Concerns over opting out of community taxon

I’ve never been tempted to use “opt out.” In real life, I’m an expert in certain plants. In fact, there have been times when maybe only 2 or 3 other people in the world were better at IDing a given species that I was and they weren’t on iNaturalist. (It’s no longer true; more and more people learn these plants and more of them are on iNaturalist.) But I never felt a need to avoid all other ID’s. Maybe one reason is simply that many of the plants I’m an expert in are of little interest to most people, so there are few other ID’s to outvote, and at least the few wrong ID’s don’t send the observation to RG with the wrong ID.

I’ve “unfollowed” one person who opted out and didn’t change ID’s even when he was unquestionably wrong, and I’d do that sort of thing again. However, whether the observation has an “opt out” isn’t obvious when using the identify tool, so that’s usually not a basis for choosing to ID or not.

I think it would be wrong to send an observation with “opt out” to Casual (using “No, it can’t be improved”) if I’m putting the first disagreeing ID on an observation!* However, I think it’s totally justified for those older observations that stay Needs ID because they’ve got 5 or 6 agreeing, correct ID’s but the observer had put a different, wrong name on it, opted out, and doesn’t change anything.

  • Except dandelions.

You could be right, simply because so few people opt out for all their observations. I ID a lot. Sometimes I find that I don’t know how to distinguish species I thought I did. I try to learn how to ID them correctly or learn not to put names on them. And sometimes the taxon concepts I used (for good reasons) are not the ones that were accepted in the end. I hope I haven’t done a lot of damage, but I know I’ve done some, and have tried to correct some of that. (Memory of spending a day or two trying to undo ID errors certainly encourages one to avoid getting into that problem again.)

[Edit: Finally there’s some concensus on that The Blackberries That Ate The Pacific Northwest are, and they’re not what I called them. Many (possibly a majority) aren’t the other name most commonly applied to them, either. I’m intimidated by the numbers involved! Maybe this winter.]

What is your evidence for this? I ask because there was some frustrating discussion on a different thread with a person who thought that the symbol indicating the CV had been used meant that it was used for the identification. Like a lot of people who do many identifications, I click on a CV name if I know it is right because that’s a lot faster than typing it in. I’m impressed with how often the CV is right, but also with how laughably wrong it is at times.

I do think using the CV can create some problems. For one thing, there are the misclicks! Olympic Grasshopper when I meant Olympic Gull, for example. More seriously, there are edge cases, where I think I know what it is but I’m not sure. Can the CV ID tip me over to adding the ID when I’m not confident? Yes. I try to resist that, but my rate of resistance varies. In a data quality sense, the misclicks are too rare to be more than a trivial problem. The other can be more consistent and therefore more of an issue. Nonetheless, this is a limited problem because most of us “power identifiers” ID mostly species we know well, and the CV does not contribute to those ID’s.

10 Likes

Is this the - I can typing, I can spell Leucospermum hyophyllocarpodendron hypophyllocarpodendron - argument against an ID that ‘used CV’
?
Lots of previous threads about that.

3 Likes

For most purposes it seems methodologically bizarre to me, to exclude RG data points from a dataset not because they are erroneous (as can be the obs by opted-in users, due to The Community agreeing on a wrong ID), but rather due to a dislike of a setting checked in the observer’s profile. Sure, no feelings get hurt along the way, and your protocol your choice.

On the other hand, some proposals for a site-wide permanent policy of hiding from view (largely barring community interactions) these people who choose to opt-out from a “submit to the majority” policy… seems a bit sad, at least coming from a website supposedly valuing social/community aspects. Prejudice against opted-out users is a thing here, and I now come to think that a removal of any and all option to “opt-out of what Our Users deem right” (CIDs, taxon changes, etc.) would save these users much trouble. Idea: upon signing-up on iNat, a refusal to submit to the rule “conformism is what we value, dissent gets the red stamp of infamy” could redirect those people “unwilling to submit to the majority” to a list of other platforms playing by different rules (there are surprisingly many, taxa- or area-specific, some worldwide e.g. ObservationDotOrg).

