Childish and slightly aggressive behaviour from users

[pmeisenheimer]

@traianbertau

And I often find it better to delete rather than withdraw, which makes the entry confusing with many crossed-out posts.

The process by which something is identified can be informative and educational, to both iNat members and the software that produces the Similar Species tabs. I’ve learned a great deal from the back and forth on observations I’ve got wrong. Best to leave it.

[pmeisenheimer]

@jimmoore

my note was in reference to using iNat data for other purposes; and in that context no one has the time to independently confirm all the iNat data for all taxa. Whether something attains “research grade” is meaningless because it only means two random people have made an ID. The only reliable indication of data quality usable in that context is identification by observers with expertise and experience in the relevant taxon.

I don’t know anybody who would use third party data of any sort without curating it first. The nature of the curation would depend on the nature of the research.

As for RG being “meaningless”, I’d be interested in seeing the analysis that produced that result.

[neylon]

@pmeisenheimer

I don’t know anybody who would use third party data of any sort without curating it first. The nature of the curation would depend on the nature of the research.

Unfortunately that happens all the time. See papers that utilize a data mining approach by downloading a dataset off GBIF. Go through just about any of those and you’ll find a bunch of questionable points that should have been removed.

@jimmoore

The only reliable indication of data quality usable in that context is identification by observers with expertise and experience in the relevant taxon.

I mostly agree here. In the datamining studies that I see using iNat data, there is often a tendency to only use RG data, but at the same time also not taking pains to ensure that reliable identifiers placed the ID’s (I’ve seen some datasets used that had 100’s of observations lacking any expert ID). What these miss is for some rare taxa, there might be limited available literature on the species, and therefore limited people on the platform who can make an ID.

An example was a paper that I looked at last year where on the RG dataset they used, I was the top identifier (lot of bumble bees), but flipping to all observations, John Ascher became the top identifier. They would have had a more interesting dataset if they had selected observations based on who made the ID, not if it was RG.

[pmeisenheimer]

@neylon

Unfortunately that happens all the time

It’s only unfortunate when the conclusions ignore the constraints of the data. It is a simple enough QA/QC exercise to generate a statistical characterization of the data’s precision which can, in some cases, allow derivation of useful insights without more labour-intensive curation. In fact, unscrubbed data can sometimes be useful as they are, as long as the limitations of the approach are acknowledged and respected.

I don’t have any doubt that bad science happens with iNat data. Bad science gets done with all sorts of data and expecting iNat data to be so excellent that nobody could ever do anything stupid with it is kind of a strange ask. On the other hand, good science does get done with iNat data and it annoys me when people suggest otherwise. It also annoys me that some people want iNat to dial back its emphasis on engagement and learning so they don’t have to be bothered doing their own data curation.

[sedgequeen]

iNaturalist gives me plenty of practice in being wrong. Sometimes I’m astonished how wrong. The errors where I clicked the wrong names are more amusing than embarrassing (the Echidna Moray is not the mammal Echidna, I learn) but some are just careless or ignorant mistakes. Embarrassing! So, I get practice in saying “Thank you for the correction,” even when most of what I’m feeling isn’t actually gratitude.

Note: I am actually an expert in some of the taxa I ID here. Pretty darn good at many of the others, too, though perhaps not as many as I think.

I remember the first time I got serious edits on a paper I wrote. I was upset and angry. After many years, I got so I could sit on the same sofa with a person editing a paper for me and smile and thank her. (The smile may not have been entirely sincere, but that’s not really important in the long run.)

So. It takes time to get to where we can accept correction gracefully. Many of us on iNaturalist aren’t there yet. This is not one of the things we expect to learn on iNaturalist, but one of the things we get the opportunity to learn.

[jnstuart]

Well said. It’s good to be wrong occasionally since you learn something new in the process. But, let’s face it, it’s so much better if the other person is wrong. ;-)

[sedgequeen]

Oh, yes!