Bias in Observation of the Day

If you are interested in plants, you might have noticed how there is an overall lack of observations that are non-animal. I decided to do a few statistics and found out it's me. 89.3% are animals and only 5.0% are plants (the next largest kingdom represented). Contrasting that with the proportion of animal observations on iNaturalist, only 59.4% are animals while 35.6% are of plants. Surprisingly, the percentages of fungi are extremely close (4.3% and 4.7%). All other kingdoms are less than 1% in both Observations of the Day and iNaturalist in general. Also, both aren't all that different in representation. Here are a couple pie charts to show the representation:

This isn't really that big a deal to me, but it is interesting. Also, I think it would be good for the person or people choosing the observations of the day to know. Lastly, I am clearly biased in my opinion of these statistics since I am a botanist but I think that, in order to encourage a more balanced perspective of the observations on iNaturalist, it would be great to have more plants.

One more thing. I just want to point out that, though the number of plant observations in proportion to animal observations is low, there are actually a huge number of plant observations on iNaturalist (over 2.3 million). This is not a plant vs. animal post. It is simply an observation about the Observation of the Day has a bias toward animals and away from plants in comparison to the proportion of organisms in all kingdoms on iNaturalist.

If you are interested in the percentages, here they are:

All observations:
Plants: 35.6%
Animals: 59.4%
Fungi: 4.7%
Bacteria: <0.1%
Protozoa: 0.1%
Kelp, diatoms, and allies: 0.2%
Archaea: <0.1%
Viruses: <0.1%

Observations of the day:
Plants: 5.00%
Animals: 89.3%
Fungi: 4.3%
Bacteria: 0.1%
Protozoa: 0.6%
Kelp, diatoms, and allies: 0.4%
Archaea: <0.1%
Viruses: 0.1%

The observations that have been observations of the day in the past can be found here. The project itself can be found here.

Posted by nathantaylor nathantaylor, November 16, 2017 20:15

Comments

Thumb

@sambiology here you go. :)

Posted by nathantaylor almost 2 years ago (Flag)
Thumb

For what it's worth, I didn't even know there was an "Observation of the Day".

Posted by aspidoscelis almost 2 years ago (Flag)
Thumb

I think it's more of the "pizazz" factor of vertebrates... I checked to see how many of those observations were of vertebrates and like half of them were!
https://www.inaturalist.org/observations?captive=any&place_id=any&project_id=5358&subview=grid&verifiable=any&iconic_taxa=Actinopterygii,Mammalia,Reptilia,Amphibia,Aves

There are waaaaay more species of arthropods than the vertebrates combined, but the observation of the day I think is chosen for the amount of likes and shares and not necessarily the ecosystem importance. I'm not totally sure though.

The way the observation of the day is chosen is by the amount of "favs" an observation gets, I believe.

Just for you, Nathan, I observed some plants today. :-D

Posted by sambiology almost 2 years ago (Flag)
Thumb

We just need to get everyone to "fav" the forbs.

Posted by aspidoscelis almost 2 years ago (Flag)
Thumb

Sam, I figured that was probably it. I thought about breaking down the kingdoms into smaller groups but realized that was more work than I wanted to put into such a fleeting idea. I might do it another day, though. I have all kinds of ideas for how I might pick out observation bias as well, but that's a lot bigger project for when I have a lot more than 15-20 minutes to do some figuring.

I'm not sure that the "favs" necessarily choose the observation of the day. This observation, for instance, doesn't have any "favs". There are some others like it too. Not sure how these are performing on social media, but I get the feeling it may be more arbitrary than that. That being said, if "favs" do have something to do with it, I'm going to use it a lot more liberally from now on.

Patrick, it would be great if more people "faved" the forbs!

Posted by nathantaylor almost 2 years ago (Flag)
Thumb

FYI, I only have one non-plant "faved". When it comes to plants, I am unabashedly biased. :)

Posted by nathantaylor almost 2 years ago (Flag)
Thumb

To get at observation bias, you just need an unbiased set of observations to compare to iNaturalist. I have a large (ca. 67,000 observations at ca. 3700 sites) set of plant observations for southern NM that are, although certainly not unbiased in a rigorous statistical sense, about as close as you'll get without someone making observations at random locations. AIM data (a BLM ecological monitoring program with randomized plots that have a plot-level plant inventory for each) would pretty well get you there in places where there are lots of plots. It would take some work, either way. I've thought about having my AIM crews upload all observations at plots. If iNaturalist got its taxonomy sorted out properly, I'd probably have the crews do it for at least a subset of plots. It might be a bit too time-consuming to do at every plot.

Posted by aspidoscelis almost 2 years ago (Flag)
Thumb

I tell you what, if you get me a list of names or send them to me when you see them, I'll start trying to sort them out. Since I'm a curator, I can switch names. I have mostly just stayed in Euphorbia, but it would probably be good to venture out.

The taxonomy here is definitely not the greatest in certain areas. I think a lot of that comes from an over-reliance on The Plant List, but there are other reasons just like with any taxonomy.

