Changes to the Agree button/functionality and addition of Markdown in comments and ID Remarks

Just to clarify, like pretty much any functionality, Constructive feedback is welcome and the functionality is subject to change.

8 Likes

As a relatively active identifier (who follows many other observers and subscribes to many different taxa and places), I have spent today watching to see how the loss of the Agree button impacts my workflow and habits. The actual impact of the button change is less than I might have predicted. I still add some agreements (manually) to RG observations, when it feels important to do so, but mostly I just skip over more RG observations than I used to. If this leaves me a little more time to review other observations that Need ID, maybe that’s a good thing.

However, after a day of working in this new system, I do find that my entire outlook on iNaturalist has changed, not so much because of the change in mechanics, but because I now understand better what iNat management considers to be important. I have been pondering Ken-ichi’s comment (#14 in the string of comments under his original blog post), that what he values is not the number of IDs, but the number of improving IDs that one makes within a taxon. Probably I should have realized this earlier. If I were to take this new understanding to its logical extreme (which I probably won’t), it would mean that:

  • I should spend more time out in the field making new observations, and less time reviewing other people’s observations. When I submit my own observation, I always get an “Improving” for my own ID, whereas when I confirm someone else’s ID on their observation, I am not “Improving” anything.
  • Never mind those 2nd, 3rd, or 4th agreements. Adding even the first agreeing ID, the one that moves an observation to RG, is not valued. Better to spend time finding observations with mistaken IDs and correcting them, or moving something from “Unknown” to “Insect” or from Family to Genus, than spend time confirming correct IDs.

I’m still processing this change of world view, and feel a bit unsettled. Even if the Agree button eventually gets restored, I probably won’t go back to exacctly the way I was doing things before, because now I realize that I could be doing more valuable work in a different way.

12 Likes

from a recent example posted:

image

My understanding:

Explicit disagreement should only be undertaken when there is evidence that it is not that species, or when the challenged identifier is not responsive on the matter and there would be general concensus amongst the active community to “bump” the observation back.

Here we can’t tell what the situation is. Does the challenger see evidence to support it NOT being that species? The explicit disagreement without any sort of context would imply so, however I think it is considerate to give an indication as to why one is explicitly disagreeing. But I have done so myself from time to time, when I just think heck, I can’t be bothered explaining this for the 100th time… especially when I know there are 4 or 5 other active identifiers that are going to review this after me and will in all likelihood (ie from past experience) put the same ID as myself. In other words, that 5 of us are putting this ID is as good an explanation as any I could make with words… and if actual learning is wanted, they can ask and I (or one of the other reviewers) will give a friendly enough explanation.

Let’s assume that it is a simple case of “not enough evidence to support the finer ID”. Let’s assume also that the observer and/or first identifier didn’t see or take into account any other information as a basis for their ID, remembering that observers can see or hear or experience other forms of evidence than is expressed in the photo. Let’s also assume that …

… actually, let’s just forget all of the issues related to explicit disagreements, and focus on the “relevant to this thread” issue of what the Agree buttons are now doing… Someone has explicitly disagreed with the finer species level ID (for whatever their reason), and that first and challenged finer ID has an available Agree button while the subsequent and “in response” explicit disagreement does not. The observation is bumped back to genus, goes back into the Needs ID pool where there is the high chance of new and inexperienced users continuing to exhibit the same problematic “click the Agree because I like the ID, am thankful for the ID, or just plain want the number in my stats to be bigger” behaviour. The challenger will presumably get notifications of differing subsequent IDs, so will hopefully get to re-weigh back in with a more assertive (ie with reason via comment) position.