4 Likes

So it is a users choice to opt out of a community ID for whatever reason but not an IDers choice to not invest time on such users? That is weird…

Applying a filter setting that makes it easier to avoid opt outs does not forbit opting out but allows to better coexist on the webite without those annoyances on both sides

8 Likes

There’s never any reason to give for not identifying or rejecting others’ id or refusing engagement. If it is a site-wide policy enforced to prevent users to invest time on OptOuts, better not allow OptOuts in the first place, that’s all :)

However, I expected that if (when) a reason is given, it would be more valued and valuable if (reasonably) sound or sensible, and not (too) disparaging - maybe not? :)

Well.. with casuals it is the same situation actually. They are allowed to upload.. however, one can exclude them from ones identify set.

8 Likes

Call me crazy. But i think the more ways to filter, exclude, include, observations while searching is better. Being able to sort by geomodel score, RG, needs ID, casual, location, date uploaded, updated, observed, etc is all great. You can even search by photo copyright.

So i see no reason we shouldn’t be able to differentiate obs with opt out.

The more tools we have to filter, the better. Theres already many search URLs you can mess with to really craft strategies for finding just what you want. Or atleast to help you do that.

The list of ways to currently filter observations really is very large. Theres many things you already can do that many probably dont realize.

8 Likes

I notice there seems to be a few comments about not wanting to see opt-outs at all. The only opt-outs that bother me are the ones where the community id disagrees with the observer’s id and the observer has passed away.

I understand there is no effective means for iNat to declare someone dead, so that cannot be part of any search criteria.
I understand there are cases where the observer is an expert and the identifiers are identifying beyond what is actually possible.
I will also add that my own interaction with people who use opt-out has generally been fine and in many cases the observer has changed their id after I provided input. In fact, I can remember only one case where an opt-out user gave me the “I know what I’m doing” response.

So my main request is that it be possible to easily see or unsee the observations where the observer uses the opt-out and the community id disagrees with the observer’s id. The only current way to figure this out is to notice three different lines of text that may be spaced far apart on the browser window: “User has opted-out of Community Taxon”, the community taxon just below that, and the display taxon at the top. From a user’s perspective, that’s hard to use.

A couple other questions:

How do these observations affect the computer suggestions especially: “Expected Nearby”?

To add to @zoology123, is it currently possible to search for cases where the user opts-out and their id disagrees with the community id?

3 Likes

I dont believe so, but there are like over 30 different URL conditions if not double that or even more. I have not looked through them all. I have only looked at what i deem useful for myself in IDing Chironomids. One example isnt even a filter, but per_page=200 is a god send in the identfy tab.

Either way, one still has tons of abilities to exclude seeing peoples observations in so many different ways. Some even quite niche.

Correct me if im wrong, i think you can even search by hour of the day now?

Edit. It would be so cool if inaturalist could implement elevation search. Would it be perfect? No. But take a global heightmap and take the coordinates, you should get the elevation. Interestingly the geomodel already includes elevation data in part of its process.

I’m personally not against filtering for observations where the observer has opted out of community ID, but identifying them doesn’t bother me in the slightest. I understand my perspective is different from most peoples’ in that I’m a paid staff member so using and contibuting to iNaturalist is part of my job, but I’m happy to add any ID to any observation, regardless of whether or not they’ve opted out.

I’ve added my knowledge and my input and maybe helped out anyone else who takess a look at the observation. Maybe it will sway the observer, maybe it won’t. Maybe my ID will be part of a group of IDs that makes the observer’s ID Maverick and thus makes the observation casual grade. Maybe someone who wants to use the observation in their research will see that someone disagreed with the the observer’s ID and will scrutinize the observation a bit more. Either way, I don’t view it as a waste of my time.

