused by occasional users or inactive accounts. They might put up something then much later it is reviewed and corrected. Part of the problem is some people don’t understand its a voting system and not last ID wins. [I’ve occasionally had people re-enter their original ID after I disagreed with it.] There are only two ways to resolve this:

  1. Recruit or train more identifiers (as it is now) - with the drawbacks as they are now.
  2. Change the ID voting system to be weighted (i.e. not everybody gets 1.0 votes). This however would be complicated to implement, seem unfair to some and also possibly have maintenance issues depending on how it would be implemented.

[I guess a voting system could apply a vote score based on how many leading identifications a person has at that level (i.e. species, family, etc). Problems are of course new experts wondering why their expert opinion has little value. Flatworm expert Leigh Winsor asked me why his initial determinations didn’t seem to be accepted by the system, which is a similar sort of thing.]

2 Likes

One suggestion for anyone interested in something like this is to make a guide that highlights distinguishing features between taxonomic groups, as guides are fairly easy to share between users and can include a variety of tags. I don’t live in New Zealand but this one always comes to mind as a neat example of how guides can be more than just a list of species present at a location:

Bumblebees of New Zealand
https://www.inaturalist.org/guides/1425

And a couple examples of how tags can be used to break up categories within the guide:

Mosquitoes of Hawai’i
https://www.inaturalist.org/guides/1958

Sandmats (Euphorbia sect. Anisophyllum, previously Chamaesyce) of the Llano Estacado
https://www.inaturalist.org/guides/2351

Edit: Forgot to mention that tags can be combined on Guides. So, for example, using the Sandmats guide I can select [Cyathia in dense leafless glomerules] + [Disturbed soil] + [Fruits lacking hair] to get just a single result: Here

4 Likes

As the non-expert, I’ve had 1 of the experts ask if I had the ability to go and take a photo of the plant from a different view to confirm or deny the id. That would be 1 solution. I was able to do so and it cleared up the id issue.
Live animals are a different issue.

4 Likes

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.