Most of the time I’ve spent on iNaturalist has been spent on identifying crayfishes, especially those in the family Cambaridae. I believe one of the primary reasons why this taxon has such a low research-grade to needs-ID ratio is because a good chunk of crayfish observations on iNat are documenting sightings via photos of only their burrows rather than of the live specimens themselves. In the U.S. and other places, the best that these observations can be narrowed down to is Cambaridae. This is when the issue occurs.
To me, it just feels kinda weird to only have the two options: to leave it at needs-ID, or to go to “Based on the evidence, can the Community Taxon still be confirmed or improved?” and mark “No, it’s as good as it can be,” where the observation will then be reduced to casual-grade. I mean, it’s technically true; the community ID on these observations can and will never be narrowed down to anything below Cambaridae. However, I don’t think it’s right to demote these observations in question to casual-grade just for this reason. I believe that documenting the presence of crayfish burrows is a completely just and helpful form of evidence to consider when evaluating the status and makeup of the fauna of an area and clearly indicates the presence of crayfish, even if all we know is that the species belongs to Cambaridae. But, if you choose the option of leaving the observation at needs-ID, that implies to the observer and anyone else coming across the observation that the community ID is in need of being narrowed down further, when it is the narrowest it can and will be. It also clogs up the “Identify” page with lots of observations of crayfish burrows that do not need any reassessment or reconsideration.
So, that’s why I believe that observations containing only evidence of crayfish burrows should be viable to be research grade. Thank you for reading, and if you have any thoughts, please share!
That works if the community ID is at subfamily level or lower, but not in this case, since the community ID on these observations is at family level. Take this observation for example; once you’re there, scroll down and check the “No, it’s as good as it can be” box, and then check the quality grade status of the observation again. When I do this on my end, it recategorizes the observation into casual grade.
(Edit: I was originally mistaken, so I switched “genus level or lower” to “subfamily level or lower.” Thanks to Elliot (egordon88) for pointing this out)
I don’t think the concept of research grade is to be based on a family ID. My argument here is an observation doesn’t need to be down to be species level to be considered research grade but family level is too broad. Take for example a single family of snake (Colubridae) have many genuses with different behaviors, habitats, and preys; I would say those need at least a genus ID to be able to make it even sensible to be considered research grade which Inat supposedly already has with the “No, it’s as good as it can be”.
“Research grade” really only means that the observation will be shared with GBIF if the user’s license allows. “Casual” is a catch-all for observations not eligible for this.
I believe the reason that observations at family level or above are excluded here is because the ID is not considered specific enough to be useful for most research purposes. It makes sense that other databases might have a threshold for what information they want to include – thousands of records for unspecified “crayfish” are probably of limited value for people studying species distribution or ecosystems. If the records are marked as showing burrows, they may be useful for people studying crayfish habitats or burrow construction, but most databases don’t systematically differentiate this kind of data from other types of evidence.
The issue here seems to be connected with “casual” being felt to be a mark of shame or trash bin for poor observations. I think more differentiated labelling and search options would probably go a long way towards addressing this (i.e. distinguishing between observations with missing information or other problems, non-wild observations, and difficult taxa where a fairly broad ID may indeed be all that is possible based on certain types of evidence).
I generally agree with the current taxonomic cutoff for Research Grade, as there needs to be one somewhere, and this level (subfamily or more specific qualifying for RG) seems to allow for the most useful data to be exported and less specific observation data to remain Casual. I think that there will always be taxa and/or use cases for which broader IDs may be useful for research, but it is hard for me to envision having a better system. Having the cutoff be less specific isn’t ideal for many taxa (like in @maxcafe’s example). Having different RG criteria for different taxa (one possible solution) seems like it would introduce a lot of confusion for little benefit in my opinion.
I agree that there are definitely some cases where family level data could be useful, and would point out that casual grade observations are still available for export and use from iNat itself. Not being in GBIF will certainly limit the number of researchers who will receive those data (which is part of the point of the cutoff), but researchers can definitely still access them if desired.
