Implement an iSpot style reputation system

In my experience BugGuide is more effective for identifying obscure insects than iNaturalist, at least for the insect groups I look at. As far as I can tell, that’s simply because they have experts for each group who is able to identify those groups, whereas iNat does not have those experts. I’m not sure though whether they are missing because something about about iNat repels them (which this proposal could potentially address) or just because nobody has asked them to join. Over time we do have more experts here. I agree that BugGuide is less user-friendly for the observer though.

3 Likes

Is it possible to cast a vote against feature requests?

1 Like

No. Only comment against it

I sometimes use Bugguide. (It is less user friendly for a host of reasons, and I favor iNat. as a result, but the mission is different–I get that). I just want to point out that I do see some of the experts from Bugguide here now. I appreciate that they are willing to add to both sites despite the differences in approach.

5 Likes

As an “expert” on bugguide (https://bugguide.net/bgimage/user/97817), I’ll state that the reason I don’t focus much on insects here on iNat is the general low quality of the observations (the majority being too blurry or distant to be of much value) and the overwhelming tendency for misidentifications to proliferate. If I come across a misidentified Sarcophaga and correct it, the observation doesn’t get fixed… it gets bumped to, say, family or order. If there are multiple incorrect IDs on an observation, my correct ID may not even bump it from that ID. Then what? I tag several more users just to get that misidentification to disappear. Tag more users so that it actually gets the correct name applied. It’s tedious and avoidable with a properly functioning reputation system. Do you, as a user, want low quality IDs without merit or an accurate ID weighted by the knowledge of the identifiers?

I’ve seen a couple comments here critiquing bugguide, but the reason it attracts so much of the professional entomological community is the curatorial quality control. It isn’t filled with misidentifications. iNat would likely see the same buy-in from experts in other fields if more importance was placed on accuracy. There’s no reason this has to sacrifice participation from amateurs. If anything, it should encourage more of it, as this would become a resource where a user can post an observation and expect to get an informed identification.

5 Likes

Oh yes. Users are constantly retraining the AI so that it switches between the Centrosema and Clitoria genus as top suggestion for some pea flowers. I would love it if when it suspected either genus it would keep the top suggestion as the taxon that encompasses both: tribe Phaseoleae.

Though I suppose even if it did only offer the LCD higher taxon (in either your example, mine, or others), people might still ignore that top coarse ID and choose the next species suggestion in the AI suggestions that looks similar to their observation (and distinctions are going to be hard to pick out in a tiny thumbnail).

I understand that it’s all related, but let’s please try to keep this topic focused on the proposed reputation system and less on photo quality, observation quality, or improvements to the computer vision suggestions, which have other topics where they can be discussed.

6 Likes

I don’t think anyone is suggesting that only “Experts” should make IDs. The goal is to weight the IDs based on the proven knowledge of the identifier. A vetted taxonomist would surely be at the top of the scale, but a knowledgeable “amateur” wouldn’t necessarily trail far behind. It’s simply far too easy for IDs to reach the nebulous “Research Grade”. This term should mean something. It should truly be the equivalent of an accurate, museum-quality identification… reserved for only the most reliably identifiable observations which have been confirmed by the most knowledgeable participants of this site. That is a far cry from the current situation.

As for how to weight the value of an ID, it should ultimately come down to: a) is the user a verified authority on a particular taxon (i.e. museum staff, academics, etc). This will naturally be a small number of individuals and their voice should carry the most weight. or b) is the user accurately providing IDs on here. What is the total number of identifications provided for a specific taxon. What is their ratio of correct:incorrect identifications. I’m sure statisticians could come up with some clever was to score this.

I’ll give an example: me. I am a professional entomologist with expertise in Diptera and Coleoptera. My IDs in these groups should carry a large amount of “weight” or “reputation”, despite the relative low number of IDs I’ve provided in these taxa. I’m also a very knowledgeable identifier of Amphiprion anemonefishes. I’ve yet to publish in a peer-reviewed journal on this group, but I would argue I know their biodiversity better than the small number of taxonomists who have. I would expect my “amateur” IDs to be weighted heavily for those groups based on the 1000+ accurate IDs I’ve already provided. And I am a true novice when it comes to angiosperms. My IDs in that group should count for almost nothing until I learn to consistently provide accurate IDs. When I post a flower, I want someone far more knowledgeable than myself to tell me what I’m looking at. I’m less interested in the opinion of some backyard gardener with a similar knowledge base as my own. I inherently weight the IDs that others give.

I suspect much of the opposition is from casual users of this site who believe a reputation system that weights IDs will nullify any voice they have in identifying observations. For those novices who want to provide IDs without any effort paid to whether they are correct, then, yes, it will intentionally limit (though not eliminate) your say in the process. That sort of user should not be given the ability to guess an ID to “Research Grade”, as is currently the case. For the dedicated naturalists who put in the effort, it should have a self-reinforcing effect. Each accurate ID gives your voice a little more strength. You’ll be able to see your own expertise grow numerically.

Reputation systems tend to encourage cliques and elitism, and do little to encourage grassroots education (which is the primary purpose of iNaturalist). So even if I agreed that the general accuracy of IDs was a major problem (which has not yet been proven), I could never support such an obviously divisive solution.

