Does 'Research Grade' actually mean anything?

Of course the name means something, and I think the fact that people are dismissing such confirms the problem I’m pointing out.

If the name could be anything, then it is functionally useless, hence a confidence meter or colour tabs or similar would at least clearly indicate the meaning and use of ID totals.

All the other things people are talking of here in terms of users reviewing and updating and debating ID’s can happen without (and in spite of) the Research Grade label.

3 Likes

I would only like to point out that the idea of the taxon and its identification as something objective and unambiguous (which I sometimes encounter among not only amateur naturalists, but also experts) is partly false. Certainly, scientific taxonomy is always based on objective facts (otherwise it would not be scientific). But the set of facts available to us is always limited, and their interpretation may be different. Therefore, the scientific concept of a taxon itself can be a matter of consensus in the scientific community.

For example, the taxon Megacopta punctatissima has 130 RG observations. But the authors of the Catalogue of Palaearctic Heteroptera are of the opinion that it is only a junior synonym of M. cribraria. Is M. punctatissima an “RG taxon”?

In fact, any identification (regardless of the number of confirmations, including by experts) may be false. Scientific taxonomy is created, in essence, by the same people, and errare humanum est (perseverare diabolicum :wink:).

Therefore, making any precise gradation of the “quality” of identification seems to me an unlikely task. Regardless of whether there are 2, 3, or, for example, 10 categories. There will always be uncertain “gray areas” and variability within the categories.

The existing category “Research Grade” is a rather clearly defined entity, a well-known and widely used term. Even if its name does not quite capture its essence and may be misleading, I would be extremely cautious about changing it. Any change of such a term (used by millions of people) can have serious and not always predictable effects.

As for increasing the number of identifications needed to achieve RG, I will only repeat myself and agree with my colleagues’ opinion - “we don’t have so many experts”. That is, this would automatically take numerous quite reliably identified observations out of this category. And it may reduce the motivation of identifiers because qualitative improvement of the data (their transition to a “higher” category) in many groups of organisms will become unattainable.

In addition, I think it is important not to forget that any change in the RG category will affect the iNaturalist Research-grade Observations dataset in GBIF. Many of the data in this dataset are already in use by researchers. Consequently, such changes would also affect their work and its results. It seems to me that really serious reasons are needed to consider this an acceptable risk.

8 Likes

I’d suggest that that is the popular interpretation of research, whereas those of us involved in it are extremely aware of just how messy it is, and (in the sciences dealing with species nomenclature) how often scientific species names change. This is a subject that @charlie has commented on often, pointing out that often the frequently maligned common names are more stable and long-lasting than the academically assigned binomials.

5 Likes

No, it can’t, you need to know when your observations are going to GBIF, you need to easily filter what you want to id or see. So no, name is just put there to be there, because everything should have a name, if there were no RG written we would make our own names, because that’s how human brain is working.

5 Likes

:joy: Agreed ~ but I’d probably be less likely to volunteer my time as an IDer if it led to more “George observations”. Contributing to make something “Research Grade” just sounds more worthwhile.

~ no offense to the Georges out there:)

4 Likes

Assuming all the basic criteria for a verifiable observation are fulfilled, it just means that more than one person gave the same ID which happens to be the majority. How they each arrived at that ID and their reasoning is more important, because it makes the difference between agreeing and independently arriving at the same conclusion

Well, that’s what it says on the tin. :-)

As a researcher, I do not consider the “research grade” marker to be useful in identifying the observations that will be informative to me. Other researchers will have the opposite opinion. There’s a lot of other context involved.

“Research grade” certainly does mean something—but it tends not to mean what people think it means.

5 Likes

Yup. Viewed in this light, we can think of “research grade” as meaning: If you’re OK with messy data and working on a problem where the volume of data is going to make the analysis pretty resilient to errors in individual observations, you can probably relax your data QA / QC somewhat.

It doesn’t mean “just trust the data and don’t do any QA / QC”, and conversely lacking the “research grade” marker doesn’t mean an observation isn’t useful for research. “Research grade” just means there’s been a little more QA / QC on the iNaturalist side, and for some research that’ll make your life easier.

10 Likes

I don’t see any issue with the terminology. Research Grade means this data is being used on GBIF, and thus being used for research. “Research Grade” in no way implies the data is flawless, because no dataset used for any research is flawless, even when it is collected with the specific research objectives in mind. Hate to break it to you, but all datasets are messy to one extent or another.

6 Likes

Good point

I think there are things we can change in the ID framework without negating or taking away from the Research Grade apparatus
For example, it might be best to remove the ‘agree’ button on the list of observations requiring ID, as it definitely confers a level of bias on those who don’t really know what they’re ID’ing but just want to help. Also, it might be worthwhile to only allow people to punt a previously given ID or give an ID of their own once and only once they have opened all available photos for that particular observation

This would undoubtedly cause a lot of backlash from the top identifiers. They are the reason iNat functions as well as it does and they don’t appreciate functions that make their actions harder.

There was even backlash when iNat tried to remove the “agree” button from observations that were already Research Grade:

6 Likes

Only if you limit the amount of photos allowed to be uploaded which I am against as well. What is an easy ID for me might be blood sweat and tears for someone else. There are a lot of really easy ones on here that only needed one photo where the photographer uploaded a dozen or more, but they didn’t know what they were supposed to photograph. I’ll glance through, but it’d be rather annoying if I had to open every one to be able to agree.

