Rename "Research Grade"? (discussion and polls)

I’m not sure there’s actually such thing as scientificly stringent accuracy. Where does that start? Field ID by a field tech? Field ID by an expert in said Taxa? A herbarium specimen? Genetic data? Most data used for scientific studies is closer to the start of that gradient than the end.

@charlie however deep you want to go with that…the point is that people do not oversimplify the identification of organisms and inadvertently minimize the role of science/trained professionals by implying a false equivalency.

you aren’t the only one who posted similar sentiments so i don’t mean to pick on your specifically… but i just don’t agree with that being an actual issue. in terms of accuracy for plant data in California or Vermont, aside from a few urban areas contaminated by ‘duress user overload’… the so called ‘research grade’ data on iNat appears equivalent in error rate to the normal baseline data most scientific studies and land management agencies use… with the added bonus that there are photos and other info, so as an ecologist i can decide for myself, and if it seems wrong, disagree with the ID. I understand the issue with the term ‘research grade’ and think changing it is a good idea, but also don’t think there’s a need to minimize the value of inat data. But maybe for other taxa or geographic areas it’s of lower quality. I don’t know.

3 Likes

I like the idea that a hover or click could reveal more information about RG/CID/whatever it’s called – something similar to the info for endangered status.

For endangered status:

Crappy mock-up for ID status:

And yes, you can find this in the “About community taxa” pop-up, but that can be way overwhelming for someone who just wants to know what RG means, not the entire algorithm for calculating it or the reasoning behind it.

3 Likes

remember that currently, RG means a whole lot more besides “2/3 identifiers agree”. Also, I very much disagree with the casual obs being labelled “no identification required” or some such. Often they are uploaded because the observer wants an ID.

9 Likes

well, along with ‘no identification required’ i would remove captive-cultivated observations from ‘casual’… if that is what you are getting at. Though i don’t feel strongly about that wording either.

Just a note that the word “consensus” doesn’t mean decided by majority vote (or more like 2/3 vote required here), but means basically everyone agrees (maybe the odd dissenting voice).

1 Like

That’s always been my perception of the term “consensus” too. As noted earlier in the thread, perceptions vary on that term, which is another reason not to use it.

1 Like

I did actually do a page search before posting my comment, but I think I must have been stricken by a case of alternative factsspellings, because I didn’t spot that :-)

2 Likes

I thought that too but then everyone here was using it differently.

1 Like

As long as we’re obsessing here over spellings and definitions (which I dearly love to do), I’d observe that the current front-runner term “(Community) Identified” can be taken as “The Community Identified this!” implying it’s correct and everyone agrees. That’s more confusing even than Research Grade, I think, especially for our many members who aren’t nuanced in English.

Oh gosh I hadn’t even thought about all the languages and how to say this in all of them. It’s so hard even to figure out one.

1 Like

In reading through the comments I believe there is consensus that the term “Research Grade” does need to be changed and that a lot of creative suggestions have been put forth and I really like the idea of polling us. But ultimately the iNaturalist staff is going to have to grapple with this issue. What to me is clear is it indeed needs to be grappled with.

2 Likes

I noticed currently the topic title is ‘Rename “Research Grade” to “Identified”’, which is different from before if I remember correctly. I don’t think “Identified” is ideal because usually the original poster of an observation already identified it as something when they posted. So adding that label is kind of meaningless in my opinion. The observation is “identified” right from the start and continues to be “identified” whether it’s captive or has a bad location, etc.

I voted “(Community) Identified” because “Community Identified” seemed like the best label to me. But I do have an understanding of how Research Grade as currently defined works, which shifts my perception.

If that gives the impression that the entire iNaturalist community agrees with the current identification (which what I intepret @janetwright’s comment to mean?), then I think something like “Majority Agreement” would be better (although it doesn’t technically align with the 2/3 rule…).

1 Like

Ok, I’ve put up a 2nd poll, for “Identified” vs. “Community Identified”, and edited the title again to reduce bias on the polls. Speaking of bias, the poll is, of course, quite biased: only the people on the forum and interested enough to read this thread have voted, meaning about 25 out of 450 users on the forum out of 65,000 active users on iNaturalist. The main function is to summarize the opinions of those who have an interest, which has been useful in keeping this thread on track.

Personally, I think the best option would be “Identified” with a tooltip or a help icon or something which would explain the > 2/3rds agree rule and the various data quality issues which can move an observation into “Casual” status (or “Doesn’t Need ID” status). I guess I’m in favour of keeping it as short, simple, and understandable as reasonably possible for new people, but providing all the details if they ever want to look into it.

In the end, no name we’ve come up with has been completely satisfying. Part of that is that several things are mixed up in “Research Grade”: > 2/3rds of identifications agreeing, no data quality issues, removal from “Needs ID”, and inclusion in data being exported. I think the majority here agrees that the identification threshold and removal from “Needs ID” are the most important part, and that that should be as clear as possible in the name. There’s not enough space in a short label to say everything! Another unsatisfying thing is that no name comes completely free of implications which don’t quite fit what we want to communicate. At best we can attempt a neutral name that doesn’t imply too much, and take steps to provide more thorough information when anyone cares to look.

4 Likes

I’ve been following this thread and wanted to add to the conversation from a data aggregator perspective (taxon specific).

I know from watching Apodea that “needs ID” gets expert attention. Maybe adjusting the algorithm so that “needs ID” is treated differently when searching for a location versus taxon perspective. “Differently” in that it is put to the top of the feed in locality based searches for a LIMITED time.

I like the concept of Research Grade. However, Most of the folks that I’ve presented iNat-GBIF (bee data) immediately questioned the data. Usually, on the basis that even with a key and synoptic set identification to species is extremely difficult.
** Downloading data from GBIF

However, there is immense value of even generic IDs being flagged as high quality or research-grade IDs. Exporting this data to GBIF Is useful . As example, most bee-experts can identify at the generic level using event metadata and the overall gestalt of the bee.

I’m suggesting that after a bit of back and forth “research-grade” (or new equivalent) is applied to the lowest consensus determination/taxon level for export. I have no suggestions as to the the power of a dissenting determination - it seems to be working as it should by removing species level determinations from research grade. Mostly by Pied Piper determiners reacting too fast with “agreed”. Incorrect determinations are just part of every data set and as long as there’s a way back I think iNaturalist is doing its part.

Finally, having that brief conversation with the other determiners on the observation page is essential. Kudos, there.

I’m in agreement with most of the proposed changes. Cheers,H

4 Likes

As suggested by @cmcheatle, I prefer the term ‘Community Consensus’ or ‘Community Consensus ID’ to any term that implies that the identification is verified.

Observations with two incorrect IDs are very common, and often result from newer users and their friends agreeing with each other. These are currently marked ‘Research Grade’, and any term implying a verified observation is misleading.

Excuse me. Can we step back? Isn’t it silly to debate the slightest of nuances in the English language on a global website?
(This is not directed at any person or reply, just at the topic under play here)

2 Likes

i don’t think so, i don’t have the stats but English is the most commonly used language on iNat as far as I know. I think instead, we should just make sure to leave space so those who speak different languages can also discuss issues in their languages… but for this particular message board, English is the default language being used. I think it might be worth figuring out mechanisms to have message boards using other languages, or maybe some sort of mutually agreed upon translation policy so we can all talk. But my guess is that the other languages have various translations of ‘research grade’ which might make it even worse due to any translation issues…

As Charlie says, and also, other languages will show their translation of what is used on the English version of the site. They might lose the nuances, as you say, but the English version won’t!

I can use google to translate a page in-situ, and I am sure other languages/countries can do similar…