Thanks for sharing all this Joss!
Great that somebody has logged and quantified the issues further to move on from the anecdotal evidence and the endless circular debates on UK Facebook groups haha.
My thoughts (lengthy …sorry! … ) …but fwiw :
Recorder names
Interesting that the recorder names were the biggest issue.
This confuses me, I see no real logic or solid argument around this still.
It makes sense historically, without photos that an observers name might carry weight.
But given iNaturalist records are with photos and most identification is done by a third party, this seems of little relevance. Especially given the privacy concerns attached to sharing real time location data associated with one´s name.
I would be curious to see the data on the verifiers and whether there was a bias here in age and gender. I would imagine there is a dominance of verifiers on iRecord who are older and/or male. Whilst I imagine those with the greatest privacy concerns would be younger and female. I know of at least four female friends (all late 20s / early 30s ) who have been subject to stalking and personally, I would not encourage friends or family to use their real names. I see this as a perception that simply needs dismantling on the iRecord side. It just seems like bad practice to me to use real names or have this as a prerequisite for records to be verified.
Identification accuracy
Great to put this argument a bit more to bed. It´s a shame you can´t find a way to obtain data on accuracy of incoming records on iRecord. When I started biological recording, I would often log species at a higher level or incorrectly - without the community to verify or guide, I imagine there are doubtless more incoming records on iRecord either inaccurately logged or simply not logged at a species level in comparison to iNaturalist records which are at least seconded, and none of which even come through if not at genus/species. My guess would be that any discrepancy between the two platforms is even less significant than verifiers claim. If they were higher overall, this would only be down to ID support on Facebook groups I expect - which is simply a less automated and less practical platform for the purpose.
Life stage and annotation
I think a big issue here is in regard to perception of records as a static datapoint - again, historically more true… but on iNaturalist they are continually being added to, annotations may be added years down the line. Personally, I often don´t add on upload for one reason or another. I am more likely to add annotation as an identifier than as an observer I think.
I would be curious to know whether you took the records from the iRecord side post-import, or if not, how you defined what age of record to subject to analysis. It must vary greatly by age of record.
As I´ve said on Facebook threads, iRecord could choose to only import data after a certain point in time to increase level of granularity in this respect if it were really a huge issue. Regardless, any verifier can presumably at least filter by age of record and choose to check older records if they so desire.
iRecord is such a black box from the outside - are records even updated after import? If it reaches RG and goes to iRecord in August, and an annotation is added in September, is the iRecord data updated if not already verified?
Licensing
I wonder if part of the issue here is that datapoints and photos need to have distinct licensing if possible - perhaps something one could feature request on the iNat side if not already requested. Though maybe it simply isn´t realistic in terms of broader dataflow to GBIF and iRecord.
Personally, I use CC-BY-NC. I have no issue with my data being openly licensed with regard to the logging of the species and location, etc. But when it comes to the photos I do wish to have control over commercial usage of extreme macro work as the photos are higher quality and equipment more costly, etc.
I wish it were easier to choose license on upload on iNaturalist though. I would always have any phone uploads on open license by default if that was an option. And I would choose to use a more open license by default if I could easily alter select photos to retain CC-BY-NC when I so desire. Again, a feature request around this may help.
Spatial precision / location names
Here again, the exact naming of a location seems to me to be more of a historical tradition than of relevance now given GPS I think (?)… Nevertheless, I accept location names can add weight to accuracy. I think there are a few issues here which could be addressed in feature requests on the iNat side, if not already logged. e.g. I´ve noticed the location name defaults to a coarse name even if you make minor adjustments of 50 metres or so. Cool to see that anyhow, this issue wasn´t as significant as verifiers presumed.
Species coverage
This just seems to me to be a trade off between an inclusive platform like iNaturalist with newer users / more photo-based vs a more exclusive platform with older users who collect specimens.
I see this as less likely to change as it seems integral to user-base.
If we do want to attempt to shift iNaturalist UK users in general, I wonder what the percentage of records are per top 50 / 100 / 500 users. Given the data has a long tail of distribution, focus on addressing any of these issues with the top users would at least see the biggest result.
I am also curious how many (if any) of these issues resonate with data users outside of the UK.