FWIW “maverick” isn’t meant to be a judgement or a negative term, just a neutral term denoting that the ID is not in agreement with the rest of the community and may deserve more scrutiny (either because it’s right or it’s wrong).

You shouldn’t need to vote in the DQA here. If the obsever’s ID is maverick then the observation should be casual regardless.

5 Likes

I have my doubts about that, since I have run into “opted out” observations with several correct ID’s (but the observer’s was wrong) when just doing regular identifying of older observations in “Needs ID.”

3 Likes

This has been requested and declined in the past, but there is this:
https://forum.inaturalist.org/t/you-can-now-access-elevation-and-other-details-from-macrostrat-for-an-observations-location/39066

1 Like

Actually terrible to hear, but thanks. That makes little sense to me seeing inaturalist literally uses elevation in the training of all the taxa for the geomodel. So the observations in the geomodel have to be given an elevation data point for that to even work. Otherwise you couldn’t factor in elevation to the geomodel.

This is why Diamesinae (a chironomid subfamily) currently has a geomodel centered largely around hills and mountains. Becuase thats where most observations are, high elevation points along mountains.

So again inaturalist literally already uses elevation data for millions of observations.

1 Like

The difference is that the Geomodel is created outside “production” iNaturalist, using data exported from iNaturalist. It therefore does not require restructuring of the iNat apps themselves, and does not create additional load on app performance by needing to query the appropriate DEM (if available) for the particular part of the globe, then re-index, each time an observation is mapped or remapped.

At least that’s my understanding of why the developers have decided not to invest the time and resources to pursue in-app creation/storage of elevation data, for what would probably be pretty niche use cases that could be addressed in other ways.

P.S. if you want to continue a discussion on elevation, I will move this side conversation to one of the other existing topics for that, so we don’t distract further from this topic.

1 Like

Me too. I wonder if that is a recent change?

1 Like

And this is fine, and I’m 100% for more filters. And more ability to hide and mute (categories of) users according to one’s very own subjective criteria.

Interestingly, the way “casuals” are currently dealt with by default by the platform also comes with its fair share of criticisms, issues, suggestions for improvements, feature requests… (because: mixed bag/vague definition, automatic votes as captive-cultivated, site-wide policy of hiding-by-default, etc.). If not removing the option completely, shifting to a site-wide policy of hiding (or casualizing, or any other “devaluing” action) opt-outs by default should ideally be given some deep thought, and announced clearly.

I realize most users may take ‘maverick’ in a non-judgmental way, at least consciously. Still, there’s the red color (implying danger or error, in some cultures), and the absence of a stamp whenever a “conformist” clicks ‘Agree’. Involuntary bias or deliberate policy – hopefully this is not enough to nudge users and have them be overly cautious before entering a dissenting ID, or even renounce (for fear of ‘going against the grain’).

edit: messed the quotes

1 Like

I had to write a journal post for my Pre-Maverick project
https://www.inaturalist.org/journal/dianastuder/77588-meet-the-pre-mavericks

5 people PMed me. Hurt, startled, a little angry to be labelled Maverick? ‘Why pick on me, what did I do??’

2 Likes

Couldn’t agree more. Thank you!

IMO, people should not be correcting observations without some explanation as to how they arrived at their disagreement. I’ve literally had people kick observations up to Kingdom, when even the family is obvious (e.g. Asteraceae). Then they ghost you, and there’s no other recourse besides opting out but to delete the observation, along with the bad correction, and try reposting it. (Ain’t nobody got time for that!). ;)

This is supposed to be as fun and interesting as it is accurate. Otherwise, folks are going to dump the platform out of frustration and we all lose the citizen science observations, which is essentially why iNat was created in the first place.

2 Likes

I deal with too many misidentifications to do that. Too time consuming, if one wants to know they can ask. Though many times they may not actually understand my explanation without me taking timr to specifically think how to write it out in “english” for somebody that doesnt know the anatomy to understand.

Like who understands, MCu not showing? Or the inferior volsella are pin shaped.

5 Likes