Require 2 "No" votes in Can the Community Taxon still be confirmed or improved? to turn an observation casual.

In the Data Quality Assessment, one vote of “No, it’s as good as it can be” in "Based on the evidence, can the Community Taxon still be confirmed or improved? " is enough to turn an observation casual. This vote is hidden away at the very bottom of the page and the observer is not notified when someone uses it, thus most people will never find out their old observations that they’re still hoping for an ID for are now casual grade.

I have seen too many people use this whenever they cannot further improve the ID, presumably under the assumption that if they can’t, no one can.
I’ve also seen people use it after they’ve added a species-specific ID, which probably means they don’t realize this function turns the observation casual, because obviously they thought the observation could be ID’ed.

To prevent abuse and accidental misuse of this feature, I’d like to see a requirement of 2 No votes, to prevent one person from having the power of turning hundreds of observations casual without anyone even noticing.

I personally think it shouldn’t be allowed to be set inside the first month the observation is up for… and I never set it unless the observation is relatively unchanged for 3-6 months at least. Another possibility is to have it auto set a vote for “can’t be improved” after 12 months of being unchanged.

The revamp of the alerts system is likely to provide alerts on changes to these, so the problem will be mitigated somewhat.

4 Likes

In some cases, this may be more effective. The subject organism is not really settled for many observations at first. There are observations that become “casual” right away, while one or more identifiers are focused on a cultivated plant, for example. Then the observer chimes in after a while (sometimes quite a long while) and wants the insect identified instead. The identifiers then agree there is an insect, and identify it, but forget to mark it as “wild.”

3 Likes

I’ve just been going through European Diptera which have been left at the rank of order.
Some are years old but interesting observations worthy of identification.
Some should have been taken out years ago as they are just a little blur.

If auto-set after a certain time, I think you might lose nice observations of less attended taxa and locations. So, I’m not totally convinced - unless there was a way to search as an identifier to specifically check those auto-logged as casual, and not have to go through all the casual obs ever.

If requiring two votes, then this could also mean more blurry dots languishing for longer in the pool of Needs ID. You can balance out the “no” vote with a “Yes, this can be improved” can’t you?
Is that not sufficient in the cases you mention?

6 Likes

You can only balance out the ‘no’ vote if you actually see the observation. And if it’s been downgraded to casual, it won’t show up in your search, so you’re unlikely to ever see it to make a disagreeing vote.

4 Likes

I like this idea. I used to use this button sometimes, but after realizing that problems could happen easily I barely touch any of the buttons at the bottom of the screen - only when people upload a variety of species to a single observation do I use it.
I’ve seen others use the button and then politely comment to let the observer know that they have made that change. I think it is great but would like to see it made automatic so that there is no chance something could change to casual without your knowledge.
But having two votes needed would also be a good thing. I don’t think a single click should make or break an observation.

1 Like

Is this not also the button that turns genus-level community observations RG? I know I’ve marked a few of my own when the photos are definitely too bad to go lower than genus, but before the community has given any ID on yet, assuming that it isn’t casual until there’s a community ID…

6 Likes

personally, i think this doesn’t really address what i think is a fundamentally flawed design. the “can the… improved?” data quality metric is currently implemented at the observation level. but really, i think it needs to be implemented at the identification level. if you make, say, a genus-level identification and want to indicate that that’s the best identification that can be made, the system should allow you to check a box on your identification that would indicate that you believe that genus is the best ID that can be made. checking that box should ideally also force you to include notes in your ID about why you think an ID cannot be improved. such a genus-level ID would then effectively implicitly disagree with species-level IDs.

by default (unless you check the box), a higher-than-species IDs should always assume that a more specific ID could be made. so then keeping the same kind of existing voting system for identifications in general, you would need at least 2 genus-level IDs with the box checked to get a genus-level research grade observation.

4 Likes

Do you see this auto-vote as being able to change an observation from Needs ID to Casual?
I’ve had several Hemiptera observations sit at order-family-genus level for at least a year before someone with local knowledge joined iNaturalist and identified them. Had the observations all been shifted to Casual before then there’s a good chance that they never would have been identified.

My impression is that locations such as the USA, Canada, New Zealand, and South Africa have a healthy community of identifiers who can go through observations and handle a good chunk of them within a year. The same does not hold true in other countries, unfortunately, and I would prefer not to see iNaturalist add features that increase the difficulty of having observations identified.

3 Likes

I actually didn’t know this button makes some observations casual; I’ve only ever seen it turn them to research grade.