4 Likes

It does not necessarily mean something but, especially if an observation has received many identifications from skilled users, it is a useful tool to distinguish observations that are reliably identified from those that, for example, cannot be identified for many reasons. In other words, it can be useful for scientific research. Not by chance such RG observations are weekly included in the GBIF database. And this is why observations that should not be eligible for such a scope should not reach RG.

“RG” is everywhere. It’s a “Really Good” ID and I’m happy with that. If you disagree, fine, maybe you’re right. Even if you’re a maverick, it will get the attention of someone doing serious research. If you explain yourself, maybe someone who knows basic botany will have starting point to learn enough to disagree as well, removing that RG status.

2 Likes

I’d be fine with the agree button disappearing altogether for new IDers or users (not sure how “new” would be defined, but there are multiple measurable/feasible options). Also, “new” might not be the best word …not sure. But if this change were to be implemented, I’d recommend also adding a withdrawal button on their own observations (on the observations of said “new” users); some new users get confused what their options are and how to execute them when a disagreement is added to one of their observations.

But removing the agree button for IDers/users who aren’t new would be surprisingly time-consuming (and annoying) …especially for IDers who ID a lot.

It’s probably already known, but there are multiple forum threads that discuss the issue of the agree button …and as mentioned in another comment, iNat has already experimented some with this issue.

1 Like

I think “Research Grade” means something, though sometimes not as much as we might wish. RG observations have been checked by somebody who is not the original observer, and agreement has been reached (though the agreed-on ID may be wrong).

For many species, RG observations are usually correct. Which species? An odd mix of well-known, easily identified species like most North American birds and taxa of interest to very few people, like the few grasses that get to RG. In the latter group, “Needs ID” observations are often misidentified but RG observations are usually right.

For some species, even RG observations are a mess. Sometimes the cause is taxonomic, sometimes it seems to be popular misunderstandings of the taxa in question. Any researcher using these data will soon figure that out, if he/she does the necessary quality control.

Like so many things on iNaturalist, Research Grade seems messy but good enough, and hard to improve given the limits of volunteer identifiers.

4 Likes

Alright, excuse me if I get a bit emotional but I have strong feelings on this topic
This is how you alienate new users. The point of citizen science is that there is no barrier to entry, as opposed to academia where you need to be able to afford and complete several years of rigerous study.

When I joined iNaturalist, I knew basically nothing. Over time I’ve become a high volume observer and skilled identifier, but when I started I wasn’t, and the reason I became such was because I liked being able to participate. It felt amazing to be part of something like iNaturalist. And I can tell you that if iNaturalist had banned me from participating because I didn’t have any credentials I probably would have just abandoned the website. And I know a lot of people who would have done the same
The point of iNaturalist is introducing new users to science. Not implementing new ways to prevent new users from participating in science.

People who use iNaturalist data in research are probably aware how it works and that the data needs extra scrutiny. Nobody is making them use this data. There is no reason to discourage people who might otherwise be fantastic users in order to maintain some mythical purity of data.

/end emotions

edit and addition

But removing the agree button for IDers/users who aren’t new would be surprisingly time-consuming (and annoying) …especially for IDers who ID a lot.

How do you suppose people will become IDers who ID a lot if they aren’t allowed to ID and learn once they join… Taxonomic knowledge doesn’t happen overnight

10 Likes

Typing an id =/= ban, it’s not like new users can’t type? It can be done in small “amounts” like make ten ids and then the button appears, that’s enough to learn how to add ids, it all should be going with the explanations popping up until the treshold is met. It would be helping new users to learn how to start, gradually, as many websites/apps/games do, where you don’t get everything possible until you’ve learnt the basis.
I don’t get why you say it twice “not allowed to id” when they literally would be allowed to do that, just not click the agree button?

1 Like

Is eagerness to click the Agree button inversely related to tenure on iNat? Perhaps, although making rash ID decisions isn’t solely the preserve of noobs - I see some iNat veterans making them all the time.

As it stands, wrong IDs can be refuted anyway, so no ID is forever. If ID was immutable I’d be more concerned about the Agree button, and might support its removal. Even though I ID a lot, I rarely use the Agree button. I know some do though.

3 Likes

I don’t get why you say it twice “not allowed to id” when they literally would be allowed to do that, just not click the agree button?

New users usually don’t start out seeing something at unknown or family level and going “oh, I know exactly what taxon this is”. You begin learning by agreeing, looking up why someone would ID something as what they did and backing them up. And, of course, anything a new user could easily ID is almost always already ID’d to species… either by the poster or by high volume identifiers. How do I know this? Again, I was a new user less than a year ago. This is how I started.

Also why? What would this accomplish other than adding a new hurdle to users and making them feel like their input is less valued? I reiterate, which I notice you didn’t really respond to, the fact that iNaturalist is for amateurs and hobbyists. If someone only wants to double-check sparrow IDs all day long, they can do that. Putting arbitrary restrictions on activity is again, how you alienate new users.

To quote @jamesjarrett00 from another thread

You hit something right on the head with this. So many people on this site are not scientists, but are just casual users. And giving them a hard time because they don’t do it quite the way you want them too is not the best route. Most are here for different reasons and different levels of learning and enjoyment.

2 Likes