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

[annkatrinrose]

@pisum

i’m not sure you need “a picture of” in front of “a bat sonogram” and “a drawing of the subject”. but also, since neither of these would be a different species situation, i’m not sure why you’ve included these here.

I have seen some of these (mostly drawings of organisms) being either ID’d as “human” (because a human made them) or as “no evidence of organism.” Not sure those are appropriate either but I take that as evidence that some identifiers are looking for ways to render these casual. So from there it’s not too far of a stretch to see the new DQA being misapplied to observations that contain these.

[annkatrinrose]

@loarie

If the observer does split or remove the unrelated photos, please remove your “No” vote and any disagreeing ancestor identifications.

This is good advice but how do identifiers know when it happens? I’ve gone through the lists of observations with comments to split that were posted and came across quite a few for my area where the observer did fix the problem but there has been no follow up. I was able to move a few of them to RG now. However, there is no notification unless the observer posts a comment or tags identifiers. Are there plans to add a checkbox or something to the identify interface to specifically search observations marked this way? That might help “recover” those that have been fixed quietly.

[SQFP]

@loarie

to understand what the subject was intended to be

I see (at least) three ways to solve the “what/where is the subject” issue:

1. • leave it to members to DQA ‘thumbs down’ because [they suppose/it’s obvious] there are “multiple species mixed in the same obs / same photo(s), i.e. there is no unique subject to be IDed here” → possibly confusing/ambiguous thus over- or under-used, and uploaders get zero notification so unhelpful/un-educative
2. • force people to systematically leave a short text describing the intent (“please tell us what your photos want to show, and where the subject is in each photo”) → annoying, and often useless if the subject is bleedin’ obvious
3. • provide an easy tool so that people, upon uploading pics, are strongly advised to (not ‘must’) “show” (color dot, arrow, frame, whatever) where and what the subject in their obs is - not everybody has time/skills/resources to launch a Photo Editor app for each new flower or bug they engage with. → difficult to implement reliably

[DianaStuder]

I hope the various - please don’t - will eventually have a dedicated solution.
Annotate each photo (for mommy, daddy and lion cub).
New - Annotate or DQA for habitat.
‘What are we looking at?’ can wait politely in the broken queue until (if and when) the observer responds.
Caterpillar and butterfly is DQA wrong date (that we have already).

And split the Not Wild obs from the broken and only the observer can fix them obs.

[wildwestnature]

@loarie

please remove your “No” vote and any disagreeing ancestor identifications.

What specific action does the community advocate that would satisfy removing the disagreeing ancestor ID: deleting it (which removes context) or adding a new ID at some specific taxonomic level? Thanks

[cthawley]

Yes, adding a photo from a later date to the original observation should not be done - this would not be a “proper” observation.

A second observation should be created for the adult which, if raised in captivity, should be ticked as captive. There are other threads which focus on this.

The reason to do this is because the definition of an observation is that it represents a human’s encounter with an organism at a time and location. Creating a new observation is a similar amount of time to editing an existing one. If researchers access photos or future features allow individual photos or parts of photos to be annotated, the adult photos in this scenario would have incorrect dates creating erroneous phenology data.

Edit: As noted above, observations like this will generally be DQAed:

@DianaStuder

Caterpillar and butterfly is DQA wrong date (that we have already).

In regards to

@wildwestnature

What specific action does the community advocate that would satisfy removing the disagreeing ancestor ID: deleting it (which removes context) or adding a new ID at some specific taxonomic level?

I would suggest withdrawing (not deleting) the ancestor disagreement ID as this keeps context. Alternatively, if a user is comfortable making a more specific ID on the amended observation, they can just add that which will automatically withdraw the disagreeing ID again retaining context.

[cthawley]

I think this is overall quite clear, and I like the examples! My only real suggestion for the proposed explanation is making it clear that multiple “incidental” species are allowed. In this sense, the key issue is not that there are

@loarie

two or more photos depicting different species in each

– this technically happens in most observations where there are different species in the background, that an organism is perched on, is a host/parasite of, species are congregated in some way, etc. This wording could lead to inappropriate usage of this DQA for observations that are of mixed species flocks, a multispecies patch of plants, a forest, etc.

To me, the key diagnostic issue is that:
The focal organism/species is not present in all media.

So I might propose wording like
“If you see an observation that does not have the same subject (for example, individual organism or species) present in all media, it’s best to vote…”

In regards to @spiphany’s suggestion, I would prefer keeping the field title as “Evidence related to a single subject” because it isn’t always immediately discernable what the species in an observation is/the observation may not have been IDed to species/the identifier may not know the species that is present.

Making the field title something like

@spiphany

same species is present in all media

could imply to users of the DQA field that they need to be able to identify the relevant species, which won’t be necessary in most of the uses (in which the pictures obviously show clearly different organisms).

One other consideration might be how the DQA field name/text would be translated though – perhaps “species” would be more easily understandable/interpretable across a variety of languages while “subject” would be vague? A related point was raised by @carnifex here:
https://forum.inaturalist.org/t/easy-way-to-mark-multiple-species-observations/278/194

On a tangential point, I might also add for the example:

@loarie

When there are different photos of the same individual at different points in their life history (e.g. photo 1 is an egg, photo 2 is a caterpillar, and photo 3 is an adult butterfly).

To downvote “Date is correct” as this is the current approach as @DianaStuder mentioned.