4 Likes

I don’t understand why the assumption is that more accurate identifications are going to make me happier here if my voice seems to be less important as a result. If you can recommend a system that allows me voice and respect and more accuracy (even though many IDs depend on photo quality) I’ll consider supporting it, but reducing me to just someone who sends photos and has a number attached to her name that indicates she can’t identify a cabbage butterfly from a monarch probably isn’t it. I know you are responding from an experts’ point of view. I’m just pointing out that the average contributor here values other aspects of the site beyond the identification. Surely that’s worth considering too. I think the number system as proposed here lessens my value here–that makes me truly unhappy. If such a system were in place, I would just not identify (which I suppose would be fine with those who think I shouldn’t be identifying anything anyway) or I would just close my account. I can use an app that doesn’t post here, have a reasonable sense of what I’m finding, follow up with books, and learn on my own. It’s what I did before iNat anyway, and the site goes on happily without me. But you are losing someone who is more willing to spend time identifying for school kids, helping people who throw their observations into “unknown,” willing to send notes welcoming others and adding tips, and dealing with fairly easy identifications (I am not guessing in those instances), so experts can engage with the more difficult ones. Yep, I’m a novice (and apparently a vocal one ;), but I’m a novice who already knows her place in the system (lowly as it is). Honestly, I find this discussion disheartening, so I’m going for a walk–probably with observations to follow :).

8 Likes

You’re conflating expertise with elitism. If you go to a natural history museum and ask a researcher to identify a specimen, do you want your answer to be crowdsourced for the sake of egalitarianism or would you rather an informed response from those most qualified to provide one? There is room for a compromise position on iNat. Surely a system can be devised that isn’t inherently “divisive”. There already is some version of “elitist cliques” here. For instance, iNat has basically 4-5 individuals who bother to identify Indo-Pacific corals. That’s not likely to change much under a reputation system… it would merely make it easier for that handful of users (which includes yours truly) to more easily correct misidentifications and give users some insight into the value of the ID attached to their observation. And perhaps if other researchers see the quality of these observations, it will encourage more to participate, further increasing the utility of this platform. Those opposed seldom bring up the legitimate merits this can bring.

I think you know perfectly well that that is a false dichotomy. There are many ways to improve the accuracy of IDs on iNat that wouldn’t require a reputation system (or anything like it).

4 Likes

I think most of the suggestions for improving accuracy could generally be lumped under the category of “non-egalitarian”. Things like increasing the number of IDs needed to reach research grade (which is similar to a reputation system in that it requires more expertise to get to that level, but in this instance it is simply brute force numbers versus a more nuanced metric of taxonomic “knowledge”). Or eliminating the “computer vision” IDs as counting towards research grade. I’ve seen it suggested that new users not have their initial IDs counted towards research grade. These all seem like band-aids to me. Perhaps there are other suggestions for improving accuracy that I’m not familiar with. Consider starting a new post on the topic if you have an alternative that makes better sense.

Hmm, what if the current Research Grade system was renamed, but otherwise kept the same, and then an additional label was made to be given to observations that have been identified by an “expert”. Sort of like the tiered system described here except that the “Research Grade” level isn’t just by any third person.
The new label wouldn’t be as in-your-face as the current Research Grade label (which has bright colours and is right at the top of every observation; very attractive). Instead it would be more hidden on the side, perhaps near the Community Taxon box, and indicate that someone recognized by the community to be knowledgeable about the taxon has identified it. Then normal users could still contribute to making the community of the ID more trustworthy, but more authoritative users can still increase the confidence more than others. Or pehaps even instead of a label, the community taxon confidence would just be affected by the level of “skill” attributed to each identifier?
I’m not sure what I think of this, just an idea I had…

The most vocal opposition, both on this thread and in the many other threads discussing this topic over the past couple years, has been from extremely dedicated users of iNaturalist who intentionally chose to contribute to iNaturalist instead of alternative citizen science websites.

7 Likes

This is essentially what the reputation system would do. An ID based on a couple of inexperienced users would be similar to the current “Needs ID” status, while an ID that crosses a certain threshold of cumulative expertise from the identifiers would meet the standards that “Research Grade” implies. I’m surprised that such a simple recommendation is so controversial. It seems a fairly minor tweak to the current standard. It simply rewards those in this community who work hardest to improve the IDs, which is why I figure that only the most “casual” users will feel disadvantaged by this. Perhaps that term is a bit imprecise.

It is absolutely superb the professional entomologist community in North America has in BugGuide a resource they feel meets their needs.

Why however folks feel they won’t get push back when they choose to interact with a site that had been founded under, and is thriving under different principles, and then tell them they need to adopt “our way” or forever be an unimportant sideshow is a mystery to me.

If someone is that upset about engaging with the great unwashed masses or losing their absolute right to be deferred to, then spend your time in a community build under those principles. They are right there if that’s what you want. You can contribute just as much to science there as here.

Hundreds of thousands of users have chosen inat because of its principles, not in spite of them. The site is growing exponentially because of them, not in spite of them.

8 Likes

i think a reputation system as most people are envisioning it here would make sense only if there was a second category of observations added – something like a model observation category that would require a higher level of evidence, etc. you could apply a reputation-weighted voting system to those model observations, but i don’t see how a reputation system helps the vast majority of observations.

1 Like

These sorts of histrionics are a bit much, no? Nobody is saying that you or any other novice cannot meaningfully contribute towards the process of identification on here. The proposal is to encourage a growth of expertise from all users, in whatever amount and whatever form it takes.

I don’t wish to dominate this thread any more than I already have, so I’ll take a rest. Please do continue the dialogue.

I’m not sure this discussion has contributed much to the many discussions that have been had regarding a reputation system, people have restated their opinions and it’s gotten heated here so I’m going to close this.

As has been discussed in the Google Group, if iNat ever does develop a reputation system, it will be based on someone’s iNat use and not outside factors.

15 Likes