1 Like

Another idea: have leader boards display percentage of each ID type. And differentiate agreement of a RG observation from an agreement that boosts something to RG. That way if someone really only verifies easy RG observations of mallards or whatever it’s really evident on there and they look absurd

11 Likes

Just got one of these on one of my observations. https://www.inaturalist.org/observations/53360773

first, i want to thank identifiers for identifying. part of the magic of iNaturalist is that the observation you submit might be identified by any one of 150,000 people from all around the world, many of whom are the best at what they do, and who all do it on a voluntary basis. thanks for your hard work.

i’ve done around 5000 identifications for others. so i don’t do anywhere near the volume of identifications that many others do. i personally don’t imagine that i’ll feel much of an impact to my own workflow because of the loss of an Agree button in some cases, but i can calculate how an extra 2 seconds each over, say, 10000 identifications can easily translate to an extra 5 and a half hours, and i can sympathize with what that means for super identifiers.

i get a feeling that many identifiers already feel overwhelmed with the volume of observations that are out there to identify. i admire the dedication to a cause and personal responsibility to take on the task of overseeing an entire taxon or an entire place. but in the grand scheme of things, with the amount of observations growing as fast as it is, i think that taking on that kind of personal burden will become untenable at some point, even if the Agree button is preserved. if you’re thinking about to new workflows, you may just also want to think about that larger picture.

for now, i would urge folks to just continue to identify as best as you can. if you would have identified 10000 observations at 5 seconds per identification in the past, and now it takes 7 seconds per identification, then just do 7000 identifications. or even better, do 3500 identifications with half the time, and spend the other half of the time mentoring others to be better identifiers. (ultimately, i think sharing the identification burden is really what’s needed to tackle the growth in observations in the long run.) or just take a break, if that helps to recharge. this is supposed to be enjoyable. so please enjoy what you’re doing.

i get the sense that some people feel disrespected by the change or perhaps the lack of communication in advance of the change. (how could they make such a change that has such an impact on a core thing that i do in the system without at least getting my input? if it wasn’t worth anyone’s time to find out how i work, then obviously all the time i spend isn’t valued very much, right?)

i think the right way to express such frustration is not simply to quit and burn bridges or to issue a “roll back the change or else” ultimatum. i think you’ve got to express your points in a way that you can make sure they get heard, and at least allow for a response. for example, i don’t think it would be unreasonable to ask that the points in this discussion here and on the journal blog be summarized into a list of concerns and proposals for action, with arguments for and against each item, and then request that iNat respond to each item. from that, everyone will get a better understanding of the situation, i think.

it may also be worth talking about ways to prevent surprise changes in the future. as iNaturalist’s user base grows and diversifies, certain wide-ranging changes may deserve a little more input and consideration before they are implemented. certainly you wouldn’t want analysis paralysis, nor would you want to have a vocal minority to drive all action. you also wouldn’t want to add too much administrative burden to the small iNaturalist staff. but there’s got to be a good balance between discussion up front and dealing with complaints and inquiries on the back end. obviously since iNaturalist is a private entity, they can do whatever they want, but maybe they might find that having a structured process for previewing and discussing big changes might be useful to at least ensure folks aren’t surprised by changes in the future? (the way some governments publish a periodic register and hold public input meetings might serve as a model.)

as for the merits of the Agree button, i’m personally ambivalent about it in general. yes, it might help speed up the process of identifying for some folks in some cases, but is the Agree button really the thing you dream about at the end of the day? really, i think most super identifiers just want a fast way to make identifications, right? the Agree button just happens to be one of the fastest ways to make IDs on certain observations right now, but i bet most would be willing to part with it if there were other ways to make IDs just as fast (or faster).

just as an example, this time of the year, i’ll often look at observations that have been identified as Rudbeckia amplexicaulis, and i’ll end up invariably doing one of 3 sets of actions in 95% of cases:

  1. ID as Rudbeckia sect. Rudbeckia + comment “probably R. hirta or something similar. R. amplexicaulis would not have hairy green parts like this.” + annotate as flowering
  2. ID as Ratibida columnifera + annotate as flowering
  3. ID as R. amplexicaulis + annotate as flowering

so instead of having an Agree button (especially since i’m often the first to make a disagreement), i would much rather be able to press 1 to do #1 above, press 2 to do #2, and press 3 to do #3. that would save way more time for me than having an Agree button, i think. so i’d encourage others fighting for the return of the Agree button to think about and be open to other (possibly better) alternatives.

the need to keep an open mind to alternatives also applies to some of the other considerations that were brought up. for example, folks said they use Agree as sort of a social lubricant or as thanks for an identification. sometimes i think that a little heart button on an identification would be equally effective, with fewer unpleasant side effects.

my point is that just as we don’t want to make changes that cause us to go backward, we don’t want to go back to simply go back to the way things were, if the same level of effort could take us forward.

and with that, just as i thanked identifiers at the beginning of the post, i’d like to thank iNaturalist staff. i know you all have to balance many priorities, and i know you have to make unpopular decisions sometimes. on balance, i think you guys are doing a great job. thanks for your hard work.