[That_Bug_Guy]

May I also ask how this affects observations including multiple infraspecies (subspecies, form, variety, etc)? Are we meant to press ‘Yes’ or ‘No’ on ‘Evidence related to a single subject’?

[pisum]

seems like there’s been a lot of good feedback so far. i wonder if it would be easier at some point to make sure all the feedback gets captured by starting to accumulate the suggestions in a wiki so that not everyone has to read every single post separately?

[SQFP]

Could someone clarify whether the “Observation of the day” from last sunday (“a nilgai and a coyote”) qualifies or not as ‘Evidence related to a single subject’?

[spiphany]

@cthawley

it isn’t always immediately discernable what the species in an observation is

I understand the hesitation about referring to a species if it is not clear what the species is. However, part of what I find potentially misleading about the current phrasing is “single subject”. This could be interpreted – and based on the responses to the new DQA is being interpreted – to mean evidence can only show a single organism (i.e., the DQA would apply to a single photo that includes multiple organisms as well as multiple photos showing different organisms). “Same species” avoids this ambiguity.

I considered and rejected “same subject” and “same organism” because this is ambiguous – it could mean the same individual, or it could mean the same taxon. If the DQA is not meant to apply to observations that include multiple individuals of the same species, then it seems reasonable to include “species” as a criterion in the DQA.

@cthawley

The focal organism/species is not present in all media.

With slight alteration, this might be a good way of formulating the DQA. I suppose “present” isn’t quite right, since there is no requirement that evidence depicts the organism itself (since tracks etc. are allowed). So maybe: “Evidence of focal organism/species in all media”?

[swampster]

Observation here: https://www.inaturalist.org/observations/199232392

@loarie

Please only vote “No” to the “Evidence related to a single subject” Data Quality Assessment condition if there are multiple photos or sounds not related to the same species (e.g. the first photo is a frog and the second photo is a flower).

@loarie

Please do not vote “No” to the “Evidence related to a single subject” Data Quality Assessment condition in these scenarios:

When there are multiple species in a photo and it’s unclear what the subject is (e.g. a photo with both a duck and a pigeon in it)

Emphasis (bold text) mine. The Nilgai is in all of the photos so it would not qualify. A separate observation was created for the coyote and it also occurred in each photo, so the DQA should not be marked for that either.

[loarie]

Thanks all for the feedback - we updated the help section here
https://help.inaturalist.org/en/support/solutions/articles/151000171680

we’re consider altering the feature (e.g. changing the name) and new features (e.g. notifications on DQA votes) separately

[SQFP]

Thanks!
From the second bullet/item, I take it that habitat pics, e.g. a broad panoramic landscape without the obs subject, or a bush where the bug was observed before it flew away, are no DQA-prone and thus allowed? How to reliably distinguish these “non-subject photos” from accidental “other species/irrelevant” photos?

[seblivia]

Having spent a lot of time recently annotating moth observations, I think the thing that would allow me to annotate them most accurately in cases like this would be if the person clarified in the description/notes that the date was for the first observation of the larvae, and that the adults are the same individual at a later date added in to allow for easier identification. In that case I’d feel comfortable annotating it as “larva” to give the correct date for that, but the image of the adult is there for reference.

(As a side note that’s what I’ve done with moths I’ve found in the garden and raised myself, but I was unsure of whether they counted as wild or captive and ended up marking them as captive anyway. The original observation was as I found it and it marked for that time and location, but the later photos are after being kept in captivity).

I’ve also had the same experience of being tripped up by pinned specimens where it’s not clear whether the date or location are actually accurate to where and when the specimen was found.

[rcavasin]

@seblivia

I think the thing that would allow me to annotate them most accurately in cases like this would be if the person clarified in the description/notes that the date was for the first observation of the larvae

Agreed, there should be some indication that the adult photos are their to corroborate the ID on the larval photo. And this was exactly the case in the example I provided. At first, we only had the larval photo to go on. It was some time later (after the adult emerged) that the observer added the adult photo to the observation with the note saying something like “see, I was right”. Yes, they could have put it in a separate observation, but not many people are good at cross referencing observations. I’ve had folks post follow up observations and simply tag me without providing a link, which then requires that I search through their observations to find the one they want me to look at.

[rcavasin]

@cthawley

The reason to do this is because the definition of an observation is that it represents a human’s encounter with an organism at a time and location.

I regard the followup photo of the adult as “supporting documentation” for the original observation of the larva. Consider a slightly different, but analogous scenario:
Observer photographs adult butterfly on a flower - photo isn’t very good, making ID problematic. A week later, they add a series of photos of the spread specimen so that we can examine it more carefully. A recent example:
https://inaturalist.ca/observations/133790528

Or perhaps the observer might add photos of dissected genitalia.
Would that not be “proper” ? In all these scenarios, the follow up photos are simply there to aid in the ID of the original observation. Presumably, there would be no objection if the photos in question were hosted elsewhere, and a link was included in the observation.

[carnifex]

now worded in the help section:
habitat shots are (still) discouraged, but not something to be addressed with the new DQA vote

[pisum]

i’m still not a huge fan of the way the last 3 bullet points in the “please don’t vote no” section are worded – not because i have a particular strong opinion about how to handle these situations, but just because you’re not explaining what you think is not great about these situations and how you would improve things. (ex. male, female, cub lion – presumably you would ideally like these to be split into separate observations so that each individual is in its own observation and so that that they can be annotated separately.)