2 Likes

Yes, me too. The only problem I had with it was when I found an observation with 3 agreeing ID’s that was not Research Grade. I was confused but then realized someone had marked it “Yes” for “Can the Community Taxon still be confirmed or improved?” I marked it “No, it’s as good as it can be” and it immediately became Research Grade.

1 Like

I’d vote for that as a feature req, if you wanted to make one!

1 Like

Yeah me too perhaps. Though I guess this also plays into all the complexities of the disagree prompt. Maybe at least something to be considered in the revamp of that.

Okay. I just tested it. I knew that this could be used to change a genus level ID into RG. I just voted “no” on an obs that was to family and too poor quality photo to further narrow down. It changed to casual. I see now how the one vote could be a problem. When something becomes casual it kind of “disappears”. I think requiring 2 votes is a good idea. Also it makes sense for the RG thing. If you need two people to get a species to research grade, you should have to have two people to make a genus research grade.

1 Like

I may be in the minority here, but I would prefer that it be easier to use this feature, not harder (with caveats, see below). My knowledge base is strongest in birds, where I realize that this is less of a problem than with other taxa, but there are certain taxa where most observations are simply not identifiable past genus from photos (by anyone), and they “clog” up the Identify pages. Having a shortcut key to be able to click “No” in the Identify page would make it so much easier to clear all of these observations. Ideally, they would be become RG at this point.

@lappelbaum, I’ve always thought that clicking “No” would make an observation RG only if there are already TWO IDs at genus level. Anything with one ID at genus level, or anything still above genus level, will become casual after clicking “No”. Perhaps this could be changed so that any taxonomic level could be come RG after two IDs for that taxonomic level? Maybe @pisum’s suggestion of having this integrated with the individual IDs would help, but I’m guessing that this would be more complex from a database/coding perspective than we realize.

I agree that users abuse this by clicking “No” if they personally can’t identify it further, which is clearly not the proper use of the button. I’ve always thought it should be used only in cases where the observation is not identifiable to a finer scale by anyone, whether because the photo is blurry/pixelated/etc, or the taxon is impossible to identify from photographs. This seems apparent by the wording of the actual button, but perhaps the wording could be improved?

1 Like

I would like to leave it working like it is now on genus/species level, I want to have control over own observations without tagging people (as it would be with 2nd vote needed). And also voting “yes” can help with misidentified obs with many ids, when your id doesn’t change Community taxon, that way you also can make experts see it without tagging.
But I’m up for that for anything higher than genus, people making stuff casual without notification is not cool.
Actually, making a better notification system would be more effective.

2 Likes

This was noted as one of the things that will be included in the new notifications design update I think.

That’s what I hoped for after writing the post, it definitely will solve many problems!.)

1 Like

Why do these observations become casual in the first place though?? To me, a casual observation is an observation ‘missing’ something - e.g. one without a location or date, or an observation of a captive organism, or one with no organism at all. The current ‘cutoff’ for this feature is at family, so anything with a community ID below that will become RG when the box is ticked, and anything at family or above will become casual when the box is ticked. This just seems very arbitrary to me and I don’t think that it is the right way to go about it.

An example: two sightings, one of which is at genus-level and cannot be IDed further, and one of which is at family-level and cannot be IDed further. Neither would be of use to someone looking for data on individual species, but both would be useful to someone looking for data on the family as a whole. If someone was just looking for information about genera or tribes, only the first observation would be useful, but the second one would not even show up in their search in the first place. What is the point in making it casual? It is still useful information and there is nothing ‘wrong’ with it, it just can’t be identified further. What is the difference between a blurry huntsman photo that cannot be identified beyond Sparassidae, and a very slightly less blurry photo that can be assigned to a genus because the pattern or shape is slightly clearer? The second sighting will be made RG, whereas the first will be pushed back into casual. In my mind there is little difference in value between the sightings, and making the first one casual just removes some of its usefulness. Is there something inherently wrong with making it RG as well??

‘Boundaries’ of identifiability and usefulness vary drastically between taxa, and I think it would be much better to let data users decide which sightings they will find useful rather than putting an arbitrary limit of family-level.

This has bugged me for a while, and consequently I do not use the “No” button unless the sighting is of an undescribed species. Even then I sometimes avoid it, because the best taxon to use for it is at family-level or above, and making the taxon casual causes more harm than good.

Anyway, rant over :P

2 Likes

My, understanding, and I could be wrong, is that iNat doesn’t want to flood GBIF with tons of family, superfamily, order, etc. observations, so they’re marked casual instead of RG. The cutoff is somewhat arbitrary, but anyone can still use these observations, they just have to get them from iNat rather than from GBIF.

1 Like