Posted by nathantaylor almost 2 years ago (Flag)
Thumb

I've heard that before -- iNat has some taxonomic issues... Very true, but I'd sure like to see ANY authority without 'taxonomic issues!' I vividly remember my time at BRIT where specimens would have taxonomic arguments through annotations on specimens. It was kinda funny, actually. :)

I think The Plant List was chosen as an authority based on global plants (opposed to USDA or FNA), but I'm not totally sure. Anyways, as Nathan mentions, there's a good group of curators here on iNat that try our best to find common grounds on taxonomic decisions. :) It's not always a win-win, but hey, our taxonomic distinctions are eternally temporary. :)

Also, just FYI -- here's the taxonomic policies on iNat from the curator guide:
https://www.inaturalist.org/pages/curator+guide#policies

Posted by sambiology almost 2 years ago (Flag)
Thumb

Well, I could become a curator and do it myself, I'm sure, but I think the long-term solution is along the lines of the following: have a big list of all the names; allow any observation to be identified, annotated, etc., as any of those names; have a default taxonomic database indicating which are the accepted names and which are synonyms; allow users to upload taxonomic data that overrides that default taxonomic database globally, or in particular taxa. For instance, if person A thinks the Rhus with three leaflets in NM is Rhus trilobata and person B thinks that taxon is Rhus aromatica... let person A call it Rhus trilobata, and see, annotate, etc., observations under that name... let person B call it Rhus aromatica, and see, annotate, etc., observations under that name. I don't think the "here's the list of correct names" approach is workable. People disagree, the system needs to be able to deal with that. There is no one right list of accepted names and synonymies, and while I enjoy arguments about what might be on that mythical list in particular instances, for a database like iNaturalist the observation management functionality shouldn't be hostage to those discussions. It should be built to work fine among disagreeing viewpoints.

Posted by aspidoscelis almost 2 years ago (Flag)
Thumb

Great points, Patrick. Also, just so you know, I'm SUPER glad that you're on iNat (and of course, helping me out with all of those NM plants!). :)

I actually think that this fits in somewhat with Nathan's observation of the day -- perhaps one of the reasons that many plants don't get observed as much as the vertebrates is because of the extraordinary diversity of plants! There's too many plants for a good 'field guide' to even a region. There are Flora but those tend to be more technical than the photo-based field guides. Whereas it's pretty dang easy to find a guide to the 'snakes of New Mexico' or 'birds of Arizona.'

We tend to focus on the stuff we know well or think that we can ID fairly easily (hence why I tend to avoid many sedges or grasses!). And I think those add up to folks 'liking' and 'faving' the stuff that they could ID easily as well. :)

Posted by sambiology almost 2 years ago (Flag)
Thumb

I'm definitely in favor of making Patrick a curator. Anyone here opposed?
All good points. I wonder if it would be good to bring one of the admins here for discussion. I was actually talking to @loarie about some taxonomic confusion in Kalmia earlier today and I think this conversation is very relevant to that one.

Posted by nathantaylor almost 2 years ago (Flag)
Thumb

I've been having a conversation along similar lines here, for what it's worth:

https://groups.google.com/forum/#!topic/inaturalist/nSXEUKwKvSI

Posted by aspidoscelis almost 2 years ago (Flag)
Thumb

@aspidoscelis , I lurk the google group all the time, and I was following that discussion. I suppose that's what I enjoy the most about iNat -- we all use it for different reasons! The platform itself is not for just one purpose, so folks use it for various things. As an urban wildlife biologist, I use it mostly for the 'engagement' factor and to influence policy change (showing city councils/park boards what critters/plants live in the public spaces and where these things are and who cares about them -- the observers).

As a guy that kinda enjoys biodiversity, I also use it to remind myself of all of the other organisms on the planet! That's how I have some fun on it -- not to mention, getting to interact with naturalists from around the world.

Also, @nathantaylor7583 -- I checked, and TWO Euphorbia made observations of the day: https://www.inaturalist.org/observations?captive=any&place_id=any&project_id=5358&subview=grid&taxon_id=51821&verifiable=any

Posted by sambiology almost 2 years ago (Flag)
Thumb

That's great! I'm excited that this issue is being discussed. I really hope that it moves forward as this could really help a lot.

Posted by nathantaylor almost 2 years ago (Flag)
Thumb

@sambiology Yeah, those African Euphorbs really have a way of making people turn their heads. :-) The Chamaesyces, not so much.

Posted by nathantaylor almost 2 years ago (Flag)
Thumb

Maybe to separate them out from the others we should call them the eueuphorbs.

Posted by aspidoscelis almost 2 years ago (Flag)
Thumb

I'd go and fav all the Chamaesyce, but I can't search for Euphorbia sect. Anisophyllum in the little box.

Posted by aspidoscelis almost 2 years ago (Flag)
Thumb

No, actually I can! I was betting that wouldn't work. Score for the iNaturalist taxonomic database on this one.

Posted by aspidoscelis almost 2 years ago (Flag)
Thumb

I made sure I was able to add sect. Anisophyllum in. ;-)

Posted by nathantaylor almost 2 years ago (Flag)
Thumb

Has anyone gone through the Observation of the Day posts on Twitter and Instagram and tabulated how many likes plants get compared to others? As said above by others, it might reasonably be a conscious bias to target what people actually want.

Posted by colinpurrington over 1 year ago (Flag)

Add a Comment

Sign In or Sign Up to add comments

Is this inappropriate, spam, or offensive? Add a Flag