This is something I’ve been thinking about for a while. Does there actually need to be a cutoff somewhere??? It makes sense in theory that things which cannot be identified further from evidence would be marked as casual, but in practice this is a good example of how it’s not necessarily suitable for every case.
But more to the point, I assume the reasoning behind having them Casual rather than RG is that they would be clogging up GBIF with useless data. I don’t think this is the case though, because there just aren’t that many observations that fit the criteria. I did a quick search to see if I could just isolate those observations that are Casual because someone has ticked that box. I couldn’t quite filter out observations which have other DQA issues, but either way there’s at most 160,000 observations to which this applies. A quick look through the first few observations suggests to me that the number is actually much lower than that as well. That’s a decent number but compared to the total number of observations on iNat, it’s tiny - less than 0.01%. So if the initial reason for having these at Casual was that they’d be clogging up external databases, I think that’s entirely negligible at this point.
Observations at Needs ID are also not exported to GBIF. The main reason these become Casual is that they get removed from the default filters on the Identify page because the community feels that they don’t need more identifiers to look at at them.
When something at family level is marked as “ID can’t be improved”, not only does it get marked as casual, but it vanishes from the maps that Inat produces, even when looking at family level. For taxa that often can’t be identified from photos even to genus level (Ichneumonid wasps are one example), it would be nice for there to be a level where something can be moved out of Needs ID because it can’t be identified further, but doesn’t end up hidden away in the casual bin, along with observations that are marked as not containing an organism or having an incorrect date or location. Would it be possible to have a designation specifically for things that are marked as can’t be improved so that they are removed from the Needs ID pool, but not marked as casual? All the other DQA options imply that something is missing or inaccurate (no media, date, location or discernable organism).
Yes! Has this been proposed in the feature request? I poked around a while ago and didn’t find anything. I’m assuming it’s because my search terms were poor.
There is a significant difference in casual due to missing meta-data and casual b/c ID can’t be made at sub-family or lower.
Yes, I agree with you here, especially with your last paragraph. I think where the problem arises is when “casual grade” is associated with poor or lackluster observations. I don’t think it’s right for these crayfish burrow observations, which clearly and validly indicate the activity of a wild organism in a wild ecosystem, to be placed into the same tier as observations of, say, a captive organism at a zoo.
My take on this is that any such categorisation should depend on the number of existing research-grade observations below genus level. For geni with many confirmed species level observations, an additional genus level observation will add little. For geni with very few observations, that are mostly unidentified on species level, any additional observation is valuable. Fairly easy to implement on an algorithm level.
I’ve added a feature request for a 4th data quality category. This category is for observations with complete meta data but cannot be IDed at the sub-family level or lower. Giving it a Complete data quality means it will be moved out of Needs ID without consigning it to Casual limbo.
Complete is a great term for this level of identification. It’s as far as it can be taken with the limited data available, but it’s still a wild organism with all its meta data.
I’d want to add at least a couple other categories beyond that to split up “Casual” haha. [higher taxa and can’t be ID’d further] seems like yet another in this list:
“Casual” is a badge of shame, a trash bin. Why? Because this is where we put observations that lack date or place or picture of an organism. Photos of the moon. Useless observations. We shouldn’t pretend it’s not a trash bin.
This is also where we put captive/cultivated observations, some of which could be useful and most of which were posted by people wanting ID’s. And now we put there observations of wild organisms that simply cannot be identified further (e.g. crayfish burrows and the 10-pixel birds). So yes, there are good things in Trash – er, Casual – but they aren’t likely to be found once they’ve been tossed there.
Some day, I’d like to see some splitting up of Casual, and also improvement in the way we deal with observations that haven’t been identified as well as they can be. (There’s no equivalent of “Needs ID” vs. “RG” in Casual, so if you try to ID observations there, you have to deal with the ones that already have six ID’s.)
And to get closer to the original topic . . . . I think Families are too broad a category to qualify for RG. However, I think it might be valuable to make some exceptions, e.g. for crayfish burrows. That’s isn’t going to happen in the foreseeable future, though, so in the meantime, setting up a project for the burrows is probably the best way to keep them available to researchers – or rather, to the small subset of researchers who will learn that there is a Crayfish Burrows project.