New DQA: Does "Evidence related to a single subject" mean the same individual is in all photos?

The new DQA, as I understand it, is for an observation with multiple photos that show different species, but what about observations in which different individuals of the same species, say one photo of an adult and another with a juvenile, are presented?

Since the use of the DQA makes an observation casual, I assume it should be used sparingly, but those type of observations are a bit of a gray area, since technically all the photos in an observation should include the same individual.

5 Likes

My stance is that it should be reserved for when a user accidentally adds an unrelated photo or photos. For example, a new user who uploads an image of a frog, along with an image of a flower and an image of stick, all in the same observation as that first photo, the frog. If it’s a supplemental image being added, with related context, I don’t think it should be marked as having more than one subject, as the DQA asks. I think it should be used for instances where users are unintentionally observing multiple species in one observation, not knowing that’s improper.

4 Likes

This means, “Photo show different species; they’re not all of the same species.” The wording needs to be improved. This is intended to deal with those observations, often produced by new observers, that have a tree in the 1st photo, a bird in the 2nd, a flowering shrub in the 3rd, maybe a fish in the 4th. We haven’t had a good way to mark these and get them out of the ID queue.

13 Likes

Before, the easy way to deal with them was to mark them at the highest level visible and say “can’t be improved”. Since this essentially results in the same outcome (casual grade) I’m curious if there’s another reason for it. I agree, it’s not very clear as written. I almost used it to annotate an observation with a single photo that contained multiple species at first, before I read the unrelated update about the observation accuracy experiment.

1 Like

An observation should have a single organism at a single point in space and time as its subject.

So observations of the same organism photographed at different times (eg a organism when it was a caterpillar and a month later when it was an adult)

Or several different organisms of the same species (e.g. photo 1 of individual box turtle 1 and photo 2 of a separate individual box turtle 2 which may have a different age or sex etc)

Should both be flagged as not having a single subject and the observer should be encouraged to separate,

Otherwise tools like that leverage annotations like life stage phenology or sex etc will be muddled by these observations (eg will return caterpillars during times of year when only adults are present or photos of juveniles in response to queries for adults etc)

In the future we hope to build tools to rescue these obs from the casual bin by indicating what the single subject is (eg a bounding box around the turtle in the first photo could indicate that that the subject is only in the first photo and the second photo is not relevant, allowing the obs to be moved back out of casual) but for now these obs should be made casual unless the user can separate the single organism subjects

12 Likes

It is not always the same result.
For example, an observation with two photos, one of Convolvulus arvensis and one of Calystegia sepium was IDed at the tribe level as Convolvuleae and « can’t be improved » makes it Research Grade at tribe level.
« Evidence related to a single subject » makes it casual.

10 Likes

Would observations showcasing plant parasites or herbivores alongside the plant they’re feeding on be flagged as not having a single subject in the DQA? I often upload pictures of things like aphids with a wide shot of their host plant alongside them in the same observation (especially if I don’t know what it is or if my pictures aren’t good enough for a species level ID, just incase it can help with identifications)

I have the same question regarding fungi, will it be marked in the DQA as having multiple subjects if I upload a fungi growing on a tree and the tree itself in the same observation?

4 Likes

It was stated that habitat shots are acceptable, thus the plant can count as the habitat for the organism living on it

8 Likes

Ok cool, Thank you!

Extending this to include multiple individuals in one observation as “invalid” seems disastrous. It’s arbitrary and difficult to enforce, and I really can’t imagine it solves the intention of avoiding confusing or muddled annotations.

It is effectively pointless to make separate observations for multiple individuals occurring in the exact location (within say 100 feet) at the same time. I say effectively because some could argue there’s value in abundance data, but iNaturalist has never been useful for this.

20 Likes

The language needs to improve, but this is intended to deal with cases where each photo shows a different species. (Or the same individual at different dates, apparently, but defining the use of this feature narrowly and clearly may be better than trying to use it to solve multiple different problems.)