20 Likes

As to this, I will give not a presumptive example, but real cases from my own experience. There are OBs, research grade, with four or more IDs - erroneous. And the case is not Gerald the muskrat. The cases are lichens, for which IDers are very few, especially outside US. So, I add correct ID for the named observation.Guess, what happens then? Especially when previous IDers have moved on or are just irresponsive?

1 Like

Interesting to hear the perspectives of my fellow identifiers.
Personally, I have mixed feelings.

I spend a lot of time in Unknowns where this change has no effect. So no comments on that specifically.
But I also spend a lot of time identifying more difficult taxa. As others have said, I’m looking for the occasional mistake but may as well add an ID to confirm a correct CID.

I usually find there are two types of observations I come across in an identifying session-- those that are obvious to me, and those that require careful thought.

For the obvious ones, I only have to take a few seconds. Removing the Agree button by default is a good idea in my eyes, but removing the ability to quickly confirm an ID easily doubles the time these IDs take. Sure, I could just mark reviewed and move on, but in many cases there are only two IDs, one by a specialist and one from a novice user (auto-agreeing or else using AI suggestion) and adding my ID really does make a difference in showing that someone knowledgeable is verifying the observations. But for many taxa, adding the third or fourth ID is pretty pointless in my opinion.

For the slower ones, it usually takes between two and fifteen minutes. For now, if it takes me longer, I skip it and cone back later when I have studied the taxon some more. Adding five seconds to type in the taxon name in this case is no big deal.
I would argue that most expert identifiers should be focusing on these sorts of IDs – ones that are not easy. Take a bit of time to tutor or create resources for new users and let them take on the easy ones.
The only big problem with this is that there is no way to filter by observations that are difficult to ID, only by taxa that are difficult.

edit:
In short:
if the taxa you are identifying consistently only take three seconds to confirm, maybe you should spend your time on something else.
BUT for the sake of reviewing taxa of mixed difficulty, the ability to quickly agree should absolutely be preserved.
When it comes to really difficult taxa, the extra few seconds makes no difference.

7 Likes

Please remember some of your users live in rural, less connected parts of the world. On bad days with our internet, I can fill in a few letters in the search box, then wait (go off make a cup of tea) while the drop-down appears (if it doesn’t time out) before I can click on the relevant taxon. With the agree button, once I have clicked, I can leave the tab open and move onto a new identification, which has been slowly loading in the background. “little” is rather subjective.

13 Likes

That is unfortunate, of course. But I still do not get why should you add 3rd or 4th ID to the RG observation which is, in your opinion, correct. Even with the Agree button, it is still seems waste of time.

I do not support this change. I understand why it was made, but I think it discourages participation, and thereby reduces the possibility of education on conservation and environmental issues. I thought (and by reading other posts I found I must have been ignorant of the purpose of IDs) that the idea was not just to track what species occur where, but to encourage people to virtually “explore” this natural world. Agreeing with observations made that exploration interactive. By making iNat as interactive as possible - and face it, searching through and identifying observations is also fun and an enjoyable pasttime - iNat encourages people to search through observations and thereby learn more about nature, and perhaps thereby more aware of the need for conservation.

I’m also curious the impact upon projects. Perhaps I’m again ignorant of the “true purpose” of projects and indentifications, but I created the Okefenokee Photography Project because I want to highlight this beautiful National Wildlife Refuge. I want people to come to the project, make id’s, and be aware what animals and plants live in the swamp. That way, if those beloved species are ever threatened, as they can be (see https://www.inaturalist.org/projects/okefenokee-national-wildlife-refuge/journal/35780-okefenokee-mining-proposal-hearing-frustrates-participants) they will have built an appreciation and possibly participate in the political realm to protect it. Now, since I often quickly ID observations that occur in the Okefenokee NWR, other people will be discouraged from coming to those observations since they are already “research grade” and there “no point” in adding an agreeing idenfification.