7 Likes

i don’t know about disastrous, but it certainly would be annoying. i put in photos of different individuals all the time in my observations because i want to show the diversity of forms within, say, a patch of wildflowers, or simply because i can’t tell the difference between ant #1 and her 1000 sisters. so it would be really annoying for someone to come through and mark all of those observations as not single subject, moving all those to the casual bucket, in an effort to encourage me to carve the photos out into different observations.

i’m sure a lot of identifiers won’t want to see, say, 5 observations, each of different individuals of the same species from the same place and time, especially if there’s no difference in any annotatable features.

i also fear that this kind of usage of this flag will encourage folks to start arbitrarily assigning annotations in cases where the intended subject is not clear. for example, if i post a photo of a mixed flock of blackbirds and just label it as blackbirds, it would already be annoying to me if someone else came and arbitrarily said that the observation is actually for that brown-headed cowbird in the corner, even when there are plenty of other kinds of blackbirds in the photo. so it would be even more annoying for someone else to come along and annotate the observation as the adult male brown-headed cowbird in the corner, even when the photo has other individuals of different sexes and ages. i would rather folks just dump it into the casual bin in that kind of situation because it has multiple subjects, but really, i would prefer that they just leave the observation alone.

that’s not what i see, based on loarie’s comments above.

the only time i’ve seen staff say this, it was always with the caveat that the subject organism needs to be in the habitat shot.

10 Likes

This is still not an ‘official’ statement (as in a help section) and also, this might rather be referring the use of the DQA vote (that this option should not be used for such cases where a habitat shot is added) rather than it is now supported to include habitat shots w/o the organism visible

2 Likes

when you say “this”, are you referring to my caveat, or your statement that habitat shots are acceptable?

Those are extra cases.
Beyond the (clear) frog - daisy - tree example we start with.

Agreed. I have observations like this one where I photograph all photogenic individuals of a plant population in a particular place and time. I usually lump them together into one observation and indicate in the comments that these were multiple plants. If they need to be separate, I predict I’d be getting a lot of “duplicate, please combine” comments on observations like this because they are basically all at the same species, location and time.

10 Likes

I’ve also already used the new DQA for “landscape” shots where there is no clear single subject. I id broadly and leave a comment about it being for the general contents of the shot. Also kosher?

Edit: Meanwhile I saw that another person had sunk a RG-level photo of one vertebrate because a different vertebrate was present alongside it in the center of the image (where id’ers had just chosen one for RG since observer had left their id blank). I presumed this was not kosher, so I countervoted. (I also left a ping on that obs inviting the voter to join this discussion.)

2 Likes

I’m hoping most people will understand that what you’re doing isn’t really what this DQA is meant for. Heck, I do that sometimes. Or observations like this https://www.inaturalist.org/observations/181605296 of mine, where I took a picture of a flock of birds and made multiple observations for the individual species.

1 Like

I think the latter should be a key qualifier. If organisms are from the same place & time and there are no differences in any annotatable features, I think it should be ok to lump them (my personal opinion, not saying that is what staff is implying). I am glad for more encouragement to split separate organisms with different annotatable features. I use sex and life stage filters regularly to find reference photos.

For example, I understand that photos of the adult can provide key clues as to the species for bird eggs and chicks. But if you have separate photos of the eggs and adult (i.e., the eggs are not present in every photo of the observation) it would great to post them as separate observations (and link them in the comments). If you don’t, one of two things happens: 1) no annotations are added because no annotations apply to both photos. This probably isn’t a big deal for things like Northern Mockingbird that have tons of photos to reference, but there isn’t a single observation annotated as egg for Gray-headed Swamphen. 2) annotations are added but the annotation doesn’t relate to both photos. Usually, I’m looking at the taxon photos when I’m looking for a reference, not the observations themselves. So if it is a taxa I’m not very familiar with, it’s very confusing if some of the reference photos are mislabeled.

2 Likes

but aren’t the system overseers the ones who decide what that DQA item is for?

loarie mentions 2 cases above that should be flagged, and the second case seems to include annkatrinrose’s example.

the new DQA doesn’t really solve the whole annotation issue. for example, suppose i make an observation for a particular individual of a plant. it has both flowers and fruit. i take one photo of the flowers, i take one photo of the fruit, i take one photo of the leaves, and i take one photo of the stem.

i legitimately annotate the observation as flowering and fruiting, but since annotations are at an observation level, if you look up flowering photos, you’ll see the 3 other photos without flowers.

loarie says that they have future plans for ways to “rescue” observations marked as multi-subject, including features which sound like they could effectively handle different annotations per photo, and that kind of thing sounds like a reasonable solution for the annotations-at-photo-level problem, but why jump the gun and use this new DQA flag to try to address the problem, even though it can’t possibly address the whole problem?

7 Likes