I feel the removal of the agree button discourages participation, will decrease the amount of people that “virtually explore” our natural world, and will have much bigger negative ramifications on conservation (just because people are annoyed by leaderboard hounds???) I think we should encourage people to participate in as many ways as possible. Even if they can’t outdoors to explore, they are becoming more aware of our planet through virtual exploration, and that should be encouraged. (And I don’t even get into the huge numbers of observations that are “research grade” based upon 1 supporting ID, but are actually incorrect.) Thanks! William

9 Likes

See my post below. I think there is value in adding a 3rd or 4th ID. It is a way for people to explore observations and learn more about our natural world. The agreeing IDs are a way to track your explorations (yes, I suppose you could hit “reviewed” as opposed to “agree”). Plus, there are so many incorrect research grade observations based upon 1 ID. Yesterday I saw a “research grade” Little Blue Heron that was actually a Great Blue Heron. Removing this feature discourages searching observations that are already research grade and could thereby cut down review of incorrectly identified observations.

5 Likes

Fully aggree with that statement!

5 Likes

I use the ‘favourite’ button for this. If one ‘favourites’ an observation they’re notified on any subsequent activity on that observation.

1 Like

This doesn’t seem like a logical conclusion. If the site wants to encourage more Improving IDs, that implies more Supporting IDs are required as well. A Leading ID needs some Supporting IDs before it can become an Improving ID. More to the point, to be genuinely improving, it needs genuine support - not just auto-agreement without due consideration. The big problem for iNat is how to distinguish the former from the latter.

1 Like

How would it discourage? I check all unreviewed observations (time allowing), except that I do not add ‘agree’ if they are correct. i do not see any value in adding my 50 cents there. I do not care about the numbers of IDs I do,only about correctness of the IDs.Ithink, this should be the aim of the IDers. Again, education comes with correct IDs and comments, not by the number of IDs.

7 Likes

I wonder, how you will achieve this with some groups of organisms for which there are only very few IDers?

1 Like

I was interested in getting an idea of how folks as a group are reacting to this change, so I calculated some basic summary stats on the commenters on the post announcing the change. Current as of about 7 am EST 7/19/20. I used this post rather than the forum because a lot of folks aren’t on the forum, and many people posting on the main page post are also here in the forum.

For each poster, I went through and determined whether they expressed support for the change, were opposed to the change, or either didn’t express a clear opinion or commented on something else (like the formatting changes for text, which have gotten a little lost in all of this). Most cases were pretty clear on whether for/against, but I did have to make a few judgment calls. I also pulled the # of observations, IDs, and curator status of each poster and calculated a ratio of Obs/IDs for each poster:

Agree Change Total N Curator N Curator % Mean Obs Mean ID Mean O/I Ratio
Pro 11 5 0.45 8129 21410 0.62
Against 42 15 0.36 11670 57722 0.54

I think that there are a couple things that stick out to me. The most obvious is that about 80% of posters that express a strong opinion are against the change. Another is that, while those against the change on average do both more observing and more IDing, they do comparatively more IDing than observing. This is also represented in the lower mean O/I ratio for this group. There’s been discussion before on the forum about how some folks differ in whether they focus more on observing or IDing, and this result may hint at a difference of opinion along those lines, with the more ID-focused users more strongly opposing this change. Users against the change also average 270% of the IDs of those for the change, a pretty large difference. Taken together, these results would make sense, as the activity of high volume IDers and more ID-focused users are more directly impacted by the change.

12 Likes

Very well written. Thanks for calculating statistics. It reveals much.

2 Likes

Then your purpose is very different to mine. I not only provide expert confirmation of identifications (as the only person who has worked on my taxa in the past years), which many people value and often tag me to confirm identifications, but I also require expert confirmation of my observations by knowledgeable people in the field. Not all identifications are the same. “Research grade” is not the same as an authoritative identification.

1 Like